LongBench v2 / 66f00f1a821e116aacb252cf / Considering both ‘Social Identity and Intergroup Conflict and Conflict…

List o
f
 Tables and Figures 
List o
f
 Contributors 
Pre
f
ace 
Contents 
Part I 
Representations of Self and Group 
Regarding the Relationship Between Social Identity and 
Personal Identity 
Jean-Claude Deschamps and Thierry Devos 
VII 
IX 
XVll 
2 
Social Representations in Personal Identity 
1 3  
Willem Doise 
3 
The Personal-Social Interplay: Social-Cognitive Prospects on 
Identity and Self-Others Comparison 
24 
Carmencita Serino 
4 
Self 
-We-Others Schemata and Social Identifications 
44 
Maria Jarymowicz 
Part II Establishing Group Identity 
53 
5 
A Developmental View of the Search for Group Identity 
53 
Stephen Worchel 
6 
The Role of Prototypicality in Group Influence and Cohesion: 
Contextual Variation in the Graded Structure of Social 
Categories 
75 
Penelope Oakes, S. Alexander Haslam and John C. Turner 
7 
Group Beliefs as an Expression of Social Identity 
Daniel Bar-Tal 
8 
Social Identity and National Identity 
Jose Miguel Salazar 
Part III Identity and GrouplIntergroup Process 
9 
Social Identity and Intragroup Differentiation as Subjective 
93 
1 14 
124 
Social Control 
124 
Jose M. Marques, Dario Paez and Dominic Abrams 


VI 
Social identity: international perspectives 
10 
Categorization and Social Influence 
142 
Juan Antonio Perez and Gabriel Mugny 
11 
Social Identity and Interpersonal Relationships 
154 
Miguel Moya 
12 
Social Identity and Aspects of Social Creativity: Shifting to 
New Dimensions of Intergroup Comparison 
166 
Steve Hinkle, Laurie A. Taylor, Lee Fox-Cardamone and 
Pamela G. Ely 
13 
Mere Categorization and Competition: A Closer Look at 
Social Identity Theory and the Discontinuity Effect 
180 
Stephen M. Drigotas, Chester Insko and John Schopler 
Part IV 
The Impact of Culture on Identity and Categorization 
199 
14 
Discrimination and Beliefs on Discrimination in Individualists 
and Collectivists 
199 
J. Francisco Morales, Mercedes Lopez-Saez and Laura Vega 
15 
Constructing Social Identity: The Role of Status, Collective 
Values, Collective Self-Esteem, Perception and Social 
Behaviour 
211 
Dario Paez, Cristina Martinez-Taboada, Juan Jose Arrospide, 
Patricia Insua and Sabino Ayestaran 
Re
f
erences 
Index 
230 
255 


List of Tables and Figures 
Tables 
4.1 The social self as a structural basis of different forms of 
social identifications 
4S 
S.1 Perceived differences in ingroup and outgroup similarity 
70 
S.2 Minority influences in groups over time 
72 
1l.l Meta-contrast ratios in self-perception and the four stimuli 
used by Moya and Peplau 
162 
13.1 Max. reI. as a function of matrix type (PDR versus PDC), 
trials and groups versus individuals, experiment 2 
190 
14.1 Summary of some differences between individualists and 
collectivists 
202 
14.2 Relation between identification with the goup and intergroup 
differentiation as a function of individualism/collectivism and 
autonomous/relational orientation 
203 
IS.1 Perceived ingroup and outgroup heterogeneity 
216 
IS.2 Perception of ingroup and outgroup by salience versus 
non-salience of social comparison 
217 
IS.3 Perceived social desirability of attributes for engineers and 
psychologists 
218 
15.4 Attributes for engineers and psychologists 
220 
IS.5 Factorial analysis of the 'most important social identity 
scale' (Garza & Herringer, 1986) 
224 
IS.6 Differences between engineers and psychologists as a 
function of their degree of collectivism 
227 
Figures 
1.1 Personal and social dimensions in psychosocial identity 
11 
3.1 Mean scores (x 100) of each content category elicited in self/ 
ingroup versus self/outgroup similarity judgements ( N  = 121) 
39 
4.1 Selfˮwe schemata distinctness and the degree of individuals' 
involvement with outgroups and dissimilar others: the 
theoretical model 
SO 
4.2 Recall of children's problems 
SI 
4.3 Motivation to solve children's problems 
51 
4.4 Readiness to host different nationalities 
SI 


Vlll 
Social identity: international perspectives 
4.5 Tolerance of German minority activity 
5 1  
4.6 Positivity of evaluation of Slovak and Irish paintings 
52 
4.7 Positivity of evaluation of Japanese painting 
52 
5.1 Stages of group development in existing groups 
58 
5.2 Nature of desired interaction with outgroup as a function of 
group life 
68 
6. 1 A categorial hierarchy of vegetation 
78 
6.2 A categorial hierarchy of Europeans 
79 
9. 1 Manipulation of group uniformity and group members' 
status: distribution of ingroup and outgroup members' 
responses 
1 34 
9.2 Evaluations of attractive and unattractive ingroup and 
outgroup members as a function of group uniformity 
1 35 
9.3 Example of ranking of characters A-F used to build modal 
and deviate members 
1 36 
9.4 Effects of norm awareness on judgements of modal and 
deviate ingroup and outgroup members 
1 37 
9.5 Effects of ingroup and outgroup accountability on 
judgements of modal and deviate ingroup and outgroup 
members 
1 38 
9.6 A two-step process of inter- and intragroup judgement 
140 
1 1 . 1  Interpersonal and intergroup demands according to 
Stephenson (1984) 
1 58 
1 1 .2 Example of a hypothetical identity taken from Deaux's 
formulation and Moya's data 
161 
12. 1 
Ingroup favouritism on different comparison dimensions as a 
function of outgroup status 
1 73 
12.2 Ingroup favouritism on different comparison dimensions as a 
function of outgroup status 
1 75 
1 3. 1  A Prisoner's Dilemma Game matrix (PDG) 
1 88 
1 3.2 A Prisoner's Dilemma max. reI. matrix (PDR) and a 
Prisoner's Dilemma confounded matrix (PDC) 
1 89 
1 3.3 A Prisoner's Dilemma Game alternative matrix (PDG-Alt.) 
195 


List of Contributors 
Dominic Abrams is Professor of Social Psychology and Director of the 
Centre of the Study of Group Processes at the University of Kent, UK. His 
main research interests are intergroup conflict and differentiation, social 
identity processes and social self-regulation. He is co-author and co-editor 
(together with M.A. Hogg) of four books: Social Identifications (London: 
Routledge, 1988); Social Identity Theory (Hemel Hempstead: Harvester 
Wheatsheaf, 1990); Group Motivation (Hemel Hempstead: Harvester 
Wheatsheaf, 1993); and Social Identity and Social Cognition (Oxford: 
Blackwell, forthcoming, 1998). 
Juan Jose Arrospide is currently Assistant Professor at the Department of 
Social Psychology, School of Psychology, Basque Country University, 
Spain. He has done research on the topic of social identity, group dynamics 
and social representions. He is co-author (together with S. Ayestaran and 
C. Martinez-Taboada) of a chapter on 'New Techniques for the Study of 
Groups' ('Nuevas Tecnicas en el estudio de los grupos') included in 
Methods and Techniques o
f
 Research in Social Psychology ( Mhodos y 
Tecnicas de Investigacion en Psicologia Social, Madrid: Eudema, 1994). 
Sabino Ayestaran is currently Professor of Social Psychology at the 
Department of Social Psychology, School of Psychology, Basque Country 
University, Spain. His research interests focus on social identity and group 
dynamics, with special attention to intra- and intergroup conflict. He is the 
editor of a volume on The Group as a Social Construction (El Grupo como 
Construccion Social, Barcelona: Plural, 1996). 
Daniel Bar-Tal is Professor of Psychology at the School of Education, Tel­
Aviv University, Israel. His research interests focus on political and social 
psychology, studying societal beliefs regarding conflict, delegitimization, 
security, patriotism and siege mentality. He is the author of Group Belie
f
s
 
(New York: Springer Verlag, 1990) and Societal Belie
f
s
 o
f
 Ethos (London: 
Sage, in press) and co-editor of Social Psychology o
f
 Knowledge (with W. 
Stroebe, A.W. Kruglanski and M. Hewstone; New York: Springer Verlag, 
1988), Stereotyping and Pre
judice (with C. Grauman, A.W. Kruglanski and 
W. Stroebe; New York: Springer Verlag, 1989), Patriotism in the Lives o
f
 
Individuals and Nations (with E. Staub; Chicago: Nelson-Hall, 1997) and 
Concerned with Security (with D. Jacobson and A. Klieman; Greenwich, 
CT: JAI Press, in press). 


x 
Social identity: international pers
pectives 
Jean-Claude Deschamps, born in France, is currently Professor of Social 
Psychology at the School of Social Sciences, Lausanne University, 
Switzerland. He has studied social identity and intergroup relations, the 
processes of causal attribution and social representations. He is co-author 
(with W. Doise) of a chapter on 'Crossed Categories Membership and 
Intergroup Relations', included in H. Tajfel (ed.), Differen
Social Groups (Cambridge: Cambridge University Press, 1978) and (with A. 
Clemence) of a volume on Everyday Explanation (L'ex
plication quotidienne, 
Friburg: Delval, 1987). 
Thierry Devos is Assistant Professor at the School of Social Sciences, 
Lausanne University, Switzerland. He has studied social identity, causal 
attributions and social representations. He is co-author (with J.-c. 
Deschamps and L. Comby) of 'Assymmetries in Judgment of Ingroup 
and Outgroup Variability', European Review o
f
 Social Psychology (1996). 
Willem Doise, born in Belgium, is Professor of Social Psychology at the 
School of Psychology, University of Geneva, Switzerland. His research has 
focused on the link between mind and society, specifically social represen­
tations and the role of socio-cognitive conflict in mental development. He 
has published several books, among them Groups and Individuals 
(Cambridge: Cambridge University Press, 1978) and Levels o
f
 Explanation 
in Social Psychology (Cambridge: Cambridge University Press, 1986). 
Stephen M. Drigotas is currently Assistant Professor of Psychology at the 
Southern Methodist University, Dallas, Texas, USA. His recent publications 
include: 'On the Peculiarities of Loyalty' (together with G.A. Whitney and 
C.E. Rusbult), Personality and Social Psychology Bulletin (1995), 'The Role 
of Identifiability in the Reduction of the Interindividual-Intergroup Dis­
continuity Effect' (together with J. Schopler et al.), Journal o
f
 Personality 
and Social Psychology (1995), and 'Willingness to Sacrifice in Close Rela­
tionships' (together with P.A.M. Van Lange et al.), Journal o
f
 Personality 
and Social Psychology (in press). 
Pamela G. Ely is Assistant Professor and Chair of Psychology at Saint 
Andrews College in Laurinburg, North Carolina, USA. She also directs the 
Human Services Program of the Saint Andrews at Sandhills program. Her 
research focuses on the relationship between individualism/collectivism and 
group and intergroup processes, including group-oriented instructional 
methods. Her most recent publication is 'Recognizing Group Diversity: 
Individualist-Collectivist and Autonomous-Relational Social Orientations 
and Their Implications for Intergroup Processes' (with R.J. Brown et al.), 
British Journal o
f
 Social Psychology (1992). 
Lee Fox-Cardamone is Assistant Professor of Psychology at Kent State 
University - Stark Campus in Canton, Ohio, USA. Her research interests 
include political activism, intergroup processes and group decision-making. 
Her most recent publications are 'Grassroots Political Actions as an 


List o
f
 contributors 
Xl 
Intergroup Phenomenon' (with S. Hinkel et al.), Journal o
f
 Social Issues 
(1996), and 'Recognizing Group Diversity: Individualist-Collectivist and 
Autonomous-Relational Social Orientations and Their Implications for 
Intergroup Processes' (with R.J. Brown et al.), British Journal o
f
 Social 
Psychology (1992). 
S. Alexander Haslam is a Senior Lecturer in Social Psychology at the 
Australian National University. His main research interests are in the social 
psychology of stereotyping and social judgement with particular emphasis on 
the role which group memberships play in those processes. This work 
addresses issues of 
social categorization, social identity and prejudice, dealing 
specifically with topics of stereotype consensus, perceived group homo­
geneity, illusory correlation, prejudice and extremism. His recent publica­
tions include Stereotyping and Social Reality (with P.J. Oakes and J.C. 
Turner; Oxford: Blackwell, 1994), The Social Psychology o
f
 Stereotyping and 
Group Lif
e
 (with R. Spears, P.J. Oakes and N. Ellemers; Oxford: Blackwell, 
1997), The Message o
f
 Social Psychology (with C. McGarty; Oxford: 
Blackwell, 1997), and Doing Psychology (with C. McGarty; London: Sage, in 
press). 
Steve Hinkle is an Associate Professor of Psychology at Miami University 
in Oxford, Ohio, USA. His main research interests are in motivational 
aspects of social identity theory, the relationship between individualism/ 
collectivism and aspects of group and intergroup processes, and the psy­
chology of political action. Recent publications include 'Grassroots 
Political Action as an Intergroup Phenomenon' (with L. Fox-Cardamone 
et al.), Journal o
f
 Social Issues (1996) and 'Recognizing Group Diversity: 
Individualist-Collectivist and Autonomous-Relational Social Orientations 
and Their Implications for Intergroup Processes' (with R.J. Brown et al.), 
British Journal o
f Social Psychology (1992). 
Chester Insko is Professor at the Department of Psychology, University of 
North Carolina, USA. Research interests are intergroup relations, inter­
personal attraction and attitude change. Recent publications include 
'Memory for and Experience of Differential Competitive Behavior of 
Individuals and Groups' (with M.J. Pemberton and J. Schopler), Journal 
o
f
 Personality and Social Psychology (1996), and 'Differential Distrust of 
Groups and Individuals' (with J. Schopler) and 'Entitativity, Loss of Per­
sonal Control, and Evolution' (with J. Schopler and C. Sedikides), both 
included in C. Sedikides et al. (eds), Intergroup Cognition and Intergroup 
Behavior (Hillsdale, NJ: Erlbaum, in press). 
Patricia Insua is Assistant Professor of Language and Thinking in the 
Department of Basic Psychology, School of Psychology, University of the 
Basque Country. She is co-author, with D. Paez and J. Marques, of a chapter 
on 'Social Cognition' in the book edited by J.F. Morales et al. entitled Social 
Psychology (Psicologia Social) (Madrid: McGraw Hill, 1994). Her main 


xu 
Social identity: international perspectives 
research interests in social psychology are social representations, attitudes, 
AIDS and drug-addiction processes, and they also embrace general 
psychological processes of language, cognition and emotional problems. 
Maria Jarymowicz has been Professor of Psychology at the School of 
Psychology of the University of Warsaw and Head of the Laboratory 
'Studies on Personal and Social Identity' since 198 1 .  Her main research 
interests focus on ego and self in social perception, self-we-others schemata 
distinctiveness and personal/social identity differentation - in relation to 
egocentric/non-egocentric perception, the individualism-collectivism rela­
tionship and the affective basis of social stereotypes. Her most representative 
publications (both as editor) are To Know Self 
- to Understand Others (Delft: 
Eburon, 1993) and Beyond Egocentrism in Self- and Social Perception (Poza 
EgocentrycznQ PerspektywQ Widzenia Siebie i Swiata, Warsaw: Wydaw­
nictwo Instytutu Psychologii PAN, 1994). 
Mercedes Lopez-Saez is Assistant Professor of Social Psychology at the 
Department of Social Psychology of the Spanish Open University at 
Madrid (UNED). Her main research interests include gender stereotypes, 
social influence and persuasion. Recent publications include 'Cultural and 
Individual Processes Involved in Gender Stereotyping: An Empirical 
Approximation to Career Election' ('Procesos Culturales e Individuales 
Implicados en la Estereotipia de Genero: Una Aproximacion Empirica a la 
Eleccion de Carrera'), Revista de Psicologia Social (1994), 'Gender Stereo­
types in the Spanish Population: Looking Toward the Future' (with 1.F. 
Morales), in L. Amancio and C. Nogueira (eds), Gender, Management and 
Science (Braga: Institute of Education and Psychology, 1995), and The 
Election o
f a Typically Feminine or Masculine Career f
rom a Psychosocial 
Perspective (La Eleccion de una Carrera Tipicamente Femenina 0 Masculina 
Desde la Perspectiva Psicosocial), Madrid: Ministry of Science and 
Education, 1995). 
Jose M. Marques is Professor of Social Psychology at the University of 
Porto, Portugal. Research interests centre mainly on the field of intergroup 
relations, specifically the role of intergroup processes in generating group 
cohesiveness and enhancement of social identification. Recent publications 
include 'The Black Sheep Effect: Social Categorization, Rejection of 
Ingroup Deviates, and Perception of Ingroup Variability' (with D. Paez), 
European Review o
f
 Social Psychology (1994), 'The Representation of 
Emotions in Groups: The Relative Impact of Social Norms, Positive­
Negative Asymmetry and Familiarity on the Perception of Emotions' (with 
D. Paez and A.I. Vergara), European Journal o
f
 Social Psychology (1996), 
and 'Collective Memory Processes Associated with Traumatic Experiences: 
Social Sharing, Emotional Climate and the Transgenerational Transmission 
of Information in the Case of the Portuguese Colonial War' (with D. Paez 
and A.F. Serra), in 1. Pennebaker et al. (eds), Social Memory and Collective 
Trauma (Hillsdale, Nl: Erlbaum, 1997). 


List o
f
 contributors 
X 
111 
Cristina Martinez-Taboada is Assistant Professor at the Department of 
Social Psychology, School of Psychology, Basque Country University, 
Spain. He has done research on the topic of social identity, group dynamics 
and social representions. He is co-author (with S. Ayestarin and J.J. 
Arrospide) of a chapter on 'New Techniques for the Study of Groups' 
('Nuevas Tecnicas en el Estudio de los Grupos') included in Methods and 
Techniques o
f
 Research in Social Psychology ( Mhodos y Tecnicas de 
Investigaci6n en Psicologia Social, Madrid: Eudema, 1994). 
J. Francisco Morales is Professor of Social Psychology at the Department 
of Social Psychology, of the Spanish Open University at Madrid (UNED). 
He has done research on leadership, social identity and its relation to 
individualism-collectivism. Recent publications include 'Three Portraits of 
Social Psychology in Europe', Proceedings o
f
 the Tenth Meeting o
f
 Cheiron 
( 1992), 'Leadership in Two Types of Health-Care Organizations' (with F. 
Molero), in J.M. Peiro et al. (eds), Work and Organizational Psychology: 
European Contributions o
f
 the Nineties (London: Taylor & Francis, 1995), 
Treatise o
f
 Social Psychology I: Basic Processes ( Tratado de Psicologia 
Social I: Procesos Basicos (with M. Moya); Madrid: Sintesis, 1997) and 
Social Psychology ( Psicologia Social) (with M. Moya et al.); Madrid: 
McGraw-Hill, 1 994). 
Miguel Moya is Professor of Social Psychology at the School of Psychology 
of the University of Granada, Spain. His main research interests are 
gender, interpersonal and intergroup relations, prejudice and stigma. 
Recent publications include Treatise o
f
 Social Psychology I: Basic Processes 
( Tratado de Psicologia Social I: Procesos Basicos), (with J.F. Morales; 
Madrid: Sintesis, 1 997), Psychology and Gender (Psicologia y Genero) (with 
M.e. Gomez; Granada: Universidad de Granada, 1997), and Social 
Psychology (Psicologia Social), (with J.F. Morales et al.; Madrid: 
McGraw-Hill, 1994). 
Gabriel Mugny is Professor of Social Psychology at the Faculty of Psy­
chology and Educational Sciences, University of Geneva, Switzerland. His 
main research interests are social influence, intergroup relations, social 
identity, cognitive development and social representations. Recent 
publications include Psychologie Sociale et Developpement Cognitif (with 
W. Doise; Paris: Armand Colin, 1997), La Psychologie Sociale, Vol. i: 
Relations Humaines, Groupes et In
fluence Sociale (with D. Oberle and J.L. 
Beauvois; Grenoble: Presses Universitaires de Grenoble, 1995), The Social 
Psychology o
f Minority Influence (with J.A. Perez; Cambridge: Cambridge 
University Press, 1991) and In
fluences Sociales (with J.A. Perez et al.; 
Neuchatel: Delachaux et Niestle, 1993). 
Penelope Oakes born in England, is currently Senior Lecturer in Social 
Psychology at the Australian National University in Canberra, Australia. 
She has published widely on stereotyping and group processes, including 


XlV 
Social identity: international pers
pectives 
Stereotyping and Social Reality (with A. Haslam and J.C Turner; Oxford: 
Blackwell, 1 994). She is also co-editor of The Social Psychology o
f
 Stereo­
typing and Group Lif
e (with R. Spears et al.; Oxford: Blackwell, 1997). 
Dario Paez, born in Chile, is currently Professor of Social Psychology at 
the Department of Social Psychology, School of Psychology, Basque 
Country University, Spain. He has done research mainly on social identity, 
emotions and health, collective memory and social representations. He has 
published 'Inhibition, Alexythimia and Health' (with N. Basabe et al.) in J. 
Pennebaker (ed.), Emotion, Disclosure and Health (Washington: APA, 
1995), and Collective Memory o
f
 Political Events (with J. Pennebaker and 
B. Rime; Mahwah: Erlbaum, 1997). 
Juan Antonio Perez is Professor of Social Psychology at the University of 
Valencia, Spain. His main current research interests are in minority 
influence and racism against gypsies. Recent publications include Influences 
Sociales (with G. Mugny et al.; Neuchatel: De1achaux et Niestle, 1993) and 
'The Conflict Elaboration Theory of Social Influence', (with G. Mugny), in 
E. Witte and J. Davis (eds), Understanding Group Behavior, Vo!' 2 
(Hillsdale, NJ: Erlbaum, 1990). 
Jose Miguel Salazar is Professor of Psychology at the Central University of 
Venezuela, where he pioneered social psychology studies in the late 1950s. 
Currently he is Editor of the Revista Interamericana de Psicologial 
Interamerican Journal o
f
 Psychology. His main interests centre on issues of 
nationalism, national identity and supra-nationalism, and he has done 
research on the relationships of nationalistic attitudes to behaviour, Latin 
American identity and value changes, and, more recently, the construction 
of national and quasi-national identities. He has published Psychological 
Foundations o
f
 Nationalism (Bases Psicol6gicas del Nacionalismo, Mexico: 
Trillas, 1987) and numerous articles and chapters on the subject. 
John Schopler is Professor of Social Psychology, Department of 
Psychology, University of North Carolina, Chapel Hill, USA. His main 
research interests are in intergroup relations, social power, interdependence 
theory and conflict resolution. His recent publications are 'The 
Discontinuity Effect in Interpersonal and Intergroup Relations: Generality 
and Mediation' (with CA. Insko), European Review o
f
 Social Psychology 
(1992), 'The Role of Communication in Interindividual-Intergroup 
Discontinuity' (with CA. Insko et al.), Journal o
f
 Con
flict Resolution 
(1993), 'Individual-Group Discontinuity: Further Evidence for Mediation 
by Fear and Greed' (with CA. Insko et al.), Personality and Social 
Psychology Bulletin (1992), 'The Survival of a Cooperative Tradition in the 
Intergroup Discontinuity Context' (with CA. Insko et al.), Motivation and 
Emotion ( 1994), 'The Role of Identifiability in the Reduction of 
Interindividual-Intergroup Discontinuity' (with CA. Insko et al.), Journal 
o
f
 European Social Psychology (1995), 'Memory for and Experience of 
Differential Competitive Behavior of Individuals and Groups' (with M.J. 


List o
f
 contributors 
xv 
Pemberton and CA. Insko), Journal o
f
 Personality and Social Psychology 
(1996), and Intergroup Cognition and Intergroup Behavior (with C 
Sedikides and CA. Insko; Hillsdale, NJ: Erlbaum, in press). 
Carmencita Serino is Lecturer in Social Psychology at the Department of 
Psychology of the University of Bari, Italy. Her main research topics are 
social categorization, intergroup relations, similarity judgement, self-others 
comparison and social representations. Her recent publications include 
Asymmetry Phenomena in the Interpersonal Comparison (with L. Arcuri; 
Naples: Liguori, 1992), 'Information Processing and Self-Others Compar­
ison: Social Cognitive Extensions of a Non-Metrical Approach to Simi­
larity', in M.C Hurtig et al. (eds), Studies on the Self and Social Cognition 
(New York: World Scientific Publishing, 1993), 'Mental Illness in Psycho­
social Research: Processes of Explanations and Implicit Personality 
Theories' ('La Malattia Mentale Nella Ricerca Psico-Sociale: Processi di 
Spiegazione e Teorie Implicite della Personalita'), in G. Bellelli (ed.), The 
Other Illness: How Society Thinks about Mental Illness (L'Altra Malattia: 
Comme la Societa Pensa la Malattia Mentale, Naples: Liguori, 1994), 
Singular Worlds: Socio-Cognitive Perspectives on Identity ( Mondi al 
Singolare: Prospective Socio-Cognitive Sull'identita) (with J.C Deschamps; 
Naples: Liguori, 1996), and 'Similarities and Differences in the Social 
Representations of the Handicapped' (,Somiglianza e Differenza nella 
Rappresentazione Sociale dell'Handicap', in G. Petrillo (ed.), Social Psy­
chology o
f
 Health ( Psicologia Soziale della Salute, Naples: Liguori, 1997). 
Laurie A. Taylor is the Manager of Community Evaluation Processes at the 
United Way of the Greater Dayton Area in Dayton, Ohio, USA. Her 
research interests include group decision-making, intergroup perceptions 
and identity formation, conflict management and applied outcome 
evaluation. She has most recently published 'Recognizing Group Diversity: 
Individualist-Collectivist and Autonomous-Relational Social Orientations 
and Their Implications for Intergroup Processes' (with R.J. Brown et al.), 
British Journal o
f
 Social Psychology (1992). 
John C. Turner, born in England, is Professor of Psychology and Head of 
Division, Division of Psychology, Australian National University. His 
research interests are in social identity, intergroup relations, group 
processes and social cognition, particularly from the perspective of self­
categorization theory. He collaborated with Henri Tajfel at Bristol in the 
1970s in the development of social identity theory and is the originator of 
self-categorization theory. Recent publications include Stereotyping and 
Social Reality (with P.J. Oakes and S.A. Haslam; Oxford: Blackwell, 1994), 
'Social Identity, Interdependence and the Social Group: A Reply to Rabbie 
et aI.' (with R.Y. Bourhis), in W.P. Robinson (ed.), Social Groups and 
Identity (Oxford: Butterworth Heinemann, 1996), and 'The Socially Struc­
tured Mind' (with P.J. Oakes), in CA. McGarty and S.A. Haslam (eds), 
The Message o
f
 Social Psychology (Oxford: Blackwell, 1997). 


XVI 
Social identity: international pers
pectives 
Laura Vega has done graduate studies in Social Psychology at the Depart­
ment of Social Psychology of the Spanish Open University at Madrid 
(UNED). She is currently working for her doctoral dissertation and doing 
research on the relations between individualism and identity differentiation. 
She is co-author of a paper on 'Individualism, Collectivism and Social 
Identity' (,Individualismo, Colectivismo e Identidad Socia!'), included in a 
monographic volume on 'Social Identity' published by the Revista de 
Psicologia Social ( 1992). 
Stephen Worchel is Dean of the College of Arts and Sciences and Professor 
of Psychology at the University of Southern Maine. His research areas 
include intergroup relations, group dynamics, conflict resolution and cross­
cultural psychology. Recent publications related to social identity include 
the following: 'The View from Below: Intergroup Relations from the 
Perspective of the Disadvantaged Group' (with H. Rothgerber), Journal o
f
 
Personality and Social Psychology ( 1997); 'A Dynamic Approach to Group 
Dynamics', in P. Nye and A. Brower (eds), What's Social about Social 
Cognition? (Newbury Park: Sage, 1996); 'Changing the Stereotype of 
Stereotypes' (with H. Rothgerber), in R. Spears et al. (eds), The Social 
Psychology o
f
 Stereotyping and Group Lif
e (Oxford: Blackwell, 1997); and 
Con
flict between People and Groups (with J. Simpson; Chicago: Nelson­
Hall, 1993). 


Preface 
This volume deals with the issue of identity, the concern over which is far 
from being something new. In fact, it can be traced to classical Antiquity. 
Identity is present in the famous 'Know Thyself' precept, inscribed on the 
frontispiece of the Delphi Temple. It is also the target of many of the 
famous Socratic questions and the stimulus for Plato's well-known reflec­
tions. It can be considered, without exaggeration, an issue which has 
occupied humankind since the beginning of civilization. 
It also focuses on categorization or, more explicitly, on the analysis of 
how we perceive other people (stereotypes) and how we decide to behave 
towards them (prejudice, discrimination), and, in general, on the under­
standing of our readiness to come to grips with the social environment. All 
these problems have a long tradition, but they cannot be taken as old­
fashioned or as belonging to the past, since they are pervasive in our 
current, so-called 'civilized' societies. 
Admittedly, it would be illusory to try to find in the Socratic-Platonic 
analysis of the Delphic formula the point of departure of a current 
psychosocial analysis of the personal and social self. Only in the last 20 
years or so have social psychologists provided such an analysis, framed in 
the recent, but already well-known theories of social identity, and social 
and self-categorization. It is the purpose of this volume to offer an updated 
and balanced view of these fairly new developments, in the expectation that 
it will foster a better understanding of some of the more urgent and central 
problems of our time. 
The latest elaborations of the theories and the most recent research 
inspired by them are especially attended to. A detailed summary of them 
would exceed the length usually allowed to any preface. However, we can, 
at least, highlight three of their main contributions, closely linked to three 
salient features of the theories mentioned. 
In this respect, it should be stressed that one of the most important assets 
of the social identity and self-categorization theories is the building of an 
original theoretical corpus which has provided social psychology with new 
theoretical insights. 
Undoubtedly, before Tajfel's formulation of social identity theory, there 
were several psychosocial approaches to the study of identity. However, 
perhaps due to their low theoretical coherence, they did not succeed in 
making social identity a crucial concept of the discipline of social psy­
chology. The same can be said of self-categorization theory. 


XVlll 
Social identity: international perspectives 
This theoretical corpus was not without its appropriate empirical content, 
since the theoretical concepts and propositions were elaborated through an 
intense dialogue with certain psychosocial phenomena demanding explana­
tion, such as intergroup discrimination, the effect of stability and legitimacy 
of status differences, the processes of social comparison and of group 
competition, among others. In providing such an explanation, both theories 
became more and more influential and began to play a leading role within 
social psychology. 
Going still a step further, they widened their scope to accommodate 
other psychosocial phenomena. Let's take the example of the group con­
cept. Tajfel himself had suggested the need for its reformulation. However, 
this was only done by Turner several years later. The covariation hypoth­
esis, formulated by Deschamps, and linking intra-
and intergroup 
differentiation, is just another example of the success of the theories in 
their efforts to cover other psychosocial phenomena. Besides the group 
concept and the covariation hypothesis, the following processes, among 
others, are worth mentioning: influence, group polarization, collective 
behaviour, the impact of minorities, prejudice, stereotypes and leadership. 
This speaks in favour of the fruitfulness of both theories. 
These three features correspond approximately to the stages of develop­
ment of the theories, but, at the same time, all three have been, and still 
are, open and subject to changes and modifications. The fact that there is a 
mutual influence between them cannot be forgotten either. Thus, the 
intensive study of intergroup discrimination has helped to reformulate the 
early concepts, throwing light on the functions of comparison dimensions. 
By the same token, the extension of the theory to the group concept has 
allowed Turner to replace the early interperson-intergroup continuum by a 
new, more accurate interperson-group one. 
In recent years, the sustained vitality of the theories has allowed them to 
gain new empirical content with the corresponding increase in the amount 
of explanations of phenomena of interest for social psychoiogy. As a result, 
they have become both more complex internally and more sophisticated in 
their theoretical formulation. The contributors to this volume, among them 
some of the original formulators of both theories, have tried to give them 
even more impulse. They raise new issues, examine critically some crucial 
points, and explore new areas of research. Their work will serve to expand 
the theories and, we hope, to persuade others to join the effort. 
The overall aim of this volume is, then, to demonstrate the breadth of 
scope of issues related to social identity. Indeed, one of the most unique 
features of the area is that it has found a warm welcome in a variety of 
camps. As such, we begin the volume with chapters addressing identity and 
the self. We move from here to contributions that deal with identity in the 
group and intergroup context. And the final contributions venture into the 
realm of cultural influences on social identity. 
But the breadth of social identity is not only demonstrated by its con­
tribution across the social spectrum. It is also a theory that has captured 


Pre
f
ace 
xix 
the imagination of investigators in diverse lands with different cultural 
perspectives. In an effort to represent this geographical scope, the chapters 
in this volume represent a mosaic from five continents (Asia, Europe, 
North America, South America, Australia) and seven countries (Israel, 
Australia, Spain, Great Britain, the United States, Poland, Venezuela, 
Switzerland). Although the specific focus of each chapter represents a 
combination of the investigator's interest and his or her cultural founda­
tion, the social identity framework provides a common language that 
eclipses geographical, political and cultural boundaries. 
We'd like to extend our appreciation to our colleagues who have com­
mented on specific chapters. Special thanks is given to Dawna Coutant, 
John Turner and William Webb. Jennifer Ford, Kim Cozzi and Jon Iuzzini 
provided invaluable editorial assistance in preparing the manuscript for 
publication. And above all, we extend our appreciation to Ziyad Marar 
and Lucy Robinson, who were brave enough to accept such an ambitious 
project and whose expertise and gentle guidance moved it from the idea 
stage to book form. 




PART I 
REPRESENTATIONS OF 
SELF AND GROUP 
1 
Regarding the Relationship Between 
Social Identity and Personal Identity 
lean-Claude Deschamps and Thierry Devos 
The problems to be discussed in this chapter can be considered the central 
issue regarding both identity and a general perspective of social psychology; 
the theme of this chapter describes the relationship between two concepts 
called social identity and personal identity. 
Few concepts are as polysemic as identity. The idea of identity appears in 
the first texts of Antiquity (the famous 'Know Thyself' of Delphi) and is 
used in both everyday language and scholarly discourses; it is also the 
concern of many scientific disciplines. There are many synonyms or near 
synonyms for identity (one can mention at random: oneself, t, the self, we, 
self-perception, self-image, self-representation or self-awareness, the ego, 
and so on). The self can refer to a familiar subject, but also to a social 
agent's action. It can also refer to active processes and mechanisms that rule 
behaviour. Therefore, our remarks are limited to specific aspects of identity. 
Social identity and personal identity 
Identity is a central concept in social psychology, probably because it is one 
of the main concerns of that field. As Codol wrote (1979, p. 424), that 
major concern - which has been the subject of very old debates regarding 
philosophy and religion as well as ideology and the humanities - simply 
refers to the conflict of affirmation and individual necessity versus affir­
mation and collective necessity, the search for personal identity versus the 
search for collective identity, what constitutes individual difference versus 
what constitutes similarity to others (same things constitute both), social 


2 
Social identity: international perspectives 
visibility versus conformity, in short the conflict of the individual versus the 
group. And that concern, which can take many forms, refers to the 
relationship between the concept of the individual and the collective, often 
viewed as conflicting. This relationship is central in social psychology. In 
other words, it is an opposition between the diverse and the homogeneous, 
or, as Robert Pages notes, it is the recurrent theme of the same and the 
other. 
Studies about self-concept and identity also deal with the opposition 
between the individual and the social; and that opposition is codified 
according to the distinction made between personal identity and social 
identity. This distinction was already made - although not explicitly - at 
the beginning of modern psychological thinking. At the end of the last 
century, William James ( 1890) wrote about the distinction between the I 
and the me and came up with the idea of duality in self-representation. 
Nowadays, this distinction continues to have a great impact on studies on 
identity. Mead ( 1934) goes deeper into the question with his idea that the 
self consists simultaneously of a sociological component (the me) - which 
would only be an internalization of the social function - and a more 
personal component (the I); the developments of that definition of the self 
have gradually led to the distinction between social identity and personal 
identity. Of course, everyone agrees that identity is both personal - in the 
sense that it is 'situated within' a person - and social, inasmuch as its 
processes of individual formation are social, but the individual is indeed 
characterized by two poles - one is psychological and the other sociological 
- that have no connection to each other. At first, the individual/society 
dichotomy seemed outmoded, particularly with Mead's idea of conversation 
between the I and the me, according to which the individual is a continuous 
creation of society, and society an unremitting creation of individuals. But 
we will see that the two poles, the individual and the social, tend to be 
considered opposite. 
Let us first make clear the concepts of social identity and personal 
identity. These concepts are based on the idea that every individual is 
characterized by social features which show his or her membership of a 
group or a category, on the one hand, and by personal features or indi­
vidual characteristics which are more specific, more idiosyncratic, on the 
other. 
The former features define the social identity of a person. Belonging to a 
group or to a given social category is the most important. Social identity is 
codified as the part of the self which refers to cognitions ensuing from 
social ecological positions (Sarbin & Allen, 1968). Those who have similar 
positions and common backgrounds have similar social identities, therefore 
social identity does indeed refer to the similarity pole. However, the feeling 
of belonging to a group and the phenomena of identification are only 
possible in connection with groups or categories one does not belong to. 
Thus social identity refers to the fact that the individual perceives him- or 
herself as similar to others of the same background (the we), but social 


Social identity and personal identity 
3 
identity also refers to a difference, to a specificity of that we in connection 
with members of other groups or categories (the them). We have then a 
double motion which combines ingroup similarities and intergroup or 
categorial differentiation. The stronger the identification with a group, 
the more significant the differentiation of that group from other groups 
will be. 
More personal features or specific character attributes of each individual 
refer to personal identity and the idea that each individual is a unique 
combination of features which make him different from others, and which 
explains his uniqueness and the fact that he is specific. Personal identity is 
not well defined. However, it indicates how an individual is aware of his 
difference with respect to others. That feeling can only be experienced in 
relation to others and personal identity refers to the fact that the individual 
perceives himself as identical to himself; in other words he is the same in 
time and in space, but that is also what specifies him and marks him out 
from others. Personal identity is what makes you similar to yourself and 
different from others. 
Of course, one could debate this postulate regarding self-permanence as a 
relatively stable and durable concept held by everyone (see, for instance, 
Goffman [1956] with his descriptions of the individual who plays different 
roles for different audiences and Gergen [1965, 1982] with his idea of 
fluidity of the self). It is therefore necessary to speak of identities in the 
plural, considering that every social agent - whether individual or collective 
- can actualize, mobilize or produce identities according to the context. 
This also leads us to the question of the historical universality of this feeling 
of identity. But that does not change fundamentally the problem that 
concerns us. The fact that the feeling of identity is socially built into every 
individual and modulated in every social position and that it is an 
ideological construction does not mean that it is not real and efficient; it is a 
guide for individuals' actions and it is essential for the functioning of our 
society. 
Now, we must see how the articulation, the organization, the synthesis of 
personal and social aspects of identity are conceptualized. Social identity 
refers to a feeling of similarity to (some) others; personal identity refers to a 
feeling of difference in relation to the same others. This distinction between 
personal identity and social identity is indeed only an example of duality 
between the individual and the collective or difference and similarity. On 
the basis of generally accepted definitions of what is conventionally called 
social identity and personal identity, we have to deal with the conflict 
between the individual and the social. This lack of differentiation on a 
certain level - in other words social identity - and this differentiation -
personal identity - are considered most of the time as two opposite poles 
between which human behaviour oscillates. This combination of similarities 
and differences is problematic. It is hard to conceive how one can feel very 
similar and very different at the same time; it seems as though one has to 
choose to be one or the other. 


4 
Social identity: international perspectives 
Social identity theory 
The studies carried out by the 'Bristol School' emphasize the link between 
social identity and personal identity. In order to understand how the con­
cept of identity is theorized and used in that perspective, first of all a certain 
number of developments regarding the process of categorization must be 
addressed. 
Categorization and social categorization 
Categorization refers to psychological processes which tend to organize the 
environment into categories or groups of persons, objects, events (or groups 
of some of their characteristics) according to their similarities, their equi­
valences concerning their actions, their intentions or behaviour (Tajfel, 
1972a, p. 272). Categorization divides the environment into groups of sub­
jects which are or seem to be similar according to certain criteria; these 
subjects are or seem to be different from other subjects according to the same 
criteria. Categorization plays a specific part in structuring the environment 
by systematizing (dividing and organizing) and thus by simplifying it. 
One of the major effects of categorization and the simplification it implies 
is that it emphasizes the differences between categories (contrast or cogni­
tive differentiation effect) and the similarities within the same category 
(assimilation or cognitive stereotype effect). Each element of the category is 
characterized by the features shared by all the subjects of that category. 
Thus, a category is a group of elements which have in common one or 
several features. Cognitive apprehension is a simpler apprehension of 
reality. The emphasis on similarities and differences is one of the aspects of 
that simplification. Dealing with a psychological process, there is no reason 
at first sight to think that this process does not also play a role in the 
perception of the others, who are characterized by the fact that they belong 
to social categories. The fact that the elements referred to are persons and 
that we deal with social categories does not matter; the process does not 
change in that perspective. Actually, a social category is principally a 
cognitive category, where individuals who are supposed to share one or 
several features are grouped together. Stereotypes are - on a social level -
the expression of the attribution of features shared by different members of 
a group without taking into account the interindividual differences. The 
classification used as a basis for the stereotypes minimizes the differences in 
features among the members of a social group. As Doise noted in 1976, 
categorization is then a psychological process that simplifies the perception 
of physical and social worlds and the way the individual organizes the 
subjective perception of his or her environment. 
But apart from the homology between the perception of physical environ­
ment and social environment, social categorization - unlike physical stimuli 
categorization - implies that subjects are themselves inside a system of 
categories. They are subjects and objects of categorization at the same time. 


Social identity and personal identity 
5 
Social categorization minimizes the in-category differences and exaggerates 
the differences between categories. In addition, these differences are used for 
evaluation. When people hold the representation of a dichotomous universe 
where you can belong only to one category, they end up having a dis­
criminating attitude towards the members of other categories. 
Social categorization and identity 
How can discrimination between groups be explained? Unlike in Sherif's 
theory of the objective conflict of interest between groups (see, for instance, 
1966), competition is not a sufficient explanation for discrimination between 
groups. For Tajfel (l 972a) the idea of belonging to two different groups 
leads to discrimination in favour of the group one belongs to. But then, one 
must refer to concepts other than competition to explain the phenomenon 
of discrimination. This is when the concept of identity is introduced. It is 
important to recall the main points of social identity theory (see in 
particular Tajfel & Turner, 1979). 
According to Tajfel ( l972a), social identity is conceptualized as being 
connected to the individual's knowledge of belonging to a certain social 
group and to the emotional and evaluative signification that results from 
this group membership. Thus, it is through their belonging to different 
groups that individuals acquire a social identity defining their specific posi­
tions in society. But belonging to a given group contributes to the devel­
opment of a positive social identity only if the characteristics of that group 
can be compared favourably to other groups. In comparing the groups, 
individuals tend to establish a difference which is in favour of their own 
group. 
From 1 974 on, Tajfel adds an a priori distinction between two opposite 
poles in social behaviour. At one extreme (which can probably not be found 
in such a 'pure' form in 'real life'), we find interactions between two or more 
individuals. These individuals are totally determined by their interpersonal 
relations and by their individual characteristics and not affected at all by the 
different groups or social categories they belong to. At the other extreme, 
there are interactions between at least two individuals (or groups of 
individuals) which are totally determined by their respective membership of 
different groups or social categories and not affected at all by inter­
individual relationships among the concerned persons. Here again, there are 
few chances of encountering such 'pure' forms of this extreme in 'real' social 
situations (Tajfel & Turner, 1979). 
The two poles refer to interpersonal attitudes on the one hand, and to 
intergroup behaviour on the other. As Turner (198Ja) noted, personal 
identity corresponds to interpersonal behaviour - which means differ­
entiation between the self and the others; and social identity corresponds to 
intergroup behaviour - which means differentiation between groups or 
between 'we' and 'them'. 


6 
Social identity: international perspectives 
When identification with a group is emphasized, one changes from the 
interpersonal pole to the intergroup pole. The introduction of this distinc­
tion between interpersonal and intergroup phenomena is important for 
things are not going to be as simple as in Tajfel's early theory, according to 
which this distinction motivates a need for a positive difference in favour of 
one's own group compared to other groups, the creation of that difference 
and the emphasis on it. What really matters it the need for self-esteem, the 
'desire' for a positive self-evaluation, the need for individual positivity. The 
basic postulate is that 'individuals tend to maintain or increase their self­
esteem; they try to reach a positive self-concept' (Tajfel & Turner, 1979, 
p. 34). Thus, the individual tries to preserve or reach a positive self-image, 
and if she cannot evaluate herself according to the interpersonal behaviour 
pole - which lies in evaluating oneself directly by comparing oneself to the 
others - she can satisfy her desire for positive self-evaluation through social 
competition between groups (which must be considered as a groups com­
parison) with this tendency to introduce a positive difference in favour of 
one's own group in relation with other groups. 
In fact, social identity theory is founded on two bases: one is cognitive 
and the other is motivational. As mentioned before, it is cognitive insofar as 
the categorization process leads the subjects to overestimate intergroup 
differences and to underestimate ingroup differences. It is motivational 
insofar as what motivates discrimination is the need for self-esteem - or 
self-respect, according to Emler and Hopkins (1990). 
Turner (1975a, 1975b), who belongs to that school of thought, concludes 
that subjects act in terms of differentiation between groups only when this 
differentiation is the only means for them to achieve a positive self­
evaluation, a positive identity. If they can act in terms of a differentiation 
between self and others, they do not resort to the intergroup 'bias'. This only 
occurs when identification with the group provides for a positive identity. In 
other words, categorization on its own would not be enough to observe a 
differentiation between groups; the tendency to establish a positive differ­
entiation between self and others underlies the social categorization process. 
To put it differently, if the individual accedes to a positive identity through 
his identification with a group, he establishes differentiation between groups 
but no longer tends to establish differentiation with the other members of his 
group. If the individual is able to differentiate himself from the others 
directly - acceding in this way to a positive self-evaluation - he no longer 
establishes differentiations with other groups which he knows. 
The intention here is not to develop the limits of these studies or to 
question aspects of social identity theory (see Deschamps, 1984, 1991, for 
more details). However, it must be underlined that on the basis of this 
asserted dichotomy between interindividual and intergroup behaviour, one 
can say that the stronger social identity is, the less important is personal 
identity, and the more prominent a personal identity is, the less the 
individual needs a social identity, since social identity and personal identity 
satisfy the same need for a positive self-image. According to the social 


Social identity and personal identity 
7 
identity model, we can predict that when the salience of belonging to a 
group increases, every subject identifies him- or herself more and more with 
the in-group and that will reduce differentiation between the self and the 
ingroup while exacerbating differentiation between groups. In other words, 
when identification with a group increases, one changes from the inter­
personal pole to the intergroup pole, and vice versa. This dichotomy solves 
the problem of differences and similarities since only one pole is possible, 
switching from one to the other. This is indeed what Turner refers to when 
he introduces the concept of 'depersonalization'. For him, the factors which 
intensify the salience of the ingroup/outgroup categorization tend to 
increase identification (similarity, equivalence, interchangeability) between 
the self and the other members of the group (and differentiations with 
members of another group). Consequently, these factors depersonalize the 
self according to stereotypes determining the group. Depersonalization 
refers to the 'auto-stereotypy' process through which individuals come to 
consider themselves above all as interchangeable models of a category 
rather than as unique individuals who are distinct from the others (Turner, 
1987, p. 50). As we can see, both analyses - the one based on difference and 
the one based on similarity - depend on each other negatively. The dicho­
tomy postulated between interindividual and intergroup behaviour, which 
actually refers to the duality of self-awareness and the opposition between 
personal identity and social identity solves - apparently - the problem of 
similarities and differences while involving two mutually exclusive poles. 
Self-de
finition and its diff
erent levels 
Turner ( 1987) tries to explain the opposition between the psychological 
aspect of identity - which refers to an individual's unity with its constel­
lation of specific features - and the sociological aspect of identity - which 
refers to the idea of group membership and similarities - by distinguishing 
three levels in self-definition with his self-categorization theory: 
(a) a supra-order level where the self is defined as a human being: it refers 
to a human identity based on comparisons between the species (simi­
larity with the human race, differences with other forms of life); 
(b) an intermediate level of self-definition where the self is defined as a 
member of a group (ingroup similarities and intergroup differences): it 
refers to a social identity based on intergroup comparisons (in-species); 
(c) a subordinate level of self-definition where the self is defined as a 
unique being (differentiation between the self and the others inside a 
group); it refers to a personal identity based on interpersonal (in-group) 
compansons. 
However, when individuals situate themselves at a level of categorization, 
they obviously have to occlude the two other levels. As a matter of fact, 
there is a fundamental antagonism between the salience of one self-categorization 
level and the other levels: the salience of one of the categorization levels engenders 


8 
Social identity: international perspectives 
in-class similarity and interclass differentiation which reduces or inhibits the 
perception of in-class differences and interclass similarities on which inferior and 
superior levels of categorization are based. (Turner, 1987, p. 49) 
Turner adds that if, for simplification purposes, the supra-orderly categor­
ization level is ignored, the (social) intermediate level and the (personal) 
subordinate level are linked by a reverse relation. Obviously, the intro­
duction of these levels does not solve the opposition between the collective 
and the individual problem. Their relations remain negatively dependent 
(self-definition in terms of categorial membership - both on the supra­
orderly and the intermediate levels - versus individual specificity). 
In fact, as a whole these studies give the impression that it is difficult to 
get rid of this opposition between the individual and the collective, and that 
we are trapped with a dichotomous analysis according to which similarity 
and difference, social identity and personal identity, are two poles which 
depend on each other negatively. Having questioned the link between 
ingroup resemblance and intergroup divergence, we shall now present 
another series of studies in order to go beyond this opposition between 
individual and collective forms. 
The categorization model, like social identity theory or self-categorization 
theory, states that the same factors increase differences between groups and 
similarity or homogeneity inside the groups. However, this especially 
applies to outgroups and the way they are perceived and treated. The 
individual's own group would be perceived in a much more heterogeneous 
way. In fact, a certain number of studies carried out over the last 1 5  years 
illustrate the fact that the individual believes that his or her own group is 
more differentiated than other groups. This means that the group one 
belongs to can be perceived as heterogeneous by its members. This tendency 
to consider one's own group as relatively less homogeneous than a group 
one does not belong to (or the fact that a group tends to be considered more 
heterogeneous by its own members and not so by individuals who are not 
part of it) has been called the 'outgroup homogeneity effect'. This phenom­
enon has been explained in many ways (some explanations are based on 
motivation, in terms of beliefs, and refer to the self as a special category, 
recalling the way information about ingroup and outgroup is processed and 
codified in the memory; see Ostrom & Sedikides, 1992). But this is not 
essential here. What is important is that the concept according to which 
differentiation between groups necessarily means convergence within groups 
has been questioned and is no longer the rule. 
Covariation between interindividual and intergroup differentiation 
hypothesis 
A number of studies carried out in the 1970s already stated that emphasis 
on similarities inside a group does not necessarily mean emphasis on 


Social identity and personal identity 
9 
differences between groups (for instance, Deschamps, 1972- 1973; 
Deschamps & Doise, 1978). These observations, which were astonishing 
at a time when categorization effects were taken for granted, indicated 
(Deschamps, 1979) that the simple representation of belonging to two 
different groups could cause discrimination between groups in favour of 
one's own group but it would also engender differentiation between the self 
and the rest of one's group. When the representation of a dichotomized 
world is not prominent or relevant in a situation, there is no discrimination 
between groups - which is obvious. Moreover, in certain conditions, the 
differentiation between self and others could be less important than that 
between the self and the rest of the membership group when belonging to a 
group and dichotomy in groups are significant criteria in defining a 
situation. 
This has led us to a new perspective on similarities and differences: the 
intergroup and in-group differences covariation. Instead of considering 
interindividual and intergroup differences as two extremes of a continuum, 
and that they are in this way mutually (at least partly) exclusive, we assume 
that in certain conditions the stronger the identification with the group, the 
more important is interindividual differentiation within the group. Codol 
(for instance, 1975) underlines this issue with what he calls the phenomenon 
of 'superior conformity of the self': the more an individual conforms to the 
standards of a group (and the more she identifies herself with that group), 
the more she will tend to consider herself different from the other members 
of the group, believing that she corresponds to the standards better than the 
others. 
Here is the central concept of the interindividual and intergroup differ­
entiation model, in a few words. A general process of cognitive centrism 
appears when individuals are induced with the representation of a dichoto­
mized world, divided into two mutually exclusive categories. According to 
this representation, both ingroup favouritism or intergroup differentiation 
(which can be called sociocentrism) and autofavouritism or differentiation 
between self and others (which can be called egocentrism) would increase 
when categorization is emphasized. 
The aim of a certain number of our studies was to test this postulate 
(see Deschamps, 1982a, 1983- 1984, 1984; Deschamps & Lorenzi-Cioldi, 
198 1 ;  Deschamps & Volpato, 1984). The experiments carried out in these 
studies are similar to the ones used in Tajfel's minimal group paradigm and 
refer to the changes introduced into that paradigm by Turner in 1975. An 
outline of the surveys will suffice here (Deschamps, 1984). One of them was 
carried out in a girls' college. Using a sample of 88 girls, aged between 16 
and 20, half of the subjects were divided into two groups (categorization 
condition), apparently according to their preference for the works of two 
modern painters (in reality, it was an arbitrary way of splitting them). The 
other half was not divided into two groups and did not belong to any 
group in terms of their artistic preferences (non-categorization condition). 
In accordance with our hypothesis, it was predicted that differentiation 


10 
Social identity: international pers
pectives 
between self and others as well as intergroup differentiation would be more 
significant in the categorization condition than in the non-categorization 
condition. 
Concerning differentiation between groups, the results confirmed our 
expectations: discrimination in favour of the ingroup in the categorization 
condition and of course no discrimination in the non-categorization condi­
tion. Concerning differentiation between self and others, in the categoriza­
tion condition, the differentiation is much more significant than in the 
non-categorization condition. In other words, when individuals are simply 
induced to the representation of an environment including their group and 
another one (which leads of course to differentiation between groups), it 
increases significantly autofavouritism 'bias' or differentiation between self 
and others, compared to a non-categorization condition (without categor­
ization induction). Emphasizing intergroup differentiation in an experiment 
(in this example, by inducing or not the subjects with the representation of a 
dichotomized world) is sufficient to increase the differentiation subjects set 
between themselves and others. 
But more observations must be addressed. The contrary is also true. If 
differences between self and others within a group are emphasized in an 
experiment, differentiation between groups increases simultaneously. 
Another survey was carried out in a high school with 1 12 boys and girls 
aged between 14 and 1 5  (see Deschamps & Volpato, 1984). As in the 
categorization condition of the experiment mentioned above, the subjects 
were placed into two groups, apparently according to their preference for 
the works of a contemporary music composer. Half of the subjects, the ones 
in the individualization condition, expected that each individual would be 
given exactly the same remuneration, according to the minimal group 
paradigm. The other half, the ones in the fusion condition, expected that all 
members of a group would receive the same remuneration, corresponding to 
an average of points which all the subjects would give to the different 
members of that group. Our expectations proved to be right. In the fusion 
condition, differentiation between self and others and discrimination in 
favour of one's own group are less than in the individualization condition. 
When differentiation between self and others is emphasized - by focusing 
not on the groups as a whole, but on the fact that individuals are char­
acterized by their group memberships - differentiation between groups 
increases in the same way. 
In addition, the covariation model seems to apply especially to indi­
viduals belonging to privileged groups. Not only do differentiation between 
groups and ingroup homogeneity not necessarily go together, but the 
simultaneous variation of differentiation inside a group and between groups 
also depends on the relative status of the groups in question (Deschamps, 
1 982b; see also Lorenzi-Cioldi, 1988). These are the conclusions of other 
studies. For instance, it has been shown (Deschamps, 1977) that when boys 
had to evaluate themselves as well as other boys and girls for different 
assignments, they discriminated between boys and girls and between 


Social identity and personal identity 
SI: x--x 
PI: x---------------x 
SI: x---------------x 
PI: x---------------x 
" 
" 
" 
" 
"
SI: x-------x / 
/ 
/ 
/ 
/ 
Social identity 
PI: x-------x 
axis (SI) 
-----Compromise -----
SI: x --x 
PI: x --x 
/ 
,, W 
. i:- / 
,,<9
‚ 
(similarity) 
n
o / 
,,ƒ
-? 
""
i-
{!>
/ 
"
'.90,,.
 
,(
..,0/ 
'- -? 
'- °
0 ' 
/ 
,, œ  
/ 
,,'.9-
Key 
SI: x ---------------x 
PI: x--x 
x--x 
restricted range of identity 
(personal and/or social) 
x -------x 
medium range 
x---------------x 
expanded range of identity 
(personal and/or social) 
Figure 1 . 1  Personal and social dimensions in psychosocial identity 
1 1  
themselves and the other boys. Girls would discriminate between boys and 
girls, but unlike the boys they would not tend to discriminate between other 
members of their own category. 
In conclusion, the results of these studies briefly mentioned underline the 
relevance of the simultaneous variation of differentiation between groups 
and between self and others. However, it does not mean that under certain 
circumstances there is an opposition between personal identity and social 
identity. Obviously, relationships between the individual and the collective 
can be examined according to situations, cultures and societies. Never­
theless, one must at least consider the possibility of simultaneity between 
similarity and difference. Therefore, the postulated axis defining the inter­
individual relations and behaviour pole versus the intergroup relations and 
behaviour pole should be reconsidered. In that prospect, similarity and 
difference, social identity and personal identity must no longer be con­
sidered as two poles of the same continuum which are negatively dependent. 
These different elements must rather be considered as two disconnected 
dimensions that can be 'orthogonalized' in some way (see Figure 1 . 1). 
Consequently, models based on Tajfel's work and the covariation 
hypothesis can be explained at the same time. 


12 
Social identity: international pers
pectives 
Note 
This chapter was translated from the French by Mehr-Afarin Khosrowdad. It takes up and 
develops arguments which have been put forth in some of the previous texts and quotes several 
passages from an article published in 1991 by the first author in Cahiers Internationaux de 
Psychologie Sociale called 'Identites, Appartenances Sociales et DiITerenciations Individuelles'. 


2 
Social Representations in Personal 
Identity 
Willem Doise 
Introduction: Self as social representation? 
Social representations can be defined as organizing principles of symbolic 
relationships. They permit the positioning of the individual in relation with 
significant social objects. The self certainly is such an object embedded in a 
network of relationships. Given its central position in communicative 
systems, personal identity is not just an idiosyncratic entity, it is one of the 
most important organizing principles of symbolic relationships. 
In order to validate the basic assumption of personal identity as a social 
representation, existence of common knowledge about personal identity 
should be evidenced together with systematic positioning of individuals in 
the frame of this common knowledge and the anchoring of such positioning 
in other socio-psychological meaning systems. Indeed, important compo­
nents of all studies on social representations are the definition of a common 
field, the systematizing of individual positioning in that field, and the 
anchoring of these positionings in related fields of representations (Doise, 
Clemence & Lorenzi-Cioldi, 1993). 
Several investigators have already examined aspects of personal identity 
in the frame of social representation theory (see, for instance, Durand 
Delvigne, 1992; Palmonari, Carugati, Ricci-Bitti & Sarchielli, 1979; 
Ravaud, Beaufils & Paicheler, 1986; Tap, 1985; Zavalloni & Louis­
Guerin, 1984; Zeegers, 1988). Nevertheless, personal identity is more often 
studied in other theoretical frameworks and this may explain why books 
containing prototypical social representation studies do not mention 
investigations of personal identity (see, for instance, Abric, 1994; Doise & 
Palmonari, 1986; Duveen & Lloyd, 1990; Jodelet, 1989). A remarkable and 
pioneering exception is the study by Codol (1 984a; see Farr & Moscovici, 
1984). In this study Codol analysed the system of relationships connecting 
representations of self, others, group and task. 
Another possible reason for the paucity of research on personal identity as 
social representation is the limited definition of the concept of personal 
identity. Generally, identity is conceived of as a very peculiar set of opinions, 
judgements, evaluations, attitudes, manifested by a person towards him- or 


14 
Social identity: international perspectives 
herself. Although research from different theoretical traditions bears on the 
general processes governing identity development and construction (Damon 
& Hart, 1982; Erikson, 1977; L'Ecuyer, 198 1 ; Marcia, 1966) and on 
situational regularities enhancing saliency of identity aspects (Codol, 1990; 
Kuhn & McPartland, 1954; McGuire & McGuire, 1988), the study of 
individual identity as socially organized still remains marginal. 
Nevertheless, identity, as other social representations, can be studied as a 
cognitive organization, oriented by a meta-system of social regulations. In 
order to illustrate this thesis I shall successively show (a) that common 
knowledge on identity exists, (b) that organizing principles rule individual 
positioning in relation with this common knowledge frame, and (c) that 
specific social insertions and experiences modulate such positioning. In 
developing my position, I draw on selected evidence from studies on social 
representation theory and a reinterpretation of traditional research on social 
identity. My main objective is to convince the reader that studying identity 
as a social representation is feasible. 
Common knowledge in personal identity 
Personal identity is often considered as unique and singular. On the other 
hand, shared group membership involves common characteristics. Hence, 
individuals would be considered similar to the extent that they belong to the 
same or similar categories. In this sense group membership would result in 
homogeneity and individual differences would correspond to personal char­
acteristics which are not shared by other group members. Such considera­
tions fit with Turner's conception of a functional antagonism between 
the salience of different levels of self-categorization, the group level and the 
individual level: 'The salience of one level produces the intra-class simi­
larities and inter-class differences which reduce or inhibit the perception of 
the intra-class differences and inter-class similarities upon which lower and 
higher levels respectively are based' (Turner, 1987, p. 49). 
Recent research, however, suggests that the problem of similarity and 
difference is not so simple: commonness could also be studied at the indi­
vidual level and specificity at the group level. Let us consider some of the 
research evidence. 
Ravaud et al. (1986) investigated school classes attended by physically 
handicapped pupils as well as classes of 
non-handicapped students. Students 
in both types of classes were asked to describe on the same scales members 
of their class (themselves included), as well as handicapped pupils in general 
and non-handicapped pupils in general. Factorial correspondence analysis 
yielded very contrasting descriptions of the general group of handicapped 
and non-handicapped students. However, self-descriptions of handicapped 
and non-handicapped pupils were very similar. In addition, when the 
descriptions targeted specific handicapped or non-handicapped individuals, 
there was a high degree of similarity. In terms of social representations, 


Social representations in personal identity 
1 5  
differences are found at the category level, but they do not manifest 
themselves when individuals are asked to describe individual members of 
the categories. Commenting on these results, the authors contest the gener­
ality and importance of the so-called 'deductive effect' of social categ­
orization. If such an effect were strong, distinctive traits of categories as 
such should be attributed to single members of each category. 
This is not the case: 
It is a rather striking observation that stereotypes are not very important in 
structuring representations of individuals considered to be bearers of the stereo­
types. The abstract individual described by the valid pupils is totally opposed to 
the descriptions of the concrete individuals and the image that valid pupils 
construct of real handicapped ones and that appears in their answers seems not to 
be mediated through stereotypical representations. This is a strong challenge for 
assumptions that consider stereotypes to be a necessary tool, a measure applied to 
all members of a stereotyped group. (Ravaud et aI., 1986, p. 1 80) 
However, it should be noted that the same research reports results showing 
that real discrimination can coexist with a blurring of stereotypes. When the 
students were asked to choose others for a party at their home or as work 
partners, non-handicapped and handicapped pupils clearly prefer non­
handicapped companions. Such discrimination does not reveal itself in the 
concrete images pupils give of themselves and of each other. 
In the same line of thought, research carried out on adolescent identity by 
Palmonari and colleagues is relevant (Palmonari et aI., 1979; Palmonari, 
Pombeni & Kirchler, 1989). Their results show striking similarities between 
descriptions of self given by adolescents belonging to different formal and 
informal groups. These descriptions reflect their openness to the social 
world, their commitment, disengagement and dissatisfaction. 
Similar results were obtained by Doise and Lorenzi-Cioldi ( 1 99 1 )  study­
ing Swiss and second-generation immigrant pupils. These adolescents gave 
highly contrasting descriptions of Swiss in general and foreigners in general. 
However, when describing themselves and their friends, Swiss and immi­
grant pupils gave quite similar 'concrete' descriptions. Moreover, variations 
between subjects of the same category in self-descriptions are not more 
important than variations between them in the description of group 
stereotypes. To state it differently, self-images are as stereotypical as group 
images. 
The results of other investigations (Nakbi, 1990; Nakbi & Arnal­
Duchemin, 1987) support similar conclusions. These studies gathered not 
only self-images and other-images of male and female students, but also 
ideal images and self-images attributed to others. Multidimensional analyses 
of these data revealed a very similar structure of these representations for 
males and females. 
This research demonstrates a high degree of similarity between self­
definitions of members of different groups. General societal norms affect 
these definitions much more than do specific group memberships. This is 


16 
Social identity: international perspectives 
also a conclusion that can be drawn from research by Cowan and Hoffman 
(1986) and by Inoff, Halverson and Pizzigati (1983). These investigators 
examined the self-descriptions of children, finding striking similarities in the 
self-descriptions of boys and girls, despite the fact that conventional gender 
stereotypes are used in the description of others. It was assumed that these 
children did not apply gender stereotypes to themselves because they tend to 
ascribe favourable traits to themselves, independently of the gender conno­
tation these traits may convey. A similar motivation underlies the self­
descriptions of Genevan pupils: Swiss and immigrant students assigned 
positive traits to themselves that are part of the stereotypes of Swiss 
nationals and foreigners. This tendency also corresponds to the inclination 
of individuals to ascribe to themselves numerous and contrasted charac­
teristics (Sande, 1990). 
The content of self-representations varies from culture to culture and 
epoch to epoch. Commonality in self-definitions in the Netherlands was 
studied by Zeegers (1 988), who analysed the contents of 'lonely heart 
advertisements' in Dutch newspapers from 1945 to 1986. Three styles of 
self-presentation appeared successively in these personal ads. During a first 
phase starting in 1945, a majority of the ads contained traits such as serious 
(degelijk), tidy (net), educated (beschaa
f
d), that is, transsituational evalu­
ative characteristics. During a second phase, from 1965 on, personality 
traits adapted to more specific situations appeared, carrying information 
about lifestyle considered important for developing commitment to a 
relationship that should above all result in friendship. After 1975, the 
advertiser's description of his or her identity became still more specific, 
including more personal traits and declarations of interest in particular 
activities. 
It would be remiss not to mention that there are individual differences in 
the presentation of self-images of members in a group. Experimental results 
also show that it is easy to entice individuals to differentiate themselves 
from other individuals within as well as between groups using characteristics 
deemed to underscore one's originality. However, such differentiations often 
involve an accentuation of conformity to prevailing norms, the so-called 
'superior conformity of the self' (Codol, 1975). 
Normative regulations strongly intervene in self-presentations. More 
generally, linguistic systems offer definite dimensions for describing the self. 
The English language contains numerous terms for individuals to describe 
their personality. Allport and Odbert (1936) identified more than 4,000 such 
terms. Various techniques have been used to systematize this set of traits. 
By asking individuals to use samples of these adjectives to describe them­
selves or other persons, or to indicate similarities between the traits, 
investigators consistently demonstrated the existence of a common structure 
(Goldberg, 1 990; Peabody & Goldberg, 1989). This structure shall be 
described in the next section, but it is important to note here that according 
to various authors this structure corresponds to a common-sense psychology 
embedded in daily language, much more than to the results of strict 


Social representations in personal identity 
17 
scientific observation (Leyens, 1983; Semin, 1987). The important point in 
this discussion is that individuals use common knowledge systems to 
describe themselves. 
Identity as individual positioning 
The images that individuals construct of themselves, although highly similar, 
nonetheless vary from one person to another. In many studies five dimen­
sions are considered necessary and sufficient to describe these interindividual 
variations. According to Goldberg (1990): 'These "Big-Five" factors have 
traditionally been numbered and labeled as follows: (I) Surgency (or 
Extraversion), (II) Agreeableness, (III) Conscientiousness (or Dependabil­
ity), (IV) Emotional Stability (vs. Neuroticism), and (V) Culture. Alterna­
tively, Factor V has been interpreted as Intellect' (p. 1217). 
In fact these organizing principles of between-individual variation deal 
with the way people behave towards each other. A person is involved daily 
in multiple social relationships and the way s/he participates in these 
interactions is being constantly evaluated. If common language, well before 
psychological expert language, contains an abundant range of terms to 
qualify individuals, this variety primarily serves the purpose of differentiat­
ing between individuals according to their behaviour in all kinds of social 
interactions. 
It is impossible to estimate the number of people who have been asked, at 
least once, to describe themselves on a personality questionnaire. If these 
individuals were able to comply with such a request it is because they 
disposed of common references, of comparison frames allowing them to 
compare their own way of behaving with the way prototypical persons 
behave. Apparently, a stable organization of the description of interactional 
systems can be traced back in ancient historical documents. 
Adamopoulos (1982) examined the descriptions of social interactions 
between figures in epic texts of various historical epochs (Homer's Iliad; an 
Old English poem of a fight with a dragon: Beowulf, and a story of a young 
recruit's adventures during the American Civil War: The Red Badge o
f
 
Courage). In each text, dyadic interactions were retrieved and the type of 
interaction between each dyad was classified. Finally, a matrix was con­
structed for each text linking types of interactions and pairs of persons. For 
each matrix a principal component factor analysis was performed. Three 
common factors were obtained for the interpersonal structures in the three 
documents: affiliation, power (status) and formality (versus intimacy), 
although the latter factor appears less clearly in the most ancient text. 
Adamopoulos and Bontempo (1 986) repeated the procedure for The 
Odyssey and The Song o
f
 Roland, and were able to retrieve the same three 
factors in both documents. 
These results can be compared to those obtained by Wish (1976; Wish, 
Deutsch & Kaplan, 1976), who analysed similarity judgements between 25 


1 8  
Social identity: international perspectives 
social relationships (for instance: between husband and wife, between prison 
guard and inmate, between two enemies, between two political opponents). 
The main aim of the study was to identify the dimensions underlying the 
conceptions of these different social relationships. 
Four dimensions described the judgements of prototypical relation­
ships. Dimension I contrasted competitive versus cooperative relationships: 
divorced couple, personal enemies, political opponents versus good friends, 
husband-wife and team mates. Dimension 2 involved unequal versus equal 
relationships: guard-inmate, interviewer-interviewee, teacher-pupil versus 
team mates, political opponents, good friends, personal enemies and cousins. 
Dimension 3 concerned socio-emotional-oriented versus task-completion­
oriented relationships: parent-child, brother-sister, husband-wife versus 
interviewer-job applicant, seller-buyer and political opponents. Dimension 
4 included intense versus superficial relationships: parent-child, husband­
wife, good friends, psychotherapist-patient versus chance encounters, 
interviewer-interviewee and employee-employer. Individual differences in 
the use of these common dimensions were also investigated. 
These representations of social relationships offer important coordinates 
for individuals who have to position themselves in self-descriptions. In this 
respect McAdams (1985) proposes a theory of self as composed of internal 
personified and idealized images. This Imago theory postulates that identity 
'is a story complete with setting, scenes, characters, plot, and recurrent 
themes. Identity formation is the process of constructing a self-defining life 
story. The main characters in the story are imagoes' (McAdams, 1985, 
p. 127). References are classical authors such as Erikson (1977), James 
(1 892), Jung (1943), Klein (1948), but also Markus's (1984) ideas on 
'possible selves'. The symbolic relationships between these prototypical 
images remind us of Adamopoulos's analyses: 
The two major thematic lines around which the content of identity can be 
organized are agency (power/mastery/separation) and communion (intimacy, 
surrender, union; Bakan, 1966). Initial investigations of life stories of college 
students and men and women at midlife suggest that the power motive (Winter, 
1973) and the intimacy motive (McAdams, 1980) are two independent personality 
dispositions - both assessed via the Thematic Apperception Test or TAT - that 
serve as significant predictors of salience in life stories of thematic lines of agency 
and communion, respectively. (McAdams, 1985, p. 128) 
Furthermore, McAdams presented a taxonomy of 12 major deities of the 
Greek pantheon according to their high or low degree of 'agency' and 
'communion', a taxonomy that he also applied to the analysis of identities 
of 50 respondents on the basis of personality tests and of a life story 
interview. 
Individuals define themselves in relation to each other, referring to rep­
resentations of their participation in social interaction settings. The research 
tradition in which I locate my own work emphasizes the importance of 
group and category membership in self-definitions. However, we have 
already seen that individuals often do not assign for themselves the same 


Social representations in personal identity 
19 
characteristics they ascribe to their groups. Notwithstanding such findings, 
it remains true that differences exist between individuals and that some 
people more readily attribute typical group characteristics to themselves. 
This is especially true for gender characteristics. Each time individuals are 
invited to express their degree of agreement with the relevance of masculine 
and feminine traits to their self-descriptions, factors opposing both kind 
of traits can be found in the data (for research with French-language 
subjects, see Lorenzi-Cioldi, 1994; Tap, 1985). This effect certainly cannot 
be explained by mere chance that this organizing male-f
emale principle 
actualizes the opposition between 'agency' and 'communion' described by 
McAdams. In his taxonomy of the pantheon, deities with high 'agency' and 
low 'communion' were masculine (Zeus, Hermes and Ares) whereas deities 
with opposed characteristics were feminine (Demeter, Hera and Aphrodite). 
The organizing principles of identity regulate symbolic relationships 
between social agents and, therefore, correspond to the definition of social 
representations (Doise, 1986). But the fact that positioning in relation with 
each other actualizes dimensions of 'agency' and 'communion' does not 
preclude the importance of other dimensions. 
Verkuyten (1990) has analysed the identity components of a nationwide 
sample of Dutch and Turkish adolescents living in the Netherlands (N = 
2,710 Dutch and 237 Turkish). Some significant differences were found in 
self-descriptions between the Dutch and Turkish youngsters, but these 
differences explained no more than 1 % of the overall variance. More 
importantly, in order to analyse relationship between self-esteem and five 
components of the self-concept (ethnic identity, academic ability, sports, 
popularity, body image), a standardized multiple regression analysis with 
global self-esteem as dependent variable was used for respondents of Dutch 
and Turkish origin separately. The results were highly similar. For the 
Turkish and the Dutch adolescents, body image had the greatest impact on 
self-esteem and ethnic identity, while academic ability and sports had the 
lowest. The similarity of results for both samples suggests that general 
norms are very important in shaping self-esteem, and that general societal 
values transcend, in this case, cultural barriers, probably as a consequence 
of the homogenizing effect of common schooling. 
Makris-Botsaris and Robinson (1991) adapted for Greek youngsters 
scales elaborated by Harter (1985) for measuring different components of 
self-evaluation and perception in the USA. They compared factorial 
structures obtained in Greece and in the USA with youth aged II to 14. In 
both countries they found a highly similar five-factor structure of scholastic 
competence, social acceptance, athletic competence, physical appearance 
and behavioural conduct. Correlations between subscales corresponding to 
these factors and global self-worth were comparable in Greece and the 
USA, the correlation between physical appearance and self-worth always 
being the highest. 
Common knowledge reflects itself in the way individuals define themselves, 
others, and their relationships. Application to oneself of this knowledge is 


20 
Social identity: international perspectives 
highly normative and results in significant interindividual resemblances. 
Nevertheless, between-individual variations are equally important and they 
are organized along dimensions which are also consensually defined and 
which refer to characteristics of the multiple relationships in which indi­
viduals take part. 
Anchored identities 
The anchoring of social representations has mainly been studied in three 
different ways. First, the relationships with general beliefs and values (such 
as belief in a just world or egalitarianism) have been investigated. Such 
values and beliefs are considered general to the extent that they supposedly 
organize symbolic relationships in various domains. Social representations 
have also been studied as anchored in the views individuals develop in 
structuring their social environment, as for instance the representations they 
hold concerning relationships between social groups and categories, such as 
gender. Finally, a third way of studying anchoring investigates how group 
memberships or social positions held by individuals influence their represen­
tations, the general hypothesis being that shared social insertions lead to 
specific interactions and experiences that modulate social representations. 
The first way of studying anchoring mainly uses correlational methods in 
order to determine if links exist between adherence to given beliefs or 
opinions. Such a method was used in the studies reported at the end of the 
previous section showing that important aspects of youth's identity, namely 
self-esteem, were linked to perceptions of their physical appearance and less 
to evaluation of their popularity, academic or sportive achievement. 
Anchoring in the views on the social environment is directly studied in 
research on psychological androgyny linking personal sexual identity to the 
actualization of a gender scheme (see, for instance, Bem, 198 1 ,  1985). 
Indeed, what else is the gender scheme if not a 'naive' theory on the 
respective positions of males and females in society? 
More generally, representations held by individuals on the nature of 
social groups intervene in the structuring of personal identity. Lorenzi­
Cioldi (1988) studied this intervention experimentally. He distinguished two 
conceptions of the social group: the collection group and the aggregate 
group. Members of dominant groups in our society tend more to adhere to 
a conception of their group as a collection of individuals, each having their 
specificity. These individuals define themselves through their personal 
characteristics, apparently not linking themselves to their group. However, 
members of more dominated groups tend to consider themselves as less 
differentiated from each other, as more homogeneous, and they attribute 
more directly to themselves characteristics of the group as a whole. 
Lorenzi-Cioldi (1988, 1994) illustrated the complexity of the links between 
collective and aggregate groups, gender groups, within- and between-group 
differentiation on common and specific dimensions. His data supported the 


Social representations in personal identity 
21 
general hypothesis of a homology between the representations of groups 
which are apparently not linked. Groups of men, of original and singular 
personalities, of competitors, on the one hand, and groups of women, 
individuals with similar personalities and cooperators, on the other, are 
respectively associated with respect to identity dynamics of their members. 
Social representations of groups, membership groups as well as other 
groups with which one interacts, intervene in actualizing personal identity. 
As social interactions often involve groups occupying asymmetrical posi­
tions in a representational field, one is easily led to the conclusion that 
identities of interacting individuals differ. This was not the starting point of 
my theorizing on personal identity as a social representation. Initially, I 
insisted on the similarity between the identities of members of different 
groups, such as autochthonous and immigrants. I will return to this point 
after discussing the relevance of the third way of studying anchoring in 
research on identity as a social representation. 
This approach to studying anchoring is still more directly concerned with 
the problem of differences between groups. Results presented by Durand 
Delvigne (1 992) can be used to illustrate this form of anchoring. She 
analysed self-descriptions of men and women, of different ages and social 
conditions, and found important similarities for different groups of 
respondents in factorial structures of answers to a French version of the 
sexual role inventory. Notwithstanding this similarity, a striking difference 
also characterized the structures of men and women, the opposition between 
the feminine and masculine poles being much stronger in the factorial 
structures of responses obtained from women than in those of men. Durand 
Delvigne summarized these results in the following way: 
This result is to be considered in relation with a hypothesis prevailing in relevant 
literature and stating that women more than men are referred to their social 
identity. Social models would be available for defining them, whereas men would 
be more easily individuated. For women, gender category would be a necessary 
marker of identity (Hurtig and Pichevin, 1986; Lorenzi-Cioldi, 1988). (Durand 
Delvigne, 1992, p. 1 82) 
Hence, differences exist in the organizing principles of male and female 
identities and they result in different self-descriptions. This is particularly 
true when comparing the self-descriptions of males and females of high and 
low social status. Masculinity scores are high for individuals of higher social 
status, independent of their gender, and femininity scores are especially low 
for men with a higher status. Androgyny scores are high for all groups of 
higher status independent of gender. Considering these results together with 
many others that are accumulating in the area of gender studies, one cannot 
but subscribe to Durand Delvigne's thesis: 
Gender, as identity, is not inherent to a sex group, but refers to social status. 
Otherwise said, gender orientation of women is related to the opportunities they 


22 
Social identity: international perspectives 
do have or do not have to overcome the social definition of themselves. In this 
case, and related to identity, gender is not a dynamic set of attributes produced by 
sexual category membership. It is a social marker linked with the asymmetric 
position of dominant and dominated individuals. (Durand Delvigne, 1 992, p. 223) 
Gender opposition does not represent sex opposition. It is a social con­
struction of hierarchical relationships linked with sex group reports and 
modulating personal identity through social differentiation processes. 
Nevertheless, the existence of similarities of personal identities across 
group boundaries is real. This apparent similarity may correspond to 
within-group differences along lines that were not analysed in the studies 
reported previously. Another explanation for the repeated findings of 
between-group similarities in identity research is suggested by some findings 
of Lorenzi-Cioldi (1988) and Amancio (1989), who report that identity 
processes are differently actualized as a function of intergroup encounter 
situations and of available comparison dimensions. In line with such 
findings, one can expect that a request to describe oneself on scales without 
explicit gender connotation or without any intergroup comparison will 
induce descriptions that privilege the individual as abstracted from his or 
her group memberships. Otherwise stated, it is possible that questionnaire 
instructions convey social representations of the 'collection' type. Depriving 
the individual of his or her representations of membership in 'aggregate' 
groups focuses his or her attention on representations of self as an indi­
vidual in search of originality and distinction. But such a search itself is 
highly normative and results in rather similar identity descriptions. 
The predominance of such normative self-definitions should not prevent 
us from studying less dominant ones. Reicher and Emler (1986) investigated 
such identities in their work on the origin of adolescent delinquency. 
According to them, behaviours such as vandalism, theft or physical 
aggressions are more often admitted by early to mid-teens than by any other 
age-group. There is evidence that such self-report scores are related to the 
actual level of involvement in delinquent activities (Singh, 1979). More 
importantly, such self-reports are analysed by Reicher and Emler (1986) as 
part of a reputation-building strategy. At this age life chances are con­
sidered to be settled by one's educational success or failure, and a young 
person's identity is characterized by its relation to his/her school career. 
Negative attitudes towards school authority are not infrequent and are 
generalized to other forms of institutionalized authorities, such as police and 
legal authorities. This means that important aspects of the social contract, 
namely that authorities will protect individuals from anarchy and redress 
violations of rights, are no longer considered valid: 
Thus delinquency can be seen in part as a strategy for coping with the lack of 
institutional protection. In the first place, delinquency acts as redress. It resolves 
grievances through attacks upon enemies, upon the police force and upon school 
and public property. Second, it offers its own kind of protection through the kind 
of reputation it creates. (Reicher & Emler, 1986, pp. 30-31) 


Social representations in personal identity 
23 
A self-presentation and social reputation is actively promoted in order to 
signify to others that one is the kind of person who stands up for one's 
rights. Another objective of the social construction of such a representation 
is to obtain support from others sustaining the same oppositional identity 
(see also Emler & Reicher, 1 995). 
Conclusion: A definition of personal identity 
Personal identity can be considered to be a social representation, an organ­
izing principle of individual positioning in a field of symbolic relationships 
between individuals and groups. This positioning is related to pre-existing 
insertions in a network of social relationships. 
Not all social relationships in which we take part are equally present in 
our minds. Depending on situations, various relationships become salient 
and give rise to different identity dynamics. Many self-descriptions only 
actualize general societal norms inducing individuals to define themselves in 
relation to each other on consensual dimensions, such as physical appear­
ance for youth. Other situations actualize social representations more 
directly anchored in group and category memberships. Memberships inter­
vene more intensively in conditions that are now well studied (Fiske & 
Neuberg, 1 990; McGuire & McGuire, 1988; Pavelchak, 1 989). Studying 
personal identity as a social representation should contribute to a better 
understanding of identity processes. 


3 
The Personal-Social Interplay: 
Social-Cognitive Prospects on Identity 
and Self-Others Comparison 
Carmencita Serino 
Introduction: 'Personal' and 'social' in social identity theory 
In his analysis of social categorization, Tajfel describes intra-categorial 
assimilation and inter-categorial differentiation as two directly related and 
simultaneously occurring processes. The minimal group paradigm (for 
example, Billig & Tajfel, 1 973; Tajfel, Flament, Billig & Bundy, 1 971) is 
intended to show that these effects usually occur even when group member­
ship is an extremely abstract and irrelevant experience for the subjects. In 
most of these experiments, the assimilation/differentiation processes manifest 
themselves by means of some form of favouritism: the mere categorization 
into different groups would elicit subjects' 'behaviours' (that is, giving 
money, rewards, and so on) which are clearly intended to favour their own 
group and to enhance the distance between ingroup and outgroup. 
Indeed, Tajfel aimed at providing a purely cognitive basis for intergroup 
differentiation (Hewstone, Hantzi & Johnston, 1 991). On the other hand, he 
also pointed out the existence of a collective level of reality which cannot be 
reduced to the sum of merely individual phenomena: under certain circum­
stances, individuals may act essentially as members of a particular group, 
even independently of their strictly individual tendencies or characteristics. 
Social identity theory (SIT) was later developed by Tajfel and Turner 
along these lines (for example, Tajfel & Turner, 1 979; Turner, 1 978). 
According to Tajfel ( l978b), 'at least in our kinds of societies, an individual 
strives to achieve a satisfactory concept or image of himself'. This 
individual will show, then, a tendency either to confirm or to change his/her 
condition of group member, depending on whether a particular group can 
contribute (or not) to the positive aspects of his/her identity. In other words, 
when his/her need for self-positive evaluation is not complied with, an 
individual will tend to leave the group (unless some objective conditions or 
particular values and norms interfere with this tendency). 
We may notice that, despite the original orientation towards a purely 
cognitive approach, motivational factors are mostly put forward by the SI 
theorists. A true 'axiom related to the individuals' need for positive self-


The personal-social interplay 
25 
evaluation' (Deschamps, Clemence & Roux, 1983) might in fact underlie 
their assumptions. 
Indeed, the idea according to which a fundamental need for self-positivity 
is at the basis of social identity and group membership could be easily 
verified in short-lived groups (such as the 'minimal' ones), wherein joining a 
group is a rather extrinsic and fortuitous (if not totally exploitable) experi­
ence. By contrast, when real social groups are involved, individual and 
social identities appear to be much more intertwined. Indeed, belonging to 
particular social groups is in general a deep-rooted aspect of persons' 
identity, a real 'way of being', which also affects the ways in which all social 
knowledges are filtered and reconstructed. ' As we will argue below, 
belonging to some particular social groups may even be the necessary 
condition allowing subjects to develop a sharper and better defined self­
image, to perceive themselves as being quite specific and highly distinctive 
individuals. 
In this vein, Tajfel and Turner's approach to the personal-social inter­
play is worth further analysis. Their idea of 'continuum' fails adequately to 
represent the dynamic connection of individual and collective processes. The 
personal-social continuum seems to imply an essential, unsolved dichotomy 
between individual and collective processes, personal and social identity, 
self
/others and intergroup comparisons. 
Alternative views on the personal-social continuum 
The close interdependence of intra-psychic and social psychological pro­
cesses in identity formation is recognized by Zavalloni (Zavalloni, 1973, 
1983; Zavalloni & Louis-Guerin, 1984). Her research suggests: (a) that even 
negative stereotypes about the ingroup are possible: these do not weaken, 
however, the individual feeling of being a group member; (b) that, on the 
other hand, the strong identification with a group does not exclude the self/ 
others differentiation. The intra-categorial differences can be as marked as 
the inter-categorial ones. Individuals can make a fine-grained distinction 
between the characteristics of their own group and the degree to which these 
characteristics refer to themselves (Serino, 1988a; Zavalloni, 1 983). Thus, 
the conditions in which group membership is made salient do not lead, 
necessarily, to an intra-categorial assimilation process. 
On the whole, the idea of a necessary and positive relationship between 
intra-categorial assimilation and inter-categorial differentiation was not 
always completely confirmed (cf. Eiser, 1983; Tajfel & Wilkes, 1963). 
By contrast, some experiments (Deschamps, 1982a) suggest a different 
hypothesis. In these studies, the greater the social identification, the greater 
is also the self/others differentiation occurring inside the membership group. 
On the other hand, when the self/others differentiation is emphasized, the 
intergroup differentiation is also found to become stronger. In other words, 
intragroup and intergroup differentiations do not appear as being 


26 
Social identity: international pers
pectives 
necessarily located at the opposite poles of a continuum. Sometimes, they 
appear as being consistent, simultaneous judgements which can 'vary 
together'. This is usually referred to as the 'covariation hypothesis'. 
The 'covariation' model was developed in the frame of a greater attention 
to comparison processes occurring in 'natural' sets, where the different 
groups are often characterized by a different social status. Natural groups 
and their functioning already represented a matter of central concern in 
Tajfel's perspective (see for instance, Tajfel, 1 982b). Yet, in his research, 
prominent attention was given to the minimal experimental conditions of 
social categorization. 
It was mostly Doise's model of the 'categorial differentiation' which led 
to a systematic employment of natural groups (for example, males/females; 
high school students/apprentices) in the experiments (Doise, 1 976). This 
yielded a special focus on the asymmetrical functioning of categorial differ­
entiation, due to the place taken by different groups in our society. In this 
vein, Deschamps ( l982b) focused on the relationships between 'dominant' 
and 'dominated' social groups. According to him, subjects belonging either 
to a 'dominant' social group or to a 'dominated' one would show important 
differences in perceiving intergroup similarities and in 'favouring' behav­
iours, as well as in the ways of representing themselves. Social identity 
would vary as a function of the 'symbolic and material capital' available to 
different subjects. In particular, the members of a 'dominant' social group 
would show a greater tendency to perceive themselves as highly distinctive 
individuals endowed with unique idiosyncratic traits. The members of a 
'dominated' group, instead, would usually show a tendency to perceive 
themselves as somewhat undifferentiated components of a 'collective object' 
(Deschamps, 1 980; Lorenzi-Cioldi, 1 988). 
In a study already mentioned above, Zavalloni suggested that there may 
be deeply different identities, depending on the social status of the mem­
bership group. In her view, a meaningful example is offered by the asym­
metrical development of men's and women's identities. The two genders are 
seen as the 'most basic identity groups, not only because they were the first 
to be learned . . . but also because they have traditionally englobed all 
others' (Zavalloni, 1 983, p. 2 1 5), thus becoming the basis of a social 
hierarchy codified in the culture. The dynamic interplay of men's and 
women's identities implies a power relationship, analysed with reference to 
its most subtle and elusive aspects. Language itself, for instance, reveals that 
only 'male categories' are usually associated with the positive cultural 
connotations of the concept of humanity (Zavalloni, 1 983, p. 217). 
There is an essential convergence in these different studies. Although 
adopting different methods and starting points, they emphasize the close 
interdependence between comparison processes, self-representations and 
individuals' concrete positions in a hierarchical social context. They suggest 
that the different levels of self/others comparison (namely intraindividual, 
interindividual, intergroup levels) are closely articulated with each other 
(Doise, 1 988). 


The personal-social interplay 
27 
In this view, Deschamps's studies on 'covariation' and those on the 
comparison between 'dominant' and 'dominated' social groups should be 
considered as a whole. The covariation hypothesis suggests that, instead of 
being opposed along a continuum, the personal 'pole' and the social one 
can be seen as 'similar' (Doise, 1 988, p. 1 05). Yet, in our opinion, this 
'similarity' is not to be understood merely as a form of isomorphism, that is, 
as a 'structural homology' between interpersonal and intergroup com­
parisons. A deeper and more concrete connection between different levels 
and aspects of self/others comparison can be underlined. This rests on the 
fact that social conditions may affect not only intergroup comparisons but 
also the forms of individuals' membership, as well as the structure and 
contents of personal identity. The links between interpersonal and inter­
group comparisons can be further analysed along these lines. 
We will briefly consider now two main questions singled out in this 
section. These refer, respectively, (a) to the relationships between similarity 
and difference and (b) to the interplay of cognitive and social processes in 
social comparison. 
Similarity relationships and self/others comparison 
Some of the studies mentioned above suggest that there may be several 
ways of connecting similarity and difference with each other in social 
comparison. This point can be further highlighted by referring to Codol's 
experiments on the 'superior conformity of self' behaviour (Codol, 1975). 
This is also described as the primus inter pares (PIP) eff
ect: the subjects of 
Codol's experiments, set in a highly normative situation, tended to present 
themselves as conforming, better than the others, to the shard norms. 
Actually, they based their own self-distinctiveness on the claim for a 'simi­
larity', deemed to be a value in this case. Considering these effects, we may 
notice that an entangled co-occurrence of similarity and difference appears 
at the intra-categorial level. This being similar and different at one and the 
same time (Codol, 1 982) sheds light on the dynamic interplay (if not, on the 
very 'genetic' relationship: see, for instance, Luria, 1 976) between personal 
and social identity. It is only by starting from some adhesion to a group, to 
its norms and values (that is, by starting from a certain conformity) that it 
becomes possible to develop the very idea of one's own uniqueness, and to 
set out an image of oneself as a highly distinctive individual. 
Further insights into the complex structure of the self/others comparison 
can be derived from a particular theoretical approach: that singled out by 
Tversky (1977) in his analysis of similarity. Tversky goes beyond the 
traditional 'metrical' approach, wherein similarity and difference are 
usually described just as proximity/distance between points along a linear 
dimension. He opposes this view with a 'non-metrical' approach: in this 
case, similarity is understood as a global judgement (that is, a J
eature­
matching process), taking into account both common and distinctive 


28 
Social identity: international perspectives 
features of the objects to be compared. This author demonstrated that 
similarity judgements usually take a specific direction. In general, one of 
the two objects is in fact 'prominent' (that is, more 'salient' or 'proto­
typical' than the other). It then becomes the reference point (the 'model') 
which the other term (that is, the 'variant') is usually referred and 
compared to. By systematical variations of the comparison reference point, 
Tversky showed that perceived similarity is higher in this standard form of 
comparison than in the inverse condition (that is, when the model is judged 
in reference to the variant). 
Following these studies on the 'asymmetry phenomena', several authors 
attempted to verify whether (and in what conditions) the self is a major 
reference point in interpersonal comparison (Holyoak & Gordon, 1 983; 
Srull & Gaelik, 1 983). Codol (1984b, 1 987)2 suggested that interpersonal 
comparison too is characterized by a cognitive asymmetry, depending on 
the different 'salience' of the terms (self, other) under consideration. Codol's 
findings show that, in general, perceived similarity is higher when the others 
are compared to the self than in the inverse condition. In practice, it is not 
the same to state: 'I am similar to the others' as to state: 'the others are 
similar to me'. A person would tend to claim others' similarity ('the others 
are similar to me more than 1 am similar to the others') as well as hislher 
own difference ('I am different from the others more than the others are 
different from me'). 
Starting from these results, several meanings of similarity and difference 
can be pointed out (Serino, 1 988b): similarity may mean 'conformity', but 
also 'representativeness'; difference may mean 'specificity' but it may refer 
to 'deviance' as well. These different meanings are the result of a social 
construction and can be 'recoded' depending on the comparison direction 
(Serino, 1 989). 
A 'non-metrical approach' to similarity may yield important conse­
quences for the analysis of social categorization. One of these consequences 
could be the greater attention to the 'productive' aspect of categorization. In 
fact, a feature-matching process involves a number of questions concerning 
the ways in which similarity judgement is performed whenever a few simple 
(and generally a priori provided) dimensions are not available as yardsticks 
for comparison. The multidimensional character of comparison is then 
emphasized: in this light, similarity and difference may even appear as 
compatible and co-occurring judgements, rather than inversely related and 
opposed along a continuum. 
A 'metrical approach' to similarity? 
The way in which concepts are operationalized may affect the way in which 
particular questions are focused on. Thus, when considering social identity 
theory, we may observe that (a) by focusing, mainly, on 'favouritism', it 
appears as shifting to a motivational analysis of group membership, and 


The personal-social interplay 
29 
that (b) by considering personal and social identity as inversely related 
levels, the true meaning and the subjective relevance of an identity group for 
the individuals risk being somewhat neglected. 
Starting from Tversky's analysis, we might even observe that Tajfel's 
approach to similarity seems to lie on some 'metrical' assumptions. Right 
from his earliest studies in this field (Tajfel, 1 959; Tajfel & Wilkes, 1 963), 
some similarity/difference judgements were inf
erred, being based on the 
perception of some 'proximity' or 'distance' between points along dimen­
sions (such as weight or length) already provided by the experimenter. The 
authors were quite aware that this procedure needed to be further 
controlled. Wilkes and Tajfel (1966) could in fact demonstrate that there 
exists a connection between the 'distance of an actor from the series of 
stimuli and its categorization as the "same" or "different from" the series' 
(Tajfel, 1 98 1 ,  p. 92). Yet, these experiments cannot explain how similarity 
judgements are performed whenever comparison is not based on a few 
simple units or dimensions which are already available and/or directly 
provided to perceivers. 
Some general issues of research on social categorization might be traced 
back to this 'metrical' view on similarity. For instance: 
(a) special attention is paid to the 'deductive' face of categorization, that 
is, to its consequences: categories are understood as a condition, at the 
origin of the categorizing activity, more than as a product of this 
activity; 
(b) judgement is split up into as many similarity relationships as there are 
comparison dimensions; 
(c) similarity and difference appear as being two opposite and mutually 
excluding forms of relationship between the terms to be compared. 
The same characteristics are also shared by social identity theory, as IS 
highlighted by the following SIT assumptions: 
As a consequence of the outstanding categorization (whether 'inter­
personal' or 'intergroup'), subjects will act either as mere individuals or 
as group members. 
2 
A more direct or explicit involvement of the self in the comparison 
would change the whole situation: the previous categorization would be 
cancelled and a different one would be introduced. Thus, the existing 
opposition between two groups is replaced by a more general (oneself
/ 
another) opposition, wherein membership of either category ends up 
becoming somehow irrelevant. 
3 
When an intergroup categorization is activated, the typical accentuation 
of intra-categorial (that is, self/others) similarity will be observed. Con­
versely, an accentuation of oneself/another difference is likely to occur 
only when intergroup categorization is not salient. 


30 
Social identity: international perspectives 
As a result, two main issues arise in the frame of social identity theory. 
First of all the close interdependence between personal and social aspects of 
identity is often overlooked. Second, self
/ingroup and self
/outgroup 
conditions are not sufficiently differentiated from each other. These are 
just described as two, more or less similar, examples of 'interpersonal' 
comparison. 
To summarize, the continuum model, based on a necessary co-occurrence 
of intra-categorial assimilation and inter-categorial dif
f
erentiation, is not the 
only way of describing the relationships between personal and social 
identity. In our view, even the analysis of 'covariation' could be further 
developed. Not only can intra- and intergroup comparisons appear to be 
'similar' rather than 'opposite' poles (as in Deschamps's experiments show­
ing that intra- and intergroup similarities can 'go together'). As a matter of 
fact, a special form of 'covariation' also occurs when the self takes a 
prominent (prototypical) role, thus appearing, simultaneously, as being 
both similar to and dif
f
erent f
rom the other ingroup members. Here, too, 
similarity and difference are linked to each other: yet, in this case, they 'co­
vary' in a quite unexpected way. 
The relationships between similarity and difference are not totally 
explained by an 'all or none' approach (Deschamps & Serino, 1 996), and 
this may yield some consequences for identity research. 
Self-categorization theory 
More recent developments of Turner's analysis were aimed at dealing with 
some of the problems pointed out above. Although an essential continuity is 
affirmed between social identity theory and Turner's self-categorization 
theory, nonetheless some important novelties are also introduced. In par­
ticular: (a) the attention given to motivational factors is reduced; (b) the 
'productive' face of categorization is emphasized; (c) similarity and 
difference are more closely connected with each other. 
Of course, an exhaustive description of this theory exceeds the scope of 
our chapter. Thus, we will just illustrate, briefly, the three points stated 
above. 
Focus on cognitive processes 
Turner integrates the prototypicality theory with the analysis of social 
categorization. 
The prototype concept was singled out by Rosch (1973, 1 975, 1978) in the 
frame of her analysis of categories as 'fuzzy' sets. According to Rosch, 
class-inclusion can hardly be conceived as a rigid all-or-none process. In 
most cases, in fact, classification appears as a probabilistic process. More or 
less typical examples are gathered together in the same category; therefore, 


The personal-social interplay 
3 1  
categories take a hierarchical structure, being organized around a central 
(prototypical) element. Categories are organized, in turn, into taxonomies 
(that is, hierarchical categorial systems), with different levels of abstraction 
and inclusiveness (Rosch, 1 978). 
Turner draws on these ideas when developing his self-categorization 
theory. According to him, the self is (also) a cognitive structure, and self­
categorizations can be described as 'a part of a hierarchical system of 
classification' (Turner, 1 987, pp. 44-45). In this view, 'all social com­
parisons with others depend upon the categorization of the others as part of 
a self-category at some level of abstraction' (Turner, 1 987, p. 48). 
Conforming to Rosch, the self-categorization process gives way to a 
number of superordinate and subordinate self-categories, with increasing 
levels of abstraction and inclusiveness. Turner explains this point by singling 
out three different levels, describing 'one's human, social and personal 
identity respectively'. A very abstract and inclusive level of self­
categorization is the one opposing 'human beings' to other forms of life 
(inter-species comparison). At an intermediate level (intergroup compar­
ison), the self-concept describes the individual as a member of particular 
social groups. The most concrete and least inclusive level of categorization 
would be based on interpersonal comparisons, wherein a person is 
distinguished from the others by means of hislher particular personality, of 
hislher unique and strictly individual characteristics. 
The productive f
ace o
f
 categorization 
The 'productive' character of social categorization is put forward by 
focusing on those processes directed to select, among different levels of 
abstraction/inclusiveness, the more pertinent ones in the given comparison 
conditions. For this reason, the question of salience is particularly 
emphasized (Oakes, 1 987; Oakes & Turner, 1 986). According to Turner and 
his co-workers, the salience of a given categorization can hardly be 
described as the effect of an automatic perceptual prominence of certain 
stimuli. Instead, it depends on the interaction between accessibility and fit 
(see Bruner, 1 957). Accessibility deals with the degree of readiness with 
which certain pieces of information are activated, while fit 'ties 
categorization firmly to reality' (Oakes, Turner & Haslam, 1991) in that 
it refers to an adequate and veridical representation of social situation. In 
this view, the salience of a given categorization is highly dependent on 
context, as well as on perceivers' purposes and goals. 
Unlike Rosch, Turner suggests that the basic categorization level (that is, 
the level at which the best 'compromise' between information generality and 
concreteness is achieved) is a 'highly contextual variable rather than fixed' 
(Turner, 1987). 
Cantor and Mischel (1979) already emphasized that the relation­
ships between different levels of abstraction is flexible rather than rigid in 


32 
Social identity: international perspectives 
categorial systems concerning social reality. These authors too applied 
prototypicality theory to the analysis of persons' perception, focusing on 
context and perceivers' purposes as crucial aspects of classification. They 
pointed out a number of analogies, but also some important differences in 
the ways categorial judgements are performed in the social versus non-social 
domain (see also Holyoak & Gordon, 1 984; Wyer & Podeschi, 1 978). For 
instance, the different levels of abstraction seem to be 'interchangeable' 
(Cantor & Mischel, 1 979) in the social domain. Moreover, class-inclusion is 
mostly based on a bipolar opposition, that is, on the incompatibility 
between different categories, rather than on merely descriptive typologies 
(Cantor & Mischel, 1979, p. 33). These differences between social and non­
social categorial systems emphasize the evaluative character of classification 
in the social domain. Elsewhere (Serino, 1990) we addressed this question 
by referring to Billig's (1985) analysis of 'particularization'. In a well-known 
example, Billig describes the process by which one general category (for 
example, Christianity) is split up into two distinct sub-categories (black and 
white Christians). Notice that the new distinction does not describe only the 
shift from a wide, superordinate category to less abstract and inclusive ones. 
The difference between black and white christians, in fact, may also entail a 
basic opposition between a lay and primitive religiousness and a deep and 
mature one. From this viewpoint, particularization yields an evaluative 
discrimination allowing the new ingroup to take its place at the top of an 
extremely general hierarchy (that opposing 'true' religiousness to super­
stition). The reference to a value system entails both particularization and 
generalization, as two interrelated aspects of categorizing. The reference to 
a value system also explains why, although the level of abstraction is not 
immutable, the 'inclusiveness of the category remains stable' (Cantor & 
Mischel, 1 979, p. 1 6). In fact, even the introduction of a very abstract level 
of classification can be aimed at enhancing interpersonal or intergroup 
differentiation! 
Our discussion converges on Turner's position. According to him, 
categorization and particularization are two interdependent aspects of the 
same process (Turner, 1 988). Turner's description of 'normative fit' as an 
important component of salience also emphasizes, as we do, the social 
meaning of categorization. He argues, for instance, that categorizing some­
one as 'Catholic' is not only the result of a comparison with other categories 
(cf. Oakes et aI., 1 991); it also deals with behavioural and normative 
contents of that person's identity. 
Indeed, social categories are highly meaningful and personally relevant 
for individuals. This might account for 'people's tendencies to perceive 
themselves and others in terms of psychologically opposing forces' (Cantor 
& Mischel, 1 979) which seem to affect the judgements of prototypicality in 
the social domain. In this light, proto 
typicality itself might carry a par­
ticular meaning: when social categories are involved, most 'prototypical' 
features would be in fact those endowed with a strong (positive or negative) 
value. 


The personal-social inter
play 
33 
Similarity and dif
f
erence as interdependent judgements 
In self-categorization theory, similarity and difference are linked to each 
other in a more complex fashion. Difference always implies a 'higher level 
identity', so that no comparative relation in terms of difference could be 
assessed between two stimuli if they were not suitable to be included in 
the 'same category' at a higher level of abstraction (Turner, 1 987). Accord­
ing to Turner, categorization follows the meta-contrast principle. This 
entails an inverse relationship between intra- and inter-categorial com­
parisons.3 Therefore, it is consistent with the idea that the perception of 
intra-categorial similarities and inter-categorial differences is at the basis 
of categories formation and functioning. In this view, the salience of a given 
category and even the prototypicality of its prominent examples can be 
understood as a function of the 
meta-contrast perceived between intergroup and intragroup differences. As the 
differences perceived between individuals belonging to different (implicit) groups 
increases, compared to the differences between individuals belonging to the same 
(implicit) groups, then the more salient will the social categorization become . . .  
i.e. members will tend to be perceived less as individually different persons and 
more as unitary social groups. (Oakes et aI., 1991, pp. 126-127) 
Turner (1987, p. 49) claims there is a 'functional antagonism between the 
salience of one level of self-categorization and other levels', assuming that 
'there tends to be an inverse relationship between the salience of the 
personal and of the social levels'. In this sense, Turner clearly draws on 
Tajfel's idea of 'continuum'. 
Prototypes: similar and different 
The meta-contrast ratio also provides a simple operational description of 
the prototype concept. As Turner suggests, a given stimulus will tend to be 
perceived as more exemplary or representative of the category as a whole 
the higher the meta-contrast ratio: the higher the mean perceived difference 
between the target stimulus and the out-group members over the mean 
perceived difference between the stimulus and other ingroup members, the 
more prototypical the target stimulus. 
Yet, by referring to the meta-contrast, only the comparative aspect (the 
'horizontal' dimension) of prototypicality is emphasized. Being aware of 
this, Turner himself recently focused on some aspects of the prototype 
which cannot be described in this way. 
In our view, the prototype concept also contains the idea that 
similarity and difference 'go together'. A prototype in fact differs from 
the other items because it better represents an abstract image or an ideal 
of the common category; an ideal which allows that category to be 


34 
Social identity: international pers
pectives 
defined and distinguished in turn from the other categories. Similarity 
and difference, conformity and uniqueness, are tightly intertwined in the 
prototype concept. 
This close connection of seemingly opposite judgements can be observed 
also when referring to the self-concept. On the whole, Codol's studies 
mentioned above seem to suggest that the self may act as a sort of proto­
type (Hardoin & Codol, 1 984) in interpersonal perception. By varying the 
order of two different description tasks (self-description, description of 
another stimulus-person), Codol (1987) also shows that common and dis­
tinctive traits carry a different weight across different conditions. Subjects 
starting with the self-description task used the same self-descriptive traits for 
describing the others. By contrast, when the other-description came first, 
subjects tended to use quite different traits in the subsequent description of 
themselves. (Different results were obtained when the other person was the 
subject's 'best friend'.) According to Codol, these results would suggest that, 
in general, 'the others' are assumed to belong to the same category as the 
self, whilst the reverse is not true. 
More generally, there would be a dynamic interplay between similarity 
and classification. As Tversky (1977) argued, similarity can determine the 
way in which objects are classified, but it also depends on the adopted 
classification. 
While developing his analysis of salience, Tversky suggests that each 
pertinent feature is weighted according to both its 'intensity' and its 'diag­
nostic value' (this refers to the 'kind of classification' introduced by that 
feature). Of course, these two different components are assumed to be both 
present and tightly intertwined in any matching process. Nonetheless, we 
may easily imagine some conditions where either intensity or diagnosticity 
is, in turn, particularly emphasized. 
In some cases, when two objects are to be compared, they are already 
assumed to belong to the same category. Comparison is just aimed at 
singling out the reasons for an overall accepted similarity. When a common 
category is involved (even implicitly), comparison might give rise, mainly, 
to the (differential) intensity of the shared (categorial) features. 
In other conditions, similarity is the final result of a more inductive 
process. Here, judges are searching for an appropriate category including 
both objects (thus making the comparison possible): greater attention would 
be paid, then, to the simple presence/absence of several (either common or 
distinctive) features, whose 'diagnostic value' is foregrounded. 
When considering in this light Turner's analysis of the different self­
categorization levels, it seems that similarities and differences are judged 
mainly by referring to the diagnostic value of particular traits. Thus, at the 
interpersonal level, individuals are deemed to make comparisons only on 
the basis of their unique, exclusive personality traits. These are obviously 
opposed to other traits (which are not recognized as pertaining to them­
selves) in a kind of judgement giving rise, mainly, to the simple presence/ 
absence of the critical features. 


The personal-social interplay 
35 
However, we are dealing here with an abstract idea of the individual. 
Indeed, when self and others are perceived in terms of 'psychological 
opposing forces', little attention will be paid to common characteristics. 
Nonetheless, interpersonal comparisons are often performed in a frame of 
already accepted common categories, relevant to each member and to his/ 
her sense of personal identity. In these conditions, greater attention would 
be paid to some characteristics which are present, to a different degree, in 
the different (more or less representative) ingroup members. 
In this way, we can also point out an important difference between self
/ 
ingroup and self/outgroup comparisons: these are conditions where a 
different focus on either intensity or diagnostic value of the relevant features 
might be observed. 
Research on intergroup relations already demonstrated that the ingroup 
members are perceived as varying along specific dimensions more than the 
members of an outgroup are (Park & Rothbart, 1 982); actually, it seems 
that 'ingroupers' are compared in terms of different intensity of the same 
features, while 'outgroupers' are not. 
At the intragroup level, the self/others comparison might be based on the 
different intensity of common characteristics, more than on the different 
salience of features unique to each individual. Again, some aspects of the 
personal-social interplay seem to escape an explanation in terms of 'con­
tinuum', and we can imagine an interpersonal level of self-categorization 
where some shared social characteristics do not stop being salient and 
important. 
In most cases, a special form of particularization is likely to occur at the 
intragroup level: it would be based on dimensional discrimination about 
some crucial categorial characteristics which are shared, to a different 
extent, by oneself and by the other ingroup members. This process seems to 
have been described, for instance, in a study by Montei! ( 1992): this shows a 
particular form of 'individualization', based on the individual appropriation 
of the positive characteristics of the group. 
On the other hand, a generalization process is required when self and 
other are (perceived as being) quite different. 'Similarity' can be assessed, in 
this case, only by shifting to a very inclusive category (see also Turner, 
1 987). At this level, people may share, at least, the fact of being 'indi­
viduals', endowed with their own personality. Thus, personal traits, which 
are usually considered as being the most genuine and concrete identity 
issues, would be typically elicited in some very general and abstract inter­
personal conditions. 
Starting from the prototype concept, personal identity and intragroup 
relations appear as an amazing, intertwined connection of similarity and 
difference, unicity and conformity, assimilation and distinctiveness (Codol, 
1 982). It would be hard to reduce this complex interplay to the frame of a 
personal-social bipolar opposition. In this light, what about the 'functional 
antagonism' defended by Turner? Is it possible to reconsider the same idea 
in a diff
erent perspective? 


36 
Social identity: international perspectives 
'Functional antagonism': different levels of categorization or different 
modes of processing information? 
According to Turner, different contents become salient at different (intra­
group, intergroup) levels of categorization. His idea of the functional 
antagonism between personal and social levels of self-categorization also 
entails the opposition of personal versus social features. This dichotomy can 
hardly be overcome just be postulating, as Turner does, a 'midpoint' of the 
personal-social continuum 'where the self-perception is likely to be located 
much of the time' (Turner, 1 987). 
We would refer to a different form of functional antagonism, where 
different modes of processing information are applied to a substantially 
unitary social reality. Cognitive research on person perception has pointed 
out a fundamental opposition between categorial and personalized 
information, top-down and bottom-up ('stimulus-related') processes (for 
example, Asch & Zuckier, 1 984; Fiske & Pavelchak, 1 986; Pratto & Bargh, 
1991). 
Brewer (1988), for instance, proposed a 'dual model', wherein which route 
is chosen depends, mainly, on the perceiver's purposes and goals. Similarly to 
Turner, Brewer suggests that 'the level of abstraction at which new 
information is initially encoded is expected to be different for category-based 
and person-based processing'. It is worth emphasizing, however, that exactly 
the same information 'can be processed in a top-down (de-personalized) 
manner (resulting in category-based cognitions), or in a bottom-up (per­
sonalized) fashion, that results in person-based representations'. Thus, 
'personalization refers to a different kind of cognitive structure in which the 
individual becomes the basis for organization of all relevant information'. 
When the personalized view is activated, perceivers are dealing with 
salient information which can hardly be reduced to strictly individual 
properties. Personalization does not lead, necessarily, to focusing on 
particular contents or on isolated individuals. Self-involvement may also be 
linked to the increasing awareness of a number of complex interactions. 
In other words, personalization could give way to an extremely sensitive 
appraisal of all pertinent variables and of their concomitant variations. This 
was suggested by a recent study (Serino & Mazzara, 1 995), bearing on 
persons' perception in particular social contexts. Subjects (high school 
students) were presented with a slide showing either a class-room (C) or a 
fast-food restaurant (F) and an ambiguous figure entering the situation. 
This figure was described either as 'a person' (N) or as a 'handicapped 
person' (H). Among other tasks, subjects had to describe themselves and the 
stimulus-person by means of a list of 1 5  bipolar dimensions. The order of 
these descriptions was systematically varied: half the subjects started by the 
self-description, later describing the other person; the other half described 
the stimulus-person first, while the self-description was performed later. On 
the whole, three independent variables were then introduced: stimulus­
person (NIH); situation (C/F) and re
f
erence point (either 'oneself' or 'the 


The personal-social interplay 
37 
other person', assuming that the term described first also serves as a 'model' 
in the subsequent description). 
According to our results, the description of the other person was mostly 
based on categorial information (that is, whether that person was presented 
as being 'normal' or 'handicapped': F(1 5, 120) = 4.87; P = .001). The self­
description was performed in a quite different way. Different effects of the 
three independent variables are observed in reference to each particular self­
descriptive trait, so that, for instance, some aspects of the self-description 
seem to be affected by the situation, other aspects depend mainly on the 
re
f
erence point, while, in other cases, the interaction of all three independent 
variables yields significant effects.4 
In short, the self-description seemed to be more mutable, supple and 
'personalized'. The special nature of this 'personalization' is worth empha­
sizing, however. It does not lie only on the reference to personal traits 
independent of the context. Personalization might also spring from an 
extremely fine-grained analysis of all variables involved in a given social 
context. 
Asymmetries and prototypes in intergroup relations 
In the previous paragraphs we challenged the idea that personal identity 
refers, necessarily, to an isolated individual and to hislher unique person­
ality traits. However, the close interdependence of personal and social 
identity can also be met from a different perspective. 
As we have already suggested, being a member of particular groups may 
deeply affect the individuals' experience, their ideas about social reality, 
others and themselves. In an intergroup context, usually characterized by 
several status differences, even self-concept, self
/others comparison and 
personal identity could be defined in different ways, depending on indi­
viduals' and groups' position in the social hierarchy. Thus, a number of new 
questions arise: What structure does the self
/other comparison take when 
the 'others' are ingroup or outgroup members, respectively? Do the nature 
and the social status of these different 'others' affect comparison modes and 
direction? What are prototypes in an intergroup frame? 
By referring to the asymmetry phenomena described above, we may even 
wonder whether there is any link or overlapping between the objective, 
social asymmetry of intergroup relations and the cognitive asymmetry of 
perceived similarity. By linking the cognitive asymmetry of similarity 
judgement to the structure of concrete intergroup relations, the close inter­
dependence between mental and social processes could be better highlighted 
(see also Boltanski, 1982). 
In a study by Hurtig, Pichevin and Piolat (1 992) this connection is 
highlighted. These authors focused on sex and age as factors affecting 
asymmetry phenomena. Their study centred on similarity judgements 
between pictures presenting two people of different sex and age. Indeed 


38 
Social identity: international perspectives 
these two variables were considered as defining different social categories, 
hierarchically related to each other (younger/older; male/female). Conform­
ing to the hypothesis, a strong asymmetry effect was found, due to the sex 
of the ref
erence point person. When the male photo was the reference point, 
the similarity perceived between the two stimulus-persons was higher. The 
authors argued that the male figure showed a great 'salience' (or 'proto­
typicality' degree) than the female figure. Moreover, when same-sex pairs 
were compared, similarity was perceived as being lower in male pairs than 
in female pairs: that is, the more salient (prototypical) stimulus also 
appeared as fitting a highly articulated and best differentiated social 
category. 
In the same vein, I focused on the self/other asymmetry occurring when 
different (advantaged/disadvantaged) social groups are involved. After a 
pilot-study, the powerful asymmetry between Northern and Southern 
Italians was taken as an example of relationships between 'dominant' and 
'dominated' social groups. The Northerners were considered as the upper, 
dominant group. 
In the first of these studies (Serino, 1 988b), 358 subjects belonging to the 
two groups (Northern and Southern Italians) had to perform a comparison 
between themselves and the Northerners or between themselves and the 
Southerners, later providing, by free explanation, the reasons for their 
previous judgement. The comparison reference point (oneself or others) was 
systematically varied, as in other experiments on the asymmetry effect.5 A 
2 x 2 x 2 design was set up, taking as independent (between subjects) 
variables the kind o
f
 judgement (in terms of similarity or difference), its 
content (comparison with the ingroup or with the outgroup) and the 
comparison re
f
erence point (oneself or the others, respectively). 
According to our results: 
A certain asymmetry (that is, an independent effect of the reference 
point variable) was observed. 
2 
This effect was found mainly in comparisons involving the group of 
higher status (that is, the Northerners). 
3 
Self
/ingroup and self
/outgroup comparisons tended to take an opposite 
'direction'. In the self
/ingroup comparison, the reference point was the 
group rather than the self (for example, 'I am similar to the Northerners 
more than the Northerners are similar to me'). An opposite (although 
non-significant) tendency appeared in comparisons with the outgroupers. 
On the whole, the different contents elicited in self
/ingroup and self
/ 
outgroup comparisons show that, at the intragroup level, subjects are com­
pared not only to other individuals, but also to a general norm. Similarity 
with the 'ingroupers' is based on a large amount of references to 'member­
ship'. Figure 3.1 suggests that group-belongingness is dramatically empha­
sized in the self
/ingroup condition. Thus, on the one hand, we can stress the 
difference between self
/ingroup and self/outgroup conditions; on the other 
hand we can observe that it is not a simple comparison between individuals 


The personal-social interplay 
1 20 
1 00 
• Self/ingroup 
D Self/outgroup 
80 
60 
40 
20 
0 
TR 
CU 
HIS 
BHV 
SEC 
MBSH D/ND 
OTH 
Key: 
TR = personality traits; CU = culture and habits; HIS = history; 
BHV = behaviours; SEC 
= social and economic conditions; 
NC 
MBSH = group membership; D/ND = difference and/or non-difference; 
OTH = other categories; NC = non-classified 
Figure 3.1 
Mean scores ( x  100) o
f
 each content category elicited in 
self/ingroup versus seljloutgroup similarity judgements (N = 121) 
39 
(just focusing on strictly idiosyncratic personal traits) that we observe at the 
intra-categorial level. 
The above-mentioned study did not provide decisive results. It suggested, 
however, that subjects belonging to the dominant group (that is, Northern­
ers) were more sensitive to the structure of the comparison: perceived 
similarity and difference varied as a function of the reference point (either 
self or group) even at the intragroup level. By contrast, Southern subjects 
appeared mostly concerned with the comparison content: in this case, sig­
nificant effects were due, mainly, to the 'type of comparison' variable (that 
is, whether the Northerners or the Southerners were involved). 
On the other hand, the free answers pointed out two different ways of 
referring to a common category. The Northerners employed 'we' mostly in 
relation to their own group (we, the Northern Italians). By contrast, the 
Southerners referred to a more inclusive category, describing the set formed 
by both groups (we all, Italians from the North and the South). Thus, it 
seems that the 'dominant' subjects focused mainly on their own category, 
while the 'dominated' subjects centred, mostly, on the relationshi
ps between 
the two categories (Northerners/Southerners). In a sense, our Northern 
subjects (that is, the 'dominant' ones) would given an example of 
comparison 
conditions wherein similarity and difference are activated simultaneously at 
the intra-categorial level. Conversely our Southern ('dominated') subjects 
would set out a kind of comparison where similarity and difference are 


40 
Social identity: international perspectives 
activated simultaneously at the inter-categorial level. A strong tendency to 
overcome the Northerners-Southerners opposition goes hand in hand with 
the clear awareness of the difference between the two categories (a difference 
recoded as unjustified discrimination, and then rejected). 
Some of the above results were worth further investigation. To this end 
new research on the same topic (Serino, 1 996) was carried out. A group of 
III subjects, equally divided into Northerners and Southerners,6 had to 
assess the perceived similarity between themselves and the Northerners or 
between themselves and the Southerners. The 2 x 2 x 2 design considered 
as (between subjects) independent variables: 
(a) the subjects' origin (either from the North or from the South); 
(b) the comparison content (comparison with the Northerners or with the 
Southerners); 
(c) the comparison re
f
erence point (either 'oneself' or 'the others'). 
Let us examine some results: 
The comparison content yields significant effects: greater similarity is 
assessed in comparisons with the Northerners than with the Southerners 
[F(l ,  101) = 7.76; p = .01]. This happens for all subjects, irrespective of 
their origin. All subjects underline that the Northerners (that is, the 
group of higher status) are 'more similar' to themselves. 
2 
On the whole, asymmetry phenomena are more marked in comparisons 
with the Northerners (tS4 = 2.80; P 
= .05) than with the Southerners 
(tSI = - l .65; n.s. ). This tendency may suggest that a group of high status 
also entail a more prominent, highly differentiated cognitive category. 
3 
Again, the subjects from the North and those from the South seem to 
construct their self
/others comparisons in different ways. It is mainly the 
subjects belonging to the group of high status (Northerners) who grant a 
prominent role to the self, which is taken as the comparison reference 
point. By contrast, the Southern subjects show a greater tendency to 
judge their own similarity in reference with 'the others'. 
4 The self
/ingroup comparisons appear to be more 'asymmetrical', which 
might suggest that ingroup refers, in any case, to a more differentiated 
and complex category than that describing the outgroup. This is true for 
all subjects, but, once again, the comparison direction is not the same 
for Northern and Southern respondents. For the Northerners, the self/ 
ingroup similarity is higher when it is judged in reference to themselves 
(t24 = 3.182; P = .004). They appear, then, as particularly centred on 
their own 'self' inside the membership group. By contrast, the Southern 
subjects tend to focus on their own similarity, by taking the 
'Southerners' (rather than 'themselves') as the comparison reference 
point ( t26 = -3.38; P = .01). 
On the whole, the different direction of these comparisons may suggest 
that different meanings of similarity are entailed by Northern and Southern 
subjects. In fact, the former seem to emphasize their own representativeness, 


The personal-social inter
play 
41 
while the latter appear as mostly concerned with their own conf
ormity to an 
external model. Cognitive asymmetry, which is more marked in those 
comparisons involving the group of higher status (that is, the Northerners), 
may shed light on the close connection between mental and social aspects of 
companson. 
Conclusion 
In this chapter, we have focused on personal and social identity by referring 
to the directional character of similarity (Tversky, 1 977), to the role of 
prototypes in person perception (Cantor & Mischel, 1 979; Rosch, 1978), 
and to the different ways of processing social information (Brewer, 1 988). 
We argued that the close relationship between individual and collective 
processes, personal and social dimensions, is not always suitably represented 
by the continuum model described in social identity theory. A dichotomic 
view of intra- and inter-categorial relations, of personal and social identity, 
still appear as the basis of Tajfel and Turner's research. More recent 
developments of the same perspective (such as Turner's self-categorization 
theory) integrate the theory under several important aspects. Yet that 
'dualistic conceptualization of macro/micro 
social relations' (Condor, 1990) 
which is typically expressed by the notion of 'continuum' does not appear to 
be completely overcome. 
When considering self/others comparisons performed in a concrete social 
context, it seems that (a) even personal identity is socially defined, 
depending on individuals' and groups' place in a given social hierarchy, and 
that (b) even categorial information can be processed in a 'personalized' 
manner. From this viewpoint, personal and social processes, individual and 
group identities, show their close connections. 
In short, the personal-social interplay is worth deeper analysis. Con­
cerning the self/others comparison, several strategies can be observed, where 
similarity and difference are linked to each other in unexpected ways which 
are beyond the classical connection of intra-categorial assimilation and 
inter-categorial differentiation. We explained this point by referring to the 
prototype concept. Prototypes are - at one and the same time - highly 
'conforming' (to the category standards) and clearly 'distinctive' (in refer­
ence to more peripheral categorial elements). A special interdependence of 
similarity and difference is then required. We have suggested that, in the 
social comparison, a similar close intricacy of similarity and difference often 
occurs at the intragroup level. Here, personal identity could be based not 
only on one's unique personality traits, but also on the different intensity of 
common characteristics, expressed to a different extent by oneself and by 
the other ingroup members. Thus, in some cases, different levels of self­
categorization could be characterized by different modes of judging (for 
example, intensity versus diagnostic value), more than by the emphasis on 
different contents. 


42 
Social identity: international perspectives 
The prototype concept implies a probabilistic view on class-inclusion and 
a non-metrical approach to similarity. We suggested that this non-metrical 
approach may yield some consequences for the analysis of social 
comparison, and we reconsidered some aspects of Tajfel's theory in this 
light. 
In the same vein, we referred to some studies on the asymmetry 
phenomena in similarity judgement, showing that these effects may also 
occur in the social domain. We must admit that it is not yet so clear what 
asymmetry really means when comparison involves such meaningful social 
'objects' as the self or an identity group. Nonetheless, these phenomena 
often reveal that some sophisticated strategies are activated in interpersonal 
perception and that respondents are sensitive to the context and to the 
(social, cultural) consequences of their judgement. From this viewpoint, 
asymmetry may even serve as an indirect measure of a certain stereotype, 
which operates independently of respondents' intention (Greenwald & 
Banaji, 1995; Serino, 1 996). 
The typical procedure of asymmetry research (based on systematic 
variations of the comparison reference point) would point out the different 
meanings of similarity and difference, which are often at the core of the 
interactions and 'negotiations' between different social groups. To illustrate 
this point, we referred to a few studies where asymmetry effects have been 
observed in similarity judgements performed by subjects belonging to 
groups of different status. Social status seems to affect the meaning of 
comparison, the ways in which self
/others similarity is assessed and the role 
of the self. 
From this point of view, asymmetry research may contribute to developing 
a typical social-cognitive level of analysis, wherein individual, interpersonal 
and intergroup processes are closely interrelated. Social categorization and 
similarity judgements are typical 'social-cognitive' processes: cognitive 
research can provide fruitful suggestions in order to understand their 
functioning. Yet the close articulation between mental and social processes 
still ought to be further highlighted. 
Notes 
I .  In this light, we may consider the extremely rich literature on social representations (for a 
review, see Jode1et, 1989). Intergroup differences in the ways of representing some particular 
aspects of social lif
e (such as mental illness) were highlighted in a series of studies carried out 
simultaneously in Italy and Spain (see Ayestanin, 1985; 8ellelli, 1987, 1994). 
2. The substantial continuity of Codol's research (from the discovery of the PIP effect to the 
analysis of its asymmetry phenomena in interpersonal comparison) is discussed in Serino 
(1992). 
3. Hogg and McGarty (1990) suggested that the meta-contrast ratio is more appropriate in 
simple unidimensional contexts than in more complex and cross-cutting social frames. 
4. For this reason, the self-description would hardly be explained by means of just one 
variable. Thus, we observe a slightly significant interaction of the three independent variables 
(F(15, 125) 
= 1 .69; p = .06) when considering all traits together. However, a number of 


The personal-social interplay 
43 
significant effects are observed when considering each particular bipolar dimension. Thus, for 
instance, some self-descriptive traits (for example, calm/agitated) vary according to the situation 
(F( 
I ,  141) = 16.46; p = .0 I); the re
f
erence point affects the description of oneself as being serious 
rather than f
rivolous (F(I ,  141) = 6.01; p < .01); the judgement of oneself as being gloomy or 
cheer
f
ul depends on the interaction of all three independent variables (F(I, 141) = 10.47; P = 
.0 
I); and so forth. 
5. For instance, some subjects had to assess 'their own similarity to the Northerners 
(Southerners), rather than the 'Northerners' (Southerners') similarity to themselves'. According 
to the experimental conditions, all kinds of comparisons were perf
ormed, by different subjects. 
6. The data in the Northern region were collected by N. Cavazza, University of Bologna. 


4 
Self 
-We-Others Schemata and Social 
Identifications 
Maria Jarymowicz 
Searching for the social self: the self and prosocial attachment 
In order to examine the social self from a social psychological perspective, 
we must refer to traditions in other areas of psychology (Jarymowicz & 
Greenwald, 1 993), and distinguish between the ego and the self (Allport, 
1 961; James, 1 890). In social psychology, this distinction is revealed in the 
view of the self as a sub
ject (or self as a knower) and the self as an object 
o
f
 knowledge (Epstein, 1980; Greenwald & Pratkanis, 1 984). In other 
words, we can describe the self as an active agent or as an object to be 
understood. 
Different conceptualizations of the social self define how the individual 
is connected to other people in the social environment. There are two 
particularly important approaches that deserve attention. On the one hand, 
the social self can be described by the individual's social identifications: 
different forms of syntony with other people and states of prosocial ego 
involvement. These states are accompanied by feelings of attachment, 
sympathy, affiliation, devotion, respect for others' expectations, common 
goals, understanding and acceptance. On the other hand, the social self 
can be associated with people's ability to perceive and recognize 
themselves, in other words, the ability to create cognitive representations 
of one's person and self schemata. The content of the self-description 
combined with the representation of other(s) constitutes the social self and 
we schemata. 
The relationship between the ego and the self, or between the ego-identity 
and self-identity (Erikson, 1956), has not been carefully examined in 
psychology. The aim of this chapter is to elaborate on the relation between 
the self and the other(s) schemata, on the one side, and the states of social 
identification, on the other. We are particularly interested in the relationship 
between the distinctness of self-we-other(s) schemata and attitudes towards 
others, ingroup versus outgroup members. 
In order to understand the process of social identifications, we can begin 
by questioning the generality of the popular thesis that states that the self­
others similarity leads to prosocial ego involvement and that the perceived 


Sel
f
-we-others schemata 
45 
Table 4. 1 
The social self as a structural basis o
f
 dif
f
erent f
orms o
f
 social 
identifications 
T
he self structure: 
internal basis of social identifications 
'T
he experiential social self' 
affect and drives mechanisms: emotional 
empathy, security, affiliation, and other 
hedonistic motives 
2 
'T
he puhlic self' 
dependence on group's reinforcements, 
the we-as-a-group concept (e.g. 'We - a 
family') 
3 
'T
he collective self' 
cognitive representations of real social 
divisions and memberships, the we-as-a­
category concept ('We - workers, women, 
etc.) 
4 
'T
he conceptual social sel
f' 
cognitive categorization based on abstract 
criteria, the attributive-we concept ('We -
honest, tolerant, etc.) 
5 
'T
he autonomous social self' 
differentiation of two self subsystems: 'me 
as an individual' and 'me as a member of 
social groups and categories' 
Ego-involvement: 
states of social identifications 
syntonic feelings, seeking for contact, 
affiliation, sensual comfort 
conformity: respect for particular ingroup 
norms and goals 
feeling of ingroup (categorial) homogeneity, 
internalization of social norms, ingroup 
favouritism 
f
eeling of we-ness across (and independent 
from) social groups and categories 
recognition of diff
erent and common values/ 
goals, enjoying communality with 
autonomous others, groups and categories 
similarity to others is the main determinant of attachment to them. This 
thesis has support in theoretical and empirical work (Byrne, 1 969; 
Reykowski, 1979). However, relationships between self-other(s) schemata 
and ego involvement are far more complex than this work suggests (Brewer, 
199 1 ;  Jarymowicz, 1987, 1 994c; Snyder & Fromkin, 1980). In order to 
understand this relationship, it is necessary to analyse different forms of 
prosocial ego involvement, ranging from the simple drives to affiliate or 
conform (when prosocial behaviour leads to affect comfort) to exocentric 
ego involvement (when prosocial behaviour is focused on other's states and 
needs, regardless of one's own needs: Jarymowicz, I 992a; Karylowski, 
1982). 
Social identifications, expressed by syntony with others and prosocial ego 
involvement, are sometimes considered to be defining characteristics of the 
social self. This view, however, creates more problems than benefits because 
there are many different forms of social identifications based on different 
social structures (Jarymowicz, 1 992b, 1994a; see Table 4. 1 ), 
On the primary level ( 1 )  drives and affective mechanisms lead to non­
specific affiliation and seeking syntony with others. The social self cannot be 
explicitly articulated by the individual, it is more like 'I have experienced it' 
or 'I f
eel it'. Identifications are automatic and strong, their objects are 


46 
Social identity: international pers
pectives 
persons who can gratify primary needs. The feelings last as long as the 
relationship with particular others endures. 
Thanks to a number of social reinforcers that can be offered by groups, 
the process of socialization determines the respect and conformity with 
group expectations (level 2). The social self may be a family, partnerships or 
classmates: we as a group is f
ormed (Jarymowicz, 1993a, 1 994b). Identi­
fications refer to the concrete group, and terminate when the person no 
longer belongs to the group. 
The ability to focus on broader social categories gradually develops, and 
the social self moves beyond face-to-face contacts (level 3): we as a category 
is created (Jarymowicz, 1 994b) based on intra-categorial differentiations. It 
gives rise to syntonical identifications with category members and results in 
group favouritism. Referents of these identifications are stable only so long 
as the cognitive map of the world reflects social reality. 
The human mind can create different types of categorizations as well 
(Brown & Turner, 1 981), for example those based on abstract criteria. 
Therefore, the attributive we appears ('We, optimists, art lovers'). This 'we' 
is formed independently of, and sometimes cross-cuts, social groups and 
categories (level 4). The identifications may refer to other people, regardless 
of their social membership. They are modified as new, cognitive, abstract 
categories of understanding the world (for example, acquiring the concept 
of 'loyalty' enables one to discern between loyal and disloyal people). 
Social categorization and social comparison processes are not the only 
ways the social self is identified. Individuals may define themselves beyond 
comparisons with others or membership to particular categories. Thanks to 
intra 
personal differentiation and formation of personal evaluative 
standards, one's self can be recognized as someone realizing the pursuit 
of certain superior values. Thus social identifications do not refer to a 
category of people but to common goals (level 5). When values are per­
ceived as common, identifications are likely to last a long time. When goals 
are common, identifications last as long as they are relevant. 
The distinction, we suggest, reflects the sequence of developmental stages 
in which people can use new cognitive categories and evaluative standards 
(Golab & Reykowski, 1 985; Hoffman, 1 989; Kohlberg, 1976; Piaget, 1 975; 
Reykowski, 1 989; Ziller, 1964). The stages also determine new areas and 
objects of people's identifications (Tap, 1 979): from social groups or 
categories of own membership (levels 1 -3), to categories going beyond 
social divisions (level 4) or categorization as such (level 5). 
How can we define the relationship between the social self, social identity 
and social identifications? In our approach the social self is considered not 
as a subjective state of identification with other people, but as a subsystem 
of self-knowledge. Prototypical elements of the social self constitute indi­
viduals' social identity (and, consequently, determine its content). On the 
other hand, social syntony and prosocial ego involvement are considered as 
states of social identifications, states of ego which depend on self-we-others 
schemata and thus on social identity. 


Self
-we-others schemata 
47 
Self-we-others schemata distinctness as a basis of identity 
Self-specific traits as personal identity attributes 
Our definition of identity is based on two groups of premises. The first 
reflects the traditional way of identity conceptualization: identity is con­
sidered as a result of recognition of the self (including language-mediated 
effects of 'talks with the self'; see Mead, 1934). Identity components are 
only these elements of the self which are explicitly indicated by a subject 
answering the question 'Who am I?' 
The other group of identity-definitional criteria refer to empirical data. 
According to the well-known studies by Markus (Markus, 1977; Markus 
& Smith, 1981), traits described by participants as (a) very important 
and (b) highly characteristic of the self are processed differently (for 
example, they are processed quicker). Markus labelled them schematic 
traits. 
2 
Our own studies on social comparison processes (Jarymowicz, 1991a) 
showed that traits of the self perceived as most clearly distinguishing the 
self from the others were processed differently (for example, egotism was 
found only for these traits of the self which were the most highly 
distinct). Our definition of identity refers to self-descriptive traits, traits 
which are not only characteristic of the self (and can be also charac­
teristic of the other), but also specific - distinguishing the self from the 
other. Such traits of the self, which are characteristic and distinguishing 
at the same time, we have labelled identity traits. 
In our definition (Jarymowicz, 1991a, p. 693), personal identity is a 
subsystem of self-knowledge, which is constituted by the traits perceived as 
the most characteristic of and the most specific (distinct) to the self. 
Self
-we-others schemata distinctness as a basis o
f
 personal versus 
social identity 
According to Byrne's studies (Byrne, 1969), interpersonal similarity leads to 
attraction. On the other hand, studies on identity have shown that, in some 
conditions, self/we-others similarity may be aversive and distinctness 
attractive (Brewer, 1991; Codol, 1979; Jarymowicz, 1982, 1987; Snyder & 
Fromkin, 1980). Seeking for explanations of the incoherent data concerning 
self/we-others similarity/distinctness and attractiveness, we found that the 
perceived dissimilarity was not a good measure of self/we distinctness from 
the other. People's estimation of the degree of similarity or dissimilarity is 
determined not only by the ability to differentiate between own person or 
own group and other people, but also by motivation to perceive self or we 
as similar or not similar to others (Suls & Miller, 1977). In order to create a 
better measure, we focused on the real cognitive distinctness of self/we­
others schemata, instead of the perceived similarity (depending often on 
subjects' wishful thinking). 


48 
Social identity: international pers
pectives 
The new operationalization was based on three procedural requirements. 
Individuals' attention on the self, the we or the other has to be focused 
without inducing social comparison processes (which are likely to 
activate social norms and evaluation). 
2 As was shown by Holyoak and Gordon (1983), the self is the usual 
reference point in social perception, and usually has stronger impact on 
perception of the other than does other on the self (see Hardoin & 
Codol, 1 984) - to focus attention on the self, the we and the others, a 
reverse order has to be applied. 
3 
To focus the subject's attention on descriptive and not evaluative 
dimensions of a particular object, nouns instead of adjectives have to be 
used (as the latter may help induce affective or evaluative associations). 
The so-called Questionnaire o
f
 Social Perception consists of a list of 70 
nouns, enumerating dimensions such as elegance, intelligence, tolerance (all 
presented in a positive manner). It contains three parts, presented to 
participants subsequently. The first part of the Questionnaire concerns the 
category of others. Participants read the list of dimensions and mark those 
which 'are on their mind when they think about other people'. Afterwards, 
out of the already indicated nouns, they mark l O  'which are considered the 
most frequent'. The second part examines the 'we' category. This category 
is first activated by an additional operation: participants are requested 'Try 
to think about those people who are on your mind when you use the word 
WE' and describe the chosen 'we' in an informal way. Then they read the 
same list of 70 nouns, and choose 10 dimensions 'most frequently used 
when thinking about people considered We'. The third part examines the 
self category: subjects choose 10 dimensions 'most frequently considered 
when thinking about oneself'. 
The three tasks refer to schematic dimensions (labelled analogous to 
Markus, 1977) of three objects. The distinctness of schemata is defined, 
however, by prototype, specific, distinctive dimensions, that is, only those 
dimensions which subjects ascribed to one and only one object. Measure­
ments built upon such material are likely to reflect the distinctness of self­
others, we-others and self-we schemata. 
Assuming that social identity may be derived from intergroup (or rather 
we-others) differentiation, in our approach the content of social identity is 
determined by distinctive dimensions of the we related to the others, that is, 
those dimensions which are marked by a subject as referring to the we and 
not marked as referring to others in general. 
Further on, assuming that personal identity may be derived from intra­
group (self-other members) differentiation (or distinguishing the self from 
similar ones; see Festinger, 1954), the content of personal identity is deter­
mined by distinctive dimensions of the self related to the we, that is, those 
dimensions which refer to the self and not to the we). 
We assume that the level of the distinctness of the self (from the we) and 
the we (from the others) is related to the stage of the identities' development 


Self
-we-others schemata 
49 
and their status, that is, their role in the regulation of individuals' func­
tioning: the higher the level of the distinctness, the greater the impact on the 
individual's functioning. 
Finally we posit (and present relevant data; larymowicz, 1 991b, 1993b): 
(a) social identity is likely to prevail in persons with a high level of we­
other schemata distinctness and a low level of self-we schemata 
distinctness; 
(b) personal identity is likely to prevail in persons with a high level of self­
we schemata distinctness and a low level of we-others schemata 
distinctness. 
By 'prevail' we mean both the higher level of development of a particular 
scheme and its relatively greater impact on information processing, evalu­
ation processes or behaviour. It is particularly clear when the social self and 
personal self do not overlap. This prediction was empirically verified and is 
presented elsewhere (Jarymowicz, 1993b, I 994a). 
Self-we schemata distinctness as a basis of social identifications 
The distinctness o
f
 the self as a precondition o
f
 understanding and 
acceptance o
f
 outgroups and dissimilar others: a theoretical model 
A number of theoretical arguments and empirical results contribute to the 
well-known thesis: similarity of an other to the self and the membership of 
the same social category facilitate identifications, and result in ingroup 
favouritism. In contrast, we focus on identifications with dissimilar others 
who do not belong to the same social group or category and study empirical 
findings concerning the relationships between self 
-we-others schemata 
distinctness and intergroup attitudes. 
The following predictions are proposed (larymowicz, I 992a, I 994c); 
The strong syntonic identification with one's own group leads to ingroup 
favouritism and, in certain conditions, to outgroup discrimination. 
2 
The lower the level of self-we schemata distinctness, the stronger the 
identification with one's own group and ingroup favouritism, and 
consequently the lower the ability to decentrate and to recognize the 
outgroup's perspective, and the lower the possibility of accepting others' 
goals and values. 
3 
The higher the level of self-we schemata distinctness, the weaker the 
influence of a reference group on individuals' functioning (including the 
tendency to ingroup favouritism and outgroup discrimination). 
We propose a theoretical model of the relationship between self-we 
schemata distinctness and possible identifications with dissimilar others. 
Two groups of premises are considered. 


50 
Social identity: international pers
pectives 
Degree of 
others-oriented 
involvement 
High 
Low 
Levels of self-we 
'--_----_e_----+-
distinctness 
Low 
Medium 
High 
0>----<0
 
Ability to recognize others' (outgroup members') states 
• 
• 
Motivation for engagement 
Figure 4.1 
Self
-we schemata distinctness and the degree o
f
 individuals' 
involvement with outgroups and dissimilar others: a theoretical model 
(a) instrumental (cognitive) premises - an ability to decentration and to 
recognize others' perspectives; 
(b) motivational premises - a readiness to focus on others' perspectives, to 
think what is beneficial to them, and to be involved. 
The model is presented in Figure 4. 1 .  
According to the model: 
The low level of self-we schemata distinctness relates to lack of ability 
to distinguish perspectives of one's own group from those of other 
groups, and to recognize others' states, needs, goals - which is a 
necessary precondition of others-oriented involvement. 
2 The high level of self-we schemata distinctness relates: 
(a) to an ability to distinguish perspectives of one's own group from 
those of other groups, and to recognize others' states, needs, goals 
- which is a necessary but not sufficient precondition of others­
oriented involvement; and 
(b) to different levels of pre 
social motivation (which depends on factors 
other than self-we distinctness) - thus, degree of others-oriented 
involvement varies in different situations. 
This model has been confronted with empirical data. Below some results 
are presented. 


Self
-we-others schemata 
6 .5 
(5.7) 
-0 
,g 5.0 /;:21 
-
eo 
o 
u 
l 
QJ 
QJ 
l 
.0 
(/) 3.5 
E E 
::> 
QJ 
(3 1) 
z :o  2.0 
0 
Ci 
L 
Low 
Medium 
High 
Self-we distinctness 
Figure 4.2 Recall o
f
 children's 
problems (Szuster-Zbro
jewicz, 
1 990) 
80 
(78 1) 
-
(/) 75  5J 
o 
(/) 
QJ 
QJ 
QJ .<: 70 
m -o  
Ol eo  
QJ 
QJ 
o 
l 
65 (68 9) 
L 
Low 
Medium 
High 
Self-we distinctness 
Figure 4.4 
Readiness to host 
diff
erent nationalities ( Cylwik, 
1 992) 
The model and empirical data 
35 
-
c 
o 0 
ill
 
'';::;
 
QJ 
eo 
l .Π
Ol +-'  
QJ 
0 
o E 
25 
(24 3) 
': " 
L 
(5 m 
Low 
Medium 
High 
Self-we distinctness 
Figure 4.3 
Motivation to 
solve children's problems 
(Szuster-Zbro
jewicz, 1990) 
90 
QJ 
U 
C 
o U 70 
c O  
6 V 50 
E $ 
<i   
30 
(81 5) 
L
(35 7) 
Low 
Medium 
(37 2) 
High 
Self-we distinctness 
Figure 4.5 
Tolerance o
f
 
German minority activity 
(Pawluczuk, 1992) 
5 1  
Participants of experiments were focused on others - presented either as 
anonymous 'others' or as 'they' (people of different social or national 
membership). In all groups self 
-we-others schemata distinctness was 
measured, as well as one of the following manifestations of others-oriented 
involvement: 
(a) recall of children's problems, mentioned prior to the experiment 
(Figure 4.2); 
(b) motivation to solve children's problems (Figure 4.3); 
(c) readiness to host at home students of different nationalities (Figure 
4.4); 
(d) tolerance of the activity of the German minority in Poland (to preserve 
their national tradition) (Figure 4.5); 
(e) positivity of evaluation of a painting, the artist being presented as a 
foreigner (Slovak, Irish or Japanese) (Figures 4.6-4.7). 
As can be seen, presented data are consistent with the model. 


52 
Social identity: international perspectives 
High 31 
(30.7) 
High 36 
(35.1) 
-
29 
-
33 
o ˆ 
o ˆ 
Q) S 
Q) 'S; 
‡ :-e 27 
(28 2) 
‡ :-e 30 
0l
Cf)
 
0l Cf)  
Q) 0 
Q) 0 
0
0.
 25 
o 
0. 27 
(25 8) 
(28 3) 
Low 
--( 
Low 
--((27 3) 
Low 
Medium 
High 
Low 
Medium 
High 
Self-we distinctness 
Self-we distinctness 
Figure 4.6 
Positivity o
f
 
evaluation o
f
 Slovak and Irish 
paintings (Sobczyk, 1992) 
Figure 4.7 
Positivity o
f
 
evaluation of Japanese painting 
(Sokolowska, 1992) 
The follow-up study established that participants with a low level of self­
we schemata distinctness were least likely to favour outgroups (similar 
results were obtained by Jarymowicz, 1994c; Kobuszewska, 1 989; 
Szuster-Zbrojewicz, 1988). The only inconsistency found was the motiva­
tional pattern presented in Figure 4.3 (Szuster-Zbrojewicz, 1 990). When 
subjects had to declare more than just good will, persons with a low level 
of the schemata distinctness were least likely to favour strangers. 
2 
The results of participants with a high level of the schemata distinctness 
are less homogeneous: in some studies a rectilinear rather than a 
curvilinear relationship is found. 
Conclusion 
There are a number of conclusions that can be drawn from this discussion. 
Let us outline them briefly. 
Egocentrism or ingroup orientation IS related to universal human 
properties. 
2 
The differentiation of cognitive schemata is a necessary precondition to 
go beyond own egocentrism: to recognize different social perspectives 
and understand others. 
3 
Social self and social identity (if related to social groups or categories -
see Table 4. 1) induce syntony and ingroup involvement. 
4 The necessary precondition of coexistence with outgroup members is an 
ability to distinguish the self from the we schemata: a formation of 
social and personal identity is necessary to enable the individual to 
switch focus and shift attention from one perspective to another. 
In summary, then, and coming back to the opening sentences of the 
chapter, we conclude by stating that we have tried to reconcile different 
traditions in the study of self in psychology. We think that from now on 
they should not be considered as separate. 


PART II 
ESTABLISHING GROUP 
IDENTITY 
5 
A Developmental View of the Search for 
Group Identity 
Stephen Worchel 
I spent several years of my life on the East Coast of the United States. 
Coming from Texas, which had only two types of weather, hot or rainy, I 
was awed by the changing seasons on the East coast. Each year, fall 
followed summer and signalled the coming of winter. I could consult my 
calendar to find the exact time of the metamorphosis, but nature was a 
better, but less predictable, source of information. Some years the change of 
season would be abrupt, easily noticeable and sudden when a winter storm 
would arrive and take control of my world for months. Other years, the 
changes of season were gradual, silent, almost secret; a few leaves might 
drop from a tree to whisper the advent of fall, but the full message took 
weeks. This four-act play would be performed every year. 
The drama of changing seasons has a message that carries far beyond the 
woodlands. The Navaho Indians believe that human behaviour is a mirror 
that reflects the forces of nature. If we wish to understand the behaviour of 
humans, we should look first to understand the natural environment. 
Indeed, there may be many messages about group and intergroup behaviour 
in the winds of seasonal change. 
I begin with this retreat into nature because it so sharply contrasts with 
the approach that characterizes much of the study of group and intergroup 
behaviour. Typically, social psychology searches for e
f
f
ects and the causes 
of these effects. Our methodological microscopes focus on a single behav­
iour, at a single point of time. The context in which that behaviour occurs 
(the person's previous actions, expectations about the present situation, or 
anticipations of the future) is seen as a pesky problem and given the 
ignominious title of 'error variance'. We speak of effects such as conformity, 


54 
Social identity: international pers
pectives 
group polarization, social loafing, cooperation, social categorization and 
intergroup competition as if they were stones in a behavioural river that 
exist in isolation. 
I want to offer a somewhat different perspective in this chapter. To begin, 
I want to question the concept of e
f
f
ect. Our 'effects' are not effects in the 
sense that we characterize them. Rather they are processes that can only be 
understood in the context of the preceding events and behaviours, the 
present interpretations of the setting, and the individual's anticipation of 
future activities. These processes are intertwined with the flow of behaviour 
over time and can best be understood and predicted in this context. They 
are the stream itself, not concrete forms that divert, interrupt and stand 
alone in the behavioural stream. This is not a new view. Lewin found that 
the effectiveness of leadership styles could be described only when one style 
was examined in light of the leadership style that had preceded it (Lewin, 
Lippitt & White, 1 939). Schachter (1951) found that reactions to a deviate 
not only changed over time, they were influenced by changes in the deviate's 
behaviour. If the deviate accepted the group position, he was quickly 
embraced by other group members. Continued disagreement with the group 
resulted in a cessation of communication directed towards him and 
rejection. Sherif and his colleagues (Sherif, Harvey, White, Hood & Sherif, 
1 961) found that hostility between groups could be reduced only after a 
series of cooperative encounters. And McGrath and his colleagues (Kelly, 
1 988; McGrath, 1 988) have lamented the neglect of time in the study of 
social psychology. 
I want to build on the foundation laid by these early studies and present 
group and intergroup behaviour as a procession of events, each related to 
the other, and each taking on meaning from the context in which it occurs. 
Groups, I believe, develop through rather predictable stages (or seasons). 
The process of group development is not one with a distinct beginning, and 
often there is no clear end. In fact, many groups 'live on' long after all the 
original members have departed. Rather, the development is a cycle that 
repeats itself. Each stage has a predictable impact on the behaviour of 
group members and on the relationship between groups. Actions that are 
accepted or encouraged at one point in the process will be censored and 
rejected at other points. Understanding and predicting group and inter­
group behaviour requires that we understand the developmental process. 
Independence versus group membership: a life-long dilemma 
Before jumping directly into issues of group and intergroup behaviour, let 
me be true to my psychology heritage and focus on the individual for a 
moment. Humanistic psychology argues that nearly from the moment of 
birth we embark on a life-long journey to 'find' our identity (Maslow, 1 954; 
Rogers, 1 95 1). The journey is a strenuous one, but at the heart of it is 
determining our relationship with the groups to which we belong. Defining 


A developmental view o
f
 the search f
or identity 
55 
this relationship throws us into two raging and competing cross-currents. 
On one hand is our desire to be accepted and valued by others. Groups 
offer us support, safety, protection and a reason for existence. While groups 
are a safe harbour in many ways, they also exact their price. They demand 
obedience and conformity, they require us to place group needs above 
personal needs, they coerce us to give up some of our personal identity. 
It is in this latter demand where the counter-currents meet with such fury, 
because paralleling the desire for group membership is the desire to be 
recognized as an independent, unique individual. In order to achieve this 
goal, we must differentiate ourselves from the groups to which we belong. 
We must resist their seductive lure to abandon our independent self and 
accept the group identity. Our lives are characterized by the constant 
struggle of deciding how much of our personal identity we will sacrifice for 
the substitute group identity. 
Tajfel and his colleagues (Tajfel, I 972b, 1 978d; Tajfel & Turner, 1986) 
anticipated this discussion in their development of social identity theory. 
They implied that we actually hold two identities (more correctly, one 
identity with two opposite poles), a personal identity that includes our 
personal characteristics and a social identity derived from our membership 
in groups. Turner (1982) argued that the identity (personal or social) which 
we stress at any moment is a function of the context in which a particular 
interaction takes place. Intimate interactions, for example, enhance the 
salience of our personal identity, while interacting with people from other 
identifiable groups may make our social identity salient (Doise, Deschamps 
& Meyer, 1978). For example, being the only woman in a group of men 
makes salient the social identity, especially of the minority person, and 
influences the nature of the interaction that will take place. Social identity 
theory further suggests that we wish to maintain a positive image of 
ourselves, and in order to enhance our social identity we will advantage our 
ingroup in relation to outgroups. 
Social identity theory offers unique insight into a host of social and 
individual behaviours. I would like to push the social identity approach a 
step further and suggest that the desire to emphasize one identity or the 
other is influenced by more than the social context in which people find 
themselves. I want to argue that the individual's own context (history, 
interpretation of the present situation, and anticipation of the future events) 
plays at least as important a role as does the social situations in determining 
the salience of personal/social identity. Further, an individual's life can be 
characterized as alternating between placing an emphasis on group identity 
and emphasizing personal differentiation. I accept the social identity posi­
tion that the social situation influences which identity is salient. However, I 
suggest that it is the individual's own context that motivates him or her to 
seek social situations that emphasize one identity or the other. For example, 
descriptions of close relationships suggest that individuals often alternate 
between periods of close intimacy with a partner followed by efforts to 
'touch base' with larger groups of people (Argyle, 1 992; Oldenquist, 1 992; 


56 
Social identity: international perspectives 
Reis, Nezlek & Wheeler, 1 980). Partners in intimate relationships are often 
threatened when they perceive the other 'pulling away' and becoming 
interested in the activities of social groups. However, the natural tide of 
relationships is a cycle of emphasizing the intimate relationship and 
emphasizing the membership in other groups. A threat to one aspect of the 
identity can lead a person to focus on and attempt to repair that identity. 
Individuals who have suffered threats to their social being, such as being 
fired from a job or going through a divorce, seek social support networks to 
reaffirm their social identity (Rook, 1 987). On the other hand, when the 
threat is to the personal identity, such as the case of experiencing an 
embarrassing event, the desire to be alone and repair the personal self is 
dominant (Morris et aI., 1976). Alternating between one identity and the 
other occurs not only in brief episodes during life. It also characterizes the 
larger pattern of human existence; greatest emphasis on establishing 
personal identity occurs during early childhood and middle age, while social 
identity themes are often most evident in adolescence and in the elderly. 
Finally, none of our behaviour escapes the embrace of culture, so it is not 
surprising that something so central as identity is a product of culture. 
Numerous investigators (for example, Hofstede, 1 980; Triandis, 1 988) have 
observed that cultures place varying emphasis on personal and social 
identities. In collective cultures such as those found in Asia, Greece and the 
Middle East, the emphasis is on the group and the individual's social 
identity is stressed. On the other hand, more individualistic cultures such as 
the United States and some countries of Western Europe force the 
individual into centre-stage, and personal identity plays a critical role. It is 
no accident that humanistic psychology with its emphasis on differentiating 
the self from the social context is a Western phenomenon. 
The development of groups 
This has been a rather long preamble to a discussion of groups and 
intergroup behaviour. However, I wanted to make the point that even at the 
level of personal and interpersonal behaviour there is an ongoing process of 
identification that involves emphasizing themes of inclusion and differentia­
tion at different points of time. The theme that is emphasized is a function 
of dynamics occurring within the individual and between individuals, as 
well as events that occur within the person's physical and social environ­
ment. My aim is now to demonstrate that similar processes characterize 
groups and the relationship between groups. My interest in group develop­
ment was initially sparked because so much of the research and theory in 
group and intergroup behaviour was not dynamic in approach, focusing 
only on specific behaviours at a single period, and ignoring the history of 
the members and the groups. I was concerned that most research on group 
dynamics began with the group already established, either in the sense of a 
laboratory group of strangers who arrived for an experiment or in an 


A developmental view o
f
 the search f
or identity 
57 
established group in a natural setting. This approach overlooks the process 
of group formation, how individuals decide to join groups, and how these 
variables affect group behaviour. These are important issues that may have 
dramatic impact on all group and intergroup behaviours. Unfortunately, 
these issues are beyond the scope of this chapter. However, although I will 
not address them, I want to raise them for future consideration. 
My second concern with traditional group research is that group devel­
opment and change is rarely addressed. Specifically, after reviewing the 
literature I was left with many questions about the dynamics of groups: for 
example, how do groups develop and change over time, and how does this 
development affect intermember and intergroup behaviour? In order to 
examine these questions, I turned to empirical accounts in social science 
literature on group dynamics and popular archival accounts of groups. My 
search of the archival literature included descriptions of small groups 
(families, athletic teams, splinter church groups, gangs), large groups 
(business organizations, political parties, nations) and social movements 
(women's rights movements, the civil rights movement, ethnic identity 
movements). I included a wide variety of groups because I wanted to 
identify both similarities and differences in the process involved in these 
groups. I was also guided by a personal interest in the processes involved in 
groups gaining independence from larger groups. This interest moved me to 
consider the area of group formation, as well as development, and carried 
me into the fascinating literature on revolution, protests and terrorism. I 
have discussed group formation and revolution in earlier papers (Worchel, 
Coutant-Sassic & Grossman, 1992), and space does not allow me to deal 
with this topic here. For the present, then, I will focus more directly on 
group behaviour from the time a group is initially formed. 
Actually, there has been a long history of efforts to examine group 
development. The t-group movement beginning in the early 1 950s offered a 
marvellous opportunity to observe groups over time and chart their devel­
opment and change. Based on data from these and similar groups, investi­
gators suggested that groups do move through rather predictable stages. 
For example, Tuckman (1 965; Tuckman & Jensen, 1977) identified the 
stages of forming, storming, norming, performing and adjourning. While 
this earlier work focused on the group, Moreland and Levine (1 982, 1988) 
examined the stages that individual members go through as they enter and 
move through the group. These studies of group and member development 
are rich and interesting, but it is difficult to determine where they have 
impacted group research. Rather than representing a coordinated effort, the 
studies and observations have been undertaken in isolation from each other 
so that it is not possible to identify a unified theory of group development. 
In addition, these fragmented observations of groups do not readily lend 
themselves to developing specific hypotheses. This is not necessarily an 
indictment of the approach, because much of the work was designed to 
describe group development, not generate research, but it does help explain 
why this approach has not been incorporated into mainstream research 


58 
Social identity: international pers
pectives 
Group identity 
Group productivity 
Individuation 
Note: During the decay period numerous members may leave the original group to join 
other groups or form new groups. While the old group re-enters the group identity 
stage, new groups formed by the spinning off of defectors will enter into their own 
stage of group identity. 
Figure 5. 1 
Stages o
f
 group development in existing groups 
efforts. Finally, many of these observations dealt with rather specific groups 
(t-groups, therapy groups), and the applications were rather group-specific. 
Most of the groups involved had specific beginning and end points, and 
they were formed for specific purposes. Therefore, the work rarely dealt 
with issues of group formation or examined cycles in development. Even the 
important work of Moreland and Levine follows the individual member 
through one group; there is no discussion of how the experience of 
membership in one group impacts the member's choice of additional groups 
or behaviour in these additional groups. In this sense, most earlier models 
are linear. As we will see, my observations suggested that group develop­
ment is cyclical and repetitive. This cyclical nature of groups has important 
implications for explaining intragroup behaviour and the continuing rela­
tions between groups. 
My work on group development used this earlier work as a starting point, 
and it was designed to deal with the limitations I perceived in these earlier 
approaches. As a beginning, I was able to identify four stages or periods 
that occur in intact groups (assuming group formation as the beginning 
point). The stages represented themes that dominated group activity during 
specific time periods. The stages are reciprocal in that they affect the 
behaviour of group members and the relation between groups, and they are, 
in turn, affected by member behaviour and intergroup relations. Beginning 
with the formation of the groups, the stages are: identification, group pro­
ductivity, individuation and decay (see Figure 5. 1). 
Identification 
Once a group is formed, its initial focus is on establishing its identity. 
Efforts are directed at defining the boundaries of the group, its 'place' in 


A developmental view o
f
 the search f
or identity 
59 
relation to other groups (and history), and the relationship between mem­
bers. This focus has profound effects on member behaviour and intergroup 
relations. The group establishes a central dogma or theme for the group. 
Attitudes on these issues are often extreme, clear and uncompromising. 
Minority positions are not tolerated and minorities are quickly excluded 
from the group. There are strong pressures on members to conform. Groups 
often adopt a dress code or uniform, a special language and other symbols 
that identify the group and mark people as being members. In many cases, 
efforts are made to 'reach back' in history to establish the legitimacy of the 
group, recognize patrons and previous founders, and locate the marks that 
will distinguish the group and identify its 'rightful' members. For example, 
the 1980s were a time of rebirth of Maori people in New Zealand. The 
Maori became especially interested in their rightful place in New Zealand 
history, their language, their customs and dress. There was a renewed push 
for recognition of the Maori within New Zealand society, a return of lands 
to the tribes, and greater powers for self-governance. 
Leadership during identification is often centralized. Members may be 
asked to demonstrate their loyalty to the group by making sacrifices or 
undergoing difficult initiation rites. The boundaries of the group are closed 
to new members. Recall that the initial stages of the doomsday cult studied 
by Festinger, Riecken and Schachter ( 1956) was characterized by 
withdrawing from public focus and closing the group to new members. In 
many cases, groups in a period of identification will physically withdraw, 
demanding that members break contact with other groups to which they 
have belonged. 
Relations with outgroups become especially strained at this point. Group 
boundaries are defined by the relation of one group to other groups. As a 
result the group either seeks to isolate itself from other groups or it invites 
conflict and competition with outgroups. Either of these actions defines 
boundaries and has a bonding effect within a group. Outgroups are 
presented as dangerous, aggressive and sinister, thereby forcing members 
together to deal with this outside menace. Similarities within the group 
membership and differences between groups are emphasized. In some cases, 
demands are made on outgroups which cannot be met, precisely to 'prove' 
that the outgroups are evil and unfair. For example, in many cases where 
ethnic groups attempt to 're-establish' themselves and incite member 
awareness, demands are made for the return of land or other resources that 
were taken from the group at an earlier period of history. Even when it is 
deemed that the resources were illegally seized, the wholesale return of these 
properties would create havoc in the larger society, so that the wholesale 
return is either impossible or will involve a slow, complicated process. The 
reluctance to return this property or the slow process is then used to 'prove' 
the lack of good will in the outgroup. 
From the group's perspective, the aim of the identification stage is to 
focus group members on their social identity, as opposed to their personal 
identity. Interactions within the group and between ingroup and outgroup 


60 
Social identity: international perspectives 
members emphasize social identity. At the same time that the group seeks to 
strip away the individuating characteristics of its members, it attempts to 
differentiate its own characteristics as a group, separate and different from 
other groups. 
Group productivity 
While efforts during the identification stage increase the cohesion within the 
group, they often interfere with group productivity. Therefore, as identity 
becomes established, group members begin to examine the goals of the 
group, especially ones involving group productivity. Although the orien­
tation is still on the ingroup, the topic of concern is productivity rather than 
identification. This focus creates numerous changes in the group and inter­
group behaviour. Overall, the emotional climate cools from the passionate 
state that characterized the identification process. Members now take a 
more analytical approach. During identification, group members were most 
concerned with the similarities of the group members. As the concern turns 
to productivity, members begin to examine the differences between them­
selves, but only those differences that may affect the group's task-related 
activities. The search is for the specific skills that members may possess that 
will help the group towards its productivity goals. Although still empha­
sizing social identity, the door is now opened for members to focus on their 
personal identity as it relates to group productivity. The boundaries of the 
group become more permeable and new members may be recruited to 
perform specific tasks. In some cases, the group may search for highly 
skilled members to fulfil high esteemed roles, but in many cases the new 
members will have a lower status and be expected to perform less attractive 
jobs, until they have 'proven themselves' worthy of group membership. This 
is often the fate of new immigrants to a country. 
The group becomes less antagonistic towards the outgroups, but its 
relationship with them remains cautious. In some cases, comparison with 
outgroups is sought in order to determine how 'productive' the ingroup is. 
Leadership becomes more task-oriented and the emotional pitch of the 
group is reduced. Minorities can influence the group on task-related issues, 
but these minorities will still meet quick rejection and censure if they 
threaten group identity. Group failures are attributed to external sources, 
while group success is attributed to internal group factors. 
Individuation 
During the productivity stage, members began to differentiate between 
themselves based on their task skills. The self-focus expands and accelerates 
during individuation. Members begin to assess their contributions to and 
rewards from group membership. They demand equity based on their 
contributions to the group rather than equality based on simple group 
membership. Subgroups appear during this stage, based first on skills and 


A developmental view o
f
 the search f
or identity 
61 
roles developed in the drive for productivity, but later expanding to include 
subgroups based on attitudes, interests and skills that may be unrelated to 
group activity. Individuals become more focused on personal needs and 
characteristics. Interactions within the group become more selective, but 
more intimate. Minority voice is not only tolerated, but encouraged. Social 
loafing (Latane, Williams & Harkins, 1979), in the sense of working less 
hard for group goals than personal goals, becomes common. Leadership 
may be questioned and failures blamed on poor leadership. 
Relations with the outgroup take on a very different flavour compared to 
the earlier two stages. Individuals begin a process of social comparison with 
outgroup members, as well as with ingroup members. They seek infor­
mation about out 
groups to use as standards for comparing their own group, 
and their treatment within the group. There is a tendency to overestimate 
the quality of life in the outgroup and the success of the outgroup. But this 
view is not necessarily used as a threat to the ingroup identity. Rather, it is 
presented as evidence that the ingroup is not meeting individual needs, and 
that changes are needed within the group. Rules, roles and norms of the 
ingroup are questioned, and there is a demand for greater personal freedom. 
Individual members begin negotiating with the ingroup to improve their 
personal standing. 
Decay 
From a group identity standpoint, the shift in focus to personal identity 
among group members poses a threat to group identity. There may be 
efforts to suppress this move towards individuality, but the group has 
become fragmented and the threat of rejection is less feared by members. 
The group becomes less important in the individual's self-concept. Indi­
vidual members actively explore defecting to other groups, or subgroups 
seek to break away from the group and form their new group. Marginal 
members who are not central to the group's identity are the first to leave. 
Later in this stage, more central members who have 'saleable' skills leave 
the group. These losses are viewed with alarm by the remaining members, 
who use these losses to demand change within the group. Although the call 
for change may result from a recognizable problem, members are often 
unclear about the specific change they desire. The result is that initial 
attempts to respond to these demands are often unorganized and ineffective. 
The easiest change is to replace the leader(s). There may be rapid succession 
of leaders as much is demanded from them, but members are unwilling to 
invest much power in leadership positions. Bitterness is expressed as 
members feel that they were lured into the group on the basis of false 
promises and expectations. Anticonformity as a way of demonstrating 
personal independence and displeasure will be common. 
Outgroups perceive the vulnerability of the group and efforts may be 
launched to encourage defection. Alarm over these incursions may be 
expressed, but the group is unable to mount a concerted defence. On a broad 


62 
Social identity: international pers
pectives 
scale, the decay of one group often creates conflict and hostility between 
outgroups as each attempts to profit from the condition. 
The cycle 
The stage of decay has a dramatic impact on the group. Many members 
leave the group and a negative emotional state predominates the group. But 
as the most disaffected members leave, the remaining members begin 
turning their attention back to the group. In the chaos, a strong leader often 
emerges, offering the remaining members the hope of re-establishing their 
group. The defections are seen as a group cleansing. Anger and hostility is 
expressed towards the defectors, and the group takes some solace in stories 
of disasters that befell these defectors. Attention is turned to rebuilding and 
redefining the group. The phoenix rises from the ashes. The group begins 
the identification stage once again. Often there is a surge of energy that 
comes from rebirth. The cleansing process may reach new heights as the 
group looks internally to define traitors or marginal members who have 
remained. These people are expelled, often with public ceremony. The 
fundamentalism that characterizes identification process grips the group. 
The reborn group begins differentiating itself from other outgroups, 
especially those that have defectors or profited form the demise of the 
group. Members' social identity is emphasized over personal identity. The 
cycle begins anew. 
The issue of inevitability 
This brief summary of the process of development raises as many questions 
as it answers. One of the first questions is whether this progression of stages 
is inevitable and must proceed uninterrupted in this order. The answer to 
this question is 'no'. Although this cycle may be the 'natural' flow of events, 
it is very clear that the process can be interrupted or short-circuited by 
numerous intended and unintended events. 
At any point during the process, events may occur that force the group to 
focus on specific issues and drive it to a particular stage. The most common 
situation is a return to group identification concerns. A clear or perceived 
threat from an outgroup, regardless of when it occurs in the development 
process, often makes group identity issues salient. An outgroup may make a 
bold effort to capture resources (land, material, status) held by a group. For 
example, it is not uncommon for employees of a company to rally to the aid 
of that company in the face of a rival's bid to buy that company. This can 
occur even when the employees are expressing dissatisfaction with their 
organization. 
A wily leader, feeling that his or her power is slipping or that the group is 
disintegrating, may seek out or manufacture a conflict with an outgroup. 
The result of this conflict is that the group will be thrown back into a stage 


A developmental view o
f
 the search f
or identity 
63 
of identification. For example, President George Bush was accused of 
accelerating the confrontation between the United States and Iraq in an 
effort to consolidate his power base and tum Americans' attention away 
from a deteriorating economy and onto the issue of national identity and 
pride. 
A sudden windfall or a dramatic increase in resources may push groups 
into a stage of individuation. Such a windfall requires that the group grapple 
with the problem of how to divide these resources among the members. 
Individual members may then begin to make their claim to these resources 
based on their skills, contributions to the group, personal needs or other 
personal characteristics. It is interesting to notice that the windfall, in this 
case, may hasten the decay of the group, in its present form. 
Finally, a scandal that arises from within the group may lead to disillu­
sionment and decay. Scapegoating and dissatisfaction with leadership can 
result in members feeling that the group cannot meet their needs, and result 
in their defection to other groups. 
While the progression through stages in order is not inevitable, I would 
like to offer two observations based on the reviews of group behaviour. 
First, it will take some unusual or unexpected event to derail the orderly 
progression. Second, when a group is pushed either back to an earlier stage 
or forward into a new stage, further group development will progress from 
that new stage. For example, if a group undergoing individuation is forced 
suddenly into the identification stage, the next step of that group will be into 
group productivity. The hands of the group clock move from the stage at 
hand, they don't 'catch up' or revert back to assume an earlier cycle of 
development. 
Transition and the nature of development stages 
A second major issue concerns the nature of the various stages and what 
forces the group from one stage into the next. These questions captured my 
interest as I read hundreds of accounts of groups. Was one stage necessarily 
longer than another? How did a group 'decide' that one stage was com­
pleted and it was time to move to the next stage? Were there clear 
distinctions between stages? 
I continually attempted to fit the literature into a concrete structure, but 
the fit was always an uncomfortable one. Like the variety found in the 
transitions between seasons, the changes in group stages followed a variety 
of patterns. In some cases, the transition was abrupt and clear, while in 
other cases it was more gradual, with the group dealing with two issues (for 
example, identification and productivity) simultaneously. Slowly it became 
clear that the stages were not necessarily distinct, bounded entities, but 
rather they were periods in which certain issues predominated the group's 
concern. Although central issues could be defined with considerable inter­
rater reliability, the fact that one issue was dominant did not mean that the 


64 
Social identity: international perspectives 
group avoided issues characteristic of other stages. Again, going back to the 
analogy of seasons, winter in most parts of the world is characterized by a 
preponderance of cold days. But even in the harshest winters, sunny warm 
days may appear for brief periods of time. Such was the nature of the group 
stages. The group discussion during a period would be ruled by concerns 
with a specific issue, but other concerns would also be addressed. However, 
introduction of these 'other' issues usually generated briefer discussions that 
often involved comparatively few members. Thus, a particular stage could 
be identified by the topic of discussion and behaviour that dominated the 
group and by the length o
f
 time s
pent on discussion and the number o
f
 group 
members involved with that issue. 
Similarly with the transition from one stage to another. It was often a 
matter of degree, but the change could be empirically identified by charting 
the topics of group discussion. Such, too, was the case with identifying the 
'trigger' for the transitions. There seemed to be no single event that moved 
the group from one stage to another. In some cases, a group member or 
members would declare that it was time the group moved on to another 
issue. This occurred both in small groups and in larger groups including 
social movements. What was interesting was that this 'trigger' propelled 
the group into another stage only when the group was ready to move. 
There were numerous examples in all sizes of groups of when a member, 
often the leader, requested that the group move on, only to be ignored by 
the group, which clung steadfastly to its present issue. In some groups, the 
trigger was a deadline or an imposed demand from the outside. Often 
members would indicate that a certain issue, generally productivity, must 
be addressed in order to meet a deadline; this moved the group out of 
identification concerns. Boredom was a frequent trigger. Several members 
might complain that the group was going nowhere or that it had already 
addressed certain issues and that they were ready for something new. 
Another sign of malaise and desire to move on was a lack of participation 
by members. 
Failure and success were some of the most interesting events in the groups I 
examined. Failure, regardless of the stage, tended to fixate members in that 
stage for a lengthy period of time. Like the classic Zeigarnik effect (Zeigarnik, 
1927), groups stayed with or quickly returned to issues that they failed to 
resolve. However, if these new efforts were not successful, groups tended to 
move into decay. 
Success was often as difficult to cope with as failure. Success spelled the 
end of a particular stage and moved the group, sometimes prematurely, into 
the next stage. For example, when workers at a computer company in 
Texas began organizing to protest decisions by management, they expected 
to meet stiff opposition. But when management quickly recognized and 
encouraged the worker organization, the group moved from concerns with 
identification to expressing their demands and seeking concessions (produc­
tivity). In this case, the identification phase of development was cut short 
and the workers failed to establish a clear organization. The result was that 


A developmental view o
f
 the search f
or identity 
65 
the union was disorganized and it soon disbanded, leaving many important 
worker/management issues unaddressed. Success and the lack of outgroup 
conflict resulted in the failure of the group to develop. 
From a research standpoint, these observations have serious implications. 
Determining the precise stage of development must be based on careful 
observations of group discussion and behaviour; in this sense, the deter­
minations must be post hoc. Groups will diff
er in the amount of time spent 
in each stage of development. Therefore, direct comparisons across groups 
using only time of existence as the independent variable will be difficult and 
must represent approximations. It is tempting to try to force a more con­
crete calendar onto the group development process. However, this would 
not adequately represent the process and such efforts would not help in 
understanding group development. While the fuzziness of the process is a 
limitation from an empirical sense, it is not fatal to research efforts, as we 
shall see. Predictions about the types of behaviour that should dominate a 
group or time period can be based on the model and these predictions can 
be tested. Analysis of group discussions and behaviours can identify stages 
and allow empirical tests of specific hypotheses. And the model can be used 
to explicate a wide range of intragroup and intergroup behaviours that 
cannot be easily explained by existing, more static approaches to group 
behaviour. 
Group identity: some empirical evidence 
Enough of this talk of 'fuzzy stages' and observation! Let's examine some 
empirical data. But before undertaking this adventure, one point should be 
made. The theory of social identity and its close cousin, self-categorization 
theory (Turner, 1 987), deal with the individual. They argue that individuals 
are driven to develop an identity, and, as a result of this desire, their behav­
iour influences group and intergroup actions. Individuals are the driving 
forces and groups are their vehicles. There is a great deal of support for this 
position, and I have no desire to attack it. However, my observations of 
groups suggest that the theories may be incomplete as they stand. For lurking 
in the background of the research on group development is the suggestion 
that groups, like individuals, strive to develop their identity, separate from 
that of other groups and apart from the unique identities of the individual 
members. 
Indeed, I would suggest that groups are engaged in a struggle with other 
groups and with individual members to establish and maintain a group 
identity. The group identity often outlives the members. It is displayed in 
symbols such as flags, anthems, laws, group structure and costumes that are 
enduring and not subject to the whims of individual members in search of 
their own identities. Groups strive for their independence from other 
groups, and they struggle with group members to keep the identity of the 
group equal to, if not more important than, individual identity. The striking 


66 
Social identity: international perspectives 
motto of the Three Musketeers drives home this point: 'All for one, one for 
all [the group].' 
Indeed, the model of group development may be viewed as a model of 
group identity. The stages represent the group's struggle to establish its 
character, which is threatened by demands from individual members and 
from outside groups. In fact, it is this conflict that ultimately gives the 
developmental process its energy, its dynamic. Where social identity theory 
would suggest that groups change and evolve to meet the demands and 
desires of their members, the group developmental approach argues that 
there is a more reciprocal relationship between the group and its member; 
the needs of individuals affect the development of the group, but the group, 
too, is often the puppeteer, guiding the actions of the members to meet the 
needs of the group. 
This approach argues that the behaviour of individual members is 
affected by the stage in which the group happens to be. Most pointedly, 
when developmental patterns or external events challenge the group to 
establish a clear identity, individual members will respond to support this 
goal. Specifically, this position predicts that in the early stages of group 
development, when the focus is on group identity, group members will 
direct both their intragroup and intergroup behaviours towards group 
identity. These forces stand apart from the need of the individuals to 
develop their personal or social identities. In essence, when the group is bent 
on crafting its identity, individual members are forced to focus on the social 
(group) side of their identities. 
During the early stages of group development, behaviours within the 
group and between groups should be directed towards securing the group's 
identity. On the intergroup side, there should be a rise in intergroup 
competition and conflict designed to create clear group boundaries. Within 
the group, individuals will be 'encouraged' to sacrifice their independence 
for the sake of creating a unified, clearly distinguishable group identity. 
Such actions as high conformity, swift rejection of deviants and perceptions 
of ingroup will foster this goal. Once the identity of the group is achieved, 
the group environment can become one that supports the goals of individual 
members to establish their own unique identities (social and/or personal). As 
a result, latter periods of group development should be characterized by 
fewer demands for personal sacrifice. Three studies, focusing on very 
different aspects of group behaviour, can be offered as evidence of this 
position. 
Preferred intergroup interaction 
Relations between the ingroup and outgroups are important and different at 
each stage of development, but these relations are most vital during the 
stage of identification. Nearly every aspect of group identity is defined in 
terms of the outgroups. As individuals, we may have difficulty defining our 


A developmental view o
f
 the search f
or identity 
67 
place in the universe, but at least we have our skin that defines our physical 
being. Groups, however, have no predetermined boundaries. Hence, not 
only are they faced with establishing their physical boundaries, but they 
must, in a sense, create their own skin. Indeed, this is often the first step in 
establishing the identity of the group. 
Establishing group boundaries not only involves determining who is in 
the group, but it also requires describing who is not a group member. One 
convenient means for making a clear designation of group boundaries is 
through intergroup conflict and competition. Such conflict plays multiple 
roles in the identification process: it forces individuals to declare their 
membership in one group or another; it marks the boundaries of the groups; 
it creates a common bond and purpose between members of the ingroup; 
and it forces the group to clarify what it stands for and what it stands 
against. Therefore, it can be predicted that group members will help their 
groups establish boundaries by seeking conflict and competition with out­
groups during the early identification stage of group development. 
On the other hand, the model of development predicts that intergroup 
conflict will be increasingly distasteful as the group continues its develop­
ment. During the group productivity phase, such conflict will distract 
members from task-oriented activities, although it may still serve the 
purpose of enhancing cohesion within the group. During individuation, 
members are concerned with making social comparisons with members of 
their own group and with members of the outgroup. In attempting to 
establish their worth to their ingroup and negotiating rewards, they must 
often demonstrate their potential value to other groups. Therefore, they 
need information from and acceptance by outgroups. Intergroup conflict 
would be disruptive to this purpose. And during decay, members are 
actively seeking alternative groups. Conflict between their present group 
and the alternative groups would clearly work against their individual goals. 
In an effort to examine these predictions from the model, groups of five 
students assembled to work together for three one-hour sessions. They were 
informed that other groups would also be working in similar settings during 
this period. The groups were allowed to choose from a variety of tasks on 
which to work and they could spend as long as they desired on each task. 
Their behaviour was video-taped during all the sessions and they responded 
to periodic questionnaires throughout. 
An initial analysis of the sessions indicated that the groups were quite 
similar in their pattern of development. During the early periods, time was 
spent getting acquainted, exploring similarities in backgrounds and 
interests, establishing norms and roles, discussing the uniqueness of their 
own groups, and agreeing to best the actions of other groups. As expected, 
the next order of business was on tasks: members examined and discussed 
the options, chose tasks and began talking about their past experience with 
similar tasks, their contributions to the group work, and their expectations 
for the next task. In several of the groups, members jokingly threatening to 
'send' productive members to the outgroups, and also in the form of jest, 


68 
Only 
cooperation 
t 
I 
Only 
Social identity: international perspectives 
competition '--
_
.L..-
_
--'--
_
-'-
_
---'--
_
----'-
_
----"-
_
---'
 
_
_
 '--
_
-'---
----l
 
Beginning 
of group 
Ending 
of group 
ā - - - - - - - - - - - - - - - - - - - - - - - - - - -
Time of group existence 
Figure 5.2 
Nature o
f
 desired interaction with outgroup as a f
unction o
f
 
group lif
e ( Worchel, Coutant-Sassic & Wong, 1993, p. 84) 
some members threatened to defect to the outgroups if their present group 
did not treat them well. During the last session, many of the groups engaged 
in a review of the group efforts, members talked about what they were 
going to do after the experiment, and closure was the topic of the period. 
Most interesting from the perspective of intergroup relations were 
responses on the written questionnaires. Several of these questions queried 
subjects about their feelings about their own group and desired relations 
with the other groups. As can be seen in Figure 5.2, subjects expressed a 
desire to engage in intergroup competition during the early periods of the 
group. As the group aged, their desires changed markedly, and by the time 
of group closing the desire was for cooperation with outgroups. Feelings 
about internal conflict showed the opposite pattern: extreme discomfort 
during the early stages and greater comfort by the final period (Figure 5.2). 
This is the pattern that would be predicted by the model on the assump­
tion that intergroup competition helps differentiate the ingroup from the 
outgroup, and such differentiation is most critical during early stages 
devoted to group formation. The topics of group discussion also fit the 
predicted pattern, focusing first on group identification and productivity 
and later on more individual concerns. It should be pointed out that the 
cyclical pattern of the stages was not observed, but this may have been due 
to the fact that these groups had a defined end point. 
These results are not only interesting from a theoretical standpoint, but 
they may also carry a message for applied group dynamics. If groups 
actively seek out competition and conflict with outgroups at certain periods, 
it follows that efforts to reduce intergroup hostility during these periods will 
not be well received. Reducing intergroup hostility may only be possible 
after groups have been able to establish their identity, and do not f
eel that 
interaction with the outgroup will threaten that identity. Sherif et al. (1961) 


A developmental view o
f
 the search f
or identity 
69 
found that intergroup hostility was reduced only after several events 
involving intergroup cooperation towards a superordinate goal. One expla­
nation for these results is that several cooperative encounters are necessary 
before hostility is reduced. Another explanation argues that the timing of 
the cooperative endeavours is the critical factor in ameliorating this 
hostility. Sherif's groups may have been most accepting of those final 
cooperative situations which occurred towards the end of the session. 
Ingroup and outgroup perceptions 
A number of studies have found that groups see greater homogeneity in the 
characteristics of the outgroup than within their own group (Jones, Wood & 
Quattrone, 198 1 ;  Linville & Jones, 1980; Park & Rothbart, 1 982). Explana­
tions for the effect include the fact that people have more and varied 
contact with ingroup members than with outgroup members. Therefore, 
they can readily see the variability of behaviour of ingroup members, but do 
not experience this variability in the limited contact with outgroup mem­
bers. Further, the contact with ingroup members generally occurs in a 
variety of situations, while contact with the outgroup is often in a single 
type of situation (Quattrone, 1 986). 
This interesting effect poses potential problems for the present model of 
group development. The model suggests that during periods of identification 
ingroup members are most interested in finding commonalities between 
their own group so as to create a group identity. However, later in the 
developmental cycle, especially during individuation and decay, the model 
suggests that members will be searching for differences between ingroup 
members, thereby enhancing their own individuality and giving them reason 
to search for other groups. The basic prediction, then, is that perceptions of 
homogeneity within groups will be dependent on the period of group 
development in which these perceptions are measured. Perceptions of 
ingroup homogeneity should be greatest during the identification stage and 
least intense during individuation and decay stages. 
In order to examine this prediction, groups of eight subjects worked on a 
series of tasks (Worchel, Coutant-Sassic & Grossman, 1992). They were 
then divided into two groups of four, and the two groups competed on 
other tasks. Subjects were asked to rate ingroup and outgroup members on 
a number of dimensions at different times during the competition (immedi­
ately at the start of competition, after one competitive encounter and after 
two competitive events). A final condition involved assigning subjects to 
two groups on arrival, never allowing them to work as a single group, and 
then measuring their perceptions after one competitive event. Based on the 
model of group development, it was predicted that members would perceive 
ingroups as more homogeneous than outgroups during the early phases of 
competition. This was the situation where groups should be most concerned 


70 
Social identity: international perspectives 
Table 5. 1 
Perceived dif
f
erences in ingroup and outgroup similarity 
Traits 
Condition 
Intelligence" 
Friendliness 
Appearance 
Combined-questioned 
immediately 
3.78b,c 
3.84 
1 .24 
Combined-questioned 
after one task 
2.97 
4.03 
0.97 
Combined-questioned 
after two tasks 
-3. 1 1  
-1 .66 
-0.67 
Never combined 
-2.69 
-2.07 
-2.00 
Talkativeness 
2.47 
1 .33 
-0.1 2  
-1 .25 
" Subjects were asked: How similar were the members of your group (outgroup) on the 
following dimensions: intelligence, friendliness, personal appearance and talkativeness. I = 
very dissimilar, 10 = very similar. 
b Scores were determined by ingroup-outgroup. Positive score indicates ingroup more similar; 
negative score indicates outgroup perceived as being more similar. 
c N = 6 groups in each condition. 
Source: Worchel, Coutant-Sassic & Grossman, 1992, p. 198. 
with identity, especially if the group had just been formed from a larger 
group, Later in the competition, there would be pressures to perceive the 
differences within the group as the members focused on task and member 
skills and contributions. 
As can be seen in Table 5 , 1 ,  the results supported these predictions and 
they illustrate the need to consider group development when examining 
group perceptions, Ingroups were perceived as more homogeneous than 
outgroups during the early stages of competition, but the ingroups were 
seen as more heterogeneous than outgroups during the latter stages of 
competition or when the groups had existed as separate units from the 
beginning of the experiment. It is also interesting that groups perceived 
greater absolute differences between the ingroup and outgroup during the 
early stages of competition than during later stages. 
Responses to minorities 
There is voluminous literature demonstrating that minorities can exercise 
considerable influence over group decisions (Moscovici, 1 976; Moscovici & 
Mugny, 1 983; Moscovici & Nemeth, 1974; Nemeth & Staw, 1 989), The 
focus of much of the research is on the most effective tactic for minorities to 
follow (Maass & Clark, 1 984) and on explaining the cognitive impact of the 
minority (Nemeth, 1 986). Surprisingly, the research has not examined many 
group factors in demonstrating the impact of the minority, In fact, minority 
influence work has been conducted outside of group settings altogether. The 
typical experiment examines how subjects respond to learning of a statis­
tically rare or unexpected position, While the minority influence paradigm 


A developmental view o
f
 the search f
or identity 
7 1  
had generated exciting data, it is important to bring the research into group 
settings. Indeed, minorities are defined by their relation to groups. 
The present model of group development suggests that the minority plays 
an important role in the group process, but that the impact of the minority 
is influenced by the stage of group development. On one hand, the voice of 
dissent represents a clear threat to group cohesion and identity. Minorities 
differentiate members within the group, and may present an image of a 
vulnerable or disorganized unit. On the other hand, minorities can offer 
valuable information and new perspectives to the group. Their unmolested 
existence signals that dissent, disagreement and dissatisfaction with the 
group are acceptable. Given these functions, it can be predicted that 
minorities will be more influential and accepted in the group during the 
stages of productivity, individuation and decay. On the other hand, 
minorities will be quickly rejected during the identification stage when the 
group is concerned with developing cohesion and uniformity. Furthermore, 
the influence of the minority will be dependent on the group's previous 
experience with minorities. If the experience has been that minorities are not 
disruptive and help move the group towards its goals, minorities will be 
more readily accepted than when the group's previous experience with 
minorities has been negative. 
My students and I have completed several studies examining minority 
influence. In one (y.I 
orchel, Grossman & Coutant-Sassic, 1993) we found 
that minorities were most influential in groups if they had previously 
demonstrated that their positions were correct and the issue of influence 
dealt with productivity, not group identity. Minorities who were perceived 
as a threat to group identity were excluded from the group; indeed, they 
were viewed as forming another outgroup. It was clear in this study that 
minorities could exercise greater influence on some issues than others and 
that the group focus or activity determined which issues were subject to 
influence. Previous work by Crano (1989) also found that minorities were 
more influential on some issues (objective) than others (subjective). This 
research extends these findings by suggesting that when examined in a group 
context the group will determine the issues on which minority opinion will 
be accepted. 
In a second study (Coutant, Worchel & Grossman, in preparation), 
groups of five subjects met together weekly over a period of six weeks. In 
each session, they were given tasks on which to work. At different points 
during the life of the group, various manipulations were executed; the main 
independent variable was the time at which the manipulation took place 
(during the first session, during middle sessions or during the last session). 
In the case of minority influence, the groups were given problems to discuss 
and required to arrive at a group decision about the solution. A confederate 
in the group played the role of the minority, and after ascertaining the 
modal opinion he proposed an extremely different position. The major 
dependent measure was the influence of the minority on the group decision 
at different periods of the group's life. 


72 
Social identity: international perspectives 
Table 5.2 
Minority in
fluences in groups 
over time 
Time 
Early 
Middle 
Late 
4.24 
4.67 
4.85 
Note: Responses represent the group decision 
on how to deal with a case at hand. Minority 
took a position of lO on a scale of 1 - 1 0, 
therefore higher numbers indicate closer 
agreement with the minority position. 
As can be seen from Table 5.2, the minority had little influence in the 
initial meeting of the group; in fact, the group moved away from his stated 
position. The minority was most influential when the influence attempt 
came during the closing group session. Experience with these ongoing 
groups indicated that the initial sessions were devoted to issues of group 
identity, while in the closing session individual members focused more on 
personal issues and the ending of the group. Hence, these results are in line 
with predictions based on the model of development: minorities will have 
little influence during identification stages and greatest influence during 
periods of individuation and decay. In addition to supporting the specific 
predictions taken from the model, this study and one finding similar results 
(Coutant-Sassic, 1 991) illustrate the need to consider the factor of group 
development when discussing a group 'effect' such as minority influence. 
We cannot speak of minority influence as a general effect; rather the influ­
ence is closely tied to the issue and to the development of the group. In fact, 
it might be interesting to speculate that the process of influence differs under 
the different conditions. The present view of minority influence (Nemeth, 
1986) is that it involves a single process: creating divergent thinking in the 
majority. I would suggest that during the stage of identification when 
the minority is viewed as a threat to group identity convergent thinking 
on the majority position results from the presence of a minority point of 
view. 
The task ahead 
It is time to take stock of where we now stand. In a general sense, the 
present model can be viewed as a model of group social identity. Groups, 
like individuals, must establish their own identity. This identity has two 
dimensions: the identity of the group (similar to the personal identity of 
the individual in social identity theory) and the identity of the group as it 
relates to other groups in its universe (similar to the social identity of 
individuals). The search for identity influences both ingroup and intergroup 


A developmental view o
f
 the search f
or identity 
73 
behaviour as one identity or the other predominates. By way of summary, I 
have attempted to establish a number of points concerning group and 
intergroup behaviour. First, individuals are engaged in a constant struggle 
of balancing their desire to belong to groups and the concern of establishing 
their personal independence and uniqueness. Events or situations that 
threaten their social identity will lead individuals to emphasize their rela­
tionship with their groups. Likewise, threats to their personal identity will 
motivate individuals to seek more intimate interactions, making their 
personal characteristics more salient. This struggle is felt at the group level 
by members alternating between being active participants in group activities 
and withdrawing themselves, physically and/or psychologically, from the 
group. 
Second, groups are dynamic units that develop through a predictable 
series of stages. The development process is cyclical in that it can repeat 
itself many times during the life of the group. The repetition occurs as the 
group loses old members and gains new ones. The amount of time the 
group spends in any single stage is dependent on numerous intragroup and 
intergroup factors. The transition between stages can be abrupt and clear or 
slow and gradual. Numerous events can trigger the movement from one 
stage to another, including success, failure, leader behaviour, time pressures 
and deadlines, and intergroup events. 
Third, stages can be identified by the issue that dominates group behav­
iour and discussion and involves the most group members. Stages not only 
influence the focus of group activity, they also impact the interpersonal 
behaviours that will characterize the group. For example, minority influence 
will be most evident during individuation and decay stages, while con­
formity will be most common during the identification stage. Relating 
developmental stages to individual and interpersonal behaviour offers fertile 
grounds for developing specific predictions. 
Fourth, the development of the group is keenly affected by intergroup 
relations, and these intergroup relations are also influenced by the group's 
development. For example, the group will be most concerned with diff
er­
entiating itself from outgroups during the identification period. This 
differentiation process will be seen in efforts by the group to instigate 
conflict and competition with the outgroup. On the other hand, perceived 
threats to the group's identity from the outgroup may throw the group into 
the identification stage, regardless of when these threats occur. 
Finally, the developmental approach to group behaviour offers insights 
that may be used in intervention strategies. Those strategies that are 
congruent with the group's stage of development will be most effective. For 
example, it was suggested that prejudice and discrimination could be most 
effectively reduced by interventions during the individuation or decay stages 
of development. 
These are just a few of the advantages that result from expanding the 
scope of social identity theory. The theory has a solid base, and it is time to 
explore pushing its boundaries. 


74 
Social identity: international perspectives 
Note 
Work on this chapter was supported in part by a research grant from the Texas Coordinating 
Board of Higher Education (T 
ARP). The chapter was completed while I was a Visiting 
Lecturer at the Department of Psychology, University of Waikato, Hamilton, New Zealand. I 
wish to express my appreciation for the support I received from my colleagues at the University 
of Waikato. I also want to acknowledge the help of several students who have worked with me 
on developing the ideas presented in this chapter: Dawna Coutant, William Webb, Frankie 
Wong, Michele Grossman, Judy Ouellete, Greg Pool and Leah Worchel. 


6 
The Role of Prototypicality in Group 
Influence and Cohesion: Contextual 
Variation in the Graded Structure of 
Social Categories 
Penelope Oakes, S. A lexander Haslam and John C. Turner 
The concept of prototypicality 
The concept of proto 
typicality was one product of the 'Roschian revolution' 
(Neisser, 1 987, p. vii) of the 1 970s, which introduced some key new ideas 
to the analysis of category structure and representation. As is now well 
known, Rosch challenged the philosophically based, classical view of 
category representation with her ideas about, and evidence of, extensive 
variation in category structure. She and her colleagues (for example, Rosch, 
1978; Rosch, Mervis, Gray, Johnson & Boyes-Braem, 1 976) found that 
comparisons between categories reveal that they vary in their relative 
inclusiveness, for example 'dalmation', 'dog', 'animal', 'living thing', are 
categories of increasing inclusiveness. Those at a higher level are called 
superordinate, compared to lower level categories known as subordinate. 
Around the middle of the hierarchy of inclusiveness is the basic level (for 
example, 'dog', 'chair', 'tree') at which, it is argued, most perceptual activity 
takes place (Mervis & Rosch, 1 98 1 ;  see also Lassaline, Wisniewski & 
Medin, 1 992). More importantly for our present purposes, Rosch also 
found that within categories, members vary in their typicality, for example 
robins are seen (by American subjects) as more typical of the category 'bird' 
than are ostriches, but both robins and ostriches are recognized as sharing 
membership in the one 'bird' category. In other words, it appears that 
categories have an internally graded structure, rather than members possess­
ing an even level of shared defining attributes, as in the classical view. It was 
this finding that led to the idea of a category prototype (a best example of 
the category), and the argument that category membership requires a 
certain level of similarity to the prototype. Rather than being tightly 
bounded, categories seem to take the form of 'fuzzy sets', much looser 
groupings which depend upon what Rosch called 'family resemblance', that 
is, members do not all share a given set of defining features, but are related 
through their similarity to a prototype. 


76 
Social identity: international perspectives 
In her most widely cited paper Rosch (1978) tried to counter some 
possible misunderstandings of her ideas. In particular, she emphasized that, 
although her work indicated that prototypes served important functions in 
categorization, the concept did not 'constitute any particular model of 
processes, representations, or learning' (p. 40). Indeed, she commented: 
To speak of a prototype at all is simply a convenient grammatical fiction; what is 
really referred to are judgements of degree of prototypicality 
. . . .  
to speak of a 
single entity that is the prototype is either a gross misunderstanding of the 
empirical data or a covert theory of mental representation 
. . . .  
In short, proto­
types only constrain but do not specify representation and process models. 
(pp. 40-4l )  
Prototypicality in self-categorization theory 
Self-categorization theory (Turner, 1985; Turner, Hogg, Oakes, Reicher & 
Wetherell, 1987) is often introduced as a theory of processes underlying 
psychological group formation, and indeed its original mission was to 
explain the psychological basis of the group. In effect, however, it attempts 
more than this, and one of its most important contributions has been the 
development of a new perspective on the functioning of categorization 
processes in social perception and interaction (cf. Brewer, 1988; Fiske & 
Neuberg, 1990). We shall not attempt to provide a comprehensive overview 
of the theory here, but it is worth setting down some of the main points 
concerning categorization in order to place our discussion of prototypicality 
in its proper context (for more detailed discussion of categorization, see 
Oakes, 1996). 
In discussing the role of categorization in social perception, the theory 
applies and develops Rosch's ideas about variation in both relative 
inclusiveness across categories and relative prototypicality within categories. 
The theory begins with the assumption that self-conception reflects self­
categorization, the cognitive grouping of the self as identical to some class of 
stimuli in contrast to some other class of stimuli. Following Rosch (1978), 
self-categories are assumed to exist at different levels of abstraction related 
by class inclusion. That is, a given self-category (for example, 'scientist') is 
seen as more abstract than another (for example, 'biologist') to the extent 
that it can contain the other, but the other cannot contain it: all biologists are 
scientists, but not all scientists are biologists. Going beyond the simple 
distinction between personal identity (self as a unique individual) and social 
identity (self as a member of a social group) presented in earlier work (for 
example, see Turner & Giles, 1981), self-categories both more and less 
abstract than personal and social identity are envisaged, but for purposes of 
theoretical exposition three levels of abstraction of self-categories are 
distinguished: the interpersonal (subordinate level of abstraction, personal 
identity, self as an individual person), intergroup (intermediate level of 


The role o
f
 pro 
to 
typicality 
77 
abstraction, social identity, self as a group member) and interspecies (super­
ordinate level of abstraction, self as a human being). These are defined not by 
specific attributes but by the level at which people are being compared and 
categorized. For instance, 'altruism' could function as a cue to individual 
identity, to a particular social category, or to being human, depending on the 
context. 
In terms of the processes underlying group formation, the central 
explanatory idea in the theory is that group behaviour is made possible by 
and reflects self-concept functioning at the level of ingroup-outgroup (rather 
than interpersonal or interspecies) categorizations. Developing TajfeJ's 
(1978d) idea of an interpersonal-intergroup continuum, self-categorization 
theory contrasts the social and personal (ingroup-outgroup and inter­
personal) levels of self-categorization, and hypothesizes that shifts between 
these levels are a varying outcome of a continual conflict between personal 
and social identity as they vary in relative salience. Shifts towards social 
identity produce depersonalization of self-perception and behaviour, that is, 
self-stereotyping, perception of increased identity between the self and 
ingroup members and increased difference from outgroup members (on 
relevant dimensions). The theory proposes that it is this process of 
depersonalization that makes group behaviour possible and produces its 
emergent, irreducible properties. 
In retaining (from social identity theory) this emphasis on the two levels 
of personal and social identity, the theory appears to be suggesting that 
there are, in Rosch's terms, two 'basic' levels in the cognitive representation 
of the self and others, the social and the personal. In truth, however, this 
distinction is maintained only for purposes of theoretical clarity in 
explaining the hypothesized basis of psychological group membership. One 
of the essential points of the self-categorization analysis is that the appro­
priate level o
f
 categorization varies with the context, and that to speak of any 
level as more 'basic' than another is misconceived. 
To develop this point further, the theory portrays categorization as a 
dynamic, context-dependent process, determined (in part) by comparative 
relations within a given context (for example, see Haslam, Oakes, Turner & 
McGarty, 1 996). This approach is formalized in the principle of meta­
contrast, which is so called because it involves a contrast between contrasts, a 
judgement of difference between differences. The meta-contrast principle 
predicts that a given set of items is more likely to be categorized as a single 
entity to the degree that dif
f
erences within that set o
f
 items are less than 
dif
f
erences between that set and others within the comparative context (cf. 
Campbell, 1 958; Rosch, 1978; Taj fel 
, 1969a; Tversky, 1 977). This principle 
encapsulates but subtly transforms the classic idea that categories form on 
the basis of intraclass similarities and interclass differences. It goes further 
than earlier work in making clear that such similarities and differences are 
not independent and additive, but aspects of the same meta-contrast (cf. 
Medin, 1 989; see below). The meta-contrast principle is explicit that 
categorization is relative to the frame of reference and hence to the contrasts 


78 
Social identity: international pers
pectives 
Figure 6.1 
A categoria! hierarchy o
f
 vegetation 
available within the salient stimulus field. It depends not just on similarities 
between stimuli (as is so often assumed), but on relative similarities, more 
similarity between certain stimuli than between others. Meta-contrast 
contextualizes categorization by tying it to an on-the-spot judgement of these 
relative similarities and differences. 
For example, consider the system of natural categories in Figure 6. 1 .  
Meta-contrast predicts that we would categorize and perceive a given piece 
of vegetation as, say, a 'tree' (rather than as 'vegetation') to the extent that, 
in the current comparative context, the differences between trees (oaks, 
birches, and so on) are perceived to be less than the differences between 
'trees' and 'shrubs' - the distinction between trees and shrubs is more 
marked, and more relevant, than are the features that trees and shrubs 
share as 'vegetation'. Alternatively, the tree might be categorized and per­
ceived simply as 'vegetation' to the extent that differences between types of 
vegetation (trees, shrubs, and so on) are seen to be less than the differences 
between vegetation and, say, animals. The salient categorization will be 
'oak' when perceived differences between individual oak trees are less than 
the differences between oaks and birches (or some other comparison 
species). 
The meta-contrast principle applies similarly in person perception. 
Consider the hierarchy in Figure 6.2. Here, we might categorize an 
individual as 'Spanish' to the extent that, in the current comparative 
context, perceived differences between individual Spaniards (Carmen, 
Isabella, and so on) are less than those between Spaniards and Germans. 
Alternatively, the salient category might be 'European' in a context 
where perceived differences between various European groups (such as 
Spaniards and Germans) are less than those between Europeans and non­
Europeans. 
These ideas can be expressed in terms of a meta-contrast ratio, that is, the 
average perceived intercategory difference divided by the average perceived 
intra-category difference. For example, this might be: 


The role o
f
 pro 
to 
typicality 
Figure 6.2 A categorial hierarchy o
f
 Europeans 
average perceived difference between Spaniards and Germans 
average perceived difference between individual Spaniards 
79 
The higher this ratio, the more likely it is that Spaniards will be perceived in 
terms of their shared national identity. Turner and Oakes (1989, p. 242) 
provide specific examples of how meta-contrast can be calculated, as do 
Haslam and Turner (1 992) and some of the papers exploring prototypicality 
in more detail that we shall discuss below (e.g. McGarty, Turner, Hogg, 
David & Wetherell, 1 992). 
Note, however, that meta-contrast provides only a partial account of 
categorization. It describes the comparative relations between stimuli which 
lead them to be represented by a category, but it is also important to take 
into account the social meaning of differences between people in terms of 
the normative and behavioural content of their actions, and the relative 
accessibility of particular categorizations (Oakes, 1987; Oakes, Turner & 
Haslam, 1 991). In general terms, the theory explains the salience of any 
given category as a function of an interaction between its relative accessi­
bility (the 'readiness' of a perceiver to use a particular categorization) and 
the 'fit' between the category and reality. Meta-contrast is the theory's 
principle of 'comparative fit' (the match between category and the com­
parative properties of stimuli), but 'normative fit' (the match between 
category and the content properties of stimuli) is also always inseparably at 
work (Oakes et aI., 1991). 
This emphasis on categorization as highly variable and context-dependent 
produces a concomitant emphasis on the context-dependence of perceived 
similarity and difference, the major outcome of categorization. People who 
are categorized and perceived as different in one context (for example, 
'biologists' and 'physicists' within a science faculty) can be recategorized 
and perceived as similar in another context (for example, as 'scientists' 
rather than 'social scientists' within a university) without any actual change 


80 
Social identity: international perspectives 
in their own positions (Gaertner, Mann, Murrell & Dovidio, 1989). This is 
the essence of categorization: it is a cognitive grouping process that trans­
fonns differences into similarities and vice versa. We need some psycho­
logically neutral tenn such as perhaps 'distances' to indicate precognized 
stimulus relations: there are distances between people, but are they 
similarities or differences? Are physicists and biologists similar or different? 
Arising from the comparisons specified in the meta-contrast principle, 
categorization subjectively transfonns 'distances' into similarities and differ­
ences, and from perceived similarities and differences flow, amongst other 
things, perceptions of attraction and dislike, agreement and disagreement, 
cooperation and conflict. In sum, categorization provides the fundamental 
basis of our social orientation towards others. Within the science faculty, 
physicists might reject and deride biologists, claiming they aren't 'real 
scientists', but in comparison with social scientists the two groups may 
present as inseparable allies. 
As changes in the comparative context produce changes in the ingroup­
outgroup relationships, they also affect intra-category structure, and this is 
where the concept of prototypicality becomes important. The meta-contrast 
principle can also be used to define the relative prototypicality of members 
within a group (see Turner & Oakes, 1989, pp. 259-265). In general tenns, 
the more a group member differs from outgroup members and the less he or 
she differs from other ingroup members (that is, the more this person 
exemplifies what in 
group members share and what they do not share with 
the outgroup), the more that individual will be perceived as prototypical of 
the group. So, for example, in a communist party, members wishing to be 
perceived as prototypical communists must differ in politics from members 
of more conservative, capitalist parties, but they must not be so ultra-left 
that they begin to differ significantly from other communists - prototypi­
cality depends upon both inter- and intragroup comparisons. One important 
point is that, because relative prototypicality depends on (amongst other 
things) intergroup comparisons, it will vary along with variation in the 
intergroup context in which judgements are made. For example, the proto­
typical communist in a comparative context including fascists will be 
dif
f
erent from the prototypical communist as compared with liberal 
democrats, or Trotskyists. Self-categorization theory thus agrees with Rosch 
(1978) that fixed prototypes are 'fictions', and emphasizes context­
dependent judgements o
f
 proto 
typicality rather than fixed prototypical 
images which represent groups as constants across changing contexts (cf. 
Brewer, 1988; Brewer, Dull & Lui, 198 1 ;  see below). 
There are further aspects of the analysis of categorization presented within 
the theory (for example, the issue of the relationship between the different 
levels of abstraction, the dependence of comparison and categorization upon 
identity at a higher level of abstraction), but these issues are not central to 
our current discussion and are covered elsewhere (for example, Oakes, 
Haslam & Turner, 1994; Turner & Oakes, 1989). The major point to 
emphasize here is the context-dependence and fundamentally comparative 


The role o
f
 prototypicality 
81 
nature of 
the categorization process, and the consequent variability of 
both 
the level and the content (including the prototypical position) of social 
categories. 
Prototypicality and group polarization 
Group polarization refers to the tendency for group interaction (for 
example, discussion or some related activity) to move the average of group 
members' responses closer to the extreme of an already favoured position. 
For example, after discussing feminism a group of moderately pro-feminist 
individuals would tend to become, on average, more pro-f
eminist than they 
had been initially. There is a large and complex literature related to group 
polarization (Turner, 1 991), and our aim here is not to review that 
literature, nor to enter into any debate about the relative merits of different 
explanations of the phenomenon, but simply to show how the concept of 
prototypicality as conceptualized in self-categorization theory has enabled 
resolution of the apparent contradiction between polarization and other 
social influence phenomena (for more detail see Turner, 1991 ; Turner & 
Oakes, 1989; Wetherell, 1987). 
Group polarization has been seen as inconsistent with a central tenet of 
the traditional informational influence model of conformity derived from 
the findings of Allport, Sherif and Asch - the idea that social norms form 
through a process of interpersonal averaging or convergence. Norms are 
thought of as reflecting what the group shares, and this has been concep­
tualized in terms of a simple conglomeration of what each separate indi­
vidual brings to the group, specifically the average of individual members' 
separate views. Given this view of norms, specialized theories have had to 
be developed to explain polarization as something more than or different 
from conformity to group norms (Burnstein, 1982; Sanders & Baron, 1 977; 
Wetherell, 1 987), because in polarization group members do not converge 
on the average of their individual positions. Something occurring within the 
group moves them towards a position more extreme than the average. 
It has long been evident that polarization could be explained as conformity 
if it could be shown how ingroup norms could sometimes be more extreme 
than the mean (for example, Singleton, 1979). To do this, one needs to 
rethink the interpersonal averaging approach to the development of group 
norms, and there are additional reasons for doing this. The interpersonal 
averaging approach is individualistic (Turner & Oakes, 1986), implying that 
social influence is simply a 'change in individuals induced by individuals' 
(Kiesler & Kiesler, 1969, p. 26). In contrast, self-categorization theory argues 
that group behaviour (including social influence) reflects and is made 
possible by a change in the level of abstraction of self-categorization, a 
qualitative shift in both self-conception and the nature of relations with 
others. In this sense, group interaction does not involve individuals whose 
idiosyncratic opinions can be averaged, but group members whose behaviour 


82 
Social identity: international pers
pectives 
will be guided by the distinctive, emergent, irreducible properties of their 
group. The group norm embodies those group properties, and reflects both 
what group members have in common (on dimensions relevant to the group) 
and the way in which the group can be distinguished from relevant 
outgroups. Group norms (and other distinctive group properties such as 
stereotypes) reflect the comparative relations within which the group defines 
itself as well as, and as much as, they reflect attributes of group members. 
More specifically, the position perceived as normative for a group in a 
given context will be the position that is most prototypical in that context, 
as defined by the principle of meta-contrast (see above). It will be the 
position which is, on average, most diff
erent from that of relevant outgroup 
members and least diff
erent from that of other ingroup members (obtained 
by dividing the individual's average difference from outgroup members by 
his or her average difference from ingroup members). Of course, the posi­
tion least different from that of other ingroup members is the mean, but the 
mean may not be the position most different from the outgroup. Thus, 
under certain intergroup conditions (see below) the prototypical position 
will be skewed away from the mean. 
The extent to which the mean does not represent the position most 
different from the outgroup, and hence the degree of discrepancy between 
the most prototypical position and the mean, depends on the comparative 
context within which the ingroup defines itself. In general, the ingroup 
prototype will tend to coincide with the mean where the mean is at the 
midpoint of the comparative context, that is, the ingroup is occupying the 
'middle ground' on some issue relative to relevant outgroups. When ingroup 
responses occupy a more extreme position in the comparative context (as a 
function of either ingroup responses shifting towards an extreme, or a 
change in intergroup comparison making moderate responses appear more 
extreme), it becomes more likely that the most prototypical position will 
tend to be more extreme than the mean. For example, insofar as a group of 
moderately left-wing students is seen to occupy the middle ground in 
student politics within their institution (that is, there are other salient groups 
to both the left and the right of the moderates), the average view in the 
group will be more prototypical. However, if either individual opinions in 
the group shift to the left, or the comparative context comes to be domi­
nated by, say, fascists and other extreme conservatives, a more extremely 
left-wing position within the moderate group will become prototypical. The 
average position would no longer adequately represent the difference of the 
moderates from the right-wingers. Examples specifying this process of 
calculating the prototypical position (and its variation from the mean 
position) through the principle of meta-contrast can be found in McGarty et 
al. (1992), and a number of other sources (for example, Turner et aI., 1987). 
Several studies provide evidence consistent with the self-categorization 
analysis of group polarization (for example, Abrams & Hogg, 1990; Abrams, 
Wetherell, Cochrane, Hogg & Turner, 1 990; Hogg, Turner & Davidson, 
1 990; Mackie & Cooper, 1 984; McGarty et aI., 1992; Spears, Lea & Lee, 


The role o
f
 pro 
to 
typicality 
83 
1990; Turner, Wetherell & Hogg, 1989; van Knippenberg & Wilke, 1 988). 
For example, McGarty et al. (1 992) present correlational evidence (from a 
computer simulation and two semi-naturalistic group discussion studies) that 
ingroup prototypes do become more polarized as the group becomes more 
extreme in the social context. They also show that the magnitude of group 
polarization is a function of the degree to which groups actually shift 
towards their prototypes together with the degree to which those prototypes 
are polarized. 
One novel prediction from the self-categorization analysis is that the 
direction of polarization can be reversed by changing the comparative 
context, for example moderate left-wingers should move to the left in a 
context dominated by the right wing, but to the right when comparing 
themselves with ultra-leftists. Consistent with this prediction, Hogg et al. 
(1 990) report findings from a decision-making study in which subjects 
perceived the ingroup norm as riskier than the pretest mean when con­
fronted by a more cautious outgroup, but as more cautious than the pretest 
mean when confronted by a riskier outgroup. 
In summary, self-categorization theory explains group polarization as 
conformity to an extremitized group norm, and explains the extremitization 
of the group norm through an account of the way in which comparative 
context influences norm formation. The concept of proto 
typicality has been 
important here in specifying the position that will be most normative for a 
given group in a given comparative context: the position with the highest 
meta-contrast (maximal similarity to ingroup positions and maximal 
difference from outgroup positions) is the most prototypical of the group 
and hence the most consensual and normative. 
Prototypicality and social attraction 
Prior to the 'cognitive redefinition' (Turner, 1982) of the social group 
represented by social identity theory and self-categorization theory, the 
sense of group belongingness and attraction within groups was understood 
to rest simply on inter
personal attraction between individual group mem­
bers. This was the essence of the traditional concept of group cohesiveness 
(Lott & Lott, 1 965), the affective and motivational 'glue' that held groups 
together. However, in the same way that the interpersonal averaging 
conception of group norms has been reconsidered in self-categorization 
theory, the interpersonal attraction account of group cohesiveness has also 
been challenged (Hogg, 1 987, 1 992, 1 993). It has long been evident that 
interpersonal attraction is neither necessary nor sufficient for group 
behaviour (for example, Hogg & Turner, 1985a; Turner, Hogg, Oakes & 
Smith, 1 984; Turner, Sachdev & Hogg, 1 983), but that, at the same time, 
there are strong bonds of attraction within groups (for example, Sherif, 
1967), including 'minimal' groups (for example, Brewer & Silver, 1 978; 
Brown & Turner, 1979; Doise, Csepeli, Dann, Gouge, Larsen & Ostell, 


84 
Social identity: international perspectives 
1972). Self-categorization theory has introduced the idea of depersonalized 
social attraction to distinguish between interpersonal liking (which seems to 
be unrelated to group processes) and the intragroup liking that accompanies 
salient group membership and group interaction. 
Hogg (1992) summarizes the self-categorization analysis of intragroup 
attraction thus: 
the core idea [is] that self-categorization depersonalizes the basis of interindividual 
attraction, such that ingroup individuals (and self) are liked in proportion to their 
perceived group prototypicality. (p. 125) 
Again, we have an emphasis on the qualitative distinction between group 
and individual behaviour. Group cohesiveness and intragroup attraction are 
based on group members' liking of each other as group members, not as 
unique individuals, and it is prototypicality that defines the extent to which 
individuals are or are not group members. Recall that prototypicality 
depends upon both similarity to other ingroup members and difference from 
the outgroup. The closer any individual group member comes to this ideal, 
maximally normative ingroup position, the more he or she will be liked by 
other ingroup members, and it is this qualitatively distinctive depersonalized 
social attraction that holds groups together, rather than the individual 
personal attraction discussed in earlier models (Hogg, 1 992, 1 993). 
Hogg and his colleagues have tested a number of predictions from this 
self-categorization model of depersonalized attraction (for example, Hogg, 
Cooper-Shaw & Holzworth, in press; Hogg & Hardie, 1991 ,  1992). Hogg et 
al. (in press, Study I )  tested the basic hypothesis that, where group mem­
berships rather than interpersonal relationships are salient, liking among 
group members is depersonalized in terms of the group prototype. They 
studied six small, mixed-sex interactive work groups within a large tertiary 
educational institution, who were asked to report patterns of intragroup 
liking under conditions that either heightened or reduced the salience of 
their work group. They also rated the cohesiveness of the group, described 
the group prototype, ranked the prototypicality of fellow members, rated 
their own prototypicality, and rated the subjective clarity of the group 
prototype. As predicted, it was found that group liking and interpersonal 
liking were independent, and that group liking was positively associated 
with the proto 
typicality of both the self and others, perceptions of strong 
group cohesiveness and a clear group prototype. Interpersonai liking, on the 
other hand, was either unrelated or negatively related to these variables, but 
it was associated with perceptions of interpersonal similarity. Overall, then, 
the study produced clear evidence of depersonalized, category-based intra­
group attraction as distinct from interpersonal liking based on interpersonal 
similarity. A second study involving 1 5  mixed-sex groups of between 6 and 
10 introductory psychology students produced similar findings. 
Hogg and Hardie (1991) report a field study of depersonalized attraction. 
Members of an Australian football team completed a questionnaire at a 
practice session which assessed subjective perceptions of prototypical 


The role o
f
 proto 
typicality 
85 
features of the team, and a rating of self-prototypicality. Team members 
then ranked each other in terms of prototypicality, social attraction (elicited 
under conditions of high team membership salience) and personal attraction 
(elicited where interpersonal friendships were emphasized). 
Team members' responses indicated significant agreement about the team's 
prototypical f
eatures. The most important were a sense of camaraderie 
(,mateship', and unqualified mutual support), task-oriented mutual encour­
agement ('revving-up') and participation in good-natured 'horsing-around' 
and socializing. As predicted, in the team-salient condition there was a more 
significant relationship between subjects' liking for others (social attraction) 
and perceived prototypicality than between interpersonal liking and 
perceived prototypicality, and this effect was most marked for members 
who identified most strongly with the team and saw themselves as highly 
prototypical. Social popularity (that is, being consensually liked as a team 
member) was also strongly related to consensual prototypicality, but the 
relationship between personal popularity (that is, being consensually liked as 
a personal friend) and prototypicality was weaker. 
In summary, Hogg and his colleagues present a strong case for distinction 
between social attraction and personal attraction, with the former uniquely 
related to group processes and involving the perception of self and others in 
terms of group prototypicality rather than individual characteristics. The 
mechanisms underlying the relationship between perceived prototypicality 
and attraction are being investigated further (see Hogg, 1992, 1993): for 
example, does mutual self-categorization produce attraction simply through 
the accentuation of prototypical similarity, or is group prototypicality in 
some way inherently attractive, and might this be because intergroup pro­
cesses work to produce ethnocentric prototypes which portray the ingroup as 
more attractive than comparison groups (for example, see Dunning, Perie & 
Story, 1 99 1 ,  for evidence of the self-serving nature of prototypes)? Clearly, 
the precise nature of depersonalized attraction is not yet fully understood, but 
the fact that there is a distinction between group-prototypical and inter­
personal liking has been established, and further emphasizes the irreducible 
nature of group processes. 
The contextual variability of prototypicality 
As we noted above, Rosch (1 978) emphasized that her analysis of the graded 
nature of within-category structure focused on judgements o
f
 prototypicality 
rather than any notion of fixed prototypes, and she insisted that prototypes 
were not to be taken as a model of category representation or processing. 
Similarly, self-categorization theory emphasizes the context-dependence of 
prototypicality, seeing it as a variable outcome of the categorization process. 
Indeed, contextual variability is built in to the self-categorization definition of 
prototypicality, in that it depends upon both within-category similarity and 
between-category difference. Thus, at the very least, prototypicality for a 


86 
Social identity: international perspectives 
given category will vary as the nature of relevant comparison groups varies, 
and as the wider context in which those groups are perceived varies. For 
example, we have already cited evidence from Hogg et al. (1990) that a group 
can perceive itself as prototypically risky or prototypically cautious 
depending upon whether it compares itself with a more cautious or more 
risky outgroup respectively. 
More recently, two experiments conducted by Haslam, McGarty, Oakes, 
Turner and Onorato (1995) directly tested self-categorization theory's 
assertion that the relative proto 
typicality of extreme outgroup members 
would increase to the extent that an intergroup definition of the judgemental 
context was made salient. This prediction follows from the principle of 
meta-contrast discussed above: in intragroup contexts individuals embody­
ing the mean group position should be more representative of the group as 
a whole (since they maximize intragroup similarity), but in intergroup 
contexts polarized members of both ingroup and outgroup should be more 
representative (as they also maximize intergroup difference). 
In Experiment I all subjects watched the same video-taped message from 
a woman who discussed the causes of road accident-induced brain damage 
and the means by which it could be reduced. They were then asked to make 
judgements about the degree to which she was representative of the group 
to which she was said to belong. In one of eight independent conditions she 
was described as either an extreme or a moderate member of either an 
ingroup (a group that wanted to improve road safety) or an outgroup (a 
group that wanted to ban the sale and consumption of alcohol in order to, 
amongst other things, reduce the road toll). The salience of subjects' own 
social identity was also manipulated so that half the subjects were simply 
given information about the target's position (that is, low salience) but half 
also had to state whether or not they themselves were in favour of banning 
alcohol/improving road safety (high salience). 
As anticipated, the perceived proto 
typicality of the target varied with her 
alleged position within her own group (moderate or extreme) and with the 
extent to which an intergroup relationship between subjects and target was 
salient. More specifically, when the target was described as an outgroup 
member she was seen as most representative of her group in the condition 
where subjects' own group membership was salient and she was described as 
an extremist. In other words, in the context of a salient intergroup division, 
an extremist was more representative of the outgroup than a moderate. 
A similar pattern also emerged from a second study in which all subjects 
watched a video of a discussion in which a group of four people debated 
issues about crime and punishment and came to a pro-authority conclusion. 
For example, the group disagreed with the statement 'criminals should be 
helped rather than punished' and agreed with the statement 'the best way to 
solve the drug problem is for the present laws to be made much stricter'. In 
one (low salience) condition subjects then simply made judgements about 
how representative of the group as a whole were (a) the group member who 
was most pro-authority and (b) the group member who was least pro-


The role o
f
 pro 
to 
typicality 
87 
authority. In two further conditions the intergroup nature of the judge­
mental context was made more salient. In the medium salience condition 
subjects were asked to stage their own (generally anti-authority) views 
before they saw the video. In the high salience condition they were given a 
lecture about the psychological basis and negative social consequences of 
authoritarianism before stating their own views, and the views of all sub­
jects were then collated on a blackboard. As predicted, to the extent that the 
experimental manipulations served to make intergroup differences between 
the subjects and the stimulus group salient, the relative prototypicality of 
the extremely pro-authority target increased so that this individual was seen 
as more representative of the group as a whole than the individual who was 
only moderately pro-authority. In higher salience conditions subjects were 
also less likely to agree with the extremist and more likely to characterize 
the group as a whole in negative terms. 
Taken together, these studies confirm the point central to self­
categorization theory that the degree to which a person is perceived to be 
representative of a group is not simply a function of properties of that 
person considered in isolation. Instead, the very same individual will be 
perceived as more or less prototypical of a social category depending on the 
social context within which he or she is defined. More specifically, both 
studies show that extreme members of outgroups will be seen as more 
representative of the groups to which they belong to the extent that 
differences between ingroup and outgroup are made salient and perceivers 
therefore conceptualize the judgemental context in intergroup rather than 
interpersonal terms. 
To what extent does other social psychological work on prototypicality 
take into account this context-dependence, and the idea of prototypicality as 
a variable outcome of categorization? In general, prototypes have been 
discussed as the basis for category representation, with a concomitant 
emphasis on identifying the prototype of a given category and the effects of 
similarity to this prototype on information processing, rather than issues of 
prototype variability. For example, Brewer and her colleagues were 
amongst the first to apply Rosch's ideas in social psychology in a series of 
studies examining prototypes of elderly people (Brewer et a!., 198 1 ;  Brewer 
& Lui, 1984). This work offers the following view of prototypes: 
Categories are represented cognitively by prototypes - actual or imaginary 
instances of the category that contain attributes most representative of items 
inside the category and least representative of items outside the category. Once a 
prototype of a category has been formed, membership in the category is assessed 
in terms of 'prototypicality' or perceived similarity to the prototypic instance. 
(Brewer et aI., 198 1 ,  p. 656) 
The assumption is that category representation is in terms of prototypes 
(cf. Rosch, 1978), and that, once formed, prototypes are the fundamental 
basis of categorization decisions, that is, it is similarity to the represented 
prototype that determines whether or not an item is perceived as a member 
of a category. In later work, Brewer (1 988) has developed this analysis of 


88 
Social identity: international pers
pectives 
prototypes further to argue that categories are represented in terms of 
pictoliteral prototypes, picture-like images of the ideal category member. 
Thus, the idea of contextual variation in prototypicality within categories is 
not raised in Brewer's work, and the suggestion seems to be that categories 
are represented by fixed visual images. 
Lord and his colleagues discuss prototypes in their leadership categoriza­
tion theory (for example, Fraser & Lord, 1988; Lord, Foti & De Vader, 
1 984; cf. Hogg, 1996). They describe leadership prototypes as cognitive 
structures (for example, Fraser & Lord, 1988, p. 292), and argue that 
perception of an individual as a leader depends upon the extent to which he 
or she is similar to a leadership prototype. Lord et al. (1984) discuss the idea 
of basic-level leadership prototypes, and suggest that these could be used as 
'a standard for appropriate behaviour' (p. 359). Clearly, then, the emphasis 
here is on the identification of relatively context-independent prototypical 
standards against which behaviour is evaluated. Similarly, in their treatment 
of leadership Rush and Russell (1 988) refer to prototypes as 'easily accessed 
schemata which serve as abstract representations of the most representative 
(stereo 
typic) member of a particular category' (p. 89). Again, the emphasis 
is on prototypes as stable, abstract representations of categories rather than 
variable outcomes of the categorization process. 
In contrast, cognitive researchers have identified a number of problems 
with the view that categories are organized around fixed prototypes. 
Barsalou (1987, 1989) has reported extensive evidence of the contextual 
variability of prototypicality. For example, Barsalou, Sewell and Ballato 
(1986) found a low level of agreement between subjects' typicality judge­
ments given twice in the same context but at two different times separated 
by a few weeks. Barsalou and Sewell (1 984) asked American undergraduate 
subjects to judge typicality from both the American and the Chinese point 
of view. They generated graded structures that were, on average, uncor­
related across categories. For example, from the American perspective the 
most typical birds were robin and eagle, but from the Chinese perspective 
they were swan and peacock. Barsalou and Sewell also correlated the 
average graded structure provided by Emory University undergraduates 
with that provided by Emory University faculty for the same categories and 
found a low overall correlation (r = .2). However, when asked to provide 
typicality judgements from faculty's point of view, the undergraduates 
produced a perfect replication of the faculty's own responses, and graduate 
students were able to reproduce exactly both undergraduate and faculty 
typicality judgements. Barsalou (1 987) concludes from this work: 
The graded structure within categories does not remain stable. Instead a category's 
graded structure can shift substantially with changes in context. This suggests that 
graded structures do not reflect invariant properties of categories but instead are 
highly dependent on constraints inherent in specific situations. (p. 107) 
Similarly, Medin (1989) identifies the treatment of categories as context­
independent as a major problem for theories which assume prototype-based 


The role o
f
 pro 
to 
typ icality 
89 
representation. He cites the work of Roth and Shoben (1 983) which again 
shows the contextual variability of judgements of prototypicality. For 
example, Roth and Shoben found that tea is judged to be more typical of 
the category 'beverage' than milk in the context of secretaries taking a 
break, but the reverse ordering obtained when the context was truck drivers 
taking a break. 
An intriguing series of 
experiments by Medin, Wattenmaker and Hampson 
(1987) demonstrates how difficult it is for subjects to work with 
categorizations based simply on relative similarity to a prototype, or family 
resemblance. For the first four experiments, stimuli were constructed to form 
two categories within which items were linked by resemblance to a prototype. 
This was done in terms of the degree to which individual items had properties 
which matched or mismatched the identified prototype. For example, the 
categories might be two groups of cartoon-like animals, and the varying 
properties might be the shape of the head, length of the tail, body markings 
and number of legs (see Medin et a!., 1987, for details). Overall, these 
experiments provided virtually no support for the idea of family resemblance­
based categorization. Asked to sort the full set of stimuli into two equal-sized 
groups, most subjects simply sorted them on the basis of one feature and 
ignored the identified prototypes completely. 
In light of these negative findings, Medin et a!. re-evaluated their assump­
tion that the simple matching and mismatching of isolated properties was 
sufficient to capture what goes on when people make categorization deci­
sions. In the remaining three experiments reported in the paper, they gave 
subjects information about interproperty relationships which could serve to 
explain why integration across separate components of the stimuli was 
appropriate. For example, in Experiment 7 cartoon-like drawings were again 
used, with variation in body size, foot type (paws or webbed feet), body 
covering (feathers or hairs), ear size and mouth type (beak or mouth). Again, 
two prototypes were identified, associated with a pattern of correlated 
attributes which suggested that one category might be related to birds and 
flying. Subjects, given the task of sorting 1 0  figures into two equal-sized 
groups, were told that half the animals were flyers and half were non-flyers. 
This instruction was designed to give the subjects a basis on which to relate 
the separate attributes to each other. It was found that although the majority 
of the sortings were still unidimensional, a significant minority (9 out of 24 
subjects) now did sort by family resemblance, that is, into categories struc­
tured around the identified prototypes. Further, subjects' sortings were 
supported by explanations which drew on the information about flying, even 
when they did not follow the predicted family resemblance pattern. Medin et 
a!. attribute their increased success in these latter experiments to the fact that 
the further information provided (for example, about flying) both made 
interproperty linkages salient and allowed subjects to associate individual 
items through their relationship to a higher order theme. 
These findings are consistent with Medin's cogent argument (for example, 
Medin, 1989; Medin & Wattenmaker, 1987; Murphy & Medin, 1 985) that 


90 
Social identity: international pers
pectives 
categorization cannot be based on judgements of similarity through match­
ing of isolated attributes. Indeed, he asserts that similarity 'is more like a 
dependent variable than an independent variable' (Medin, 1989, p. 1474) in 
the categorization process (cf. McGarty & Turner, 1992). He has suggested 
that perceivers have wide-ranging 'background theories' about their world 
and how it works, and that these theories play a crucial role in 'category 
cohesiveness', that is, in defining the way in which categories 'hang together' 
as coherent wholes, and therefore in determining why we have the 
categories we do rather than others. For example, Murphy and Medin 
( 1985) point out that on the basis of simple attribute matching we could 
identify several similarities between a plum and a lawnmower (both cannot 
hear, both have a smell, both can be dropped . . .  
) which might suggest they 
should share a category membership, but this is not consistent with our 
general theories about the world, so we understand that the 'similarities' 
identified are not sufficient to determine categorization. On the other hand, 
we can construct meaningful categories which bring together items with 
very little apparent similarity. Barsalou (1983) provides the example of a 
category comprising children, money, photo albums and pets, which 
becomes sensible when one considers it in the context of a knowledge base 
which includes the fact that the category represents 'things to take out of 
one's house in a fire' (see also Rips, 1989, for further discussion of the idea 
that similarity is neither necessary nor sufficient to determine category 
membership). 
These arguments about the insufficiency of similarity have been used to 
challenge the classical view of categorization, but Medin (1989) points out 
that the more recent prototype theories (and exemplar-based theories) are in 
fact just as dependent upon similarity as a determinant of categorization. It 
is similarity to the prototype (or a specific exemplar) that is assumed to 
determine categorization (for example, Brewer et aI., 198 1 ; Lord et aI., 
1984). Medin argues for knowledge-based rather than similarity-based 
categorization, although he sees a role for similarity in contributing to what 
he calls 'psychological essentialism', which functions to constrain the 
'deeper substance' (Medin, 1989, p. 1479) of the knowledge-rich theories 
which drive the categorization process (see Medin, 1989; Medin & Ortony, 
1989, for further discussion of psychological essentialism). 
In summary, cognitive psychologists have argued that theories which 
assume that categories are represented by fixed prototypes, and that 
categorization is a matter of attribute matching to a prototype, cannot deal 
with evidence of contextual variability in category content and structure. 
They also fall foul of evidence that attribute matching (simple similarity 
judgement) is insufficient to account for category cohesiveness and cat­
egorization decisions. Medin (1989) comments: 
Prototype theories . . . fail to reflect the context sensitivity that is evident in 
human categorization. Rather than getting at the character of human conceptual 
representation prototypes appear to be more of a caricature of it. (p. 1472) 


The role o
f
 pro 
to 
typ icality 
91 
In this context, it is important to emphasize that self-categorization 
theory is not a 'prototype theory' in the sense referred to in Medin's work. 
As noted above, the theory sees proto 
typicality as a wholly context­
dependent outcome of the categorization process, not the basis for category 
representation, and agrees with Medin that this categorization is a product 
of perceivers' background knowledge in interaction with current stimuli (see 
Oakes et aI., 1994; Turner, Oakes, Haslam & McGarty, 1994). The simi­
larity and difference judgements involved in the meta-contrast principle are 
made in the context of background theories about the ways in which 
category members tend, in general, to be similar and different - this is 
embodied in the 'normative fit' principle discussed briefly above (Oakes et 
aI., 1991). One of the major contributions of the theory has been its analysis 
of perceived similarity and difference as outcomes (rather than simple 
determinants) of categorization, and we see perceived proto 
typicality as one 
aspect of the way in which categorization contextualizes simple attribute 
judgements, gives them meaning in relation to other stimuli being judged 
(Turner & Oakes, 1989; Turner et aI., 1995). 
Conclusion 
This chapter has considered the concept of prototypicality from the per­
spective of self-categorization theory. We have reviewed ideas and evidence 
from two main areas of self-categorization research in which prototypicality 
has played an important role. In polarization research, it has been shown 
that polarization can be explained as conformity to the prototypical position 
within the group, a position extremitized through the influence of com­
parative context on ingroup definition. Research on intragroup attraction 
has established a distinction between social and personal attraction, the 
former reflecting attraction within groups based on depersonalization and 
prototypicality. In both these research areas, prototypicality is seen as a 
context-dependent outcome of the categorization process, as determined by 
both comparative (meta-contrast) and normative (fit with the content of 
background theories) factors. Because the principle of meta-contrast allows 
accurate specification of the most prototypical attributes within a given 
comparative context, precise predictions about conformity and attraction 
within groups have been made possible. 
Recognition of the internally graded, variable structure of categories was 
one of the revolutionary contributions of Rosch's work. However, many 
subsequent developments of that work have transformed the dynamic idea 
of proto 
typicality judgements into a concept of idealized 'caricatures' 
functioning as relatively fixed representations of categories, and determining 
the categorization process through judgements of item-prototype similarity. 
Self-categorization theory maintains a more Roschian perspective in this 
regard, with its emphasis on prototypicality rather than prototypes, and its 
rejection of the view of categorization as a matter of 
judging similarity to a 


92 
Social identity: international perspectives 
prototype. In predicting categorization through the interaction of com­
parative and normative factors, the theory shares with the work of Medin 
and others an emphasis on the intellectual (Neisser, 1 987) basis of 
categorization. 
Whilst the theory's conceptualization of prototypicality has proved very 
fruitful in the research areas reviewed here, it is nonetheless true that more 
work remains to be done in order to demonstrate the inherently variable 
nature of category representation (cf. Barsalou, 1 987, 1 989). This represents 
one significant direction in which our current work is advancing (Haslam, 
Turner, Oakes, McGarty & Hayes, 1 992; Oakes et aI., 1994; Oakes, Haslam 
& Reynolds, in press). 


7 
Group Beliefs as an Expression of Social 
Identity 
Daniel Bar-Tal 
Social identity 
One of the seminal theories of social psychology which contributes to the 
understanding of intragroup and intergroup processes is social identity 
theory (Tajfel, 1978d, 1981 ,  1982b). The theory posits that individuals not 
only adopt a personal identity as unique persons, but also form a social 
identity which reflects their membership in various groups to which they 
belong. Specifically, social identity is defined as 'that part of an individual's 
self-concept which derives from his knowledge of his membership of a social 
group (or groups) together with the value and emotional significance 
attached to that membership' (Tajfel, 1 978b, p. 63). Social identities, in the 
form of such social categories as nationality, religion, gender, profession, 
ethnicity or political orientation, are internalized and constitute a poten­
tially important part of the individual's self-concept. They provide meaning­
ful and significant self-references through which individuals perceive 
themselves and the world around. 
Of special importance is the assumption stating that individuals have a 
need to achieve positive social identity. Specifically, it is assumed that indi­
viduals strive for positive self-image and that in the course of identifying with 
a group this need translates into a tendency to view one's own group 
favourably. This view is achieved not necessarily through positive 
perception, but rather through comparison on appropriate dimensions. 
The comparison allows one to perceive the ingroup as 'better' than relevant 
outgroups. 
The process of social identification is further elaborated in self­
categorization theory (Turner, 1 987), an extension of social identity 
theory. The theory suggests that individuals categorize themselves as 
members of social categories, and then define, describe and evaluate 
themselves in terms of these categories. In the process of self-categorization, 
a set of cognitive representations of self is formed. Part of the cognitive 
representations reflect the perceived membership of individuals in groups. 
Since individuals belong to various groups, they possess multiple rep­
resentations of self. The categorization of the self results in a perceptual 


94 
Social identity: international perspectives 
accentuation of intra-category similarities and inter-category differences on 
dimensions believed to be associated with the categorization. 
Theories of social identity and social categorization focus on the defini­
tion of 'we-ness' by group members in terms of 'We belong to a group'. 
Group members in this process perceive themselves as group members, 
identify themselves as such, and differentiate between their own group and 
other groups. This primary process of 
cognitive-emotional nature shapes the 
social identity of individuals. Nevertheless, on the basis of social identity 
and social categorization theories, a question should be asked: What are the 
other cognitive representations, in addition to social categorization of the 
self, which underlie a person's social identity? The answer to this question is 
provided in an analysis of group beliefs (Bar-Tal, 1 990). The purpose of the 
present chapter is to present group beliefs as a framework for analysing a 
group member's social identity. Specifically, first the chapter will delineate 
the conception of group beliefs; second, it will outline their characteristics; 
third, it will describe group beliefs' formation; and, finally, implications will 
be presented. 
The conception of group beliefs 
'Group beliefs are defined as convictions that group members (a) are aware 
that they share, and (b) consider as defining their "groupness'" (Bar-Tal, 
1 990, p. 36). The first part of the definition suggests that two beliefs have to 
be shared by group members. One belief pertains to any content that is the 
subject of the group belief and the other pertains to specific knowledge 
saying that the former is shared by group members. The second part of the 
definition suggests that group beliefs are those beliefs that group members 
perceive as defining the essence of their group. It is assumed that they 
provide the cognitive basis that group members view as uniting them as one 
entity. In fact, group beliefs serve as a foundation for group formation and, 
later, as a bond for group existence. 'We are exploited', 'Communism is the 
best system for human beings', 'Jesus is God's son', 'Iraq is our enemy', are 
examples of possible group beliefs. 
In this framework, the belief 'We are a group', expressing social identity 
of group members, is called the f
undamental group belief However, in 
addition to the fundamental group beliefs, group members share group 
beliefs of various contents pertaining to a variety of other subjects. These 
beliefs are part of cognitive representations which together with self-social 
categorization (that is, the fundamental group belief) define social identity 
of group members and thus underlie their we-ness and uniqueness. 
Individuals regard group beliefs as characterizing them as group members 
and as defining the boundary of the group. On their basis individuals may 
categorize themselves and may be categorized by others as group members. 
Acceptance of group beliefs is one of the important indicators of group 
membership. Sharing group beliefs implies, most frequently, that a person is 


Group belie
f
s
 as an expression o
f
 social identity 
95 
a group member. Moreover, being a group member signals to people that a 
person shares group beliefs. A person joining a group must, at least exter­
nally, express acceptance of group beliefs. Thus, on the one hand, group 
beliefs unify group members and, on the other hand, differentiate them 
from outgroups. The unification is reflected in the recognition of similarity 
(Bar-Tal, 1 993). Group members are aware that they share the same unique 
beliefs that define their 'groupness' and thus become aware of the similarity. 
Group members who may differ with regard to many characteristics are 
aware that they share the same unique beliefs that define their 'groupness' 
and thus become aware of the similarity. 
Also, since in many cases groups try to be differentiated, group beliefs 
often provide a criterion for differentiation. Group beliefs draw the line 
between the in 
group and outgroups. As Sherif (1951) pointed out: 'From 
the point view of inter-group relations, the most important consequence of 
group structuring is the delineation of in-group from out-groups. The devel­
opment of in-group and "we-experience" is accompanied by the demarca­
tion and setting of boundaries from out-groups' (p. 395). 
An experiment by Allen and Wilder (1975) demonstrated that group 
formation (including the formation of a group belief) is a more potent 
factor in discrimination against outgroup than the awareness that the 
outgroup members have some similar beliefs. Also, belief similarity about 
the basis for group classification (that is, formation of a group belief) had a 
significant effect on discriminative behaviour. Special ingroup favouritism 
was displayed when the in 
group had similar beliefs. 
Contents of group beliefs 
Group beliefs, as implied by the definition, are held by group members, who 
share these beliefs and consider them as defining their group. Thus, group 
beliefs should be considered as a category of beliefs and should be 
characterized by the same features as other beliefs. First of all, the contents 
of group beliefs will be discussed. 
Beliefs in the present conceptual framework, viewed as propositions to 
which individuals attribute at least some truth, constitute units of knowl­
edge. This is a wide-scope definition which considers any ideas, thoughts or 
opinions about any topic as beliefs. They can thus cover various contents 
which can be classified to such categories as 'values', 'ideologies', 'inten­
tions', 'goals', and so on. Group beliefs may also refer to different topics or 
ideas that a human mind can think. They may concern such subjects as 
religious doctrines, political ideologies, philanthropic goals, human rights, 
identity characteristics, prescriptions of behaviour (that is, norms), and 
others. As examples, four categories of group beliefs will be presented -
those which refer to group norms, values, goals and ideology. 


96 
Social identity: international perspectives 
Group norms 
Group norms, defined as 'an idea in the minds of the members of a group, 
an idea that can be put in the form of a statement specifying what the 
members or other men should do, ought to do, are expected to do, under 
given circumstances' (Homans, 1950, p. 123), regulate group members' 
behaviour and provide criteria for judging it. Norms, as defined, do not 
necessarily have to be group beliefs. Norms may regulate group members' 
behaviour without being considered as characteristic of the specific group. 
In such groups, group beliefs may pertain to contents other than norms. In 
other groups, at least part of the norms may define their characteristics and 
be group beliefs. Nevertheless, when norms function as group beliefs, group 
members believe that their patterns of behaviour are unique to them and 
characterize their membership. 
In an extensive description of the Amish society, 
I Hostetler (1 968) points 
out several norms that can be considered as group beliefs. That is, Amish 
people believe that particular prescriptions of behaviours define their 
'groupness' and contribute to their social identity. For example, Amish 
beliefs prescribe the necessity of separation from the world. Therefore, an 
Amish man does not marry a non-Amish woman, does not do business with 
an outsider, and even does not enter into intimate relationships with a 
person outside the Amish community. In addition to norms of separation, 
the Amish group also perpetuates norms of non-resistance. Group members 
are forbidden to take part in violence or war, and they withdraw from 
conflicts or resistance. They refuse to swear any oath, to bear arms, or to 
hold public offices. 
In another example, Ardener (1983) described the group beliefs of 
Canadian Doukhobors who emigrated from Russia at the beginning of the 
twentieth century. They believe that the Perfect Life is attainable on earth 
through the rejection of materialism. They have not recognized human 
authority structure and therefore are opposed to state legislation or oath 
taking. Their uniqueness has been reflected in deprivation, suffering and 
poverty, which became, according to Ardener, part of their self-definition, 
corresponding to their identification with Christ himself. In this vein, to 
form their own identity, Doukhobors resorted to nudity as a social artefact 
for identifying their group and to the use of arson against institutions and 
then against their own property. Ardener suggested that both practices 
helped to maintain the boundary of the group and to enhance solidarity in 
the face of outgroups. 
Group values 
An analysis similar to the one presented about norms applies to group 
values. Groups may hold values, but they do not necessarily have to 
function as group beliefs. A value is defined as 'an enduring belief that a 
specific mode of conduct or end-state of existence' (Rokeach, 1 973, p. 5). 
The former is called an instrumental value, while the latter is called a 


Group belie
f
s
 as an expression o
f
 social identity 
97 
tenninal value. Values guide the selection of the means and ends of specific 
actions, and serve as criteria by which objects, actions or events are 
evaluated. Examples of values are freedom, truth, individualism or equality. 
They reflect the ideals to which individuals and groups aspire. 
It is possible to characterize groups by the values that their group 
members share (Kluckhohn, 1 95 1 ;  Parsons, 1 968). Sociologists have focused 
on shared values, referring to them as value systems in a given society. 
Nevertheless, a given shared value is not necessarily held by every person in 
the group. When a sufficient number of group members subscribe to the 
value, it may not only characterize the group, but may also become an 
important detenninant of the group members' behaviour. 
In some groups the values, like nonns, may be fonnally fonnulated (that 
is, in writing) and maintained as group beliefs, while in other groups they 
are never fonnally defined, but may be considered as group beliefs and 
carried latently through the social processes of socialization and influence. 
However, in both types of groups, when values are shared by group mem­
bers and viewed by them as characterizing the group, they are group beliefs. 
In this case, the value or values define the uniqueness of the group and 
differentiate it from outgroups. 
To specify values that characterize a nation is a complex task. Never­
theless, social scientists commonly have suggested, on the basis of either 
collected data or their own observations, that self-sufficient individualism 
is one of the salient characteristics of American society (for example, 
Diamond, 1976; Gillin, 1 955; Williams, 1970). This value emphasizes the 
responsibility and the right of each individual to pursue personal goals with 
as much freedom from restraints as possible. Americans are socialized to 
internalize this value and to believe that it characterizes their society. 
Many groups present values as part of their credo. In fact, many of the 
group nonns and goals functioning as group beliefs are based on values. 
This stress on values can be observed in religious groups, as well as in many 
political groups. An example of a political group that was founded on the 
basis of a value is Amnesty International. In its statute, the value of human 
rights underlies more specific goals. The beginning reads: 
I .  Considering that every person has the right freely to hold and to express his 
or her convictions and the obligation to extend a like freedom to others, the 
objects of Amnesty International shall be to secure throughout the world the 
observance of the provisions of the Universal Declaration of Human Rights, 
by: 
a) irrespective of political considerations working towards the release of and 
providing assistance to persons who in violation of the aforesaid provisions 
are imprisoned, detained or otherwise physically restricted by reason of their 
political, religious or other conscientiously held beliefs or by reason of their 
ethnic origin, sex, color, or language, provided that they had not used or 
advocated violence (hereinafter referred to as 'Prisoners of Conscience'); 
b) opposing by all appropriate means the detention of any Prisoners of 
Conscience or any political prisoners without trial within a reasonable time or 


98 
Social identity: international perspectives 
any trial procedures relating to such prisoners that do not conform to 
internationally recognized norms; 
c) Opposing by all appropriate means the imposition and infliction of death 
penalties and torture or other cruel, inhuman or degrading treatment or 
punishment of prisoners or other detained or restricted persons whether or 
not they have used or advocated violence. (Amnesty International pamphlet, 
London, 1978) 
Although the Ku Klux Klan is vastly different from Amnesty International, 
its group beliefs are also based on values. The principal beliefs of the Ku 
Klux Klan are presented in the following way: 
We invite all men who can qualify to become citizens of the Invisible Empire, to 
approach the portal of our beneficent domain, join us in our noble work of 
extending its boundaries, and in disseminating the gospel of Klankraft, thereby 
encouraging, conserving, protecting and making vital the fraternal relationships in 
the practice of an honorable clannishness; to share with us the sacred duty of 
protecting womanhood; to maintain forever the God-given supremacy of the 
White Race; to commemorate the hold and chivalric achievement of our fathers, 
to safeguard the sacred rights, privileges and institutions of our civil government; 
to bless mankind and to keep eternally ablaze the sacred fire of a fervent devotion 
to a pure Americanism. (Cited in Vander Zanden, 1960, p. 459) 
Thus, the values of womanhood, white supremacy, mankind and American­
ism are the values that the group intends to protect. 
Group goals 
Group goals are defined as beliefs of valued or desired future specific states 
for the group (Cartwright & Zander, 1968; Etzioni, 1 975). On the basis of 
this definition, we can assume that with few exceptions almost all groups 
have goals. In most of these cases, goals may serve as group beliefs in 
addition to other beliefs. Individuals form groups to achieve certain goals. 
Subsequently, goals are often considered as a raison d'etre for group 
formation, frequently keep group members together, provide a basis for 
solidarity, and give direction for activity. These functions are particularly 
explicit in certain voluntary groups that are formed to advance specific 
goals (Sills, 1958). Thus it is not surprising that group goals frequently serve 
as basic group beliefs. They increase the identification of the group 
members with their group and define the boundary for group membership. 
In this vein, March and Simon (1958) noted that: 
The greater the extent to which goals are perceived as shared among members of 
a group, the stronger the propensity of the individual to identify with the group 
and vice versa. (p. 66) 
Two examples of group goals may provide illustrations of how they 
function as group beliefs. Trans-Species Unlimited (TSU) is a national 
animal rights group 'dedicated to the total elimination of animal abuse and 
exploitation'. This group belief leads TSU to direct actions and legislative 


Group belie
f
s
 as an expression o
f
 social identity 
99 
campaigns to end such practices as the use of leghold trap and live bird 
shoots and commercial exploitation of rabbits for meat, fur and research 
The Richard III Society was founded in England in 1924 with the 
following aims: 
1 .  To promote in every possible way historical research into the life and times of 
King Richard III. 
2. To secure a reassessment of the historical material relating to this period and 
of the role in English history of this monarch. 
3. To circulate all relevant historical information to members of 
the Society and 
to educational authorities. (copied form the pamphlet of The Richard III 
Society, New Orleans, undated) 
The principal goal of this group is to present the truthful history of King 
Richard III in view of past distortions and biases. 
Groups differ with regard to the goals that they try to achieve. There are a 
variety of goals that can be classified into different categories. One type of 
category has received special attention by political scientists - namely, goals 
to influence governmental decisions. Groups that have these goals are called 
interest groups (Salisbury, 1975), and then the term interest denotes a 'shared 
attitude toward what is needed or wanted in a given situation, observable as 
demands, or claims upon other groups in the society' (Truman, 1 95 1 ,  p. 33). 
These shared demands or claims, which reflect goals, serve as a basis for 
group formation and function as group beliefs. The interests can pertain to 
various contents involving either tangible or intangible claims, and concerns 
of 
either a particular policy objective or wide societal goals (Salisbury, 1 975). 
Interest groups can be based on either a few or many group beliefs. In 
some groups, the group beliefs refer to one broad topic that is relevant to 
several issues. For example, the Consumer Federation of America reflects a 
viewpoint that favours the protection of the rights of individual consumers 
through the regulation of corporate activities, and the US Chamber of 
Commerce reflects a strong belief in the character of the American free 
enterprise system, opposing excessive government regulation of business. 
Other groups' interests are formed on the basis of narrow group beliefs 
reflecting single issues. For example, the National Rifle Association has a 
strong belief in the American citizen's unrestricted right to bear arms; the 
National Right to Life Committee believes in the sanctity of the foetus. 
Group ideology 
One of the most discussed concepts for describing a group's uniqueness is 
ideology. Ideology is often referred to as the mental characteristic of 
a group. Indeed, various definitions of ideology refer to its commonality - a 
state indicating that ideological beliefs are shared by group members. It 
reflects their common experience and serves as a basis for group members' 
cooperation, morale, order and as a rationale for their behaviour. Ideology 
refers to an integrated set of beliefs constituting a programme, a theory of 
causes and effects, and premises on the nature of humanity and societal 


100 
Social identity: international perspectives 
order (Apter, 1964; Lane, 1 962; Shils, 1968). For example, Toch (1965) 
defined ideology as: 
a set of related beliefs held by a group of persons. The ideology of a social 
movement is a statement of what the members of the movement are trying to 
achieve together, and what they wish to affirm jointly . . .  
the ideology of a social 
movement defines the movement, and contrasts it with other movements and 
institutions. (p. 21) 
When group members hold an ideology, it frequently serves as a group 
belief. By definition, it consists of a set of ideas that characterize the way in 
which a group posits, explains and justifies the ends and means of its 
organized social actions. It usually provides an identity to group members, 
defines the group cohesion, and describes its exclusivity; in most cases, an 
identification with a specific ideology indicates a membership in a particular 
group (Lane, 1962). 
The contents of an ideology can be political, social or even religious. An 
ideology usually refers to images of the desired society and the means and 
conditions needed to achieve it. For example, the ideology of the Fasci 
Italiani di Combattimento (the Italian Fascist Party) placed a special 
emphasis on a nation or a state - the fascist state. The individuals and 
groups are secondary to the state, which embodies the culture and spirit of 
the people and is considered to be the highest and noblest value. The state 
provides an orientation for the individuals - it is only there that they can 
fulfil their identity. According to the fascist ideology, life is a continuing 
struggle for national supremacy. In order to achieve the nation's goals, the 
state has to exercise complete authority over all areas of public and private 
life, from the education of children to control of the economy. The state has 
the right to control and restructure the society. Accordingly, the state 
requires total loyalty. There is no place for groups or individuals to object, 
since the state is the source of political, economic and moral action 
(Mussolini, 1 935). These principles served as group beliefs of the Italian 
Fascist Party, and, with its ascendance to power in 1 922, the party tried to 
turn them into group beliefs for the entire nation. 
For seven decades, the principles of communist ideology served as group 
beliefs for the members of the Communist Party of the Soviet Union. This 
ideology was the official creed and the most accessible characteristic in 
every domain of life in Russia. The basic principles of communism were 
inculcated in Soviet citizens from the early years of education and main­
tained through all the state communication channels. The ideology 
advocates a classless social system of equality for all members of society, 
which provides everyone with material and cultural benefits according to 
their growing needs (Triska, 1 962). 
In sum, this limited review of the contents did not exhaust all possible 
categories of group beliefs. As indicated, group beliefs can have any 
contents, and the above categories serve only as relevant examples of the 
ones that are frequently used by social scientists. Group members share 


Group belie
f
s
 as an expression o
f
 social identity 
101 
group beliefs that characterize them and differentiate them from other groups. 
As illustrations of possible contents, this review was not based on group 
members' reports identifying their group beliefs, though it is possible that 
they actually serve as group beliefs. Future research should treat the formal 
statements as hypotheses that need to be validated through empirical 
evidence. 
Characteristics of group beliefs 
Group beliefs can be characterized by the same features as other beliefs. 
Degree of confidence and level of centrality are among the most important 
characteristics of beliefs (Bar-Tal, 1990; Bem, 1 970; Kruglanski, 1 989). 
These characteristics are of special relevance for group beliefs and therefore 
discussed at some length. 
Con
fidence 
Group beliefs are usually held with great confidence because they are con­
sidered to be facts and verities. This occurs because group beliefs define the 
essence of the group and thus the reality of the individuals who view 
themselves as group members. In fact, in most cases, individuals voluntarily 
join groups, such as religious denominations, political organizations or 
professional associations, on the basis of group beliefs. The act of 
joining a 
group indicates that group members accept, at least formally, group beliefs 
and hold them as valid. 
A reduced confidence in group beliefs may shatter group members' 
reality, especially in cases when group membership is very important for 
individuals or when the change is sudden. The questioning of group beliefs, 
especially central ones, may have significant implications for group life. 
Moving group beliefs from the status of verities to the status of hypotheses 
by the majority of group members may change the essence of the group or 
its raison d'etre, or even cause its division or disintegration. It is, therefore, 
not surprising that groups often use various mechanisms to maintain high 
confidence in group beliefs. 
It is recognized, though, that groups differ with regard to their insistence 
on confidence. While orthodox groups insist on high confidence in group 
beliefs (Deconchy, 1 984), other groups may be less strict on this matter. The 
former groups do not allow any doubts or sanction their expression. The 
other groups are more tolerant of changes and therefore their group beliefs 
may be gradually altered. 
Centrality 
The centrality of group beliefs implies that they are often accessible in 
group members' repertoire and that they are relevant for consideration 


1 02 
Social identity: international pers
pectives 
in making various evaluations, judgements or decisions, including behav­
iours (Bar-Tal, 1 986). The centrality characteristics of group beliefs has at 
least two different aspects. First, the characteristic may refer to the 
importance of the set of beliefs for the group members, and, second, it may 
refer to the importance of the specific belief within the set of group beliefs. 
The former aspect is based on an assumption that some groups are more 
important for their members than others. Individuals are usually members 
in a number of groups and not all of them are of equal importance to 
them. The importance of the group for group members is reflected in the 
centrality of group beliefs. That is, when the group is important for group 
members, in most cases the group beliefs are central. For example, it can be 
assumed that group beliefs are very central in groups such as the Amish 
(Hostetler, 1 968) and Lubovitcher Hassidim (Levy, 1 975). Group 
membership in these groups is the most important characteristic for the 
members and the group beliefs are frequently accessible and often taken 
into consideration. In other groups, group beliefs may be unimportant for 
most of the group members. 
The other aspect of characterizing group beliefs as central refers to the 
assumption that not all beliefs are of equal centrality for group members. 
Some group beliefs may be of greater importance than others and thus may 
be more central. It means that when a group has a set of elaborated group 
beliefs, some of them contribute more to the group characterization and its 
definition than others. The more central group belie
f
s
 are considered as 
prototypic in group characterization and there
f
ore are called basic group 
belie
f
s. Their weight in defining the essence of the group is crucial. Beyond 
the simple self-definition 'We are a group', these beliefs provide the group's 
credo. For instance, in certain groups, group beliefs that refer to group 
goals or identity may be more central than group beliefs that refer to values 
or group history. In a specific example, it may be assumed that while the 
Mormons have numerous group beliefs, one of the basic ones is the 
conviction that the Book of Mormon is another Testament of Jesus Christ, 
a religious and secular record of ancient American civilization (O'Dea, 
1 957). 
The centrality characteristic is not a given, but often is maintained by 
external factors. Availability and saliency of group beliefs in the group 
members' environment may influence their centrality. The first element 
refers to the frequency with which the content of group beliefs is exposed 
to group members. The other element expresses the prominence and 
vividness of the group belief when it is presented to group members. 
When group beliefs are repeatedly presented to group members in a 
prominent and vivid manner, they are often accessible in group members' 
repertoires (Higgins & King, 1981). Groups often keep group beliefs 
accessible in the minds of the group members by constantly repeating the 
contents of group beliefs. Moreover, they even sometimes develop cultural 
and educational mechanisms to maintain group beliefs as central in group 
members' minds. 


Group belie
f
s
 as an expression o
f
 social identity 
1 03 
Formation of group beliefs 
Formation o
f
 the f
undamental group belief 
The crucial phase in the f
ormation of group belie
f
s
 is the emergence o
f
 the 
f
undamental group belief indicating group existence (that is, ' We are a 
group'). This phase is preceded by various social perceptual and cognitive 
processes. According to the present view, numerous reasons can cause 
individuals to start believing that they constitute a group. Individuals usually 
have to f
ormulate at least one belief which provides the antecedent f
or the 
f
undamental group belief stating that a group exists (that is, ' We are a 
group'). Such an antecedent belief may later become a group belie
f 
Any belief can serve as an antecedent f
or individuals to begin considering 
themselves members o
f
 a group. The belief 'We are a group' may be instigated 
by any of the following beliefs: 'We act interdependently', 'We have the same 
fate', 'We have the same characteristic', 'We live in the same place', 'We have 
the same goal', 'We believe in the same religious doctrine', 'We have the 
same ancestors', 'We accept the same ideology', 'We are treated in the same 
way', and others. The antecedent belie
f(s) may be f
ormed on the basis o
f
 
actual experience and perception or on the basis o
f
 reliance on inf
ormation 
which comes f
rom sources that propagate these belie
f
s. In the first case, 
individuals may actually become aware of the similarity in what may be 
called natural evolvement, whereas in the other case, an epistemic source 
may indicate to the individuals that they are similar and point out that 
similarity. 
For example, Epstein's (1978) work indicates that ethnic identity is based 
on shared beliefs in common attitudes or property, which differentiate the 
group from the outgroups. The shared beliefs are products of common 
perceptions of a group's past and interaction between forces operating within 
a group and those coming from outgroups. Tajfel (1981) hypothesized that 
minority groups tend to emerge as a result of common attitudes or treatment 
by the outside groups which, on the one hand, facilitates the perception of a 
common fate and, on the other hand, indicates the boundaries operating 
between the ingroup and the outgroups. Several other conceptions stress 
individuals' shared needs (for example, Killian, 1 964; Toch, 1 965), such as 
common feelings of frustration, alienation, deprivation, exploitation and 
injustice, any of which may serve as a basis for group formation. 
An experiment by Zander, Stotland and Wolfe (1960) may be seen as a 
demonstration of how the basic group belief is formed. In this experiment, 
the investigators created a group composed of female college students. To 
foster a group entity, the experimenter provided several antecedent beliefs 
regarding similarity, proximity, common goals and identity. Specifically, the 
experimenter systematically manipulated the perception of 'groupness' by 
varying seating arrangements (creating proximity), by pointing to similari­
ties among group members (creating commonality), and by assigning a goal 
for the group (creating competition against another group). Group members 


1 04 
Social identity: international pers
pectives 
were then asked to choose a name for their group. These antecedent beliefs, 
together with the name of 
the group, produced the fundamental group belief 
'We are a group'. The study also showed that these individuals behaved as 
group members. Thus, the investigators concluded that when a group is 
formed, large proportions of the self 'become involved in the group and are 
affected by identification with the group' (p. 475). 
The effects of group experiences on the formulation of the fundamental 
group belief are underlaid by the process of perception. It has been sug­
gested that principles of perceptual organization contribute greatly to the 
self-perception of individuals as a group (Asch, 1 952; Campbell, 1958). 
Specifically, Campbell (1958) proposed the use of four Gestalt principles 
(proximity, similarity, common fate and pregnance) to explain why indi­
viduals begin to see themselves as one entity. These principles, which lead 
discrete elements to be perceived as parts of a whole, can be viewed as 
characteristics of individuals, and therefore allowing definition of the group 
and differentiation between the ingroup and the outgroups. Indeed, 
Brewer's (1979) review of group formation under limited conditions indi­
cates that factors such as interdependence, intragroup similarity or shared 
fate affect the feeling of 'groupness'. Moreover, these factors not only 
determined group formation, but also influenced group members' behaviour 
in the direction of ingroup bias. 
A somewhat similar approach to group formation was recently suggested 
by cognitive social psychologists. They placed their focus on the knowledge 
that individuals possess (for example, Pryor & Ostrom, 1 987). Thus, infor­
mation about the collective of individuals (that is, their attributes) serves as 
a basis for their self-classification as group members. This information 
denotes the similarity of group members and their uniqueness in com­
parison to outgroup members. 
Moreland (1987) extended the perceptual-cognitive perspective to four 
different bases for self-classification of individuals as group members, and, 
therefore, for the formation of the fundamental group belief 'We are a 
group'. One basis derives from the conditions and resources in the environ­
ment. Factors such as money, time, propinquity or social networks may 
affect group formation. For example, people who live or work in close 
proximity may form a group which can be based on common goals or 
common interests. Also, a social network of relationships may facilitate 
group formation, as happens among friends or members of professional 
organizations. 
Behavioural basis is evident when people become dependent on one 
another for the satisfaction of their needs. Various theoretical approaches, 
such as evolutionary perspective, social exchange theory, social evaluation 
perspective or psychodynamic theories, suggest that individuals form groups 
in order to satisfy their various needs. Needs of survival, defence, social 
support, predictability, world understanding, uncertainty reduction, anxiety 
avoidance or personal adjustment are only a few examples of needs which 
can be satisfied through group membership. 


Group belie
f
s
 as an expression o
f
 social identity 
1 05 
Groups may also be formed on the basis of affection as reflected in 
shared feelings. The positive feelings can be expressed in interpersonal 
attraction of group members, attraction towards group goals or activities, 
and love and respect of the group's leader. 
Finally, the cognitive basis indicates that groups are formed when people 
realize that they share important personal characteristics such as attitudes, 
values, interest or goals. Summarizing this basis, Moreland (1 987) pointed 
out that: 
Research on personal factors suggests that small groups will form when people (a) 
have acknowledged their shared characteristics before, or (b) have done so 
recently, or (c) have found it useful to think of themselves in that way. Research 
on situational factors suggests that small groups will form when (d) people are 
reminded of shared characteristics, or (e) their outcomes seem to depend on those 
characteristics, or (f) the characteristics that they share are unusual in some way. 
(p. 103) 
On the basis of the present analysis, it becomes obvious that social 
categorization is the psychological process which underlies group f
ormation in 
the minds o
f
 the individuals who are group members. As noted earlier, 
Tajfel's theory of social identity and Turner's theory of self-categorization 
elaborate on this process (Tajfel, 1978d, 1981 ;  Turner, 1987). 
In this vein, several studies demonstrate that individuals form a psycho­
logical group solely on the basis of shared social category, in the process of 
comparing themselves to others. In these experiments, performed within the 
framework of Tajfel's social identity theory and Turner's self-categorization 
theory, individuals were categorized as group members and they acted 
accordingly (for example, Billig & Tajfel, 1 973; Hogg & Turner, 1985b; 
Tajfel, Flament, Billig & Bundy, 197 1 ;  Turner, Sachdev & Hogg, 1983). 
In a study by Tajfel et al. (1971), individuals were divided into groups on 
the basis of a fairly trivial task (expressing a preference for abstract 
paintings of Klee and Kandinsky or guessing the number of dots in rapidly 
projected clusters). Later, they were put in individual cubicles and asked to 
divide points worth money between two persons who were identified as 
either belonging to their group or not. This minimal group situation was 
enough to produce ingroup favouritism. Individuals tended to award more 
points to members of their group. These results were replicated, even in the 
experiment by Billig and Tajfel (1973), when the division into groups was 
done on the basis of the toss of a coin. 
The important contribution of these studies to the understanding of group 
formation is the demonstration showing that the perception of the group as 
an entity does not have to be based on actual experiences, but a mere piece 
of information may be a sufficient condition for self-labelling as a group. 
Individuals may be told that they are given the same label (or category), 
and this information may cause them to form the belief 'We are a group' 
(for example, Billig & Tajfel, 1973). 
In this respect it was analysed how groups form the fundamental belief 
which underlies the formation of a group. Common experience and 


106 
Social identity: international pers
pectives 
perceived information may influence individuals to form a new shared 
reality. The appearance of the fundamental belief 'We are a group', which 
was the centre of the work of Tajfel and Turner, reflects the emergence of 
new social identity and new self-categorization. This is a psychological 
beginning of group existence. 
But the present conception suggests that social identity is not based solely 
on the fundamental group belief which reflects self categorizations of 
individuals as group members. As was presented, additional beliefs called 
group beliefs play a determinative role in the formation and maintenance of 
social identity. These group beliefs establish the raison d'etre for group 
existence and provide the rationale for individuals to be group members. 
Formation o
f
 additional group belie
f
s
 
Once the fundamental group belief 'We are a group' is formed, other group 
beliefs are usually added, though the contents of some additional group 
beliefs may be formed as antecedent beliefs prior to the formation of the 
fundamental group belief. As already indicated, the contents of group 
beliefs can be of various categories and of unlimited scope. Groups differ 
with regard to the repertoire of group beliefs that they form. They differ in 
the contents and in the quantity of group beliefs that they have. Each group 
has its own set of group beliefs. Although groups may have similar group 
beliefs, two separate groups must have at least one group belief which 
makes them different from each other. With regard to quantity, groups may 
have few group beliefs or a very long list of group beliefs. The selection of 
group beliefs depends very much on a group's goal(s), common experiences, 
ideas regarding symbols, emerging norms and values, constructed rational­
izations for group formation, perceived similarities, perceived uniting forces, 
and so on. 
One of the few studies of small groups from which the emergence of 
group beliefs can be inferred is the study of norm formation by Sherif 
(1936). He found that a group of strangers facing an ambiguous stimulus 
tended to converge their judgements into a uniform norm. That is, common 
experience in a group situation served as a basis for establishing a common 
social reality. Sherif concluded from these results that individuals who are in 
contact, in a defined place, form common values or norms as a basis for 
group formation. It may be further assumed that some of these norms and 
values may serve as group beliefs, since they may characterize the group 
and underlie its uniqueness. In this vein are findings obtained by Festinger 
and Thibaut (1951), who observed that 'belonging to the same group tends 
to produce changes in opinions and attitudes in the direction of establishing 
uniformity with the group' (p. 92). 
Another line of research by Sherif and his colleagues that contributes to 
the understanding of group belief formation was the Robbers' Cave 
experiments, which investigated intergroup conflict and cooperation (Sherif, 


Group belie
f
s
 as an expression o
f
 social identity 
107 
Harvey, White, Hood & Sherif, 1 961). The experiments also studied the 
formation of a group. In one of them, 22 fifth graders were randomly 
divided into two groups and brought to separate camps so that group 
formation would be the consequence of the bonding process within the 
ingroup without the contact of an outgroup. Within seven to eight days the 
formation of the two groups emerged and stabilized. As a point of interest, 
it can be noted that the two groups formed various group beliefs which 
defined their uniqueness and character. Thus, for example, the first group 
selected the name 'Tom Hale Rattlers', chose a flat, referred to several 
places as their territory, and developed a norm of 'toughness' (reflected in 
cursing and avoidance of reporting injuries). The second group called itself 
'Eagles', selected their own song, and selected places in their territory for a 
campfire and swimming hole. 
As Sherif et al. (1961) pointed out, these beliefs were produced from 
scratch. They were formed as a consequence of the interaction processes 
found in intragroup relations. They served as unique characteristics of the 
group and defined the substance of the 'groupness'. The particular group 
name, the flag, the norm or the territory, in addition to the belief 'We are a 
group', defined the essence of the group. For group members, these group 
beliefs served as a cognitive basis for their feeling of 'groupness'. The belief 
that they are a group, as well as other group beliefs, became part of their 
reality. It can be said that although the mere division into groups created 
the fundamental group belief, the addition of group beliefs not only 
strengthened the group identity, but also provided meaning to the we-ness 
reality. Group beliefs, thus, are an inseparable part of social identity. They 
provide the essence to the mere feeling of belonging to a group. 
It should be noted though that group beliefs do not have to be formed 
after the group was founded. Already it has been indicated that certain 
beliefs may function as antecedents for individuals to form a group and 
later become group beliefs. Also, individuals formulate beliefs which may 
serve as a basis for group formation. 
Numerous political parties, religious groups and volunteer organizations 
are based on group beliefs which were formulated before the group was 
formed. In this process, the belie
f
s
 are first f
ormulated, and then an attempt 
is made (sometimes it is even unintentional) to f
orm a group in which 
these belie
f
s
 serve as group belie
f
s. The following examples illustrate this 
process. 
The Jehovah's Witness group was founded in 1 872 by Charles Taze 
Russell in a suburb of Pittsburgh, Pennsylvania. Russell, who was brought 
up as a Presbyterian, was exposed to ideas of Adventism and Second 
Adventism until he formulated his own beliefs. These beliefs, which were 
later developed and modified by his successor, Joseph Franklin Rutherford, 
originally served as a basis for the formulation of a new denomination 
(Stroup, 1945). 
Cannon (1973) described how the Black Panthers formulated their group 
beliefs: 


1 08 
Social identity: international pers
pectives 
One afternoon in October, 1966, Huey told Bobby Seale, 'We need a program. 
We have to have a program for the people. A program that people can under­
stand. A program that the people can read and see, and which expresses their 
desires and needs at the same time.' That night they sat down in the North 
Oakland Poverty Center, where they were working, and wrote out the l O-point 
program which is still the basis for all Panthers' actions. (p. 339) 
The programme concerned freedom for the black community, full 
employment, the end of exploitation, decent housing, true education, the 
exemption of blacks from military service, the end of police brutality, free­
dom to black prisoners, establishment of black courts and self-determination 
(Cannon, 1973). 
In the present conception, an important question should be asked: Which 
common belie
f
s
 (that is, shared belie
f
s
 by group members) become group 
belie
f
s? In principle, many beliefs can become group beliefs in the group 
members' repertoire. In reality, however, relatively few beliefs achieve this 
status. Three conditions are suggested as an explanation as to which shared 
beliefs become group beliefs. One condition refers to the functionality of 
group beliefs. Belie
f
s
 which are f
unctional f
or group f
ormation and main­
tenance become group belie
f
s, since it is assumed that the need to belong and 
the need to f
orm social identity are important needs which underlie human 
beings' cognitions and behaviours. In this context, of special importance are 
beliefs which differentiate between the ingroup and outgroups, since they 
allow demarcation of the group boundary and provide significant infor­
mation about the group. 
The differentiating beliefs indicate the similarity of group members which 
overrides the differences between them. That is, through the psychological 
process of categorization, group beliefs simultaneously make the group 
members uniquely similar, yet distinctively different from other groups. 
They provide the individuals with a system of orientation towards their 
own group and other groups. This system leads to the perception and 
judgement of individuals of the same category as more similar to one 
another than they really are and enhances the perceptual and judgemental 
contrast between individuals not belonging to this category (Tajfel, 1 978d, 
1 981). An experiment by Allen and Wilder (1 979) demonstrated this 
phenomenon. They created two groups, ostensibly on the basis of artistic 
preferences (which served as group beliefs), and then had group members 
complete an opinion survey in the manner they thought another member of 
their own (or the other) group would respond to the items. They found that 
individuals assumed that another member of their group would express 
beliefs more similar to their own (previously assessed) opinions than would 
an outgroup member. This difference occurred even for belief items 
irrelevant to art. 
In addition, those beliefs which provide information about the group 
become group beliefs. It is no accident that group beliefs consist of contents 
which contain such elements as group history, group goals and common 
characteristics of the group. These contents provide information for group 


Group belie
f
s
 as an expression o
f
 social identity 
1 09 
members about themselves and their commonalities. They provide the 
raison d'etre for group formation and, later, for group maintenance. Also, a 
fulfilment of additional functions by a belief may especially strengthen its 
status as a group belief. Thus, for example, beliefs that help to achieve 
group goals, raise the self esteem of the group members or strengthen 
feelings of security are likely to become group beliefs. 
Another condition pertains to the saliencey o
f
 the belie
f
s
 which become 
group belie
f
s. Belie
f
s' saliency re
f
ers to those belie
f
s
 which draw special 
attention o
f
 group members because o
f
 their prominence and distinctiveness. 
In combination with the first condition, these beliefs are efficiently and 
swiftly absorbed and demarcate the boundary between the ingroup and the 
outgroups. In some groups the salient group beliefs may pertain to physical 
appearance, the most perceivable cues. Special clothing, ornamentation or 
even skin colour may serve as a salient basis for defining one's own group. 
In other groups, salient beliefs may refer to individual's needs (Toch, 1 965), 
group goals or common experiences. 
Finally, in order f
or belie
f
s
 to become group belie
f
s, epistemic authorities 
of group members have to consider them as such, and support them. 
Epistemic authorities are those knowledge sources who exert determinative 
influence on the formation of knowledge (Bar-Tal, Raviv & Brosch, 1 991; 
Kruglanski, 1989). Group members attribute high confidence to beliefs 
coming from epistemic authority, consider these beliefs as truth, and adopt 
them as part of their own repertoires. In our case, the concept applies 
mostly to leaders who perpetuate beliefs to become group beliefs. Political, 
intellectual, religious, social and cultural leaders determine to a 
considerable extent which beliefs become group beliefs and influence 
group members to accept them as such. They frequently select the group's 
goals which may serve as group beliefs, formulate ideologies or religious 
doctrines, decide what events in the past should be remembered as group 
beliefs for the future, select the symbols for the group, and decide on the 
attributes that characterize group members (for example, Y. Bar-Tal, 
1989). 
It is usually in the formative phase of group development that group 
beliefs are established. This is a period of malleability during which group 
beliefs are formulated and reformulated. With time, group beliefs emerge in 
a relatively stable formulation, although always with a possibility of change. 
In any event, as long as the group exists, it will always be faced with the 
talk of disseminating group beliefs and maintaining them among group 
members. 
These processes are necessary, since group beliefs serve as a foundation 
for group existence. The contents of group beliefs provide the raison d'etre 
for a group's existence and delineate its uniqueness vis-a.-vis other groups. 
The fundamental group belief indicates that the group exists. In addition, 
group members may add other group beliefs of a wide scope of contents. 
Group members share group beliefs, are aware of this sharing and believe 
that group beliefs define their 'groupness', that is, social identity. 


1 10 
Social identity: international pers
pectives 
Implications 
The proposed conception implies that group beliefs, as all beliefs, exist in 
the individual's mind. It does not suggest a view of group beliefs as a special 
entity of the 'group mind' category. Beliefs are related to individuals. 
Groups, organizations, societies or any other collectivity do not hold beliefs 
on the collective level - only individual members in the aggregate groups 
do. But the present conception describes a widely recognized phenomenon 
that group members share beliefs and these beliefs may be viewed as 
providing the basis for group members' social identity and as defining the 
essence of that group. Moreover, the presented approach suggests that 
group beliefs are more than a mere sum of group members' personal beliefs 
(see also Bar-Tal, in press). Sharing of a belief by group members and a 
recognition that it characterizes them provide group beliefs with 
distinguished properties. 
First of all, group beliefs serve as a basis for group existence. They 
provide a common basis with which individuals can identify and through 
which they can define their membership in the group. All groups have to 
have group beliefs, and a necessary condition for a formation of any group 
is formulation of group beliefs. This is in order that individuals should feel 
that they have something in common which distinguishes them from 
outgroups. They need at least one group belief - fundamental group belief. 
Tajfel's minimal group paradigm has demonstrated this effect. Using the 
minimal group paradigm, numerous studies have shown that the mere 
categorization of persons as ingroup or outgroup - that is, formation of a 
fundamental group belief - is sufficient to establish group identity and to 
influence group members' behaviour in the direction of triggering ingroup 
favouritism and outgroup bias (for example, Billig & Tajfel, 1973; Tajfel et 
aI., 1971). But in a real life situation group beliefs are needed in order to 
provide individuals with reasons for their group membership. Group beliefs 
supply the basis for cementing social identity. 
Second, group beliefs tend to arouse high confidence in their content. The 
perception of group members that a group belief is shared and characterizes 
them causes them to ascribe it high validity. The recognition that a belief is 
shared by group members reduces the subjective feelings of uncertainty and 
increases the perceived validity of this belief (see Festinger, 1954; Jones & 
Gerard, 1967). Therefore, group beliefs are usually considered by group 
members as facts, truths or verities. Another factor which contributes to this 
view is the perception that group beliefs define the essence of the group. 
Beliefs of this type cannot be held with low confidence, since low confidence 
may shatter the basic foundation of group existence. Therefore, shared 
beliefs which define the essence of the grouping are held with high 
confidence. 
One consequence of holding group beliefs with high confidence is their 
relative freezing in the group members' repertoire. Group beliefs do not 
change easily, since freezing implies closure. That is, group members do 


Group belie
f
s
 as an expression o
f
 social identity 
I II 
not easily entertain alternative hypotheses to group beliefs, but tend to 
collect information which validates them. The change of group beliefs is of 
special difficulty when they are central. The centrality of group beliefs 
indicates their importance for the group's life. Central group beliefs may 
become part of the group's culture and tradition, and then they are espe­
cially enduring. Group members pass them from generation to generation 
and rarely change them. 
Third, since group beliefs define the essence of the group, group members 
tend to organize their personal beliefs in accordance with them. Group 
beliefs provide the frame of reference for other beliefs. This process takes 
place especially in cases when group beliefs are central in group members' 
repertoire. Then, group members draw implications from group beliefs and 
add new personal beliefs consistent with them. In extreme cases of an 
orthodox group, group beliefs serve as anchors around which other beliefs 
are organized. 
Fourth, group beliefs provide the group members with special powers. 
The mere perception that group members share group beliefs indicates 
strength. Converse (1964) suggested in his classic analysis of beliefs that the 
number of people associated with a particular belief system is an important 
factor in political system: 
claims to numbers are of some modest continuing importance in democratic 
systems for the legitimacy they confer upon demands; and much more sporadi­
cally, claims to numbers become important in nondemocratic systems as threats 
of potential coercion. (p. 207) 
In addition, the power of group beliefs is derived from the perception of 
unity and commonality that may characterize group members who are 
aware of sharing beliefs. Knowledge that group members share a given 
belief expresses a unity of the group and indicates common fate. Group 
members derive strength out of this knowledge. They feel united, belonging 
and integrated within the group. One consequence of these feelings of 
strength is that group beliefs may serve as a basis for demands, desires or 
goals of group members. Leaders of the group usually take into con­
sideration the group's beliefs when they make decisions that affect the 
group's life. Being aware that group members share a belief, and are 
influenced by the belief in their group behaviour, leaders pay special 
attention to group beliefs. Group beliefs reflect the direction that group 
members desire to take in their behaviour. Therefore, leaders frequently 
make decisions regarding group behaviour that correspond to group beliefs. 
Fifth, group beliefs may determine the attitudes and behaviours of an 
outgroup towards the group. Group beliefs are important information for 
outgroups about any given group. They characterize the group and may 
imply even the behaviour that the group may take. Group beliefs may 
indicate possible goals, ideology, values, history, norms or characteristics of 
the group. This information enables acquaintanceship with the group and in 
turn influences the type of intergroup relations that may develop. 


1 12 
Social identity: international pers
pectives 
Finally, if one believes that group members act according to their beliefs, 
then group beliefs should be considered as an important source of under­
standing group behaviour. As Krech, Crutchfield and Ballachey (1962) 
suggested: 
Man acts upon his ideas. His irrational act no less than his rational acts are guided 
by what he thinks, what he believes, what he anticipates. However bizarre to 
the behavior of men, tribes or nations may appear to an outsider, to the men, to the 
tribes, to the nations, their behavior makes sense in terms of their own world view. 
(p. 1 7) 
Group beliefs provide the cognitive basis to many group behaviours. They 
may serve as reasons, goals, explanations or justifications for group behav­
iour. In other words, group beliefs may function as guiding forces for a 
group, and, therefore may determine the direction, intensity and persistence 
of group behaviour. This is one of the most important implications of group 
beliefs. 
Conclusion 
Group belief conception provides cognitive and social perspectives to the 
study of a group. The existence of a group is a social reality for group 
members. Group beliefs provide the basis which allows group members to 
view the group as a social reality. The group, then, is a product not only of 
structural characteristics, environment, situational conditions, motivational 
tendencies or social influence, but also of personal cognitive processes. 
These processes determine the essence of the group, since group beliefs 
provide the contents which serves as a basis for group formation and group 
maintenance. 
This conception carries the notion of social identity one step beyond the 
Tajfel and Turner theories. It suggests that social identity is not based 
merely on the categorization process, but also on beliefs which provide the 
rationale for the group existence. The fundamental group belief indicating 
'We are a group' and additional group beliefs of different contents which 
define the nature of the groupness allow individuals to feel, think and act as 
group members. These beliefs underlie the uniqueness of the group and 
diff
erentiate it from outgroups. 
It is important that behavioural scientists study the contents of group 
beliefs and the process through which group members acquire them, become 
aware of sharing them, and are affected by them in their behaviour. This 
line of interest will, on the one hand, add a cognitive aspect to the study of 
a group and, on the other hand, liberate beliefs from their individual closet 
to extent their social meaning. In this respect, the present approach provides 
the bridge between the individual's level of analysis and group level. Group 
members, as individuals, are the ones that acquire beliefs which shape their 
reality. But when group members become aware of the fact that their beliefs 
are shared by other group members, these beliefs acquire special quality. 


Group belie
f
s
 as an expression o
f
 social identity 
1 1 3 
That is, they may have important cognitive, affective and even behavioural 
implications for the group members as individuals and for the group as a 
whole. 
Note 
I .  Most of the examples provided in the following review are drawn from formal statements 
prepared by the groups. Although it is possible that they reflect group beliefs, group members 
were not examined to indicate their responses. Therefore, the presented 'group beliefs' should 
be treated with caution. They only serve as illustrations for presenting the conception. Further 
research among group members may determine the actual status of these beliefs. 


8 
Social Identity and National Identity 
Jose Miguel Salazar 
Of all the variants of social identity, national identity is perhaps the one 
that has had the most dramatic impact on historical events, with both 
positive and negative consequences for humankind. 
The great romantic-nationalist movements of the nineteenth century, as 
well as the movements of national liberation in the first half of the twentieth 
century and the most recent break-up of multinational European states, are 
clear examples of the past and present role played by this type of identity. 
Identity with national or ethnic categories is evidently also a reference point 
used by Tajfel in the development of his theory, as inferred from the 
character of some of his early work (Tajfel, I 
969b, I 
970a). 
Even though the eighteenth and nineteenth centuries have been called the 
era of nationalism (Kohn, 1 962), this designation could also apply to the 
present century, as there is ample evidence of creative manifestations and 
excesses of the national idea in our time. In the aftermath of the Second 
World War, nationalism and independence were bywords of the movements 
to reaffirm identities in countries which were colonies in the previous 
century, particularly in Asia and Africa. Yet not only in the so-called third 
world is evidence found of the ascent of this type of identity; it is increasingly 
noticeable in the awakening of subdued nationalisms in European countries, 
and in the breaking-up of previously functioning multinational states. Also 
in the Americas there are clear manifestations of its increasing importance 
from Canada, to Mexico, to Peru; and even in the Pacific area with the 
awakening of the aboriginal populations of Australia, silent for centuries. 
Whoever suggests that the era of nationalism is finished and that we are 
in a new era of universal humanity must be disconnected from his or her 
surrounding reality. The allegations of 'proletarian internationalism' which 
proposes that the proletariat has no country, as well as of 'cosmopolitan­
ism', which indicates that as a consequence of transnationalization now it is 
the bourgeoisie that has no country, are both questionable. On the contrary, 
the so-called process of globalization and the development of the 'global 
village', based on technological improvements, constitutes a good example 
of the dialectics of the historical-cultural process, since it has had the 
curious effect of reinforcing ethnicity and national identities. 
It has been claimed that nations and nationalism only acquire reality in 
relation to the state and that historically these are intimately connected with 


Social identity and national identity 
1 1 5 
the ascent of the bourgeoisie, and should be considered transient phenomena 
only referred to as a stage of historical development (that is, Kedourie, 1 960; 
Kohn, 1 944; Seton-Watson, 1965). Yet it is possible to differentiate between 
nationalism, or more precisely 'modern nationalism', clearly associated with 
the state, and 'national sentiment', which is more ancient and of a more 
general and less political character (Hertz, 1 944). 
The emphasis on the nation-state is important from the perspective of 
elaborating a political history, but in the study of national phenomena from 
a socio-psychological perspective it is impossible to ignore this other facet. 
Hence in this chapter we shall begin by considering the affective back­
ground of national phenomena. After that, in order to gain a better under­
standing of national identity as such, some developments within the social 
identity approach will be examined in order to answer some questions about 
the problem in hand. 
Considerations about national sentiments 
The definitions of what a 'nation' is have been very varied. The term 
appears in a fifteenth-century Bible translating a Hebrew word that means 
at the same time 'ascendance' and 'posterity'. It is also used to refer to a 
natural quality, a native quality. In the medieval universities a 'nation' 
referred to a group of students from a common region. And the idea of 
belonging to a group based on descendance, differentiable from others, was 
present and important in history long before the appearance of modern 
nationalism (Smith, 1 99 1 ;  Chap. 2). 
Even in the case of empires as ancient as Egypt and China, national 
sentiment was important, at least for members of some strata. Evidence of 
this is found in some hieroglyphs about a Theban prince who expelled the 
hyksos from Egypt around 1 580 Be; they say: 'I will grapple with them, and 
cut open their belly! I will save Egypt and overthrow the Asiatics' (cited by 
Smith, 1 991, p. 46). Even in Greece, being an Athenian or a Spartan had 
emotional and behavioural implications; the feeling of being an alien, 
dramatized in Euripides's Medea, must surely reflect a reality understood by 
the audience of the drama. 
On the other hand, the biological basis of national sentiment, and more 
concretely of ethnocentrism, has been considered by social scientists 
interested in socio-biology. Campbell (1965) in his provocative essay 'Ethno­
centric and Other Altruistic Motives', elucidates the thesis that ethnocentric 
motivation and behaviour is directed to the protection and survival of the 
members of the ingroup, leading to sacrifices and even death. This act by 
definition is an altruistic type of behaviour, going against personally selfish 
interests. More recently Reynolds, Falger and Vine (1987) have published an 
interesting book, The Sociobiology o
f
 Ethnocentrism, that gathers more up­
to-date contributions from this same perspective. 


1 16 
Social identity: international pers
pectives 
There has been great insistence that 'nationalism' is a phenomena of 
modernity, and that it is incorrect to consider it something more trans­
cendental. This is a way of dodging considerations derived from the 
biological condition of humankind and the fragmentary but unquestionable 
evidence of the antiquity of 'national sentiments'. There is no doubt that 
feelings of attachment to the nation-state are a relatively recent phenom­
enon since the objects of attachment ('modem' nations and the 'nation­
state') are also recent, but to deny the attachment to what is perceived as 
one's own (be this a city-state, as in Greece, or a tribe in the Amazonian 
forest) seems exceedingly restrictive. To do this does not take into account 
some basic considerations about the human condition. 
In speaking of a 'nation', reference to the fact of 'birth' is important. 
(Although in some languages this relationship has been lost or never existed, 
it is very clear in Spanish [nacer = to be born; nacion = nation] and in other 
Romance languages.) There is a determination given by belonging to a 
group in which one was born, an ascribed role that confers an important 
anchor to the individual. So although the term 'nationalism' in effect 
appears in common usage around the period of the French Revolution, the 
term 'nation' is much older. The concept of nationalism is endowed with 
clear political references to the nation-state, which is not necessarily the case 
for the concept of nation. 
There are groups that in the accepted sense constitute nation-states, 
but there are other conglomerates, based on the same principle of birth, 
that may generate loyalties and attachments as strong as, or sometimes 
even stronger than, the nation-state. The following three cases may be 
mentioned: 
(a) the positive affect towards a regional group within an existing state 
('regionalism'); 
(b) the attachment to a grouping that ignores existing geographical borders, 
as in the case of the Kurds of the Middle East or the Waiyu in the 
Goajiro Peninsula that Venezuela and Colombia share (,stateless 
nationalism'); 
(c) the positive feeling towards supra-national groupings, as in the case of 
Latin America ('supra-nationalism'). 
We consider that there are certain commonalties in the psychological 
contents of the sentiments expressed in these cases, as well as in the case of 
state nationalism. And it is this broadly defined national referent which is in 
our mind when we examine the question of 'national sentiment' in this 
chapter. 
So what is the basis of this national sentiment? If we take a close look, we 
find that it is based on four elements: (a) territoriality, (b) shared culture, (c) 
historical memory of genealogical communality and (d) the existence of a 
nation-state. 


Social identity and national identity 
1 17 
Territoriality 
Ardrey (1967) extensively develops the point of the importance of territory 
in human behaviour, going back to the biological base of animal terri­
toriality and making use of the concept of territorial imprinting. He defines 
what he calls the 'biological nation' as 'a social group containing at least 
two mature males which holds as an exclusive possession a continuous area 
of space, which isolates itself from others of its kind through outwards 
antagonism, and which through joint defense of its social territory achieves 
leadership, cooperation, and capacity for concerted action' (p. 1 91). His 
argument is centred on an evolutionary position and observations of 
phylogenetic ancestors, which from his point of view give evidence of the 
antiquity of the phenomenon 'nation'. Of course, like many other animal 
analogies, this may be questioned, especially if it is sustained that language 
and culture introduce a qualitative break with evolutionary continuity. Yet 
the observations continue to have interesting implications. 
The excesses of the approaches centred on geographical determinism, in 
vogue in the nineteenth century, and its later rejection, have made us lose 
sight of the importance of the geographical environment in relation to 
behaviour. Even though the geographical referent is more tangible and 
objectifiable than other referents of nation, there are few authors who give it 
the relevance it deserves, one exception being Doob (1964), who bases his 
studies about patriotism in Tyrol on the attachment to the 'Heimat'. Also 
there has been some recent rethinking about the theme amongst environ­
mental psychologists (Altman & Low, 1992; Proshansky, Fabian & 
Kaminoff, 1983). 
There are multiple cases in history of the fight to hold on to territories by 
people who were born in an area, which may not be the best from the point 
of view of economic rationality; and there are many well-known examples 
of the feeling of ostracism of migrants or displaced peoples who must live in 
a territory which is not their own. 
But while it is very concrete, the territorial referent is more evident in the 
regionalism variant of national loyalty. Given geographical variability 
within countries, it is logical that the role of territory and all that is con­
nected to it should be more marked in more restricted loyalties. 
Altogether there is no doubt that territoriality plays an important role in 
the development of national sentiments, 
Culture 
Sharing a culture involves collective ways of solving the problems con­
fronted in interacting with the surrounding world, developed and trans­
mitted from generation to generation. This level of shared experience is also 
important in the development of national sentiments. 
Language is the supreme creation of culture, so much so that for many 
anthropologists it is considered its defining trait. In any case it permits the 
transmission of collective solutions and their preservation. Language is also 


1 1 8 
Social identity: international pers
pectives 
a very early acquisition in human ontogenesis, and so it becomes a basic 
point or matrix for the development of feelings of attachment. 
There are other elements of culture that become obvious referents of 
attachment: music, different elements of the so-called 'material culture', 
customs, shared values. They may all be very powerful in awakening 
positive feelings and loyalties towards the national ingroup. 
Ethnicity 
Given the conceptual weaknesses and dangerous political use of the term 
'race', the concept of ethnicity has become a partial substitute. Smith 
(I 
984a, 1984b) develops the idea of 'myths of descent' as the basis of ethnic 
feeling and ethnic nationalisms. Through these myths, the members of a 
community perceive they have a common origin, usually connected to a 
glorious past. An ethnie is structured around these mythological socio­
cultural elements, the association with a common land, elements of 
common culture, and something of great relevance: a 'collective proper 
name' (Smith, 1 991, p. 21). It must be clear that in talking about myths no 
attempt is made to disprove the importance of the phenomena, since myths 
are not to be ignored. As Mariategui (1925), a very important Peruvian 
thinker, said: 'Myths move mankind through history. Without a myth, the 
existence of mankind has no historical sense' (p. 22). Hence considerations 
about ethnicity cannot just be discarded, and obviously are connected to the 
development and maintenance of national sentiments. 
State 
It is indisputable that the existence of a national state stimulates an 
important type of national sentiment through the systematic use of ideology 
producing institutions partially in the hands of the state like the school 
system and the mass media. This is so pervasive that for some the only 
really important national sentiments are those related to the state. But this 
is too restrictive, as regional, ethnic and more recently supra-national 
identities may have as much reality, from a psychological perspective. 
There is an additional problem in restricting the study of national 
sentiment to what occurs at the level of the nation-state, and it is that the 
idea that a nation equals a state is not a viable equation in our present 
world. It constitutes, at most, an ideal type, but the great majority of states 
are really multinational. Some contemplate it explicitly, as Canada or 
Spain, where there is recognition of the existence of diverse nationalities 
within the state. However, in states such as Venezuela the fiction of a 
culturally homogeneous state is disturbed with news of massacres of 
indigenous tribes, such as the Yanomami. Some states that are explicitly 
multinational apparently function well, although surprises occur when the 
latent centripetal forces gain ascendance over the forces of integration, as in 
the painful case of Yugoslavia. 


Social identity and national identity 
1 19 
In defining a nation, there is implicit the existence of one or several non­
nations and the existence of limits separating them. When a nation is 
defined in terms of a state, it is usually clear what the limits are. But the 
same doesn't happen when you base your definition on any of the other 
elements: the territorial, the cultural or the ethnic. Hence it becomes 
necessary to look for theorizings capable of suggesting ways of solving this 
problem. 
Social identity theory frames the problem of 'borders' in socio­
psychological terms. Consequently this theoretical approach is essential in 
any attempt to understand national phenomena. Hence we shall consider 
very briefly some elements of this approach which may help us to 
understand better the question in hand. 
The social identity approach and its relevance to understanding 
national phenomena. 
The identity theory approach is relevant to the problem of defining group 
borders since it places it at the irrefutable level of human subjectivity. This 
theoretical approach, which in its beginnings Tajfel sometimes referred to as 
CIC theory (categorization-identity-comparison) (Tajfel, 1 98 1 )  has 
stimulated multiple theoretical and empirical developments, examples of 
which can be found in the other chapters of this book. 
The three original basic elements, the existence of a category accepted 
internally and externally, the identification with it, and the process of social 
comparison, are still very fruitful concepts. The first element becomes 
concrete in the form of a label; the second element implies the acceptance of 
the category for oneself with all the affective, cognitive and behavioural 
consequences. Regarding the third element, the process of social com­
parison, it is frequently a positive evaluation of the assumed category, but 
this is not always the case. Tajfel (1981, Chap. 1 3) himself presents evidence 
of studies with children in which there is evidence of devaluation of their 
own ethnic group. Several studies have produced evidence of devaluation of 
the ingroup (Salazar, 1 983a). The postulation of the need to maintain a 
positive group evaluation seems to have been taken from Sumner's (1906) 
ethnocentrism theory, considered by many as constituting a very basic 
means of understanding intergroup relations (Le Vine & Campbell, 1 972). 
Many developments have occurred on the basis of the original elements 
of social identity theory, among these accounts regarding the basis of 
categorizations (Billig & Tajfel, 1 973; Turner, Hogg, Oakes, Reicher & 
Wetherell, 1 987), meta-contrast ratios (Turner et ai., 1 987), the acquisition 
of categories, the hierarchy of identities; the process of depersonalization, 
the salience of categories, and so on, are very relevant to understanding 
national identity. 
Making use of some of these developments, we have attempted answers 
to some questions related to the understanding of national identities. 


120 
Social identity: international pers
pectives 
Questions about national identity 
1 .  What is the motivation that underlies the acquisition o
f
 national identities? 
Kelman (1969) considers that it is based on both perceived instrumentality 
and a need for transcendence; the latter would give a particular force to the 
social identities related to the national. With a bit of poetic licence, it may 
be said that the existence of national communities is based on the desire for 
eternity. So the idea of some to conceive of nationalism as a form of religion 
(Hayes, 1 926) is not far-fetched; nationalism may be conceived as a lay 
religion that offers, like the others, the promise of eternity. 
The nation as a community links people not only with a present inhabited 
by other people they do not and could not possible know, but also with a 
past and a future. This is possible through depersonalization, derived from 
the process of self-categorization. In reality that which transcends, that 
which is eternal, is the group, but when this happens this 'eternity' gets 
transferred to the depersonalized individuals that are part of it. 
This may be seen as an attempt to postulate a basic religious motivation 
in human beings. But whether religiosity is something intrinsic to human­
kind or is a historical phenomenon is not important. What is argued here is 
that national identity is based on an aspect of such religiosity, basically the 
need to transcend. 
In more mundane terms, the 'transcendence' of a nation may be con­
ceived of as attainable by the power it exerts or has been able to exert 
within the international context. Hence national identities have their more 
powerful manifestations related to political states, which are by definition 
conceived of in power terms. 
2. On what basis do the perceptions o
f
 similarities and dif
f
erences arise in 
the case o
f
 national identities? 
It is known that a dialectical relationship exists between categorizations, 
expressed in a label and the perception of similarities. The mere existence of 
a label is not enough, for a base must exist to give justification or backing to 
assuming the category. It is also evident that the dimensions around which 
the comparisons take place vary in terms of the identities considered. 
Humans' capacity for interaction is limited. Dunbar (1992), making a 
projection on the basis of the size of grouping in higher primates, suggests 
that 1 50 is the limit of a viable human group, something like the natural 
size of a human group. Nevertheless, humankind lives in communities much 
larger than that, hence necessarily we live in 'imagined communities' 
(Anderson, 1983). An ethnic community or a nation, is hence one of these 
'imagined communities', but even these must be construed around certain 
elements or dimensions. 
When the question of the basis of national sentiments is analysed, four 
elements come to the fore: territory, culture, ethnicity and state. These 
elements may be considered as dimensions along which meta-contrast ratios 
of similitude-difference may be established. Particularly in the cases of 
culture and ethnicity, and in a lesser degree in the case of territory, it is 


Social identity and national identity 
121 
possible to conceive of continuums in which the meta-contrast ratios 
gradually change. In the case of the state the categories are discrete and 
variations in meta-contrast ratios are less applicable. 
3. Could we conceive o
f
 levels or hierarchies in the case o
f
 nation-related 
identities? 
In order to encompass the similitudes between state nationalism, region­
alism and supra-nationalism, we may speak of 'nation-related identities'. 
These identities tend to occur in concentric circles. The idea of concentricity 
of 'national-related' identities could be considered as a particular case of 
hierarchies within the intermediate level, taking as the higher one that of the 
individual as member of humanity, and the lowest that of the individual 
self. 
We suggest that there exists a common psychological substratum to 
identities centred on a region or a specific regional group, and the one 
centred on a specific nation-state or the emergent supra-national identities. 
We see this as a matryoshka of identities, one inside the other. 
The relative importance of these identities is not permanent and situ­
ational factors play an important role in determining their role or effect on 
behaviour. 
4. How does the process o
f
 social comparison f
unction in the case o
f
 
national identity? 
In the case of national identities the question of social comparisons is 
particularly important. Social comparisons will affect both the salience of 
the category as well as its evaluation, and influence the development of a 
positive social identity. The social comparisons that are established involve 
other nations or ethnic groups. These comparisons may take place along 
different dimensions: size of the countries, political systems, state of the 
economy, 'well-being' or 'happiness' of the inhabitants, and so on. In 
different cases the activated meta-contrasts will vary. 
But in the activation of 'nation-related' concentric identities, a basic 
element to be considered is the degree of transcendence that identity may 
guarantee without losing the essence of the identity. It may be argued that 
the more extensive the reference group, the more positively valued will be 
the identity, since the possibility of transcendence is greater. But it is also 
true that the smaller the unit, the greater the homogeneity, and hence the 
more solid the meta-contrasts. There must exist an equilibrium point 
between both elements, possibly of transcendence and solidity of identity, 
which constitutes an interesting empirical problem. 
There is evidence that when Latin Americans and Venezuelans (or 
Peruvians or Colombians) are evaluated, the social identity is more positive 
in the first case (Salazar, 1 983b), and it may be that this is due to an 
underlying perception of greater possibility of transcendence at that level 
than at the more restricted ones. Something similar seems to be occurring in 
Europe. 
But as has been frequently pointed out, positive social identities may be 
connected to negative manifestations of aggressive ethnocentrism. The 


1 22 
Social identity: international pers
pectives 
development of an identity within a wider context redefines the outgroup, 
and here the European case is illustrative. The strengthening of the sense of 
being European, a psychosocial correlate to the development of the 
European community is evident (Echebarria, Elejabarrieta, Valencia & 
Villareal, 1 992; Fells & Niznik, 1 992). This has led to the channelling 
towards non-Europeans of feelings of rejection that in previous periods had 
been directed to other groups within Europe. The racist manifestations 
against Turks in Germany, against North Africans in France, are cases in 
point. In Spain the development of 
the feeling of being a part of Europe has 
been accompanied by a rejection of Dominicans and other 'South 
Americans' that in another definition of the situation, that is, in terms of a 
Hispanic community, were considered less objectionable. 
The varying processes of social comparison point to the importance of 
exploring the consequences that may be derived from placing the same 
individual in varying referential frames, be these regional, national or supra­
national. 
5. In what conditions do dif
f
erent 'nation-related' identities become salient? 
To begin with, it must be accepted that nation-related identities become 
activated by field factors, situational factors that affect the salience of a 
particular category. In the case of the European community, economic and 
political developments activated a supra-national identity that has been 
latent, and submerged under a history of conflicts. Obviously identities are 
politically manageable. However, for this to occur they must exist in a 
latent form. 
In defining the ingroup, there exists an implicit definition of outgroups. 
Yet these definitions are not static. In emphasizing one level of the nation­
related identities, a group that shares the same identity at another con­
centric level could be defined as an external group. An example makes this 
clear. The identities of a Caraquefio (from Caracas), a Venezuelan and a 
Latin American are nation-related and subsist in concentric circles. When 
my Caraquefio identity comes to the fore, I consider people from other 
regions of my country as external groups, with possible behavioural conse­
quences. This ceases to be so when my Venezuelan identity gets activated, in 
which case a Colombian, or another national group, gets defined as the 
external group. This externality would disappear when Latin American 
identity is invoked, since it encompasses both groups at a higher level of 
abstraction. 
For a Latin American living in Europe, the Latin American identity is 
easily activated; in a Latin American country different from ours the 
Venezuelan identity becomes frequently activated; in a gathering taking 
place in a small provincial town, the regional identity of coming from the 
capital city may come to the fore. It is here that the questions of accessi­
bility and ad
justment come into play. 
But there are qualifications that must be made in relation to the thesis of 
concentricity. It is erroneous to think that there exists perfect balance or 
concordance within the concentric circles. On the contrary, it must be 


Social identity and national identity 
123 
recognized that sometimes these may appear as contradictory and in effect 
enter into conflict. In this sense the Basque case is particularly significant, 
where for many this identity is independent of the Spanish one, and hence 
does not function in concentric circles. According to Linz (1986), a high 
percentage of the inhabitants of the Basque provinces do not consider 
themselves Spaniards. Nevertheless the idea of concentric circles probably 
functions with regard to the European identity (Torregrosa & Ramirez, 
199 1). 
Another problematic situation concerns the conflict between national 
identity and other types of identity, such as those related to partisan politi­
cal loyalties. Here these may take precedence and lead to dramatic situ­
ations as those of a civil war, as has happened so many times in history. 
National identity may also come into conflict with human identity, as 
would be the case with a soldier who refused to comply with the order to 
kill someone on the basis of their national label. 
This brings to the fore the question of 'globalization'. The interrelation 
between the different countries in the world has always existed; what is new 
is the development of technologies that bring us close enough to find out 
what is happening in other places in the world with a speed non-existent in 
previous times. But while it may be true that this facilitates the imitation 
and copying of foreign models, usually from more developed countries, it 
also has created consciousness of similitudes and differences, and does not 
necessarily lead to homogenizations. A more probable outcome, it seems to 
me, would be the development of wider communities, on the basis of 
common elements, in which the sense of national identity is not eliminated, 
but transferred to a higher level. 
In my view, homogenization is not a desirable goal. Rather, I would 
argue that the recognition and acceptance of differences is preferable to the 
universal assimilation of all to a single pattern. 


PART III 
IDENTITY AND GROUP/ 
INTERGROUP PROCESS 
9 
Social Identity and Intragroup 
Differentiation as Subjective Social 
Control 
Jose M. Marques, Daria Paez and Dominic Abrams 
In analysing collective behaviour associated with religious movements III
 
sixteenth-century Europe, the historian 1.1. Delumeau (1993) wrote: 
The Protestant uprising naturally provoked an increasing aversion to heresy, 
which was considered as the ultimate evil . . .  
: the traitor inside is worse than the 
enemy outside. He must be made to leave his hide-out, he must be eliminated as a 
priority and no other punishment is severe enough for him. (pp. 514-51 5) 
This statement illustrates the fact that the deviate members of a large 
community may be more harshly treated than outsiders, because, while both 
depart from the community's values, the former threaten these values more 
severely than the latter. Deviates' behaviour and beliefs endanger the 
collective value standards more to the extent that these deviates cannot 
simply be expelled to the outgroup. 
Punishment of deviates and traitors appears to be an important com­
ponent of a community's social regulation system. Another historical illus­
tration of this phenomenon is witch-hunting in the Middle Ages (for 
example, Hamilton & Rauma, 1995; Yamagishi, 1 995). Witch-hunting, as 
well as more recent instances of collective persecution - like the Inquisition 
in the fourteenth- to sixteenth-century Iberian countries, McCarthyism in 
the United States in the 1 950s, or political purges in the Stalinist period 
in Soviet Union, to mention but a few - often find support from the general 
population, or, at least, from those who identify with the value-system that 
legitimates the persecution, even when the persecuted do not represent an 
immediate or material danger. This support is sometimes tangible but, most 


Social identity and intragroup dif
f
erentiation 
125 
often, emerges in the form of derogatory attitudes and judgements. It likely 
stands for the individuals' commitment to the community's standards and 
self-definition in light of these standards (cf. Doise, Chapter 2, this volume). 
Concomitant with this social self-definition is the effort to ensure and to 
legitimize beliefs on the distinctiveness and relative superiority of the 
community's standards as compared to the outgroup. This process may also 
be viewed as a reaction against social change and as an immediate psycho­
logical response against threatening minorities cf. Perez and Mugny, 
Chapter 10, this volume). It may be a major factor in the maintenance of a 
clear-cut and secure social identity as well as a guarantee of the individual's 
psychological well-being as a group member. As sociologist Georg Simmel 
(1955) put it, in groups structured around their members' attitudes. 
the opposition of a member to an associate is no purely negative social factor, if 
only because such opposition is often the only means for making life with actually 
unbearable people at least possible. . . . It allows us to prove our strength 
consciously and only thus gives vitality and reciprocity to conditions from which, 
without such corrective, we would withdraw at any cost. Opposition achieves this 
aim even where it has no noticeable success, where it does not become manifest 
but remains purely covert. Yet while it has hardly any practical effect, it may yet 
achieve an inner balance . . .  
, may exert a quieting influence, produces a feeling 
of virtual power, and thus save relationships whose continuation often puzzles the 
observer. (p. 19) 
In the present chapter, we argue that identification with ingroup norms 
and values may lead individuals to subjectively engage in processes analog­
ous to social influence upon actual or represented ingroup instances. These 
subjective social influence processes may allow individuals to restore their 
confidence in the group and, hence, to re-establish a positive social identity. 
The social psychological analysis of small groups and large social 
categories 
Social psychological research has focused on different factors involved in 
the explanation of group processes. Traditionally, this research evolved 
along two separate lines which, for simplicity, we may designate as the 
'small-group approach' and the 'social identity framework' (including self­
categorization theory). The former has focused mainly on processes like 
productivity, decision-making, role and status differentiation, cohesiveness 
or socialization (cf. Brown, 1 988; Cartwright & Zander, 1 968; Levine & 
Moreland, 1994). Most of this work has been conducted in the realm of 
face-to-face groups. The latter has emphasized the effects of social cat­
egorization in the psychological dynamics involved with individuals' sense 
of group membership. This work has often been conducted with large social 
categories and their representations (for example, Tajfel, 1978d). Only 
recently have these two lines been subject to integrative efforts (Hogg, 1992; 
Hogg & Abrams, 1 988; Marques & Paez, 1994). 


126 
Social identity: international pers
pectives 
Group affiliation, social influence and intragroup differentiation 
Research in small groups was strongly influenced by the work of Festinger 
and colleagues. Festinger (1950) proposed that affiliation in groups fulfils a 
social reality f
unction, that is, uncertainty reduction through selective 
affiliation with those who share similar beliefs, and a locomotion function, 
that is, facilitation of goal achievement through the association with those 
people whose goals are similar to the individual's. Indeed, people in small 
groups generate consensual opinions (Festinger, Schachter & Back, 1950) 
and reject deviates who resist group pressures to comply with the group's 
position on a relevant matter of opinion (Schachter, 1951), or who inten­
tionally fail to contribute to the group's goal achievement (Jones & 
DeCharms, 1967; cf. Levine, 1980). Most importantly, evidence shows that 
increases in external pressures upon the group, in group uniformity and in 
the discrepancy between majority and deviate positions increase the 
majority's pressures towards uniformity (Cartwright, 1968; Janis, 1968; 
Levine, 1 980; Moreland & Levine, 1982; Shaw, 1976). Evidence also suggests 
that social reality and group locomotion are ensured within the group 
through two forms of social control, traditionally known as inf
ormational, or 
private, and normative, or public, influence (Jones & Gerard, 1967). How­
ever, research in small-group processes has rarely addressed itself - at least 
conceptually - to the role played by intergroup relations in processes 
occurring within groups (Hogg, 1992; Hogg & Abrams, 1988; Marques & 
Paez, 1994). 
Social categorization, social identification and intergroup 
differentiation 
In tum, research on large-scale categories has emphasized intergroup more 
than intragroup processes. This may be due to the influential work of Tajfel 
and colleagues (for example, Tajfel, 1978d; Tajfel & Turner, 1979; Turner, 
1975a). This work started from the assumption that social behaviour may 
be conceived of as occurring along one continuum, whose poles are inter­
group and interpersonal behaviour: 
What is meant by 'purely' interpersonal is any social encounter between two or 
more people in which all the interaction that takes place is determined by the 
personal relationships between the individuals and by their respective individual 
characteristics. The 'intergroup' extreme is that in which all behaviour of two or 
more individuals towards each other is determined by their membership in 
different social groups or categories. (Tajfel, 1978a, p. 41) 
Whereas interpersonal behaviour involves the perception of features 
specific to the individuals involved, intergroup behaviour implies self- and 
other-stereotyping. In categorizing themselves as members in a social 
category, people focus on those features that they share with all category 


Social identity and intragroup dif
f
erentiation 
1 27 
members. Brown and Turner (1981) advanced this idea under the heading 
of the self-stereotyping hypothesis. 
In brief, the social identity framework proposes that group representa­
tions correspond to prototype-like representations, because intergroup 
similarities and intragroup differences are irrelevant in intergroup situations. 
Clearly, this assumption restrains a view of intergroup and intragroup 
processes as simultaneously occurring in social life. Although this is not 
entirely true (Hogg, 1 992; Turner, 1 99 1), it nevertheless remains that the 
social identity framework (including self-categorization theory) has diffi­
culties in dealing with the simultaneous emergence of intergroup and 
intragroup behaviour. 
Group prototypes as denotative norms 
Among other well-known assumptions, the social identity framework states 
that, in social life, people attempt to achieve cognitive clarity by exagger­
ating intergroup differences and intragroup similarities (Hogg, 1 992; Doise, 
Deschamps & Meyer, 1 978; Tajfel, 1 969a, 1 978d; Tajfel & Wilkes, 1 963). 
Basic to this process is comparative fit (Oakes, Haslam & Turner, 1 994, and 
Chapter 6 this volume; Oakes & Turner, 1 990). Comparative fit stands for 
the actual differences and similarities between the stimuli present in the 
judgement setting that allow the establishment of correlations between 
continuous stimulus properties and discrete categories (cf. also Tajfel, 
1969a). 
According to Turner, Hogg, Oakes, Reicher and Wetherell (1 987; cf. also 
Hogg, 1 992; Hogg & McGarty, 1 990), comparative fit generates the cog­
nitive construction of group prototypes. This process has been formalized in 
terms of a meta-contrast princi
ple (Turner et al. 1987; cf. also Hogg, 1 992; 
Hogg & Abrams, 1 988; Hogg & McGarty, 1 990). Given a distribution of 
group member positions on a relevant intergroup dimension (for instance, a 
personality attribute or an attitude dimension), the meta-contrast principle 
allows perceivers to ascertain whether intergroup differences are, on the 
average, higher than intragroup differences. Formally, a ratio can be com­
puted for each group member, so that the group's most prototypical 
member is the one who shows the highest ratio and thus occupies that 
position which simultaneously maximizes intergroup differences and 
minimizes intragroup differences. This position corresponds to the group 
prototype. From this point, in 
group and outgroup members are subjectively 
assessed in terms of their respective prototypes. As a result intergroup 
differences are emphasized to the detriment of interpersonal differences 
among the members of the same group and interpersonal similarities among 
ingroup and outgroup. 
In addition, according to self-categorization theory, individuals expect 
group members to match their respective groups' prototypes. Once social 
categories become salient, perceivers may judge the 'normative fit' (Turner 


1 28 
Social identity: international pers
pectives 
et aI., 1987), that is, the correspondence between the activated categories and 
the content properties of stimuli, in the situation. Lack of normative fit 
would lead perceivers to switch to more adequate categorization criteria 
(Oakes et aI., 1994), that is, to alternative intergroup dimensions which best 
allow the generation of clear-cut intergroup boundaries, and more eff
ectively 
account for the judgemental situation. Different criteria of categorization 
will be used to the extent that, depending on the stimulus properties of the 
judgemental setting, they succeed or fail in cognitively maximizing 
intergroup differences and intragroup similarities (Haslam & Turner, 1992; 
Turner et aI., 1987; cf. also Bruner, 1957). To give an example, in a political 
rally, comparative fit could entail a 'liberal versus conservative' categor­
ization. But if in that rally one noticed that, regardless of their political 
affiliation, female speakers supported women's rights whereas men addressed 
themselves to other issues, one would likely switch from the initial 'liberaV 
conservative' dimension to an alternative 'female/male' categorization. 
In brief, for self-categorization theory, group prototypes fulfil structuring 
functions. The meta-contrast generates cognitive clarity in that it allows 
individuals to perceive the social world in terms of meaningful, differ­
entiated units. Concomitantly, individuals deal with emerging intragroup 
differences by changing the criteria of categorization to alternative dimen­
sions which hinder such differences in order to steadily maximize intergroup 
distinctiveness. They are strongly motivated to maximize intergroup distinc­
tiveness, and this requires that social situations be cognitively construed so 
that intragroup differences are minimized. 
But, according to self-categorization theory, group prototypes also fulfil 
emotional functions (Hogg, 1992). Following the above-mentioned self­
stereotyping hypothesis, self-categorization theory states that individuals 
identify with groups to the extent that they perceive a match between the 
self and the ingroup prototype. When self-stereotyping occurs, the ingroup 
becomes fully representative of the self, and the self becomes perceptually 
interchangeable with other ingroup members (Turner, 1 981). According to 
Hogg (1992; Hogg & Abrams, 1988), a major implication of this fact is that 
social identification must be logically conceived of not as an outcome of 
interpersonal attitudes or similarity among group members, but, rather, 
as the result of a positive orientation towards the group prototype. As he 
put it, 
when a specific social identity is the salient basis for self-conception, self­
perception and conduct become ingroup stereotypical and normative, perceptions 
of relevant outgroup members become outgroup stereotypical, and intergroup 
behaviour acquires - to varying degrees, depending on the history of relations 
between the groups - competitive and discriminatory properties. (Hogg, 1992, pp. 
90-91) 
In short, when the ingroup is cognitively salient, a positive orientation 
towards the self becomes intrinsically equivalent to a positive orientation 
towards the group. Therefore, we may see social identification as a positive 
orientation towards the ingroup prototype, as it materializes in the 


Social identity and intragroup dif
f
erentiation 
1 29 
stereotyped self (Hogg, 1 992; Hogg & Abrams, 1 988). Furthermore, as 
individuals assimilate themselves to an ingroup prototype, their thoughts 
and behaviour conform to this prototype. The prototype is thus internalized 
as a norm defining the appropriate features and behaviour standards of the 
social self (Turner, 1 99 1 ). 
Denotative versus prescriptive group norms in inter- an intragroup 
judgements 
The above idea is important to our present concerns. From it, we may accept 
a definition of 'group norm' as a descriptive principle of intergroup 
differentiation, or cognitive clarity concomitant with social categorization. 
But norms may also be viewed as prescriptive standards, which are closer to 
overt or cognitively exerted social influence (Kenny & La Voie, 1 985). In this 
case, norms may be viewed as standards on which ingroup members anchor 
their judgements (Forsyth, 1 995) about other ingroup members in order to 
ascertain, not intergroup distinctiveness, but rather the legitimacy of the 
ingroup's superiority (Marques & Paez, 1994). In this vein, we shall speak of 
denotative norms to refer to the former, and prescriptive norms to refer to the 
latter kind of process. Notice, however, that prescriptive norms should 
function as a prerequisite for intergroup differentiation. Accepting this point, 
the major question is through what kinds of mechanisms do individuals step 
from a descriptive norm, which generates intergroup differentiation and 
identification with the ingroup, to a prescriptive norm, according to which 
such identification is regulated and reinforced? We now elaborate on this 
question. 
The 'black sheep effect' 
Our main empirical argument for the distinction between denotative and 
prescriptive group norms comes from evidence on the so-called 'black sheep 
effect' (Marques, 1990; Marques, yzerbyt & Leyens, 1 988). This evidence 
(Marques, 1 990; Marques & Paez, 1994, for reviews), shows that people 
differentiate more strongly between likeable and unlikeable ingroup members 
than between likeable and unlikeable outgroup members. For instance, 
Marques et al. (1988) asked their Belgian subjects to judge attractive and 
unattractive Belgian and North African students on a series of traits. The 
results showed that attractive Belgian students were judged more favourably 
than attractive North African students, whereas unattractive Belgian 
students were judged more un 
favourably than unattractive North African 
students. Similar results were found by Marques and Yzerbyt (1 988). These 
results showed that law students evaluated a good performance of another 
law student more favourably than a similar performance by a philosophy 
student, and evaluated a poor performance by a law student more unfavour­
ably than a similar performance by a philosophy student. Also, Marques, 


1 30 
Social identity: international perspectives 
Robalo and Rocha (1992) found that high-school pupils evaluated more 
extremely, both negatively and positively, likeable and unlikeable pupils of 
their own school than similar pupils of a rival high school. 
Taken at face-value, these findings might suggest an interpretation based 
on normative fit. That is, when faced with likeable and unlikeable in 
group 
and outgroup members, subjects would switch from the initial ingroup­
outgroup categorization to one which proved more informative given the 
stimulus-material they were provided with (that is, likeable and unlikeable 
characteristics or good and poor performance). However, other findings 
raise difficulties with this explanation. For instance, Marques et al. (1988, 
Experiment 2) showed that subjects over-rated likeable in 
group members 
and derogated unlikeable ingroup members, relative to their outgroup 
counterparts, only on dimensions relevant to the definition of the ingroup as 
a whole. On other equally socially desirable or undesirable dimensions, but 
which did not contribute to the distinctiveness of ingroup and outgroup, 
subjects were equally extreme in their evaluations of ingroup and outgroup 
members. In another study, Marques (1990) showed that subjects differ­
entiated more strongly between normative and counter-normative ingroup 
than outgroup members, but only when they considered the normative 
dimensions as directly relevant to the ingroup's image. Moreover, contrary 
to an interpretation based on a switch from the original ingroup-outgroup 
dimension to another one (distinguishing, for example, between likeable and 
unlikeable persons), Marques and colleagues found that ingroup bias (that 
is, positive differentiation of ingroup from outgroup as a whole) generally 
co-occurs with higher differentiation between likeable and unlikeable 
ingroup rather than outgroup members. 
In discussing evidence on the black sheep effect, Marques and Paez 
(1994) drew from the general assumptions of the social identity framework. 
However, they raised a problem with this framework as regards the 
explanation of the entire set of results. Research on intergroup relations has 
long shown that ingroup members are expected to be more likeable than 
outgroup members (for example, Brewer, 1979). Therefore, it is true that an 
explanation based on a search for intergroup differentiation and clarity 
of group boundaries may partly account for more negative judgements of 
undesirable (that is, outgroup typical) than of desirable (that is, in 
group 
typical) ingroup members. A simple reading of the social identity frame­
work would lead one to state that socially undesirable (that is, non­
prototypical) ingroup members are disliked because they contribute nega­
tively to the clearness of group boundaries. In tum, socially desirable (that 
is, prototypical) ingroup members would be liked, since they bolster such 
clearness (Hogg, 1992). However, this reasoning would fall short of 
explaining why unlikeable (that is, outgroup typical) outgroup members are 
not preferred to likeable (that is, ingroup typical) outgroup members. 
Indeed, in assuming that the subjects' basic motivation is to enhance inter­
group differentiation, one should expect them to express more unfavourable 
feelings towards undesirable ingroup and desirable outgroup members. 


Social identity and intragroup diff
erentiation 
1 3 1  
These feelings would ensue from a negative reaction to those who threaten 
the intergroup boundaries. By the same token, subjects would express more 
favourable feelings towards desirable in 
group and undesirable outgroup 
members, who, by definition, bolster such boundaries. But, clearly, this was 
not the case. To Marques and Paez (1994), the social identity framework 
could account for the black sheep effect if it encompassed the idea that 
people derogate socially undesirable ingroup members as part of a cognitive­
motivational strategy to purge from the group those ingroup members who 
negatively contribute to positive intergroup distinctiveness. In other words, 
derogation of unattractive, or deviate, ingroup members, and upgrading of 
attractive, or normative, ingroup members relative to outgroup members 
supports, or at least is concomitant with, intergroup differentiation. Based on 
this idea, they proposed an explanation based on what we can designate as 
'subjective group dynamics'. 
The notion of 'subjective group dynamics' 
The notion of 'subjective group dynamics' was inspired by an articulation 
between self-categorization theory and theory and research on small-group 
processes (Marques & Paez, 1994). In brief, following self-categorization 
theory, self-stereotyping generates full interchangeability between ingroup 
members (including the self) as regards ingroup-defining standards. In other 
words, self-stereotyping would generate 'promotional interdependence' 
(Deutsch, 1968; Lott & Lott, 1965), or subjective interdependence between 
the self and other ingroup members. As a result, the value assigned to the 
social self concept depends on the value assigned to the salient ingroup 
members in the judgemental context. In this vein, the self should become 
fully interdependent from other ingroup members. 
The above idea states a cognitive parallel between rejection of deviates in 
small groups and negative evaluations of unlikeable ingroup members in 
social categories which people conceptualize in terms of subjective belong­
ingness (Hogg & Abrams, 1988; Hogg, 1992; Tajfel, 1978d). Moreover, the 
effects occurring in this latter case should be strengthened relative to those 
observed in face-to-face groups. In face-to-face groups, deviates are 
downgraded because they endanger the subjective validity of social reality 
and the likelihood of group goal achievement. In groups characterized by 
subjective membership, social reality exclusively amounts to self-definition. 
The only relevant social reality is the positive value assigned to the social 
self. Any potential jeopardy of this special aspect of social reality would, 
most likely, be experienced as more threatening than contingencies arising 
from other aspects of social reality. 
The subjective full interdependence between ingroup members allied with 
the fact that self-definition engulfs the social reality of the cognitive group 
would have logical consequences at the judgemental level. Ingroup members 
who are perceived in line with expectations supporting the ingroup's relative 


1 32 
Social identity: international pers
pectives 
superiority should generate a state of subjective validity (Turner, 1 991) of 
the positive social self. This would explain praise of likeable in 
group 
members as occurs in the black sheep effect. But other ingroup members 
whose perceived characteristics or behaviour hinders positive intergroup 
distinctiveness would spur strong hostility. This would explain derogation of 
ingroup deviates. 
'Subjective group dynamics' and the cognitive construction of 
deviance 
Recent research on perceptions of group variability indirectly supports the 
idea that derogation of ingroup deviates emerges as a way to subjectively 
secure perceived group cohesiveness and hence to validate the superiority of 
the ingroup relative to the outgroup as a whole. First, there is evidence that 
when social identity is threatened or dominated, individuals tend to perceive 
the ingroup as more homogeneous than the outgroup. Lorenzi-Cioldi and 
colleagues (for example, Lorenzi-Cioldi, 1988, 1993; Lorenzi-Cioldi, Eagly 
& Stewart, 1995), for instance, have shown that members of dominated 
groups tend more to see themselves as similar to an ingroup prototype (they 
see their group as an 'aggregate'), and focus more on the abstract features 
that characterize the ingroup. In turn, members of dominant groups tend to 
see themselves more as individualized instances of the category (they see 
their group as a 'collection') and focus more on interpersonal differences 
between ingroup members (Doise & Lorenzi-Cioldi, 1990). Consistently 
with this phenomenon, evidence collected by Simon and colleagues (Simon 
& Brown, 1987; Simon & Hamilton, 1994) shows that members of minority 
or dominated groups consider themselves to be more similar to other 
ingroup members and judge the ingroup to be more homogeneous than do 
members of majorities or dominant groups. In the same vein, Doosje, 
Ellemers and Spears (1 996) have found that members of low-status groups 
perceive the ingroup to be more homogeneous than the outgroup when their 
social identification is strong. In turn, members of high-status groups, or 
members of low-status groups whose identification is low, perceive the 
ingroup to be more heterogeneous. 
Derogation of ingroup deviates as a function of lack of perceived 
ingroup cohesiveness 
Clearly, the above-mentioned research allows us to parallel findings in 
small-group settings and evidence found using larger social categories. In 
both cases, threats to the group generate a striving on the part of group 
members to increase group cohesiveness, probably as a means to reinforce 
the positive aspects of their social identity (Marques & Paez, 1994). In small 
groups, derogation of group deviates allows undesirable members to be 


Social identity and intragroup diff
erentiation 
1 33 
purged from the group. One may thus logically reason that a similar process 
would occur when the basis for judgements is membership of cognitive 
categories. 
Marques, Brito, Correia and Serodio (1996, Experiment 2) have recently 
explored this idea. They divided their subjects in minimal groups according 
to a bogus test of imagination, and asked them to fill in an attitude ques­
tionnaire. The questionnaire was composed of several statements ordered 
from highly socially desirable to highly socially undesirable, and subjects 
were asked to choose, among those statements, the one with which they 
agreed the most. One week later, subjects were informed of their group 
membership (imagination type), and examined responses given by five 
ingroup or outgroup members to the attitude questionnaire. Figure 9.1 
outlines the information provided to subjects about ingroup or outgroup 
members in the homogeneous and heterogeneous group conditions. As can 
be seen, in the homogeneous ingroup condition, all group members but one 
(the socially undesirable member) gave exactly the same response as the 
subject (item 6). When the ingroup was heterogeneous, two ingroup 
members gave the same response as the subject, two other members gave 
neighbouring responses, and one gave a socially undesirable response (item 
2). This was reversed in the outgroup condition. 
The subjects were then asked to judge one socially desirable member and 
one socially undesirable member (respectively, members X and Y, in Figure 
9. 1). Marques and colleagues (1996) predicted that subjects would derogate 
the unlikeable ingroup member more strongly in the heterogeneous ingroup 
than in the homogeneous ingroup condition, because, in the former condi­
tion, the socially undesirable ingroup member would be more threatening to 
the validation of the group's identity. In turn, no differences would emerge 
from evaluations of outgroup members. The data supported this idea. As 
can be seen in Figure 9.2, subjects derogated the undesirable ingroup 
member more than the undesirable outgroup member in the heterogeneous 
condition, and derogated the former member more strongly in this con­
dition than in the homogeneous ingroup condition. 
These results indicate that rejection of ingroup deviates may be a direct 
function of lack of ingroup cohesiveness around valued ingroup standards. 
When the ingroup is cohesive - that is, majority members uniformly adopt 
a valued position so that the group's norms are secured - group members 
seem to feel the presence of deviates as less threatening to the group than 
when the majority is more heterogeneous, and hence the group's normative 
system is more easily impaired by this presence. 
Derogation of ingroup deviates and upgrading of outgroup deviates 
as a prescriptive means to legitimize ingroup norms. 
In more general theoretical terms, the above-summarized study suggests that 
intragroup differentiation is concomitant with intergroup differentiation 


1 34 
Social identity: international pers
pectives 
Homogeneous group 
Heterogeneous group 
Social desirability continuum of 
Ingroup 
Outgroup 
Ingroup 
Outgroup 
statements 
1 .  'Homosexuals should be exiled to 
appropriate facilities from which they 
A 
should never be released so as not to 
endanger normal people' 
2. 'Homosexuals should be housed in 
institutions and allowed to leave only 
Y 
A B C  
Y 
B 
on certain special occasions, as long as 
X 
X 
they do not endanger other people' 
3. 'Homosexuality is an i llness and 
homosexuals should be forced to 
C 
undergo medical treatment' 
4. 'Homosexuality is shameful and gays 
should at least try to disguise their 
tendency' 
5. 'Although they are normal, gays should 
C 
be helped to find a better way' 
6. 'Gays, as everyone else, have the right 
X 
Y 
X 
Y 
to choose their own sexual life' 
A B C  
B 
7. 'Homosexuality is adapted to modern 
society and, therefore, gays should be 
A 
appointed to decision-making 
positions' 
Notes: 
1 .  Statement '6' = subject's stand on the issue. 
2. Distance of ingroup deviate member and outgroup deviate member (D) from their 
groups' modal positions = 14.001. 
Figure 9. 1 
Mani
pulation o
f
 group unif
ormity and group members' status: 
distribution o
f
 ingroup and outgroup members' responses 
rather than just the outcome of a switch from one categorization level (for 
instance, 'type of imagination') to another (for instance, 'tolerance towards 
sexual minorities'). Obviously, the notion of 'norm' used in this context is 
akin to that proposed by Turner and colleagues to designate group proto­
types, since it would correspond to the definition of the acceptable ingroup 
position on a social comparison dimension. However, we also propose that 
such norms encompass a prescriptive component: individuals would be 
normative not only as regards the 'denotative' prototypical definition of their 
group (a cognitive parallel, for instance, to Jones and Gerard's [1967] notion 
of 'informational influence'), but also in terms of a 'prescriptive' demand for 
compliance with this prototype (a cognitive parallel to Jones and Gerard's 
[1967] notion of 'normative influence'). 


Favourability 
7 
Social identity and intragroup dif
f
erentiation 
Member: 
-- Likeable 
1 35 
6 
7 
6 
5 
4 
3 
2 
- - - Unlikeable 
5 
4 
3 
2 
Ingroup 
Outgroup 
Homogeneous group 
Ingroup 
Outgroup 
Heterogeneous group 
Figure 9.2 
Evaluations o
f
 attractive and unattractive ingroup and outgroup 
members as a f
unction of group unif
ormity (Marques, Brito, Correia & 
SerOdio, 1996, Experiment 2) 
The legitimizing role of outgroup 'converts' 
In another set of studies, the present authors attempted to show that this 
process is indeed the outcome of an interplay between denotative and 
prescriptive norms, as defined above. Specifically, whereas denotative norms 
serve intergroup differentiation purposes, prescriptive norms serve ingroup 
legitimization purposes. As a result of prescriptive ingroup norm salience, 
subjects would not only downgrade ingroup members who fail to attune to 
these norms, but also upgrade outgroup members who attune to them. 
In one study, Marques, Abrams and Paez ( 1996, Experiment 3) asked 
their subjects to examine a case of murder, purportedly as part of a study 
on jury decision-making. Subjects were informed that two opposed patterns 
of decision-making existed (patterns X and Y) and that their membership of 
one pattern would be detected on the basis of their criteria in judging the 
crime. Subjects were also to rank the six characters involved, from most to 
least responsible. One week later, subjects were informed that, according to 
their reasoning about the murder (but not according to their ranking), they 
belonged to one pattern. In one condition (norm condition), subjects were 
told that a norm existed according to which those who belonged to their 
pattern should rank the characters exactly as they had done, and that the 
opposed pattern should show the reverse ranking. In another condition (no 
norm condition), subjects did not receive this information. In both con­
ditions, it was made explicit that the ranking did not indicate to which 
pattern individuals belonged. The only valid criterion was the way the 
rankings were justified by them. Subjects were then presented with the 
responses of five ingroup or five outgroup members. In the ingroup con­
dition, these responses were constructed such that: (a) each subject's 


1 36 
Social identity: international perspectives 
Modal member 
Deviate member 
Rank 
Subject's 
Ingroup Outgroup 
Ingroup Outgroup 
A 
A 
B 
D 
C 
2 
C 
C 
D 
B 
A 
3 
F 
F 
E 
E 
F 
4 
E 
E 
F 
F 
E 
5 
D 
D 
C 
A 
B 
6 
B 
B 
A 
C 
D 
Figure 9.3 Example of ranking o
f
 characters A-F used to build modal and 
deviate members 
response fully matched the ingroup norm; (b) four ingroup members 
displayed exactly the same response; and (c) one ingroup member (deviate 
member) displayed a response similar but not identical to the outgroup. In 
the outgroup condition, the pattern of information was reversed. To clarify, 
Figure 9.3 exemplifies the way information was constructed to manipulate 
ingroup and outgroup members' modal or deviate status relative to the 
group norm. 
Among other results, judgements of ingroup and outgroup as a whole 
were not affected by the norm information. In line with the classical 
ingroup bias, subjects evaluated the ingroup more favourably than the 
outgroup. That is, the categorization according to the X and Y pattern 
entailed intergroup differentiation. However, evaluations of modal and 
deviate group members varied according to whether subjects were or were 
not aware of the ranking norm. In the no norm condition, modal and 
deviate ingroup members were always judged more favourably than out­
group members. But, in the norm condition, modal ingroup members and 
deviate outgroup members were judged more favourably than deviate 
ingroup members and modal outgroup members. In other words, when no 
prescriptive norm existed, subjects behaved according to the same 
intergroup differentiation principle, in judgements of the whole groups, in 
judgements of modal members and in judgements of deviate members. 
However, when they were provided with information about such norms, 
their judgements diverged in the direction of the legitimization of the 
ingroup prescriptive norm (see Figure 9.4). 
These results suggest that denotative and prescriptive norms simulta­
neously operate in judgements of groups and their members. They could not 
be accounted for simply in terms of the denotative character of the meta­
contrast principle. Indeed, this principle would state that subjects would 
prefer normative ingroup members and normative outgroup members to 
counter-normative ingroup or outgroup members, because the former 
reinforce intergroup differentiation whereas the latter hinder such differ­
entiation. If, however, another quality of norms (that is, prescriptive norms) 
served to legitimate the values underlying social identity, then subjects 


Social identity and intragroup dif
f
erentiation 
Favourability 
7 
6 
5 
4 
3 
2 
Ingroup 
Outgroup 
Norm conditions 
Member: 
-- Modal 
- - - Deviate 
Ingroup 
Outgroup 
No norm conditions 
Figure 9.4 
E
f
f
ects o
f
 norm awareness on judgements o
f
 modal and deviate 
ingroup and outgroup members 
1 37 
would prefer outgroup members, who, by subscribing to ingroup norms, 
would hinder intergroup differentiation, to outgroup members, who, by 
subscribing to outgroup norms would reinforce such differentiation. 
Whereas denotative norms serve to differentiate between ingroup and out­
group as a whole, prescriptive norms serve to legitimate such differentiation. 
In support of this interpretation, we found that, in the norm condition, 
identification with the group was associated with the derogation of the 
deviate ingroup member (r = 0.32) and to the upgrading of the deviate 
outgroup member (r = 0. 1 8; meta-analytical average r regardless of 
correlation valence = 0.26). Instead, in the no norm conditions, identi­
fication with the ingroup was associated with a more favourable evaluation 
of the modal outgroup member (r = 0.68, p < .02). This result suggests that, 
in a minimal group situation, subjects' judgements follow the meta-contrast 
principle. They judge group members depending on their membership rather 
than on their contribution to the legitimization of ingroup norms, and value 
ingroup distinctiveness to the detriment of prescriptive aspects of normative 
ingroup conformity. However, when prescriptive norms enter the picture, 
subjects appear to simultaneously value intergroup distinctiveness and 
group members' normativity. 
In another study (Marques, Abrams & Paez, 1996, Experiment 2), instead 
of manipulating subjects' awareness of a prescriptive norm, we manipulated 
the normative components of the judgemental setting: whereas some 
subjects were informed that their judgements of modal and deviate members 
would be scrutinized by typical ingroup members (ingroup accountability 
condition), others were informed that these judgements would be scrutinized 
by typical outgroup members (outgroup accountability condition). We 
reasoned that, if judgements uniquely depended on the meta-contrast 
principle, outgroup accountability would make intergroup comparison more 
salient. As a result, subjects would upgrade ingroup members relative to 


1 38 
Favourability 
7 
Social identity: international pers
pectives 
6 ~ 
5 
4 
3 
/ 
/ 
2 
Member: 
-- Modal 
- - - Deviate 
Ingroup 
target 
Outgroup 
target 
Ingroup 
target 
Outgroup 
target 
Ingroup accountable 
Outgroup accountable 
Figure 9.5 
E
f
f
ects o
f
 in 
group and outgroup accountability on judgements o
f
 
modal and deviate ingroup and outgroup members 
outgroup members, regardless of their normative status, and would show a 
stronger ingroup bias in judgements of the entire ingroup and outgroup. 
However, if judgements were guided by a prescriptive principle, ingroup 
accountability would enhance the prescriptive status of the judgements. As 
a result, subjects would derogate deviate ingroup members and upgrade 
deviate outgroup members more in this condition, while simultaneously 
showing a stronger ingroup bias in judgements of ingroup and outgroup as 
a whole. The results supported this idea. Subjects upgraded the ingroup as 
whole relative to the outgroup more in the ingroup accountability condition 
than in the outgroup accountability condition. At the same time, they 
derogated the deviate ingroup member and upgraded the deviate outgroup 
member more than when they were accountable to the outgroup (see Figure 
9.5). 
Again, correlational data showed that, in the ingroup accountability 
condition, the less subjects identified with the ingroup, the less they 
upgraded the deviate outgroup member. In the outgroup accountability 
condition, the correlation was also negative but non-significant (r = -0. 1 1 ,  
n.s.). Together with the whole set of results, these data indicate that, in a 
strongly normative context (ingroup accountability or high normative 
awareness), subjects appear more motivated to legitimate in 
group norms by 
upgrading outgroup members who endorse, and by downgrading ingroup 
members who oppose, ingroup norms. 
A model of intergroup and intragroup differentiation 
The results of the studies summarized above suggest that, as predicted by 
self-categorization theory, and, specifically, the meta-contrast principle (e.g. 


Social identity and intragroup dif
f
erentiation 
1 39 
Hogg & McGarty, 1990), social categorization motivates individuals to 
engage in intergroup differentiation. But these results also suggest that the 
prescriptive context in which intragroup judgements arise strongly affects 
the subjects' responses. Ingroup members were downgraded when the 
ingroup lacked cohesiveness, as well as when prescriptive norms were made 
salient or the situation explicitly evoked the anticipated evaluation of rele­
vant others (that is, ingroup members). It thus seems reasonable to suppose 
that derogation of ingroup members is a part of a more general process of 
'subjective group dynamics'. 
'Subjective group dynamics' and 'referent informational influence' 
The relationship between the observed patterns of judgement and the 
constructs of social identity as a group's social reality as well as of absolute 
interdependence among ingroup members seems straightforward. Social 
categorization entails a sense of interdependence between the self and the 
representation of other ingroup members as a means to maintain a positive 
social identity. In turn, this social identity is built around cognitively rep­
resented group characteristics and behaviour standards whose maintenance 
is pivotal to positive intergroup differentiation. As a result, group members 
should be motivated to exert subjective control over ingroup members in 
order to sustain a sense of legitimacy of the ingroup's superior position 
relative to other groups. A threat to such legitimacy, and hence to the self, 
would arise with any perceived lack of conformity within the ingroup in 
light of the relevant ingroup standards. Subjectively control would be 
exerted especially over those ingroup members who threaten the overall 
positiveness of the group and its relative superiority over the outgroup. 
In brief, we propose that group judgements may, in circumstances more 
complex than strictly minimal intergroup situations, arise from an 'implicit 
theory' of group processes in the individual's mind. With Figure 9.6, we 
have attempted to sketch the proposed process. 
The basic idea underlying the process depicted in the figure is that 
derogation of some ingroup members occurs when, following social cat­
egorization individuals know the normatively appropriate behaviour to be 
adopted by ingroup members in order to protect a positive intergroup 
differentiation (Abrams, 1990). This corresponds to re
f
erent inf
ormational 
influence (Hogg & Turner, 1987; Turner, 1991; cf. also Hogg & Abrams, 
1988). Ingroup members who are perceived as not observing these behaviour 
requirements will trigger 'subjective group dynamics'. They will enhance 
subjective social control in order to secure the positive ingroup differen­
tiation. In the absence of perceived norm violation, the activated contrasting 
category labels will suffice to generate intergroup differentiation. This 
phenomenon can be explained in terms of a 'denotative' meta-contrast, and 
corresponds to the intergroup level, which represents the classical process of 
intergroup differentiation and social discrimination. 


1 40 
Social identity: international perspectives 
1 .  Social categorization and intergroup salience 
STIMULUS 
SETIING 
Fit 
(Comparative) 
(Normative) 
secure social 
identity 
(ingroup bias) 
inter-category 
salience 
social identity as 
'social reality' and 
superordinate goal 
norm 
violation 
? 
yes 
denotative 
cognitions 
M ETA-CONTRAST 
self-stereotyping and 
perceived 
interchangeability 
between self and ingroup 
2. Intragroup salience and subjective control 
perceived absolute 
interdependence and 
'common fate' 
legitimization of 
positive social 
identity 
Figure 9.6 
A two-step process o
f
 inter- and intragroup judgement 
But when an intragroup distinction becomes salient and threatens inter­
group differentiation, subjects will derogate ingroup members who do not 
comply with the ingroup standards. This corresponds to the intragroup 
level. In this vein, the intragroup level would correspond to a subjective 
effort intended to legitimize the intergroup differentiation, and to warrant 


Social identity and intragroup diff
erentiation 
141 
the ingroup's unifonnity with regard to the prescribed standards that sustain 
ingroup's superiority. 
Conclusion 
We propose that in most circumstances nonnative intragroup differentiation 
may function as a foundation for intergroup differentiation and a positive 
social identity. This process can be viewed as the psychological equivalent 
of public actions of nonn enforcement, like those we mentioned at the 
outset of this chapter. In this sense, nonn violation is functional to group 
life, to the extent that the social control mechanisms it evokes help to define 
relevant criteria of group membership. Because the prescriptive nonn only 
appears as such once it has been violated (Forsyth, 1995), deviance plays a 
basic role in defining the group's objectives and interests by generating 
(subjective or objective) social control (cf. Hamilton & Rauma, 1995). 
Moreover, active participation in the punishment of deviance may reinforce 
the individual's commitment to the nonn, thus reinforcing social identi­
fication. In purely psychological tenns, this process might function more or 
less as predicted by Bern's (1967) self-perception theory: when they lack a 
definite attitudinal position about an object, individuals may infer their 
attitude from their behaviours ('if I have punished a deviate person I infer 
that I have a clear attitude in favour of the nonn'). Moreover, punishment 
rituals towards deviates may be viewed as part of an initiation process, to 
reinforce the positive evaluation of the group (cf. Zimbardo & Leippe, 
1991 
). 
It is obvious that the presently proposed process is still in need of 
theoretical and empirical scrutiny. But we hope that the ideas outlined in this 
chapter may serve as a valuable contribution to the study of the processes 
occurring in small and large groups alike, and a starting point for research­
ing into the cognitive processes underlying the social construction of nonns 
and deviance in daily life. 


1 0  
Categorization and Social Influence 
Juan Antonio Perez and Gabriel Mugny 
A relevant dimension in studies of social influence is whether or not the 
targets of influences see themselves as sharing the same social category as 
the influence source. The most commonly advanced hypothesis is that the 
effects of categorization (whether at the cognitive level of accentuating 
intragroup similarities and intergroup differences, or at a more motivational 
level of ingroup favouritism and/or outgroup discrimination) should facili­
tate the influence of an ingroup source and diminish that of an outgroup 
source. However, it appears to be difficult to draw any general conclusion 
given that in practice experimental effects have not all consistently been in a 
single direction, especially if manifest and latent influences are distinguished 
(Moscovici, 1980). The first objective of this chapter is to review those 
studies which have varied the degree of identity between source and target, 
and the second is to present the principal conceptualizations of the links 
between categorization and influence. 
The effects of categorization on influence 
There is first of all a series of studies which have manipulated the cat­
egorization (in 
group or outgroup) of sources having a majority status 
(Doise, Gachoud & Mugny, 1986; Perez, Mugny & Moscovici, 1986; 
Volpato, Maass, Mucchi-Faina & Vitti, 1990). Ingroup majorities have 
more (direct and immediate) influence than outgroup majorities. In those 
studies which have also considered indirect dimensions of influence, this 
ingroup majority advantage disappears in interaction with other factors. 
Studies in which subjects are simultaneously exposed to minority and 
majority positions do not contradict and indeed confirm these findings. Thus, 
in the study by Perez and Mugny (1987), the in 
group majority achieved 
greater direct influence than the outgroup majority, but at the indirect level 
this simple effect disappeared in interaction with the categorization of 
the minority. In the Clark and Maass (1988b) study, categorization of the 
majority was not varied - it was an in 
group - but it could be seen that it had 
more influence at the public level than either an ingroup or an outgroup 
minority. At the private level an interaction could again be observed with the 
categorization of the minority as either ingroup or outgroup. 


Categorization and social influence 
143 
In studies where the categorization involved is based on arbitrary criteria 
(such as whether one over- or underestimates the number of pupils attend­
ing secondary school, whether one is in the same classroom, and so on) the 
in 
group has more influence at the public level than the outgroup. This 
'simple categorization' effect can even be observed when the source has a 
minority status (cf. two studies by Martin, 1988b, 1988c, and one by 
Papastamou et aI., cited in Mugny & Perez, 1991a, pp. 24-27), as well as 
when the minority or majority character of the source is not made salient 
(Wilder, 1990, study I). 
In two studies on attitudes towards abortion, in which gender was used as 
the basis for categorization, it was found that a male minority group had 
more influence than a female minority group. In the Maass, Clark and 
Haberkorn (1982) study they constituted an ingroup source (although when 
the variance was decomposed the effect did not reach significance), while in 
the Perez and Mugny (l 985b) study they represented an outgroup. This 
gender effect may be specific to the theme of abortion because in a task 
involving estimation of increases in study grants, Martin (l 988b) found that 
a female minority group produced more public influence among other 
women than did a male minority group, while men were more influenced by 
a male source. 
If one turns to consideration of 
another variable, the distance between the 
attitude of the target and that of the source, one finds similarly variable 
results. In the Clark and Maass (1988a) studies on homosexual rights, an 
outgroup (homosexuals) produced more influence than an ingroup source 
(heterosexuals) among those subjects most favourable to homosexual rights; 
the ingroup source, however, produced more influence than the outgroup 
source among subjects who had the least favourable attitudes. Mackie, 
Worth and Asuncion (1 990) observed that the ingroup had a stronger 
influence than the outgroup but only among subjects who were most 
opposed and when the message contained a 'strong' argument. Using the 
theme of xenophobic attitudes, Mugny, Kaiser and Papastamou (1983) 
found the opposite interaction: the outgroup (a group of foreigners) had 
more influence than the ingroup (a group of fellow compatriots) on the 
most xenophobic subjects; an in 
group minority had more influence than an 
outgroup minority on those subjects whose own position was already closest 
to the anti-xenophobic message. 
In four studies in which perception of the source was 'individualized' or 
'personalized', ingroup and outgroup achieved the same degree of influence 
(Doise et aI., 1986; Martin 1988a; Papastamou et aI., cited in Mugny & 
Perez, 1991a; Wilder, 1990). 
Negotiating style (cf. Mugny, 1982) tends to interact with the identity of 
the source. In one study on 'green' attitudes, Mugny and Papastamou 
(1982) found that when influence targets are led to see themselves as similar 
in identity to the source of influence, the latter has more influence when the 
argument is made in a rigid style than when it is made in a more flexible 
style, when subjects are led to see themselves as dissimilar, it is the source 


144 
Social identity: international perspectives 
with the flexible style that is more influential. In a study on attitudes 
towards foreigners, Mugny and Perez (1985) observed the same interaction 
effect. In the same vein, Mugny, Kaiser, Papastamou and Perez (1984) 
discovered that a source sharing the same nationality arguing in favour of 
rights for foreigners had more influence if the argument was based on 
radical socio-political principles, while a source consisting of foreigners had 
more influence when the same case was argued from humanitarian 
principles. 
Another problem concerns the effects of categorization in different 
influence conditions. In studies in which the school to which the source 
belongs is varied (same school as the subjects or a rival school), results when 
they have been significant have shown that the in 
group has more impact at 
the public level than the outgroup (Mackie et aI., 1990; Martin, 1 988a; 
Wilder, 1990). Martin observed that at the private level the outgroup had as 
much influence as the ingroup, both kinds of source producing more change 
than that of a control condition without a persuasive message. 
In some studies measures of influence at both a direct and an indirect 
level have been available. Thus, in a paradigm based on musical prefer­
ences, Aebischer, Hewstone and Henderson (1984) found that an outgroup 
majority (from a different and devalued education establishment) had more 
indirect influence than a valued ingroup majority source, while there was no 
difference in direct influence. In a study of environmentalist attitudes (Perez 
& Mugny, 1985a), it was found that when the persuasive message was seen 
to have derived from a single individual selected at random and not 
representative of the group, direct influence was greater than when the 
source was seen to represent the opinion of the entire group. However, at an 
indirect level the effects were reversed: the message representative of the 
group had more impact than the unrepresentative message. In a study 
involving attitudes on abortion and contraception (Perez & M 
ugny, 1 986a), 
the connotation of a source was manipulated (the source was characterized 
in terms of entirely positive attributes or entirely negative attributes), while 
the source was also an ingroup minority (same age-group as the subjects: 
young people) or an outgroup minority (different category: adults between 
30 and 50 years old). The minority outgroup with positive connotations was 
more influential at the direct level than the outgroup with negative 
connotations, and more influential also than the ingroup with a positive 
image. At the level of indirect influence the most significant effect was that 
of the negative-image in 
group, as compared to the positive-image ingroup. 
In two studies Volpato et al. (\ 990) measured the effects of categorization 
on creativity. They led subjects to believe that a committee either in their 
own town (ingroup) or in another town (outgroup) had produced two 
communiques, one by the majority and the other by a minority, respectively 
opposing or favouring the establishment of a final exam at the completion 
of baccalaureat studies. They were exposed only to the argument in favour 
of this proposal. The results showed that a smaller number of subjects 
indicated they were in favour of the source's proposition when this was 


Categorization and social in
fluence 
145 
presented as originating from an ingroup minority. The proposItion was 
more widely accepted when it was attributed to an ingroup majority or 
indeed an outgroup minority. Although the direct influence of an ingroup 
minority was shown to be less than in these other two conditions, in these 
two studies this same condition produced a higher proportion of subjects 
who invented new solutions to solve the problem. 
Finally, on the matter of differentiated influence, Mugny et al. (1 983) 
reported that under certain conditions an outgroup arguing on humanitarian 
grounds had greater delayed influence than the same outgroup using a socio­
political argument. 
These, in summary, are the effects which have been observed of categor­
ization on social influence. It must be acknowledged that their diversity and 
complexity is such that the most prudent conclusion would be to admit that 
there is no systematic effect associated with categorization as such, nor any 
straightforward relation between categorization and type of influence. The 
hypothesis that there is an ingroup bias in influence has therefore to be 
regarded as inadequate; the variables moderating the effects of categoriza­
tion appear to be as numerous as the paradigms. Perhaps this disparity of 
manipulations reflects a belief that 'simple categorization' (cf. Tajfel, 
Flament, Billig & Bundy, 1971) is sufficient to produce intergroup effects, 
and, by analogy, differences in social influence. This leads us to consider 
what are currently the most elaborate explanations in this field, and to ask 
to what extent they can account for this diversity. 
Independence and heterogeneity of the source 
Wilder invoked independence and heterogeneity of the source (1 977, 1978, 
1 990; cf. also Harkins & Petty, 198 1 ,  1983, 1987) to account f
or conformity 
in restricted groups, and more recently to explain why one should expect an 
ingroup to have more influence than an outgroup. The idea is that the 
individuals who comprise a group have more influence if they are perceived 
as independent entities than if they are perceived as a collective entity. The 
reasons advanced for this are, on the one hand, that information provided by 
a group is treated with less care (for example, in a less personalized and less 
specific manner), and, on the other, that members perceived as a group are 
considered less credible because they are attributed less independence. 
Recently, Wilder (1990) has used this argument to predict that an ingroup 
will have more influence than an outgroup. His reasoning is that the target 
starts from the presupposition of a greater homogeneity among outgroup 
members than among the members of his or her own group. The target 
would recognize finer distinctions within the ingroup which would lead to a 
perception of members of the ingroup as more independent of one another. 
The same author argued that his is not a motivational explanation of 
ingroup superiority (it does not appeal to a motive of ingroup favouritism or 
of discrimination against the outgroup), but rather a cognitive explanation, 


146 
Social identity: international perspectives 
based on perception of the homogeneity or heterogeneity of members of the 
influence source. A series of four experiments illustrates this model (Wilder, 
1990), and to these may be added two studies by Mackie et al. (1990) which 
show that the message attributed to an ingroup is given greater attention 
than one attributed to an 'other' group. 
One difficulty posed by this analysis is the implication that an ingroup is 
nothing but an aggregation of separate individuals. Thus the comparison 
made is between an aggregate of individuals and an outgroup rather than 
between an ingroup and an outgroup. This does not take into account recent 
developments in the theory of levels of self-categorization (Turner, Hogg, 
Oakes, Reicher & Wetherell, 1987), according to which three such levels may 
be distinguished: a superordinate level at which individuals see themselves as 
identical to all other human beings; an intermediate level at which the social 
identity of the individual is defined by the assertion of differences between 
groups or categories of membership and non-membership; and finally a 
subordinate level which defines personal identity in terms of categorization 
of the self as a unique being, distinct from all others including members of 
one's own group. From this perspective, 'personalized' perception is not 
synonymous with ingroup categorization. If this argument is conceded, it 
would appear that Wilder's model, rather than explaining the ingroup­
outgroup difference, is more suited to accounting for the superior influence 
of a personalized source (whether belonging to an ingroup or an outgroup) 
as compared to that of a source perceived in more 'depersonalized' or 
homogeneous terms. 
The results of a study (Perez, Mugny & Navarro, 1991) justify our stress 
on the importance of these distinctions. In this study subjects were required 
to react to a message either in personal terms (subordinate level of se1f­
categorization), or in 'social terms', which is to say at the intermediate level, 
in one of two ways: either to react as a representative of their membership 
group (in this case young people) or as a member of the outgroup (adults). 
The results show that direct influence is greater when the message is treated 
in either personal or ingroup terms. These findings are consistent with the 
predictions of Wilder's model. However, at the indirect level, while influence 
remains strong in the personalized condition, it disappears in the ingroup 
condition, an effect which confirms the importance of not confounding 
the personalized level with the intergroup level. Additionally, indirect influ­
ence was found in reaction to the message in intergroup terms. This set of 
results is interpretable in terms of the notion of ingroup paralysis: under 
the conditions provided subjects conform to positions prototypical of the 
ingroup, which has the effect of paralysing examination of other positions 
from a personal point of view or from any other point of view which is not 
that of 
the ingroup. In this case the ingroup introduces a resistance to change 
and innovation. In a series of studies Sanchez-Mazas, Perez, Navarro, 
Mugny and Jovanovic (1993) have confirmed that when subjects are led to 
the self-perception that in terms of their attitudes they are identical to their 
own group, this paralyses any change in their attitudes. 


Categorization and social in
fluence 
147 
Influence according to self-categorization theory 
Self-categorization theory treats influence as an intergroup phenomenon 
(cf. Turner, 198 1  
a, 1991). The fundamental factors are the level of self­
categorization and the ways in which individuals seem to operate to 
determine the validity of their opinions and behaviour. In the first place 
individuals begin by anticipating that their judgements and actions in a 
given task will coincide with those of other members perceived as similar to 
the self and who are regarded as a comparison group appropriate to the 
situation. Individuals will be confident that their own judgements are valid 
to the degree that similar others are observed to express the same opinions 
or behave in the same way in the situation. 
Uncertainty and loss of confidence in their judgements and behaviours 
increase when individuals find themselves in disagreement with other group 
members whom they perceive, despite the disagreement, as similar to 
themselves, which is to say as an ingroup. Insofar as this disagreement 
cannot be attributed to differences between the self and the group this 
theory predicts a change in judgements or behaviours in the direction of 
those of the group or of those which best represent the ingroup consensus; 
those which are most prototypical. 
No such expectation exists when the others are perceived as different, 
which is to say as an outgroup. Disagreement in this case can be attributed 
to differences in group membership, according to the principle of categorial 
differentiation (Doise, 1976). Consequently, disagreement with the outgroup 
produces little uncertainty, and it is not anticipated that individuals in these 
circumstances would change their judgements or their behaviour. 
In the application of his theory to minority influence, Turner (1991) 
claims in addition that for a minority to exert influence it must present 
itself, or be perceived, as forming part of the ingroup (and not an out­
group), which is to say a subgroup that defends a pro-normative position 
within the ingroup. He predicts a reduction in direct influence if the 
disagreement the minority introduces in the group is polarized to the point 
of turning it into an intergroup, and not simply an intragroup, conflict. 
This theory tends also to account for differences between direct and 
indirect influence in terms of the level of self-categorization which obtains at 
any particular moment. Thus a minority categorized as an outgroup will 
have less influence, whether direct or indirect. In order for minorities to 
produce direct influence they must be perceived as part of the ingroup. On 
the other hand, if a minority produces a conversion effect (indirect influ­
ence), this will be because the comparison is made within a broader context 
in which the minority is perceived as an integral part of the ingroup, and 
because at this superordinate level of categorization, its alternative norm is 
congruent with the norms and values of the ingroup. 
This theory offers one of the most complete explanations for the tendency 
of an outgroup to have less influence than an ingroup. An initial problem, 
however, is that the theory tends to predict both a direct and an indirect 


148 
Social identity: international perspectives 
influence of an ingroup majority and does not take account of the latent 
paralysis effect which appears to be precisely the effect of overt conformity 
(Sanchez-Mazas et aI., 1993). A further difficulty is that it does not allow 
for the fact that an ingroup minority produces less direct influence than an 
outgroup minority or one which is not explicitly categorized as an ingroup. 
Additionally, the explanation it offers for conversion does not seem to be 
entirely convincing. In particular, in several studies, it has been found that 
in the same situation the same source which does not produce a direct 
influence does produce indirect influence. Is one supposed to believe that in 
this situation the influence source is perceived simultaneously at two levels 
of categorization and that, for example, when faced with an item reflecting 
direct influence subjects see the source as an outgroup, but when they 
encounter an indirect item they perceive the same source as part of the 
ingroup? There is a similar difficulty in seeing how it is possible that at a 
public level the influence source could be categorized as an outgroup but at 
a private level as an ingroup, or how a public context could prime an 
intergroup context while in a private context priming a superordinate 
intragroup context. These possibilities certainly cannot be excluded but no 
research has yet supported such dynamics. 
Dissociation theory 
The need to relate different levels of influence to one another is taken into 
consideration by dissociation theory, which we will now present. Briefly, 
this theory makes a general distinction between two levels of influence: first, 
manifest influence, which includes public, immediate and direct influence; 
and, second, latent influence, which covers influence at the private, delayed 
and indirect levels. It provides a basis for relating processes that account 
for manifest influence (and which the subject can control) to the processes 
underlying latent influence (which the subject is less able to control 
consciously). 
Manif
est in
fluence and positive social identity 
Manifest influence is a function of normative and informational dependence 
(Deutsch & Gerard, 1955). The importance of normative dependence relates 
to the individual's motivation to acquire, to present or to preserve a positive 
self-image or self-concept. Given that one function of manifest influence is 
protection of a positive social identity, one is influenced at this level by 
sources able to exercise the strongest normative pressures or in situations in 
which modifications in judgements allow one to avoid costs or gain more 
credit than if one maintained one's independence. Susceptibility to manifest 
influence also flows from the uncertainties felt by the subject, such that the 
more the source is able to engender uncertainty in the subject (by virtue of 
the former's credibility, ingroup identity or majority status), the g-:eater the 
source's influence. 


Categorization and social influence 
149 
Latent influence and social constructivism 
By latent influence we do not have in mind only private influence but all 
influence reflecting constructivist activity. Thus, for example, it includes the 
case in which subjects change on dimensions which can be inferred through 
the operation of a given mental activity. This concerns the attention the 
target pays to what the source says (to its message, its judgement, its 
attitude, and so on), the deployment of an innovative form of thought 
which allows the target to derive or discover new functions and ramifi­
cations on the basis of the information provided, to establish new connec­
tions between the content given and other contents, as well as the 
elaboration of a new representation of the question. Some authors have 
discussed these kinds of activity under the heading of validation processes 
(Butera, Legrenzi, Mugny & Perez, 1 990-1991; Moscovici, 1980), others 
under the heading of divergent thinking (Nemeth, 1986). Huguet, Mugny 
and Perez (1990-I 
991) also use the concept of decentration. Within the set 
of paradigms dealing with attitudes and opinions, these activities have been 
described as inferences about the organizing principle underlying a position 
(Perez & M 
ugny, 1986b). In all these cases, however, the same basic 
hypothesis underlies these various designations: mental activity is required if 
a latent mode of influence is to arise out of social interaction, and it will 
occur only in the context of certain kinds of social relationship and not 
others. 
Postulates o
f dissociation theory 
To the degree that manifest influence is directly observable by others or is, 
ultimately, something of which subjects themselves are conscious, then 
subjects' identity is at stake whenever they are influenced by any source, and 
dissociation theory recognizes that identification is more probable in the 
face of an ingroup source. This hypothesis indicates that a message 
emanating from an ingroup is more likely to be accepted, even if it is not 
read or processed, than a message originating with an outgroup source. 
This more positive attitude towards whatever originates in the ingroup 
can, especially when the subject is highly involved or motivated, give rise to 
a higher degree of attention to a message attributed to an ingroup and thus 
a more elaborated processing of the message as compared to one from an 
outgroup. Faced with an ingroup, because identification is more intense the 
message can then be accepted with a less defensive attitude and the subject 
can feel more motivated to pay attention to it. One may therefore predict 
that in certain conditions there will be direct but also indirect influence. This 
will only happen, however, if the situation does not induce a paralysis and if 
identification with the ingroup is positively valued and poses no conflict of 
identification. 
In order for the ingroup to have indirect influence, one basic condition 
must be satisfied: subjects must perceive a divergence between their own 
position and that of their group. One risk for an ingroup is that if 


1 50 
Social identity: international perspectives 
individuals perceive the group to be in agreement with their own position, 
and this accords with an anticipated homogeneity of attitudes, they will 
develop a socio-cognitive paralysis, the consequence of which is to inhibit 
the activity necessary for the emergence of 
indirect influence. The attitudinal 
proximity which individuals often establish through their manifest 
conformity to the ingroup can therefore by this same process limit the 
latter's impact to a pattern of influences similar to compliance (positive 
manifest influence combined with zero latent influence), as has been 
observed in a number of studies (cf. Perez et aI., 1991; Sanchez-Mazas et aI., 
1 993). One way of motivating subjects to process the ingroup message is 
therefore to lead them to perceive a normative conflict between their own 
position and that of their group, which flows not from a simple self­
categorization but from an intragroup conflict. 
Another difficulty for the ingroup arises in the case of identification 
conflict (cf. Mugny et aI., 1983; Mugny & Perez, 1991b), in which indi­
viduals are aware of belonging to a group which has little positive value and 
may even be negatively valued, or in which they risk being identified with 
such a group. In these cases, identity with or similarity to this ingroup 
threatens the maintenance or construction of a positive social and personal 
identity. Would individuals let themselves be publicly and manifestly 
influenced by a socially discredited group posing such a conflict of identi­
fication? In fact it is not enough to respond that in all probability they 
would not let themselves be influenced; it is also important to recognize 
what would happen at a latent level in such a situation. A basic postulate of 
dissociation theory is that the more that socio-cognitive activities and 
attention are occupied in creating a disidentification or social differentiation 
from a source of influence, the less these are concerned with the influence 
message. In the event that individuals come to give any attention to the 
message and to analyse it, this is undertaken on the basis of a motivation to 
construct a differentiation so that a negative processing of the message tends 
to predominate. 
Let us imagine a situation in which the target of influence does not wish 
to be identified with the source of influence and that in one case this 
involves an ingroup source and in another an outgroup source. 
In the situation where the influence source is perceived to be an ingroup 
with which the influence target does not wish to be seen to be identified (for 
example, the case of a discredited minority), the available data suggest that 
in 
group favouritism is indeed not the dominant strategy. Studies by 
Marques (1990) on what he has described as the black sheep phenomenon 
provide a direct illustration of the way in which a negatively valued ingroup 
can be the object of more discrimination than an equally negatively valued 
outgroup. The reason for this is that the more intense the identification is 
(for example, by virtue of a strong identity or similarity between individual 
and group), the more effort and activity will be required, should the 
occasion arise, to try to disidentify. This is the reason why ingroup influence 
may be zero at the latent as well as at the manifest level. 


Categorization and social influence 
1 5 1  
It can thus prove to be more positive for social identity to discriminate 
against an ingroup rather than an outgroup, or to favour the outgroup to 
the detriment of the ingroup. Indeed, an outgroup source can paradoxically 
have a net advantage, particularly at the latent level, over an ingroup 
source. Given that the ties of identification are less close with an outgroup, 
it seems reasonable to predict that disidentification with such a group 
requires less activity and effort, because in this case a ready-made differ­
entiation is provided; it is founded on differences of group membership. The 
mental activity generating latent effects will no longer be confounded with 
the activity of differentiation, the two being immediately dissociated. 
A major hypothesis of dissociation theory is that even in situations 
involving intense conflict a source categorized as an outgroup will be more 
likely to produce dissociation, to the degree that social differentiation is 
assured by the existence of a pre-established categorization, or that the 
differences are so great that to shift in the direction of the source's responses 
carries no risk of a 'categorial confusion' (Lemaine, 1974). This dynamic 
is also involved when the space for comparison of subject with source is 
multidimensional, such that there is no negative interdependence between 
source and target of influence (Perez & Mugny, 1990). But this does not 
imply that an outgroup source will have direct influence. In reality, at the 
direct level it is the categorization processes that operate, as it were quite 
naturally. However, to the degree that this social differentiation is dissoci­
ated from message reception, conversion effects towards outgroup positions 
can appear. If for one reason or another the subject does not come to 
dissociate reception of the message from construction of this differentiation, 
the outgroup will no longer have any significant indirect effect (Butera, 
Huguet, Mugny & Perez, 1994). 
Conversion dynamics and the nature of the task 
Let us acknowledge that at the psychological level the influence situation 
produces a dissociation of the relation with the source from the processing 
of the message. One question remains then unanswered, namely: what can 
motivate the subject, particularly when faced with an outgroup, to process 
the content of the message? 
A first possibility is that epistemic expectations intervene which are 
specific to the nature of task to which the outgroup's divergent judgement 
relates. For example, in objective, non-ambiguous tasks (of the type used by 
Asch, 1956) for which only one correct response, whatever its provenance, is 
assumed to be possible, the influence mechanism of an outgroup generally 
rests upon the motivation of the subject to re-establish uniformity of 
judgements. In such a situation and given the impossibility of changing the 
source, it is the subject who ends up changing, always at the latent level, in 
order to reconcile social differentiation and epistemological beliefs con­
cerning the necessity of consensus (Butera et aI., 1994). 


1 52 
Social identity: international perspectives 
When tests of ability are involved (for example, estimating distances) and 
subjects presuppose that there do exist objectively correct answers but do 
not know what they are and so feel uncertain of their ability, outgroups can 
achieve an influence by the triangulation effect already described: 
outgroups' judgements can be perceived as valid by virtue of their socially 
independent origins (cf. Goethals, 1972). 
When the tasks involve opinions, differentiation from the outgroup at the 
level of attitudes itself constitutes an epistemological requirement (Perez, 
Mugny, L1avata & Fierres, 1993) which accounts for manifest discrimina­
tion. The indirect effects of an outgroup (in particular when this is a 
minority) arise from the fact that dissociation allows a centration on the 
contents and initiates a debate about ideas (Billig, 1985). Examination of a 
contrary argument, even if this is in order to resist it, implies in particular 
that one increases one's exposure to information from the other. Here then, 
as in the case of denial (Falomir, Mugny & Perez, 1996), a process of social 
cryptoamnesia may be engaged through which, paradoxically, ideas may be 
internalized whose social origins may be forgotten as a result of dissociation 
(cf. M 
ugny & Perez, 1991 
b). 
Certain social minorities (for example, black people) may also constitute 
outgroups protected by the Zeitgeist in such a way that discrimination 
towards them is socially censured, at least for a significant proportion of 
society. In this case, a double effect can be produced. At a manifest level 
there may be on the one hand, a greater influence (reflecting social 
desirability) which does not, however, generalize to the latent level, and, on 
the other, an effect close to socio-cognitive paralysis. Paradoxically it is 
when subjects discriminate against minorities that they experience a conflict 
the resolution of which can provoke a change in latent attitudes (cf. Perez, 
Mugny, L1avata & Fierres, 1993). The occurrence of racist, xenophobic or 
sexist acts can therefore be in conflict with norms and values of justice and 
social equality accepted more generally by subjects. The self-reproach 
arising from this discrimination can thus act as a mechanism for change in 
latent attitudes. 
Conclusion 
It is clear that intergroup dynamics, at least with respect to social influence, 
are not confined to the habitually recognized effects of social categorization 
(accentuation of intragroup similarities and intergroup differences). They 
also involve strategies for constructing or maintaining a positive social 
identity, and can additionally disfavour the in 
group or favour the outgroup, 
in particular as a function of whether or not the social context threatens the 
targets' personal and social identities in their relations with salient entities in 
the categorial field. 
Of the two factors that are fundamental to understanding social influence 
in an intergroup context, one certainly is the dynamic of identification or 


Categorization and social in
fluence 
1 53 
differentiation, but the other is the conflict created by the divergent position 
of a source of influence. Simply categorizing a source as ingroup to produce 
identification is insufficient to guarantee influence, whether at the manifest 
or the latent level. It is the divergence and conflict which determine matters 
in one direction or the other and these alone are able to account for latent 
processes of constructivism. 
In all cases, divergence creates a conflict by reason of epistemological 
presuppositions about the task, in particular expectations about consensus 
or dissensus. Categorization therefore is not alone in playing an active role, 
processes of influence have their own autonomy by virtue of the relational 
but also socio-cognitive conflicts which they introduce. From this point of 
view, dissociation theory seems to provide a tool which allows us, to a 
greater degree than models offering a single-process view of intergroup 
effects, to organize the multiple dynamics which can arise from the inter­
action between categorization and social influences. 
Note 
This text was prepared with the financial support of the Swiss Fond National de la Recherche 
Scientifique, the Spanish Comision Interministerial de Ciencia y Tecnologia (CICYT, SEC95-
0628) and the French Centre National de la Recherche Scientifique. 


1 1  
Social Identity and Interpersonal 
Relationships 
Miguel Maya 
Attraction and interpersonal relationships 
The emotional-sexual relationships established between men and women 
resulting from attraction (from here on referred to as 'relationships between 
couples') can, at first, be considered as being situated close to the inter­
personal end of the interpersonal-intergroup continuum, already mentioned 
in previous chapters of this book. This can be deduced both from the 
classical research carried out in social psychology and the formulations of 
some authors classified as social identity theorists. 
The traditional approach towards studying attraction and interpersonal 
relationships has always placed special emphasis on considering them from 
the perspective of the individual, without bearing in mind the social or 
intergroup dimension. This is reflected not only in the most commonly used 
phrase to define this study area - 'interpersonal attraction' - but also in the 
approaches and theories used to explain these phenomena. 
The main theoretical focal points used to explain attraction are the 
theories of cognitive consistence, exchange and interdependence (Kelley & 
Thibaut, 1978) and those based on the principles of association and 
reinforcement (Byrne, 1971). All these, as well as the factors used to 
measure attraction (proximity, similarity, reciprocity, among others), refer 
fundamentally to peculiar characteristics of individuals involved in the 
interaction. 
Some authors working within the social identity theory framework have 
also argued that this particular type of male-female relationships can be 
placed close to the interpersonal end of the continuum (Hogg & Abrams, 
1988; Huici, 1984). There is a tendency to consider male-female relation­
ships as intergroup when the treatment shown towards people belonging to 
another sex group is undifferentiated, which implies minimal or almost non­
existent knowledge of them. In this case it is likely that stereotyped percep­
tion of the outgroup will appear, in addition to intergroup conflict and 
ingroup favouritism. On the other hand, interpersonal male-female rela­
tionships tend to be characterized by a one-to-one relationship, a married or 
heterosexual couple. 


Social identity and interpersonal relationshi
ps 
1 55 
Relationships between couples as intergroup relationships 
This heading could also be termed: 'Interpersonal relationships as inter­
group relationships', which would be no mere play on words. Our intention 
here is to show that despite the fact that relationships between couples have 
been traditionally referred to as 'interpersonal relationships', and have thus 
had to be placed close to the interpersonal end of the continuum, there in 
fact exist a number of powerful intergroup aspects within them. Indeed, 
these have been acknowledged by the above-mentioned authors. 
For example, Huici (1984) argues that the existence of power and status 
differences between women and men in society, the consensual agreement 
on differential perceptions of each sex and the virtual universality of the 
discrimination based on sex tend to locate interactions between members of 
gender categories close to the intergroup end of the continuum. Hogg and 
Abrams (1988) or Williams and Giles (1978) come to similar conclusions. 
Here we would like to emphasize that the consideration of these power 
differences between sexes or the existence of outgroup discrimination 
(features closely related to the intergroup extreme) does not necessarily need 
to appear associated with anonymous or collective interactions or those 
which occur in the professional or public spheres. An interaction between 
two people, although charged with positive affection, with completely idio­
syncratic characteristics and situated within the sphere of intimacy, may still 
reveal powerful intergroup features. 
The definitions of intergroup behaviour given by both Sherif (1 967) and 
Tajfel (1978a), clearly show that what distinguishes interpersonal from 
group or intergroup behaviour is not the number of people implied. Take, 
for example, the hypothetical case of the businessman X who turns down 
the woman Y's application to work in his office because he considers that 
women are less intelligent and capable than men. If this case involved only 
a man and a woman, it would appear to be clearly based on an intergroup 
relationship, given that the businessman's behaviour (male) with regards to 
the candidate (female) was exclusively concerned with her belonging to the 
above group and not with her personal qualities. As Brown (1 988) shows, in 
order to establish at what point of the interpersonal-intergroup continuum 
is situated an interaction between two people from different social 
categories (such as male and female), we need to study in close detail the 
nature of that very interaction. If the behaviour and attitude of the man 
towards the woman (or vice versa) is that stereotypically associated with 
each gender, we can then begin to talk of intergroup behaviour. If this turns 
out to be impossible, the interaction's idiosyncratic features will emerge in a 
more interpersonal relationship. 
In most societies (although our data refer only to the Western world), 
boyfriends and girlfriends, wives and husbands and fathers and mothers 
behave differently, although coinciding to a great extent with the stereo­
types and roles that society associates with men and women (Peplau, 1983). 
A good example of this is that commonly heard saying: 'I can't do that, 


1 56 
Social identity: international perspectives 
because I'm a married man (woman) .' In this way, we can safely say that 
couples, even those characterized by high intimacy levels, also reveal 
important intergroup aspects. We will now go on to discuss our approach in 
more detail, referring to two areas in which uniformities can be seen in both 
male and female behaviour in a relationship between couples. 
(1) Conception o
f
 a couple. The majority of studies, carried out with middle­
class American college students, reveal great similarities in the expectations 
and conceptions that boys and girls have of couples and heterosexual 
relationships in general, although some differences have been found. While 
women from all cultural and economic backgrounds generally grant a great 
value to affection and companionship in their relationships, only male 
college students - but not those from a working-class background -
coincided with these feelings (Peplau & Gordon, 1985). Likewise, women 
grant a great importance to intimate communication in their relationships 
with men (Parelman, 1983). In general, men tend to have more traditional 
beliefs with regard to their relationships with women, in keeping with male 
roles and stereotypes (Peplau & Gordon, 1985). This result is as expected 
given that the traditional concept of a couple would appear to benefit men 
as opposed to women. 
(2) Pro
f
essional development and economic res
ponsibility. In accordance 
with traditional gender stereotypes, economic responsibility for both the 
couple's and family's welfare falls on the shoulders of the male. Greater 
priority is therefore given to his professional career. This is clearly reflected, 
along with other signs, in the proportion of men and women who work 
outside the home. 
Kotkin (1 983) found that 70% of married women at university had, in his 
study, sacrificed something (for example, economic support, change of 
address, renouncement of own career) to aid their partner's university 
studies or professional life, while only 10% of men had made the same 
sacrifice for their female partners. It will, therefore, come as no surprise that 
women intending to study at university or work hard at a professional career 
are often less committed to their couple (Rubin, Peplau & Hill, 1981). 
However, aside from the priority given to men's professional aspirations 
over women's, we are also concerned with the fact that simply getting 
married and having a family can act as an obstacle for women's professional 
success. For example, when Harrison, Moore and Rucker (1985) studied the 
biographies of more than two thousand eminent men and women who had 
appeared in various editions of Who's Who, they found that it was less likely 
for successful women to marry and have children, and more likely that they 
would get divorced. In the same way, Etaugh and Riley (1983) have shown 
that single women have higher income levels, from both employment and 
education, than their married counterparts, while the reverse model is true in 
the case of men. 


Social identity and inter
personal relationships 
1 57 
Another research area which would appear to reinforce the notion that 
women retreat from professional life within a couple deals with the image 
men have of successful women (in keeping with the traditional gender 
stereotypes), especially those working in typically male fields. In general, 
these studies portray a negative image, meaning that, to male eyes, these 
women would appear to be less attractive as partners (Pfost & Fiore, 1990). 
Thus, it would seem that although couples can be categorized as inter­
personal (as they involve two people, possess strong levels of intimacy and 
affection and can vary greatly in nature), they also possess strong intergroup 
elements: they emphasize the similarities of male attitudes and behaviours, 
on the one hand, and the similarities of female attitudes and behaviours, on 
the other, and in doing so they follow the traditional gender stereotypes. The 
fact that stereotyped concepts of gender are used in couples to evaluate and 
establish their behaviour (more so than other concepts) is hardly surprising, 
given that a great many of these stereotypes refer to male-female rela­
tionships. Thus, it would seem perfectly acceptable for a traditionally 
minded male to ask his wife to 'take the initiative' with regard to domestic 
problems and their children's education, or even in the realm of her work 
outside the home, but not in her direct dealings with him. 
Relationships between interpersonal and intergroup aspects 
There exist three fundamental ways of connecting the interpersonal and the 
intergroup: (1) as orthogonally related; (2) as qualitatively different aspects 
of reality; and (3) as parts of a continuous flow. We shall examine each in 
turn. 
1. Orthogonality o
f
 both dimensions 
Stephenson (1984) was one of the first researchers who, from within social 
identity theory, established the possibility that the interpersonal and inter­
group need not be placed at opposite ends of a single dimension but rather 
act as two independent dimensions. In this way, a relationship could be 
placed at a given point in a bi-dimensional space, resulting from the 
intersection of both dimensions, as shown in Figure 1 1 . 1 .  
It is possible that there exist certain emotional-sexual relationships 
between men and women in which neither interpersonal nor intergroup 
aspects are of great importance. This would be the case in an incipient 
relationship in which contact between the two members of the couple is 
rather low - the fact that both interpersonal and intergroup requirements 
are low does not imply that they are of equal importance. In this case, it is 
likely that the intergroup aspect would predominate over the interpersonal. 
Other relationships can reflect a high intergroup requirement and a low 
interpersonal one: marriages of convenience provide a good example, or 


1 58 
Social identity: international pers
pectives 
INTERGROUP DEMANDS 
High 
'Compromise' 
'Conflict' 
High 
INTERPERSONAL 
DEMANDS 
'Collusion' 
'Apathy' 
Low 
INTERGROUP DEMANDS 
Low 
INTERPERSONAL 
DEMANDS 
Figure 1 1 . 1  Inter
personal and intergroup demands according to Stephenson (1984) 
those marriages typically found in the Victorian age, in which intimacy was 
virtually non-existent and almost the entire relationship and the behaviour 
of the couple had been already established by social criteria. Relationships 
with a high interpersonal and a low intergroup component would be those 
in which the couple challenges established social criteria concerning behav­
iour and attitudes expected from men and women involved in a love affair 
and create a complete different and unique relationship. Finally, it should 
also be possible to find relationships in which both the interpersonal and 
intergroup aspects are present. A graphic example of this could be the 
relationship portrayed in Liliana Cavanni's film The Night Porter. This 
story concerns an ex-Nazi official and a former female prisoner who main­
tain a passionate relationship that reflects (to a great extent) intergroup 
requirements since a great part of their behaviour is stereotypical (domi­
nance, aggression, and so on, in the male and submission, passivity, and so 
on, in the female). However, at the same time, interpersonal aspects are 
equally important: the relationship is deeply idiosyncratic and strange, 
considering they belong to entirely separate groups, their romance thus 
being distinguished from the millions of male-female relationships estab­
lished during the Second World War. 
An example of the simultaneity of interpersonal and intergroup aspects 
appears in our study (Moya, 1987) on social episodes. Following Forgas's 


Social identity and interpersonal relationshi
ps 
1 59 
approach (1 979) a number of subjects (60 female and 36 male psychology 
students, 9 male and 13 female student nurses and 23 housewives) were 
asked to indicate the perceived similarity among a set of situations rep­
resenting usual interactions with members of the opposite sex and rate these 
same situations on 1 5  bipolar scales. Both the situations and the scales had 
been obtained in a previous session and were all representative of the 
heterosexual interactions of each group. Examples of these interactions 
include 'talking on the telephone with a boyfriend/girlfriend', 'having sexual 
contact', or 'studying with colleagues'. Here, our interest lies in those 
episodes in which the member of the opposite sex is a partner. For example, 
in the case of the housewives, these episodes included: 'speaking with my 
husband', 'having a drink with my husband', 'making love with my 
husband', and in the case of the student nurses, 'eating with my partner', 
'having arguments with my partner', 'making love with my partner' and 
'waking up next to my partner'. 
The results show that the subjects perceive these episodes as intergroup 
interactions, in the sense that they consider their behaviour here as being 
very similar to that of members of the same sex (remember that uniformity 
is one of the fundamental characteristics of intergroup behaviour). Thus, on 
the scale of 0 to 9 points whose extremes were 'I behave differently to how a 
member of my sex would' and 'I behave in a similar way . .  
.', the scores 
obtained in those episodes involving a couple were the following: 7.3 in 
male psychology students and 7.7 in female; 4. 1 5  in male student nurses and 
6.3 in female; and 6.92 in housewives. At the same time, however, these 
interactions were perceived as being highly interpersonal, since the subjects 
saw themselves as expressing great intimacy, emotional involvement and 
dignity. They also believed that they expressed themselves spontaneously in 
the relationship and felt treated as persons. This is confirmed by the results 
of a multidimensional scale analysis (INDSCAL by Carroll & Chang, 1970) 
of the ratings of the interaction episodes on the 1 5  bipolar scales. See, for 
example, the episodic space obtained by the housewives, where two 
dimensions proved sufficient to represent the space. 
The first dimension was defined by the scales 'passive-active' (-.86), 
'intimate-non-intimate' (-.79), 'boring-interesting' (-.71), 'intense-super­
ficial' (-.58) and 'frustrating-rewarding' (-.51). To our way of thinking, this 
dimension clearly refers to the emotional aspect of the episodes - which 
could be termed intimacy - and is related with the person's degree of 
emotional involvement in the situation. The episodes 'chatting with our 
child', 'making love with my husband', 'playing with our child' and 
'interactions with friends' were situated at one extreme, while 'speaking with 
the children's teacher' and 'going to the doctor' were placed at a midway 
point on the dimension. Finally, at the other extreme were positioned the 
shortest interactions with virtual strangers (shopkeeper, baker, gas man). The 
second dimension was defined by aspects of dignity, equality and control 
perceived by the person in the episode, the scales defining this dimension 
being: 'relaxed-tense' (-.89), 'pleasant-unpleasant' (-.8), 'cooperative-


1 60 
Social identity: international perspectives 
competitive' (-.8), 'free-coerced' (-.79), 'I feel treated as a person-I don't 
feel treated as a person' (-.77), 'I know how to behave-I don't know how to 
behave' (-.75), 'I express myself as I am-I don't express myself as I am' 
(-.74) and 'reciprocated-not reciprocated' (-.71). At one extreme of this 
dimension were located the following episodes: 'a man is pursuing her', 
'quarreling with our child', 'going to the doctor', 'speaking to the children's 
teacher' and 'speaking to an office worker'. At the midway point more 
superficial episodes were placed ('going to the hairdressers', 'chatting with a 
neighbour') and at the other extreme were situated 'making love with my 
husband' and the relationships with children and neighbours. 
As can be seen, there is no interpersonal-intergroup dimension along 
which the interactions of these housewives with men can be placed sub­
jectively so that interpersonal episodes could be placed at one extreme -
relationship with partner or with child - and the intergroup ones at the 
other - a man is pursuing her, going to the hairdresser. An interpersonal 
relationship (for example, with their child) is perceived in an entirely 
different way depending on whether the housewife is playing or arguing 
with himlher. 
2. The interpersonal and intergroup as qualitatively diff
erent aspects 
o
f identity 
Kay Deaux (1992, 1993) has recently established that social and personal 
identities, rather than being clearly distinguishable, are in fact deeply inter­
related. According to Deaux, social identities can be defined as those roles 
or categories within which a person feels represented. Personal identity 
would refer to those features or forms of behaviour which a person feels 
adequately describe himlherself; these characteristics are often intimately 
associated with one or more social identities. In this way, a male X can 
regard himself as being very much a typical man, similar to the majority of 
men in his social surroundings (social identity), and at the same time 
consider himself as 'affectionate' (personal identity). This latter character­
istic is probably more associated with his belonging to the category of 
father, grandfather or spouse than to that of a military officer (imagine that 
this male belongs to all these categories or plays all these social roles). In 
this way, personal identity is defined, at least in part, by our membership of 
a group, at the same time as these group memberships will, in themselves, 
be charged with personal meaning. 
It should also be borne in mind that identity can be perceived as a 
hierarchical structure in which groups of identities are related to categories 
of characteristics or attributes. For example, the identity of a female Y 
would be exemplified as shown in Figure 1 1 .2. The upper part of the figure 
shows social identities and the lower - as if it were a mirror - the com­
ponents of a personal identity. 
One of Deaux's primary concerns surrounds this hierarchical structure of 
social identity. The higher a determined identity, the more behavioural 


I 
Woman I 
1 Maintains interpersonal 1 
relationships 
1 
Professional & 1 
independent 
1 
Close 1 
1 Superficial 1 
1 Professional 1 
1 
With a progressive 1 
ideology 
I 
1 Involved in 1 
Housewife 1 1 Popular 1 
1 University 1 
1 
Boss 1 
1 
Feminist 1 1 
Lesbian 1 
a couple 
with men 
student 
- Worried about 
- Devoted to others 1- Attractive 1 
1- Ambitious 1 1- Ambitious 1 
- With social 
- Independent 
her appearance 
- Kind 
- Sexy 
- Emotional 
- Aware of her 
- Attractive 
feelings 
1- Kind 
1 
- Committed 
1- Nice 
1 
- Good-looking 
1- Helpful to 
others 
1 
- Works hard 
- Competitive 
worries 
- With many 
I 
I 
- Ambitious 
girlfriends 
I 
I 
1- Works hard 
1 
- Follows her own interests 
1- Fights against injustice 1 
- Active 
1- Follows her 
own interests 1 
- Sensitive 
- Practical 
- Affectionate 
- Responsible 
Figure 1 1 .2 
Example o
f
 a hypothetical identity taken f
rom Deaux's f
ormulation and Maya's data (Maya, 1 989; Maya & Peplau, in 
preparation) 
R 
'" 
is' 
--
) 
 
S. 
I:l 
;::
 
:::.
 
S· 
ő 
~ 
(1) 
' 
;::
 
I:l 
--
€ 
::r 
T. 
;::
 
} 
.;:;. 
'" 
0\ 


1 62 
Social identity: international perspectives 
Table 1 1 . 1  Meta-contrast ratios in self-perception 
and the f
our stimuli used by Moya and Pep/au 
UCLA 
Granada 
Person involved in a couple 
1 .09 
1 .25 
University student 
1 .22 
1 . 1 3  
Professional 
1.03 
0.88 
Committed to their family 
0.95 
1.04 
repercussions this will have for an individual. Although it is likely that 
certain exceptional identities occupy high positions in almost all individuals 
in a society (such as being male or female), it is also possible that there exist 
great individual or subcultural differences. Maintaining her position in a 
couple may, for a certain woman be thought of as belonging to an entirely 
secondary category included in, or subordinate to, her membership of 
another group, such as that of a high social status. On the other hand, for 
another woman, being part of a couple could be considered as the most 
general category, into which all other memberships or social roles fall. It is 
worth remembering that any characteristics associated with one social 
identity can be associated with others and that this association between 
social identities and personal characteristics can vary from one individual to 
the next or from one social group to another. 
Some of these points were verified in recent research (Moya & Peplau, in 
preparation). A total of 204 subjects (95 men and 1 09 women) took part 
in the first test, all students of an introductory course in psychology at the 
University of California at Los Angeles (UCLA). In the second test, 45 men 
and 1 25 women participated, all third-year psychology students at the 
University of Granada. All subjects were asked to rate themselves on 22 
bipolar scales (9 represented instrumental attributes, 9 expressive and 4 
neutral). They were then asked to rate four different types of people of the 
same sex on the same 22 scales (university student, person involved in a 
lasting relationship, a committed family man/woman and a professional). 
With these scores, 4 meta-contrast ratios were calculated, one for each 
stimulus, following Turner, Hogg, Oakes, Reicher and Wetherell's proposal 
(1987). If the meta-contrast ratio is lower than I, the stimulus in question is 
perceived as being a member of the outgroup. The higher above I, the 
greater the stimulus is perceived as a member of the ingroup. The meta­
contrast ratios obtained by the American and Spanish students appear in 
Table 1 1. 1 .  
UCLA students only considered persons committed to their family as 
outgroup members (RMC lower than 1), while for those from Granada, 
only the professional was considered as such. Likewise, both sets of students 
perceived themselves as being closer to the person involved in a relationship. 
An analysis of variance was performed using as the dependent variable 
the meta-contrast ratio. The independent variable was subjects' current 
involvement in a relationship and had two values: involvement versus non-


Social identity and interpersonal relationshi
ps 
1 63 
involvement. For UCLA students involved in a relationship, the meta­
contrast ratio with the person committed to the family was .89. For those 
not involved, the corresponding figure was 1 .01 F(I ,  1 55) = 7.26 was 
significant at p < .01 .  No similar differences appeared in the Granada 
sample. 
The importance of taking into account the specific contents that 
individuals associate with each social identity has also been shown in other 
results of this research. According to our hypothesis, those involved in an 
emotional relationship with a member of the opposite sex rated themselves 
as higher in the expressive items than those not involved. On the other 
hand, and contrary to expectations, this was also the case with instrumental 
items (although here the differences did not reach statistical significance). 
These tendencies were found in both the American and the Spanish sample. 
In our judgement, these results reveal the existence of different social 
identity configurations in two distinct social mediums (the American and 
the Spanish) since being a university student or a committed family member 
seems to occupy different positions in identity structure, while the relation­
ship between these identities in each sample also presents certain differences 
(as in the opposition present in the American sample between those 
involved in a relationship or not). 
3. The continuous flow o
f
 identities 
Turner and coIleagues (1987) conceive of the existence of different levels of 
abstraction in the categorization of '1', separated from each other and 
discretely formed, in such a way that the individual passes from one to 
another according to how the situation varies. It is therefore the dynamic 
character of identity which is emphasized here, an aspect that has been 
present in social identity theory but which is rarely reflected in practical 
research carried out under its inspiration (Condor, 1989). Tajfel (1982a) 
emphasized how social identifications could be subject to temporal fluctu­
ations: 'the psychological existence of a group for its members consists of a 
complex sequence of appearances and disappearances, of factors emerging 
and blurring into thin air' (p. 485). Turner considered his model of identity 
as neither static, fixed, global or corrected, but totaIly the reverse: 'the 
categories of "Self" are not absolute, but rather "Self" is dynamic, relational, 
comparative, fluid, context-specific and variable' (Turner, 1988, p. 1 14). 
Thus, it is possible that the interactions maintained by members of a 
couple change continuously from some category memberships to others (as 
conceived by self-categorization theory) or fluctuate incessantly along the 
interpersonal-intergroup continuum. For example, deciding what to do one 
Sunday afternoon, it might occur to a boy to invite to the cinema the girl he 
has just started seeing, to then pay for her ticket and attempt some physical 
or sensual contact (foIl 
owing certain stereotypical criteria). He suddenly 
finds out that the girl cannot stand going to the cinema and would much 


1 64 
Social identity: international perspectives 
rather have an afternoon chat in a quiet cafe, he thus rapidly changes his 
plans. In this way, a form of behaviour which commenced with certain 
intergroup components has quickly found itself being affected by the 
particular characteristics of the people involved. 
Final comments 
This chapter's purpose has been two-fold. On the one hand, it has 
attempted to show that the study of intergroup relations and the processes 
and structures related to identity can be particularly enhanced if the 
interpersonal relationships, in which the individuals are involved, are taken 
into account at the same time (for example, relationships between couples). 
On the other, it has indicated how the study of interpersonal relationships 
can be enhanced if we simultaneously bear in mind identity and intergroup 
aspects. We believe that an articulation of intergroup and interpersonal 
processes could be more useful in the understanding of certain phenomena 
than considering these two aspects as antagonistic or contrary. 
However, this articulation between interpersonal and intergroup aspects 
seems relatively complex. We have made reference to some approximations 
that contain interesting suggestions regarding how these two components 
should be articulated. Yet, rather than coming up with definitive answers, 
these approximations only suggest paths down which one can advance, thus 
granting an important role to empirical research. 
Focusing now on the case of couples, the arguments we have put forward 
show that a greater understanding of them could be attained if, in addition 
to considering the specific dynamics of the relationship, other factors were 
also taken into account. These could include the social identity structure of 
those involved in a couple, as well as the content of their different social 
identities. Nevertheless, the study of these identities demands approxima­
tions which consider their relational character (Rosenberg & Gara, 1985). 
So, although identities of a general character exist (such as 'being a man' or 
'being a woman') and guide the behaviour of individuals without involving 
other people (for example, they serve the individual when deciding how to 
dress or what hobbies to take up), there are others (such as 'being in a 
couple' or 'being a father') which cannot be understood without taking into 
account a relationship with other people (for example, his/her partner or 
the specific composition of the family as in the mentioned examples). 
Consider, for example, the different conception of self felt by a 'woman 
in a relationship' before and after her partner had been paralysed in an 
accident. 
In conclusion, this approach merely connects up with an old tradition of 
social psychology (Morales, 1989) which has included: the battle over 
structural and process aspects of identity (for example, Manford Kuhn and 
Ralph Turner versus Blumer and Goffman within symbolic interactionism); 
the extraordinarily discerning analysis of social and personal identity carried 


Social identity and inter
personal relationships 
1 65 
out by, for example, Goffman (1 968); and establishing the impossibility of 
understanding 'self' or the identity, disregarding the specific social inter­
actions maintained by individuals. 
Note 
I .  The content of Figure 1 1 .2 has an illustrative purpose. The structure of the identity 
within has been taken from Deaux (1993) and the content of the different identities has been 
elaborated from the results of research carried out by Moya (1989) and Moya and Peplau (in 
preparation). 


1 2  
Social Identity and Aspects of Social 
Creativity: Shifting to New Dimensions of 
Intergroup Comparison 
Steve Hinkle, Laurie A. Taylor, Lee Fox-Cardamone and 
Pamela G. Ely 
Most accounts of social identity theory (SIT) outline three response patterns 
which groups and group members can make to unfavourable social 
identities: individual mobility, social competition and social creativity (see 
Hogg & Abrams, 1988; Tajfel & Turner, 1986). Perhaps the earliest dis­
cussion of social creativity can be found in two of Tajfel's 1978 publications 
concerning group differentiation (l 978d) and the social psychology of 
minorities (l978e). In these works, Tajfel discussed low status or minority 
groups' response to their unsatisfactory social identities. When social cir­
cumstances supported leaving the ingroup, social mobility or moving on to 
another, presumably high-status, ingroup was proposed. When mobility was 
not sociologically or psychologically likely, Tajfel proposed three 
possi 
bili 
ties. 
(a) to become, through action and reinterpretation of group character­
istics, more like the superior group; 
(b) to reinterpret the existing inferior characteristics of the group, so that 
they do not appear as inferior but acquire a positively valued distinc­
tiveness from the superior group; 
(c) to create, through social action and or diffusion of new 'ideologies' new 
group characteristics which have a positively valued distinctiveness 
from the superior group (Tajfel, 1978c, pp. 93-94). 
Later, this list was elaborated to include a fourth possibility: 
(d) group members may also seek positive distinctiveness by 'changing the 
out-group (or selecting the out-group) with which the in-group is 
compared - in particular, ceasing or avoiding to use the high-status 
out-group as a comparative frame of reference' (Tajfel & Turner, 1979, 
p. 43). 
In more recent discussions of SIT, the first of these is typically referred to as 
'social competition', while the latter three are, generally, considered as 


Social identity and aspects o
f
 social creativity 
1 67 
aspects of 'social creativity' (see Hogg & Abrams, 1988; Tajfel & Turner, 
1986).' 
The purpose of this chapter is to take a detailed look at one of these 
aspects of social creativity: the establishment of new group characteristics or 
dimensions of intergroup comparison as a means to institute positive 
distinctiveness for the ingroup. First, theoretical issues pertinent to this 
aspect of social creativity will be summarized. The chapter will then review 
past research pertinent to establishing positive ingroup distinctiveness 
through the use of new dimensions for intergroup comparison. New, as yet 
unpublished, research we have conducted with several other colleagues will 
then be presented. The chapter will conclude with some evaluative com­
ments on this aspect of SIT and SIT research along with discussion of 
applications and issues for future research. 
Some theoretical issues 
Legitimacy and stability 
At the core of many SIT predictions are the notions of status legitimacy and 
stability. The relationships between groups can be perceived as legitimate or 
illegitimate, stable or instable (Tajfel 1978a). In general, Tajfel saw legiti­
macy and stability as covarying. That is, according to SIT, legitimate and 
stable intergroup relations tend to co-occur as do illegitimate and instable 
intergroup relations (Tajfel, 1978a; for contrasting perspective, see Turner & 
Brown, 1978, who treat legitimacy and stability as independent aspects of 
the intergroup situation). SIT goes on to associate social competition with 
illegitimate/instable intergroup relations. Hence, the social creativity 
strategies, including establishing positive ingroup distinctiveness on new 
dimensions of intergroup comparison, are postulated by SIT as being most 
relevant to legitimate and stable intergroup contexts (Tajfel, 1978b). Here, 
where groups with threatened or negative social identities see the status quo 
as justified and/or not likely to change, direct social competition to alter the 
status quo on prevailing comparison dimensions is not apt to be viewed as 
likely to succeed or justifiable. Hence, the creation of new comparison 
dimensions as a path to positive social identity is a strategy of obvious 
motivational significance. 
Group status 
Another theoretical issue posited as relevant to the use of new intergroup 
comparison dimensions is group status. While SIT is not as clear as it could 
be, it appears as though Tajfel believed that the use of new comparison 
dimensions was most likely for low-status groups in legitimate/stable status 
circumstances, but most likely for high status groups when they perceive 
their advantaged position as illegitimate and/or instable (Tajfel, 1978c). 
However, not all of the early SIT theorizing is consistent on this. For 


1 68 
Social identity: international perspectives 
example, Caddick (1982) sees illegitimacy and instability as instigating the 
search for new dimensions of distinctiveness for both high- and low-status 
groups. It also appears that, irrespective of illegitimacy and instability, early 
SIT research viewed the use of new comparison dimensions as being more 
characteristic of low- than of high-status groups (Brown & Ross, 1982). 
Selection o
f
 new comparison dimensions 
Perhaps one of the most useful and detailed early theoretical discussions in 
SIT of the use of new dimensions for intergroup comparison is that of van 
Knippenberg (1978). In particular, he offers some valuable ideas regarding 
the selection of new comparison dimensions. Assuming legitimacy and 
stability, existing status differences will continue and be difficult to deny. As 
a result, van Knippenberg proposed that new comparison dimensions must 
be capable of enhancing a group's social identity, but without affecting 
already existing status dimensions. That is, in selecting new comparison 
dimensions to effect a more favourable social identity, the new dimensions 
will afford the ingroup positive distinctiveness while being uncorrelated with 
the existing status dimension where the ingroup does not compare 
favourably. 
Lemaine, Kastersztein and Personnaz's (1978) position is somewhat 
different from those of Tajfel (1978c) and van Knippenberg (1978). The 
latter two authors take no position regarding the relative importance of new 
versus old dimensions of comparison. In contrast, Lemaine et al. clearly 
hold that groups will strive to make new bases for intergroup comparisons 
dominant. For Lemaine et a!., groups will not be content to simply 
acknowledge their own inferiority on existing status dimensions while 
claiming superiority on a new status dimension. Rather, the ingroup will go 
further and strive to make the new basis for comparison of greater 
importance and, thus, the dominant basis for status comparisons. In a later 
paper, van Knippenberg (1984) modifies his earlier stance in a manner 
consistent with the views of Lemaine et a!., at least with respect to the 
classic conflictual intergroup situations emphasized by SIT. He also pro­
poses socially cooperative intergroup relations where groups coexist in a 
climate of mutual appreciation. Here, van Knippenberg (1 984) proposes 
that groups can and will acknowledge both their own superiority and 
inferiority relative to other groups on various comparison dimensions and 
that all such dimensions can be valued by all groups. He does, however, 
note that ingroups and outgroups may weight the relative importance of the 
sundry dimensions differently. 
New comparison dimensions versus other responses to un
f
avourable 
identities 
Another important theoretical issue concerns the prepotency of turning 
to new bases for intergroup comparison relative to other responses to 
unfavourable identities such as mobility, social competition and other social 


Social identity and aspects o
f
 social creativity 
1 69 
creativity strategies, including shifting the locus of comparison to new 
outgroups and re-evaluating the valence of existing, unfavourable com­
parisons. Tajfel (1978c) clearly sees mobility as the strategy of choice in 
circumstances where mobility is viewed as possible and desirable (that is, 
with a social mobility ideology; Hogg & Abrams, 1988; Tajfel & Turner, 
1979). However, there is little in his writings to clearly order social com­
petition and social creativity into a hierarchy of strategic choice. Certainly, 
he says virtually nothing concerning the relative strength of the various 
forms of social creativity. In contrast, Ball, Giles and Hewstone (1984) 
propose that mobility attempts will precede social creativity strategies, 
which, in tum, will take place prior to social competition. However, Ball et 
al. say nothing about the ordering of the various forms of social creativity. 
In general, the lack of theory regarding use of the various social creativity 
strategies is an unfortunate omission in SIT. 
Process issues 
Finally, what does SIT say relevant to the process of creating new bases for 
positive intergroup distinctiveness? First, group members must somehow 
develop a shared perspective on the attributes serving as new bases for 
intergroup comparisons and the ingroup's standings on these new attributes 
must, in fact, be positive (Tajfel, 1978c). The next step involves persuading 
the outgroup to accept the new bases for intergroup comparison and the 
ingroup's positive standing on these attributes. That is, consensual accept­
ance by both the in- and outgroup is necessary for the new dimensions of 
intergroup comparison to fully validate the ingroup's positive social identity 
(Tajfel, 1978c). Lemaine et al. (1978) add to SIT theorizing about such 
processes. They hold that the initial intragroup aspect of this process will, 
ideally, take place without the knowledge of the outgroup. Such seclusion 
shields the ingroup seeking new bases for intergroup comparisons from 
criticism and helps reduce uncertainty regarding the new comparisons. 
Past empirical research: a selected review of the literature2 
In spite of the complex SIT theorizing concerning the use of new com­
parison dimensions as a response to negative social identities, most of the 
relevant empirical literature has yet to address the theoretical postulates 
outlined above. As we shall see, the dominant characterization of this 
literature is simply the demonstration that groups with threatened social 
identities will, at times, tum to new bases for making intergroup com­
parisons. 
In all likelihood the most frequently cited empirical work is that of 
Lemaine and his colleagues (Lemaine et aI., 1978). In one study, children at 
summer camps participated in a building competition where some groups 
were relatively disadvantaged by being given inadequate materials. While 


1 70 
Social identity: international perspectives 
unable to build as large or elaborate huts as other groups, disadvantaged 
groups began developing gardens, thereby establishing a new basis for 
intergroup comparisons more favourable to their ingroups. 
In a review, Hinkle and Brown (1990) cited 12 papers, using both 
laboratory and field contexts with subject populations from nurses to 
university lecturers; all demonstrated that the same group when making 
intergroup comparisons on multiple dimensions will evince in 
group 
favouritism on some of the dimensions, outgroup favouritism on others, 
and no favouritism at all on still other dimensions. The simultaneous 
occurrence of ingroup and outgroup favouritism on differing dimensions is 
consistent with the notion of using new bases for intergroup comparisons to 
establish positive social identities when comparisons on other dimensions do 
not favour the ingroup. In addition to the studies reviewed by Hinkle and 
Brown, other similar results are reported in Capozza, Bonaldo and 
DiMaggio (1982). The diversity of subject populations and paradigms 
across these various researches points to the robustness of this form of 
social creativity. 
Perhaps one of the most interesting early SIT studies bearing on the use 
of new dimensions for making intergroup comparisons is Brown and Ross 
(1982). The study's primary emphases are on both social competition and 
social creativity. In particular, Brown and Ross's consideration of creativity 
emphasizes the re-evalution of the importance of various comparison 
dimensions rather than the topic of the present chapter, the use of new 
comparison dimensions. Nonetheless, while the latter issue is not focal to 
Brown and Ross, they do present some important relevant data. Before 
discussing their findings, some background regarding the paper's hypotheses 
and design is necessary. 
Brown and Ross discuss illegitimacy and instability as inducing a 'threat 
to identity' or insecure social identity which, in response, produces either 
social competition or social creativity ('the strategies . . .  
were not thought 
to be mutually exclusive', p. 1 62). Their study's primary manipulation was 
presented in terms of high, moderate or low threats to identity, though 
careful examination of their procedures indicates that the study's primary 
independent variable can just as easily be termed 'status illegitimacy'. A 
second important independent variable in this research was a manipulation 
of ingroup status as being either superior or inferior. One of the 
experiment's primary predictions was 'that increases in differentiation along 
. . .  
alternative dimensions would . . .  
vary in proportion to the threat to 
identity experienced, and that these changes would be most prominent in 
the lower status groups' (p. 1 62). 
When ingroup bias on alternative dimensions of intergroup comparison is 
compared before and after implementation of the status illegitimacy 
manipulation, the data clearly show that inferior status groups in the high 
and moderate illegitimacy (that is, threat to identity) conditions manifested 
substantially more ingroup favouritism on the alternative comparison 
dimensions. This shift towards greater ingroup bias on the alternative 


Social identity and aspects o
f
 social creativity 
1 7 1  
dimensions did not occur for inferior status groups in the low illegitimacy 
condition, nor did it occur for superior status groups in any of the three 
conditions of status legitimacy.3 
These results are partially supportive of early SIT theorizing in that the 
use of new comparison dimensions to attain positive group distinctiveness is 
more characteristic of low- rather than high-status groups. On the other 
hand, Brown and Ross's results contradict Tajfel's (1978c) view that this 
and other forms of social creativity would be most prominent for low-status 
groups when there are legitimate/stable status relations. Similarly, their 
findings are also inconsistent with Caddick's (1982) prediction that the 
various social creativity strategies would occur for both high- and low-status 
groups when the intergroup status structure was both illegitimate and 
instable. It appears from Brown and Ross's data that the most empirically 
tenable theoretical position for SIT is that social creativity, at least in the 
form of using new dimensions for making ingroup-favouring intergroup 
comparisons, is most likely to be observed in low-status groups under 
conditions of status illegitimacy and/or instability (though the specific 
effects of illegitimacy and instability as distinct variables remain to be 
tested). 
Other interesting data from Brown and Ross bear on the issue of the 
valuing of new dimensions of intergroup comparison. Recall Lemaine et 
al.'s (1987) and van Knippenberg's (1984) hypotheses that groups shifting to 
new dimensions of intergroup comparison to attain positive distinctiveness 
would also see these new dimensions as of greater importance than previous 
bases for intergroup comparison. In general, Brown and Ross found that 
while inferior status groups were, with status illegitimacy, quick to use 
alternative comparison dimensions to evince ingroup favouritism, there was 
little evidence to support the notions that these new dimensions' importance 
increased or that they were viewed as being of greater importance than the 
original intergroup comparison dimensions. This result is consistent with 
similar findings reviewed by Hinkle and Brown (1990). Indeed, the existing 
evidence seems, if anything, more consistent with the view that low-status 
groups are willing to salvage damaged or unfavourable social identities by 
shifting to or emphasizing new in 
group-favouring dimensions of com­
parison even while acknowledging that the latter are of lesser importance. 
An example of this might be the losing sports team that emphasizes its own 
superiority in terms of better sportsmanship (see Lalonde, 1992) or more 
enjoyment of the game, even while willingly owning that it is more 
important to win. 
Caddick (1982) presents similar findings with respect to illegitimate status 
differences and the occurrence of ingroup favouritism on new dimensions of 
intergroup comparison. An illegitimate status difference led to increased 
expression of ingroup favouritism on several comparison dimensions other 
than the original basis for intergroup comparison. Interestingly, and in 
contrast to Brown and Ross, ingroup favouritism on the new comparison 
dimensions occurred for members of both low- and high-status groups. 


1 72 
Social identity: international perspectives 
Caddick develops a thought-provoking analysis of this result based on the 
fact that in his experimental procedure the high-status group is clearly 
responsible for the illegitimacy of the initial comparison between the 
groups. Caddick suggests that this may have compromised the high-status 
group's social identity through guilt feelings induced by violation of values 
such as fairness or justice. To compensate for their compromised favourable 
social identities, the high-status groups adopted new bases for making 
ingroup-favouring intergroup comparisons. 
Some new empirical findings 
Our own investigation of social creativity issues began rather unexpectedly 
through an effort to replicate findings from a study by Condor, Brown and 
Williams (1987). One component of Condor et al.'s experiment involved 
groups of unacquainted subjects working on a laboratory task to generate 
ideas for recruiting students to their university under an instructional set 
emphasizing task issues such as efficiency and productivity. Following this 
exercise, subjects watched a short video of another group engaged in the 
same endeavour and then made in- and outgroup ratings on task-oriented 
dimensions such as the group's competence and energy and socio-emotional 
dimensions such as the groups' friendliness and warmth. Condor et al. 
found much greater ingroup favouritism on the task-oriented dimensions of 
intergroup comparison than on the socio-emotional dimensions. This was 
expected since the task-oriented dimensions are more closely associated 
with the experimental instructional set than the socio-emotional dimen­
sions. 
Hinkle and Crook (1 987) attempted to replicate Condor et al. as closely 
as possible and found a complete reversal of their findings. That is, signifi­
cantly greater levels of ingroup favouritism occurred on the socioemotional 
dimensions than on the task-oriented dimensions. Why? There were only 
two important differences between the procedures of Condor et al. and 
Hinkle and Crook. First, the subject populations differed, with the latter 
study using American university students while the former's subjects were 
students from a British university. While it is virtually impossible to rule out 
differences in the populations as an account for the reversal of findings, one 
of the results from Hinkle and Crook points to the comparability of the two 
subjects populations. Their subjects saw the task-oriented dimensions as of 
greater importance than the socio-emotional dimensions. Thus, Hinkle and 
Crook's findings appear to be a social creativity effect in that their subjects 
are manifesting the most ingroup favouritism on a secondary, less important 
dimension of intergroup comparison. 
There was a second difference in the procedures of Condor et al. and 
Hinkle and Crook. Rather than presenting the outgroup via a video, the 
latter simply asked subjects to think of an average group of university 
students working on the same task and rate this 'imagined' outgroup. 


E 
(/) 
'." 
'§ 
0 
> 
† 
a. 
:::l
 
e 
CJ) 
.E: 
Social identity and aspects o
f
 social creativity 
35 
30 
25 
20 
1 5  
1 0  
5 
0 
-5 
Below 
average 
Comparison dimension 
• Task-oriented 
o Socio-emotional 
Average 
Outgroup status 
Well above 
average 
Figure 12. 1 
/ngroup f
avouritism on dif
f
erent comparison dimensions as a 
f
unction of outgroup status (Hinkle, Fox-Cardamone, Taylor & Crook, 
/988) 
1 73 
Assuming a rather high level of achievement motivation amongst these 
university student subjects, it may be that their vision of an 'average' group 
is of one whose performance does indeed threaten that of their own group. 
This interpretation corroborates the view that Hinkle and Crook's subjects 
are engaging in a social creativity process. If the outgroup's task perform­
ance is imagined to be quite good relative to ingroup performance, subjects' 
shift to the less important socio-emotional dimensions to manifest ingroup 
favouritism may be seen as an effort to offset the unfavourable intergroup 
comparison on the task-oriented dimensions. 
Experimental demonstrations o
f
 social creativity e
f
f
ects 
To follow up on the interpretation outlined above, Hinkle, Fox-Cardamone, 
Taylor and Crook (1988) replicated Hinkle and Crook's experiment and 
extended it by adding two additional experimental conditions. While some 
groups of subjects were asked to think of an average group of university 
students as the outgroup stimulus, others were asked to think of below 
average or well-above average groups. It was predicted that below average 
outgroups would not threaten the distinctiveness of the ingroup on the 
important, task-oriented dimensions. Hence, subjects in this condition were 
expected to manifest greater ingroup favouritism on the task-oriented 
dimensions than the socio-emotional dimensions. In contrast, the prediction 
for the average and well-above average outgroup conditions was that such 
outgroups would threaten the ingroup's distinctiveness on the important task­
oriented dimensions. As a result, social creativity would lead to greater 
ingroup favouritism on the secondary dimensions of intergroup comparison, 
that is, the socio-emotional dimensions. As can be seen in Figure 12. 1 ,  this is 
exactly what was found. 


1 74 
Social identity: international pers
pectives 
Interestingly, Hinkle et al. collected data on the perceived importance of 
the task and socio-emotional comparison dimension a
f
ter subjects had made 
their ingroup and outgroup ratings. There was no evidence for any change in 
the importance of the dimensions. The task-oriented dimensions were still 
seen as being significantly more important than the socio-emotional 
dimensions in all three experimental conditions. Thus, the social creativity 
seen in the average and well-above average outgroup conditions does not 
appear to actually compensate for unfavourable comparisons on the task­
oriented dimensions. In these conditions subjects are not saying, 'Your group 
is better at the task, but it is more important to get along well and we are 
better at that.' Rather, they appear to be saying, 'While your group is better 
than ours at the task, at least we get along better.' The social creativity seen 
in this study does not appear to compensate for an unfavourable social 
identity, but simply to take a step towards salvaging the ingroup's social 
identity. As with earlier research reviewed above, this result is inconsistent 
with Lemain et al.'s (1978) and van Knippenberg's (1984) views that the shift 
to new, ingroup-favouring dimensions of intergroup comparison would 
necessarily be accompanied by perceiving the new dimensions as of greater 
value or importance than those upon which the ingroup did not compare 
favourably. 
The Hinkle et al. research can be criticized for the use of an imagined 
manipulation of outgroup status. Recently, two of us (Laurie Taylor and 
Steve Hinkle) conducted an experiment similar to those reported above, but 
incorporating a more concrete manipulation of outgroup status. Again, 
groups worked under a task-emphasis instructional set to generate ideas for 
the recruitment of students to their university, and the ingroup and out­
group were rated on both task-oriented and socio-emotional dimensions of 
intergroup comparison. However, rather than asking students to imagine 
outgroups of varying status levels, outgroup status was manipulated by 
providing subjects with a list of some of the recruitment ideas generated by 
the other group. These lists were prepared in light of pilot test data such 
that they contained only the lowest (below average outgroup status) or 
highest (above average outgroup status) rated ideas from the pilot study. In 
the below average outgroup condition, ingroup favouritism was greater on 
the task-oriented dimensions than on the socio-emotional dimensions. In the 
above average outgroup condition, a social creativity effect was again 
observed: greater ingroup favouritism on the socio-emotional comparison 
dimensions than on the task-oriented dimensions (see Figure 12.2). 
Conclusions, applications and directions for future theory and 
research 
General conclusions 
The strongest and most consistent finding across numerous studies using 
various populations and paradigms is that social creativity in the form of 


E 
If) 
ok 
0;::
 
::J 
0 
> 
† 
Q. 
::J 
2 
CJ) 
.E: 
Social identity and aspects o
f
 social creativity 
30 
25 
20 
1 5  
1 0  
5 
0 
_
_
_
_ 
--j Comparison dimension 
Below average 
• Task-oriented 
D Socio-emotional 
Above average 
Outgroup status 
Figure 12.2 Ingroup f
avouritism on dif
f
erent comparison dimensions as a 
f
unction of outgroup status (unpublished data f
rom Taylor & Hinkle) 
1 75 
responding to a negative social identity with ingroup favouritism on new 
comparison dimensions is a reliable and robust phenomenon. Furthermore, 
the experimental investigations of Hinkle et al. and Taylor and Hinkle 
demonstrate that low ingroup status causes greater ingroup favouritism on 
new comparison dimensions than on the original dimensions upon which 
the ingroup does not fare well. 
Status, legitimacy and stability 
Past this broad conclusion, the existing research findings bearing on the 
theoretical issues raised at the outset of this chapter are limited. The issue of 
group status in conjunction with its legitimacy and stability has received the 
most attention and the early SIT theoretical formulations have received 
mixed support. The evidence from studies such as Brown and Ross (1982) 
and Caddick (1982) supports the notion that low-status groups are more apt 
to manifest group favouritism on alternative dimensions of intergroup 
comparison than high-status groups. How legitimacy and stability enter the 
theoretical equation has been far less clear. Caddick's results indicate that 
illegitimacy increases ingroup-favouring comparisons on new dimensions 
for both high- and low-status groups. Brown and Ross found this to be the 
case for low-status groups only. Clearly, other variables must be moderating 
the relationship between illegitimacy, group status and the occurrence of 
ingroup favouritism on new comparison dimensions. Relevant theory and, 
more importantly, research are needed. Caddick's speculation concerning 
the responsibility of a high-status group for the existence of an illegitimate 
intergroup status structure appears to be a potential moderator worthy of 
further empirical attention. 
At present, there appears to be no support for Tajfel's ( I978c) early 
proposal that low-status groups would be particularly likely to use new 


1 76 
Social identity: international perspectives 
dimensions of intergroup comparison when status relations are legitimate 
and stable. Perhaps with a legitimate/stable status structure it is very 
unlikely for low-status group members to perceive any change in the status 
quo relative to superior groups, including the use of new comparison 
dimensions. It seems plausible that another social creativity strategy, a shift 
to different comparison outgroups, may be the more likely response to an 
unfavourable social identity under a legitimate/stable status structure. A 
superordinate issue to any SIT discussion regarding legitimacy and stability 
is the need for more thoroughly elaborated theorizing and research on the 
relationships between legitimacy and stability. 
Importance o
f
 comparison dimensions 
The other theoretical issue which has received some empirical attention has 
to do with the importance or value placed on new intergroup comparison 
dimensions. Recall that both Lemaine et al. (1978) and van Knippenberg 
(1984) hypothesized that a shift to ingroup favouritism on new intergroup 
comparison dimensions would be accompanied by the new bases of com­
parison acquiring greater value or importance than the original comparison 
dimensions. The intuitive strength of this postulate makes it quite difficult to 
abandon. However, there is substantial evidence to the contrary. Groups 
appear to frequently evince ingroup favouritism on new, less valued com­
parison dimensions while acknowledging that they do not measure up to the 
outgroup in other, apparently more important, comparisons. This is an 
intriguing result calling out for more research to further illuminate this 
finding and perhaps identify moderator variables specifying when new 
comparison dimensions will or will not be seen as being of greater import­
ance than the initial comparison dimensions. 
Van Knippenberg's (1984) discussion of socially cooperative intergroup 
relations in contrast to SIT's traditionally competitive intergroup relations 
seems to offer a good deal of theoretical promise with respect to this as it 
seems plausible that the psychological importance of placing greater value 
on new intergroup comparison dimensions may only be relevant to com­
petitively oriented intergroup relations. 
Similarly, Hinkle and Brown's (1990) distinction between autonomous 
and relational processes of social identity construction may also be theor­
etically relevant (see also Brown, Hinkle, Ely, Fox-Cardamone, Maras & 
Taylor, 1992; Hinkle, Brown & Ely, 1992). Groups with autonomous 
identity construction orientations may establish entirely satisfactory social 
identities through seeing themselves favourably on relevant comparison 
dimensions independent of whether other groups rank below or above them. 
As a result, the relative value or importance of various comparison dimen­
sions may not be psychologically relevant with the autonomous orientation. 
Another potentially useful approach to developing more research and 
theory on the importance of different comparison dimensions takes a 
temporal perspective. It seems plausible that the immediate response to an 


Social identity and aspects o
f
 social creativity 
1 77 
unfavourable intergroup comparison is, to use a term introduced earlier, 
to salvage a positive social identity. If our research group fares disas­
trously in the latest government league tables of research quantity and 
quality, the first step is to establish a toehold towards re-establishing a 
favourable social identity. One way to do this is by claiming high status 
for the ingroup on any dimension, even an inferior one. 'Yes, the quality 
and quantity of our group's research is not on par with others, but we are 
doing some limited, innovative work and our teaching is very good. And, 
yes, we recognize that research innovation and teaching quality really are 
not the sine qua non for research groups.' This salvages the group's esteem 
in the short term and provides the foundation for subsequent develop­
ments such as actually improving the group's research productivity or 
convincing relevant entities in the social structure that research innovation 
and teaching quality ought to be the more important bases for intergroup 
comparIson. 
The processes o
f
 social creativity 
There is, to our knowledge, no research on Tajfel's ( l978c) and Lemaine et 
al.'s (1978) proposals that selection of new comparison dimensions proceeds 
first as an intragroup process and, according to Lemaine et aI., surrep­
titiously and, then, as an intergroup process where the ingroup attempts to 
persuade the outgroup regarding the importance of the new comparison 
dimension and the ingroup's superior status. The comments in the previous 
section on temporal aspects of social identity processes again seems 
relevant. It may be that groups with unfavourable identities must first take 
steps to salvage positive distinctiveness, then work within the group to 
develop genuinely favourable intergroup comparisons, and, finally, shift to 
the issue of persuading other groups as to the validity and importance of the 
new bases of comparison. Research paradigms without a longitudinal 
component are unlikely to detect or shed light on such processes. Unfor­
tunately, longitudinally oriented studies are virtually non-existent in the SIT 
tradition. 
Hinkle and Brown's (1990) notion of autonomous orientations may also 
be pertinent. For autonomous groups, since they are interested in the 
ingroup in comparison to abstract standards rather than in comparison to 
other groups, the need to persuade external audiences of the validity of new 
intergroup comparisons is apt to be irrelevant. 
An interesting and closely related issue is raised by van Knippenberg 
(\ 984), who argues that responses to unfavourable social identities are not 
always directed towards immediate enhancement of the group's social iden­
tity. Specifically, van Knippenberg discusses the role of self-presentational 
processes in the expression of ingroup or outgroup favouritism. Self­
presentational issues seem equally relevant to social creativity processes 
including the selection of new bases for making intergroup comparisons. 
Abrams and Hogg (1988; also Hogg & Abrams, 1990) have outlined similar 


1 78 
Social identity: international pers
pectives 
ideas which go well beyond social identity and self-presentation as motiva­
tional issues relevant to the expression of ingroup or outgroup favouritism. 
Strategy choice in response to unf
avourable social identities 
While there are a few studies on mobility versus either social competition or 
social creativity (or both, for example, Ellemers, Wilke & van Knippenberg, 
1993; Jackson, 1993; Wright, Taylor & Moghaddam, 1990), there is no 
research and virtually no relevant theory on selection of the various social 
creativity strategies. These are important theoretical issues. SIT's inability 
to predict the specific form of response to an unfavourable social identity is 
a significant omission and diminution of its utility. While there may be good 
historical reasons for why these issues were not well addressed in the early 
years of SIT's development, the contemporary status of SIT would be 
enhanced by progress on this front. 
Potential applications 
We are unaware of any genuinely applied literature on the use of new 
intergroup comparison dimensions. However, existing studies and theor­
etical discussions are rife with possibility. Perhaps the richest potential for 
application resides in issues of group morale, cohesion and motivation. 
How might poorly performing sports teams, business organizations and 
other similar entities deal with attendance problems of low group morale 
and motivation? The generation of new, favourable dimensions of inter­
group comparison offers one strategy. While the existing research indicates 
that a losing team's development of the view that they enjoy what they are 
doing more than their competitors does not fully psychologically 
compensate for its poor won-lost record, which such a team is still apt 
to see of premier importance, it still may serve as a starting point for 
rebuilding team spirit and motivation. While the theoretical basis for this 
has not yet been tested, the hypothesis that social creativity results in 
enhanced cohesiveness and, subsequently, better group performance is an 
intriguing one. Were notions such as this begun to be examined in applied 
contexts, other interesting issues would certainly come quickly to the fore. 
For example, in all previous research of this area, the creation or ingroup­
favouring use of new comparison dimensions has always arisen spon­
taneously from the studies' inherent intragroup dynamics. In contemplating 
applications, it is quite likely that the instigation to the use of new 
dimensions of intergroup comparison would result from the intervention of 
a group's leadership such as managers, administrators, coaches, and so on, 
or outside sources such as consultants. There is little if any helpful existing 
research or theory. Links to work in organizational psychology and the 
leadership literature would likely be beneficial in determining whether or 
not leadership efforts to stimulate social creativity would be effective or how 
such endeavours could be best implemented. 


Social identity and aspects o
f
 social creativity 
1 79 
Final comments 
In conclusion, research on social creatIVIty in the fonn of selecting new 
comparison dimensions for making ingroup-favouring evaluations has 
played a small, but important, role in SIT research, a role which appears to 
be increasing in its significance and potential applications. The robustness of 
this fonn of social creativity is well documented and some important com­
plexities have begun to be addressed. The existing empirical findings provide 
support for certain SIT precepts, but not others. Our own view on this is 
that in instances where the literature has not supported SIT it has served the 
valuable function of precipitating modifications of the theory that have 
enhanced its power and viability. However, as should be clear fonn earlier 
discussion in this chapter, there are many interesting issues still in need of 
both conceptual and empirical attention. SIT can only be enriched as these 
issues stimulate further work in the study of intergroup processes. 
Notes 
1 .  Note that other authors label or group responses to unfavourable social identities 
somewhat differently (for example, Taylor & Moghaddam, 1987). In preparing the present 
chapter, the authors have endeavoured to follow as closely as possible the conceptual 
organization offered in Tajf
el's early statements of SIT. 
2. Turner and Brown (1978) is a frequently cited paper on legitimacy and stability and their 
relationship to social competition and social creativity. However, in the present context, their 
measures are of limited relevance. The creativity indices concern the number of new possible 
comparison dimensions generated by the SUbjects. However, subjects did not actually make any 
ingroup or outgroup ratings on these new dimensions. Hence, the study's results do not bear 
directly on the use of new bases of comparison to establish favourable social identities. 
3. Note that Brown and Ross themselves do not emphasize this pattern in their results. In 
fact, they do not discuss it at all. However, the described pattern is very clear from examination 
of the data in their Tables 2 (p. 168) and 5 (p. 1 7 1 ). 


1 3  
Mere Categorization and Competition: A 
Closer Look at Social Identity Theory and 
the Discontinuity Effect 
Stephen M. Drigotas, Chester Insko and John Schopfer 
The word 'group' can have many meanings. It can be used to represent 
anything from face-to-face interaction to aggregates of unacquainted 
people. Its imprecision is further enhanced by the fact that one may be 
assigned to many groups (based on sex, race, nationality, family, and so on) 
without any feelings of belonging or particular affinity for any one of them. 
Because of this apparent amorphousness, researchers through the years have 
attempted to delineate when an aggregate of individuals becomes a group. 
Horwitz and Rabbie (1989) credit Lewin (1948) for early theoretical interest 
in perceived collectivity as the defining nature of groupness. Lewin main­
tained that interdependence of fate is the determinant of who belongs to 
which group, not simply degree of similarity-dissimilarity of features. 
Campbell (1958) called the shift from perceived aggregate to perceived 
collectivity 'entitativity' and suggested various Gestalt mechanisms for the 
perception of many as one. At this point we feel that it would be premature 
to settle on what precisely is meant by the term 'group'. 
The behaviour of individuals within groups can be as diverse as the 
considerations of what constitutes a group. Historically some social scien­
tists have focused on the supposedly animalistic nature of crowd behaviour 
as providing possible evidence for a group mind unfettered by moral 
constraints (cf. Allport, 1924; Durkheim, 1 898; LeBon, 1 896; McDougall, 
1920). On the other hand, after extensive observation McPhail (1991) argues 
that the notion of 'madding' crowd is a 'myth'. Still within contemporary 
social psychology there is evidence that under some circumstances 
individuals within groups behave in a manner that is less than praiseworthy. 
For example, individuals within groups are more likely socially to loaf 
(Latane, Williams & Harkins, 1979), give more shocks to a 'learner' 
(Zimbardo, 1969), and not help in an emergency situation (Latane & 
Darley, 1970). 
Our approach to the general problem of group behaviour has been 
specifically to compare interindividual behaviour with intergroup behaviour. 
Consistent with Roger Brown's (1954) suggestion of a discontinuity between 
individual and group behaviour, we have adopted the term 'discontinuity' 


Mere categorization and competition 
181 
for the tendency of intergroup behaviour to be more competitive, or less 
cooperative, than interindividual behaviour. Within the context of certain 
matrix games there is now an abundance of evidence for a descriptively 
large discontinuity effect (for a review see Schopler & Insko, 1992). What is 
less clear is the reason for the discontinuity effect. 
I While different theorists 
have suggested divergent causes for the competitiveness of groups, the 
purpose of the present chapter is to compare and contrast two of the major 
approaches to understanding the phenomenon: social identity theory 
(Tajfel, 1970b, 1978a, 1978b, 1982c; Tajfel & Turner, 1979, 1986; Turner, 
198 1  
b) and the research tradition which has specifically investigated the 
discontinuity effect (Insko & Schopler, 1987; Insko, Pinkley, Harring, et aI., 
1987; Insko, Pinkley, Hoyle, et aI., 1987; Insko, Schopler, Hoyle, Dardis & 
Graetz, 1990; Insko et ai., 1988; Insko et aI., 1992, 1993; McCallum et ai., 
1985; Schopler & Insko, 1992; Schopler, Insko, Graetz, Drigotas & Smith, 
199 1 ;  Schopler et aI., 1993). 
Social identity theory 
An overview 
Tajfel's social identity theory (SIT) is based upon the assumption that we all 
have a need for favourable self-identity, and that the status of the groups to 
which we belong helps us achieve such positive self-identity. Moreover, 
Tajfel and his associates claim that such positive evaluations are made on a 
comparative basis - we nourish our positive self-identities when our group 
does better than another group (Tajfel, 1978b, 198 1 ,  1982c; Tajfel & Turner, 
1979, 1986; Turner, 1981b). Thus, intergroup competitiveness is directly 
linked to the maintenance of self-identity. The SIT theorists also claim that 
mere categorization of subjects into 'groups' results in the orientation 
towards own-category favouritism. An abundance of research (Billig & 
Tajfel, 1973; Tajfel & Billig, 1974; see Brewer, 1979 and Diehl, 1 990, for 
reviews) has demonstrated that when subjects are categorized into one of two 
mutually exclusive groups and respond to a non-correspondent allocation 
task, they typically make choices favouring an own-group member over an 
outgroup member - ostensibly in the pursuit of bolstering self-identity 
through the relative advantage for own group. 
A discontinuity perspective on social identity theory 
A serious problem inherent in social identity theory is its inability 
theoretically to account for the diff
erences between intergroup interaction 
and interindividual interaction. Research involving the discontinuity effect 
has long demonstrated that group interactions can in many circumstances 
be more competitive than individual interactions. However, as noted 
elsewhere (Insko & Schopler, 1 987; Schopler & Insko, 1992), individuals 
should also receive positive bolstering of self-esteem from doing better than 


182 
Social identity: international pers
pectives 
another individual. But if this is true, why from a social identity theory 
perspective should individuals not be just as competitive as groups? It is true 
that social identity theory was developed to address intergroup relations. 
The point is, however, that there is nothing in social identity theory itself 
that indicates why the processes inherent in two-on-two relations should not 
also be present in one-on-one relations. If besting another group should 
bolster one's self-esteem and one's self-definition, the same should also be 
true of besting another individual. Why then, from a social identity per­
spective, are not interindividual relations just as competitive as intergroup 
relations? Furthermore, if it is true that individuals receive a self-definitional 
boost from besting another individual, by implication social identity theory 
does not adequately account for between-group competitiveness. Social 
identity theory's self-definitional premise is plausible, but does not by itself 
account for between-group competitiveness. The theory is not so much 
implausible as incomplete. 
The minimal group paradigm 
Investigations of social identity theory's prediction that mere categorization 
is sufficient to evoke relativistic social comparisons have used a minimal 
group paradigm in which subjects are categorized and then asked to allocate 
outcomes to own- and other-category members. The most common tech­
nique for categorization involves dividing subjects into sets, or groups, on the 
basis of their fabricated preferences for the paintings of Klee versus those of 
Kandinksy. Subjects are asked to rate pairs of Klee-Kandinsky slides and are 
then assigned to different rooms based upon their supposed preferences. 
Once in the separate rooms subjects are asked individually to select one 
column of payoff points from a matrix, and thus assign a payoff amount to 
an own-group member and an other-group member. While Tajfel and his 
associates use a somewhat complicated 'pull' procedure for scoring (see 
Turner, 1978a, for a complete description of the assessment procedure), only 
the four resultant orientations are discussed here. They include (a) maximum 
ingroup profit (MIP), (b) maximum difference in favour of the ingroup 
(MD), (c) fairness (F) and (d) maximum joint profit (MJP). It is the 
maximum difference choice that represents the greatest relative advantage 
and best exemplifies the need for ingroup favouritism. Their research has 
generally concluded that MD is the most powerful predictor of choice, 
although evidence for the other four choices does exist. In fact, Turner, 
Brown and Tajfel (1979) summarized their results as indicating that the MD 
choice is only non-significantly more influential than MIP and MJP. 
Nevertheless, the SIT theorists hold strong to their contention that between­
group competitiveness is the result of the need for positive social identity 
through the relative advantage of ingroup performance. 
As indicated below, there is a question whether or not mere categoriza­
tion is sufficient to produce relativistic social comparison. Beyond that there 
is a definitional matter as to whether aggregates of individuals should be 


Mere categorization and competition 
1 83 
defined as groups to the extent that they engage in relativistic social 
comparisons. One could interpret Tajfel and his associates as believing, first, 
that entitativity is created (on the independent variable side) by categor­
ization, and, second, that entitativity is indexed (on the dependent variable 
side) by relativistic social comparisons. While Tajfel and his associates do 
not use the term 'entitativity', they do believe that mere categorization does 
induce 'genuine awareness of membership in separate and distinct groups' 
(Tajfel, 1 978a, p. 35). 
The alternative allocation task experiments 
A series of experiments using the Multiple Alternative Matrices (MAMs) 
were conducted during the early eighties to assess directly the maximizing 
difference orientation exhibited within the minimal group paradigm. 
Bornstein et al. (1 983a) developed the procedure which enables the meas­
urement of seven different social orientations. The labels for the orientations, 
taken from the analyses of Kelley and Thibaut (1978), are differentiated 
based upon whether the orientations favour own group, other group or 
neither group. The orientations favouring own group include maximizing 
own gain (max. own), maximizing relative own gain (max. reI. own), and 
maximizing joint gain favouring own group (max. joint own). The orien­
tations favouring the other group include maximizing other's gain (max. 
other), maximizing other's relative gain (max. reI. other), and maximizing 
joint gain favouring other (max. joint other). Finally, the orientation 
favouring neither group minimizes the difference between group outcomes 
(min. diff.). 
The obvious advantage of this approach is its delineation of interpersonal 
motives. Each orientation represents a mutually exclusive motivation for 
interaction. Thus, the overlapping orientations problem inherent in the 
minimal group paradigm is easily resolved through use of the MAMs and a 
better test of SIT is possible. The max. reI. own choice best represents SIT's 
maximum difference (MD), max. own is similar to maximum ingroup profit 
(MIP), min. diff. is similar to fairness (F), and max. joint best represents 
maximum joint profit (MJP). 
Point values for own versus other group can be assigned to represent the 
seven orientations without motivational overlap. In other words, the rep­
resentations of the seven orientations are mutually exclusive as to the 
motivation for choosing them. Therefore, the specific predictions of SIT 
concerning categorization and self-esteem gained through maximizing 
relative advantage could be tested in an unconfounded manner through the 
use of the Multiple Alternative Matrices. 
A series of studies using the MAMs was conducted to test the viability of 
SIT in predicting actual motivational behaviour in the minimal group 
paradigm. Because the details of these studies have been presented in 
previous articles and a chapter (Insko & Schopler, 1 987), only the highlights 
are presented here. 


1 84 
Social identity: international pers
pectives 
Bornstein et al. (1983a) conducted a series of three experiments using the 
MAMs that represented a substantial challenge to social identity theory. In 
the first two studies, college students were given a cover story concerning 
their participation in an experiment as part of a larger investigation of 
artistic preferences. The rest of the experiment followed the Tajfelian pro­
cedure exactly except that subjects made monetary choices for one own­
group and outgroup member using the multiple alternative matrices. The 
results indicated that the subjects were not choosing max. reI. own more 
than would be expected by chance (across the seven orientations and 10 
trials), and they actually chose it significantly less than would be expected 
by chance. Recall that max. reI. own is analogous to MD, the main strategy 
believed to be evoked by mere categorization. Instead, subjects significantly 
tended to choose both max. joint own and min. diff. (with significantly 
more min. diff. for females than males). 
In the second experiment subjects were also asked to go back through the 
matrices and indicate a second preference on each matrix, and then a third 
preference. Subjects did not select max. reI. above chance for either the 
second or third preferences. Furthermore, the second experiment also 
included a condition in which subjects responded using the Tajfel matrices 
instead of the MAMs. Results using these matrices indicated support for the 
MD choice, as in previous studies using the less differentiated matrices. 
The third experiment was conducted using junior high students (9th and 
10th grade) from a middle-class US suburb. No cover story was used. 
Results again indicated no significant deviation from chance for max. reI. 
own. As in the first two experiments, min. diff. was selected significantly 
more than chance. However, unlike in the first two experiments, max. own 
was also selected significantly more than chance. 
Given the slightly different results for junior high students, Wittenbraker 
(1983) conducted an experiment using 5th, 9th and 12th grade students 
from the same school system used in the previously mentioned third study. 
He used the same mere categorization procedure except that coloured bar 
graphs were used to represent outcomes for Klee versus Kandinsky 
members. Once again, results indicated a significant effect for min. diff. 
There was also a replication of the max. own effect for 9th grade males. 
Finally, there was a significant effect for max. joint own for both 5th and 
12th graders. 
2 
Social identity theory's response 
In response to the Bornstein et al. series of experiments, Turner (1983a, 
1 983b) raised three objections. The first two were procedural in nature while 
the third constituted a major theoretical challenge. The two procedural 
objections were that the cover story distorted the results and that in the 
MAMs format subjects allocated money to just one own group and one 
outgroup. Two different studies (Bornstein et aI., 1983b) were conducted in 
response to these objections. The first compared the old cover story with a 


Mere categorization and competition 
1 85 
new cover story and a condition with no cover story at all. The results 
indicated no effect for the three-level manipulation of cover story. The 
second experiment contrasted a procedure using allocation to one own 
group and one other group to one using allocation to different own group 
and outgroup members. The results indicated non-significant effects for 
constancy of recipient. Furthermore, both studies again found significant 
effects for both min. diff. and max. joint own and a non-significant effect 
for max. reI. 
The major theoretical objection to the use of the MAMs concerned what 
Turner called the maximizing assumption. Turner maintained that the 
MAMs differentiate between social orientations only to the degree that 
individuals wish to maximize a value reflected by a particular orientation. 
Instead, subjects may make a specific choice not because of its social 
orientation as reflected by the maximizing assumption, but because of a 
different orientation altogether. For example, a subject may choose max. 
own not out of a desire to maximize own profits but instead out of a desire 
to win to a lesser extent than is possible through max. reI. own. Thus, the 
MAMs would reflect social values only to the degree that subjects adhere to 
the maximizing assumption. 
If one is able to cast doubt on whether subjects actually adhere to the 
maximizing assumption, reinterpretation of previous results using the 
MAMs allows for possible support for social identity theory. Recall that 
research with adult subjects found an abundance of evidence for max. joint 
own but no evidence for max. reI. own or max. own. Obviously, max. reI. 
own and max. own both involve advantaging of own group. But what 
about max. joint own? If the choice is interpreted to reflect a concern with 
both own-group and outgroup outcomes, the evidence of desire for 
superiority is ambiguous at best. However, Turner argues that subjects may 
choose max. joint own not out of concern for maximizing both own and 
other outcomes but out of a desire to win to a lesser extent than is possible 
in max. reI. own. Thus, if the maximizing assumption is rejected, both the 
max. reI. own and max. joint own choices support the social comparison 
prediction of social identity theory. 
Insko, Pinkley, Harring, et al. (1987) conducted a study designed to 
determine whether subjects actually invoked the maximizing assumption 
when choosing either max. own or max. joint own. An identical format was 
used to test the subjects except that the MAMs were presented one at a 
time. If and when a max. own or max. joint own was selected the subject 
was given a questionnaire containing an open-ended item and a closed­
ended item. The open-ended item simply asked subjects to state their reason 
for their immediately prior choice. The closed-ended item forced subjects to 
choose between the maximizing assumption and the Turner rationale. For 
max. own these items were 'to gain the most amount of money possible' 
and 'to win by less than the biggest margin'. For max. joint own these items 
were 'to gain the highest combined amount of money for everyone, with my 
group receiving slightly more', and 'to win by less than the biggest margin'. 


1 86 
Social identity: international perspectives 
The results for both the open-ended and closed-ended items indicated 
overwhelming support for the maximizing assumption. For example, 1 8  of 
the 21 subjects who chose max. joint own had open-ended responses 
consistent with the maximizing assumption (p < .0006 by binomial test). 
The MAMs versus the Ta
j
f
elian pull procedure 
If subjects do accept the maximizing assumption of the MAMs, why do the 
MAMs yield different results from the pull procedure? Currently, we do not 
have a definitive answer to this question. We can, however, point to two 
obvious differences in the assessment procedures. First, because the array of 
alternatives on the MAMs are unconfounded, it should be relatively clear to 
the subject that he or she is selecting among distinctive choices. 
Second, unlike the pull procedure, the MAMs simultaneously present 
subjects with the full array of possible preferences to be assessed. Thus a 
subject is able to make a choice with complete knowledge of all of the 
possibilities that are available. We find it plausible that in such a situation 
the subject's ability to indicate his or her clear preference will be optimized, 
that is, that the choice will be an informed choice. 
Consider, for example, a subject who selects max. reI. own on the 
MAMs. It should be apparent to the subject that the choice is clearly 
different from max. own, max. joint own, max. other, and so on. The 
subject should clearly understand that a max. reI. own choice is not a choice 
to maximize own category outcomes, not a choice to maximize joint 
outcomes, not a choice to maximize other category outcomes, but rather a 
choice to maximize the relative difference between own- and other-category 
outcomes. If it is advantageous to use an assessment procedure which 
maximizes the subject's understanding of what is being chosen, why should 
investigators continue to rely on the pull procedure? 
SIT summary 
Taken together, this series of experiments using the MAMs provides a very 
dissatisfying picture of the contention that mere categorization is enough to 
instil relativistic competition through social comparison mechanisms. First, 
five different studies using a variety of populations, including adults, found 
that mere categorization does not lead to maximizing relative choices when 
differentiated social orientation matrices are used instead of the traditional, 
motivationally confounded, Tajfelian matrices. Instead, subjects tend to 
choose max. joint own. Furthermore, research has shown that subjects do 
appear to use the maximizing assumption inherent in the interpretation of 
the MAMs. In other words, with closer inspection of social motives, mere 
categorization does not produce the relative advantage competition that 
Tajfel believed to characterize real groups. Given this revelation, the 
question then becomes: what exactly is needed beyond mere categorization 
to produce intergroup competition? 


Mere categorization and competition 
The discontinuity effect 
The possibility o
f
 schema-related belie
f
s
 concerning the 
appropriateness o
f
 competitive behaviour 
1 87 
As mentioned earlier, the discontinuity effect refers to the tendency for 
intergroup interactions to be more competitive than interindividual inter­
actions. Within the domain of investigated matrices, the effect appears to be 
robust and reliable. Moreover, research concerning the discontinuity effect 
has helped identify what is needed beyond mere categorization to produce 
intergroup competition. We specifically hypothesize that all of the evidence 
triangulates on the suggestion that competition flows from circumstances 
implying the believed appropriateness of competition. We, furthermore, 
speculate that a set of such beliefs in the appropriateness of competition 
constitute a schema. By a schema we mean a set of beliefs serving to organize 
and guide memory for past events and expectations regarding future events. 
In postulating such a schema we are clearly going beyond the evidence. More 
definitive evidence for a schema would be provided by data indicating that 
perceived effects are magnified or altered in memory. We have no such data; 
rather, we have data which suggest that experienced competitiveness can 
produce further competitiveness - even when the further competitiveness 
involves a different opponent. 
Competitive experience with one opponent and subsequent 
competitive relations with another opponent 
Insko, Pinkley, Harring, et ai. (1987) conducted an experiment comparing 
intergroup interactions versus interindividual interactions in a Prisoner's 
Dilemma Game (PDG; see Figure 13.1) followed by the usual mere cat­
egorization procedure. Previous research using the PDG had indicated that 
groups behaved competitively while individuals behaved cooperatively in 
the identical setting (McCallum et aI., 1985). Thus, the behaviour of 
subjects who experienced competition (in group-on-group interaction) or 
did not experience competition (in one-on-one interaction) on a previous 
task could be compared in the mere categorization task. It was hypothesized 
that the prior experience would carry over to the categorization task, 
resulting in increased levels of competitiveness (max. reI. own or max. own). 
The results indicated that there was more competition between groups than 
between individuals with the PDG and that this behaviour carried over to 
the mere categorization procedure in the form of increased max. own (but 
not max. reI.) on the MAMs. 
Reciprocation of competitive responses, or escalation of conflict with the 
same opponent, has been documented by others (for example, Kelley & 
Stahelski, 1970). It is important to realize, however, that the Insko, Pinkley, 
Harring, et ai. result occurred despite the fact that the subjects were led to 
believe that the subject composition of the Klee and Kandinsky categories 
in the mere categorization situation was not necessarily the same as the 


188 
Social identity: international pers
pectives 
Cooperation 
Competition 
Cooperation 
Competition 
$0.57 
0.660.39 
$0.66 
Figure 13.1 
A Prisoner's Dilemma Game matrix (PDG) 
subject composition of the groups in the PDG situation, and thus that the 
obtained results were not a simple result of reciprocation. Why, then, did 
competitive behaviour carry over from one situation to the other? One 
possibility is that there was a perseveration of competitive behaviour. We 
believe, however, that such an explanation is at best an oversimplification. 
Rather we interpret the results as suggesting that subjects shifted their 
beliefs regarding what was appropriate or normative. As a result of prior 
competitive group-on-group interaction, subjects believed that when sub­
sequently categorized in a different group it was appropriate to behave 
competitively towards another category or group. 
To some readers it may appear that in postulating that competitive 
experience results in a change in beliefs regarding the appropriateness of 
competition we are going beyond the evidence, and indeed we are. Why do 
we need to postulate anything other than simple perseveration of com­
petitive behaviour? There are two answers to this question. First, the reader 
should note that there is a wealth of published research indicating that 
behaviour change can result in attitude and belief change (for a review of 
this literature see Chapter 1 1  in Eagly & Chaiken, 1993). In addition the 
evidence for this linkage is particularly strong in those situations in which 
the choice to engage in the behaviour is 'free' or apparently unconstrained, 
and that appears to be the case in the above-described research. Second, the 
readers should note that competitive experience also involves competitive­
ness in the opponent, and a series of studies by Insko et aI. (1 992) relating 
to max. reI. indicates that it is an opponent's competitiveness which is 
particularly important in producing subsequent competitive behaviour. 
The design of the Insko et aI. (1992) studies used a three-choice matrix 
(labelled a PDR matrix; see Figure 1 3.2) which reflected three different 
social orientations: max. joint, max. own and max. reI. The max. own 
choice on this matrix is suggested by Campbell's (1965) realistic conflict 
theory, which implies that the dominant social motive in intergroup conflict 
is maximizing 'real' outcomes. The max. reI. choice obviously reflects the 


Max. joint 
Max. own 
Max. reI. 
Mere categorization and competition 
Max. joint 
Max. own 
$0.99 
1441 . 1 4  
Confounded 
max. rei. 
1.441 . 1 4  
0.840.84 
144Max.rei.0.84 
0.690.99 
1.14Confoundedmax.reI.1 44 
0.841 . 1 4  
0.990.99 
$0.99 
1 89 
Note: The third choice in the PDR matrix is max. rei. The third choice in the PDC 
matrix is confounded max. reI., which is max. rei. that is confounded with max. own. 
Figure 13.2 A Prisoner's Dilemma max. reI. matrix (PDR) and a 
Prisoner's Dilemma con
f
ounded matrix (PDC) 
Tajfelian notion of maximizing relative advantage to bolster social identity. 
The results of the first study supported both Campbell and Tajfel's theory 
concerning group behaviour: relative to individuals, groups made both 
more max. own and max. reI. choices. There was some hint, however, that 
the max. reI. behaviour only occurred after the completion of the practice 
trials, which were played for points instead of money against the other 
group (or individual). 
One possible reason for this initial reluctance to choose max. reI. might 
relate to tactical considerations. Some groups may try tactically to 'set up' 
the other group by not choosing max. reI. until the later trials. Therefore, in 
experiment 2 we eliminated this confound by having subjects play their 
practice trials against an experimenter, and invoked a procedure whereby 
each group (or individual) played one trial with each of three other groups 
(or individuals). This procedure was designed to eliminate possible tactical 
considerations because subjects knew they would only interact once with 
each other group (or individual). The procedure had the added benefit of 


190 
Social identity: international pers
pectives 
Table 13.1 
Max. rei. as a f
unction o
f
 matrix type (PDR versus PDC), 
trials and groups versus individuals, experiment 2 
Trial I 
Trial 2 
Trial 3 
Groups 
Individuals 
Groups 
Individuals 
Groups 
Individuals 
PDR 
PDC 
0.08 
0. 12 
0.08 
0.63 
0.17 
1 .38 
0.18 
0.28 
0.50 
1 .06 
0.00 
0.27 
Note: Because four groups (or individuals) interacted on any given trial, max. reI. scores have a 
potential range from 0 to 4. 
eliminating a possible interpretation of the results as being due to recipro­
cation of competition, or the escalation of conflict with an opponent. 
It is likely that the max. reI. that occurred after the practice trials reflects 
the fact that a number of interactions are needed to instil the belief that 
competition is appropriate within the context of the PDR matrix. A further 
complication is that subjects may have been initially unwilling to choose 
max. reI. out of a reluctance to reduce own outcomes. The max. own choice 
produced higher own outcomes than the max. reI. choice (see the PDR 
matrix in Figure 1 3.2). Experiment 2 addressed this problem by using, in 
addition to the PDR matrix, a new matrix, the PDC matrix. On the PDC 
matrix own outcomes are identical (or confounded) for the max. own and 
max. reI. choices (see Figure 13.2). Thus, for this matrix the selection of 
max. reI. imposed no penalty on own outcomes. 
The results, once again, supported both Campbell and Tajfel: groups 
selected max. own and max. reI. more often than individuals. However, 
there were significant max. reI. sequence effects for both the PDR and PDC 
matrices (see Table 1 3.1). On the PDC matrix, the tendency for groups to 
choose max. reI. more than individuals was non-existent on the first trial, 
increased on the second trial, and then levelled off on the third trial. On the 
PDR matrix, the max. reI. effect was non-existent on the first and second 
trials, but was there on the third trial. There were no sequence effects for 
max. own on either matrix. There was a cooperation effect inverse of that 
for max. reI. for both matrices, but this effect can be regarded as a 
consequence of the shifts in max. reI. 
The findings for both studies are consistent. Max. reI. selections did not 
commence until some experience with the matrices had occurred. We should 
note here that, at least in study 2, this could not have been the result of 
tactical considerations that arise when repeatedly facing another group. 
These effects occurred even when subjects knew they would face each other 
group only once. 
It is also important to recognize that the increased competitiveness cannot 
be the result of simple reciprocation of conflict. Again, subjects knew they 
would face each other group only once. Therefore, there was no chance for 
retaliation against a group that competed. Reciprocation and escalation of 
conflict seem unlikely explanations of the competition sequence effects. 


Mere categorization and competition 
191 
A further analysis of the tria1-by-trial patterns in the data indicated that 
the tendency to chose max. reI. (confounded with max. own) on the PDC 
matrix on the second trial occurred most obviously for those groups whose 
opponents on the first trial had chosen max. own. Having been hit with a 
competitive choice from the first opponent, groups, but not individuals, hit 
their second opponent with a more competitive (but not own outcome 
reducing) max. reI. choice. On the other hand, the tendency to chose pure 
max. reI. on the PDR matrix on the third trial occurred for those groups 
who had engaged in mutual max. own responses with their second trial 
opponents. Having exchanged mutual max. own choices with their second 
opponent, groups, not individuals, hit their third opponent with pure max. 
reI. (max. reI. that did reduce own outcomes and was thus unconfounded 
with max. own). Such findings are consistent with the general supposition 
that prior competitive experience can activate a schema regarding the 
believed appropriateness of a highly competitive, max. reI. response. The 
evidence suggests that after groups engage in mutual max. own responses 
they 'reach for a bigger gun' (pure max, reI.) to use against their next 
opponent. At that point groups are willing to reduce their own outcomes 
just to increase the likelihood of doing relatively better than their new 
opponent. 
Recall the above-described Insko, Pinkley, Harring, et aI. (1987) experi­
ment in which participants interacted first in the context of the PDG and 
second in the mere categorization situation. In that experiment there was 
the possibility that the max. own choices in the mere categorization 
situation were a simple result of response perseveration. Note, however, that 
in the present experiments (Insko et aI., 1992) the response-perseveration 
possibility can be ruled out for the simple reason that the max. reI. choice 
did not follow prior own max. reI. choice. For the PDC matrix the max. reI. 
response occurred following the max. own choice by the opponent., and for 
the PDR matrix the max. reI. response occurred following mutual max. own 
responses. The key thus appears to be the experience of competitiveness by 
the opponent. 
The discontinuity research paradigm 
Studies investigating the discontinuity effect are typically laboratory experi­
ments that observe interaction using a Prisoner's Dilemma outcome matrix 
(or some variation thereof). The research is usually conducted in a 
laboratory suite in which same-sex individuals or groups are located in 
different rooms connected to a central room. After extensive training 
concerning the interdependent nature of the PDG, a trial or series of trials is 
conducted. Subjects are given time to examine a version of the PDG, then 
have some time to discuss possible action with the other group (often face to 
face with representatives but sometimes over a communication system), and 
then return to own-group discussion and actual decision-making. In group 
conditions the decision is typically made by the group as a whole or based 


1 92 
Social identity: international perspectives 
upon a majority vote. The decisions are then collected and the subjects paid 
accordingly. The discontinuity effect has been shown across a large number 
of studies regardless of: number of trials; size of payoff; gender; intergroup 
contact using representatives or all group members; whether the low values 
in the matrices are positive or negative; the type of matrices used (PDG, 
PDG-Alt., Mutual Fate Control); and whether there is equal or non-equal 
power. 
Groups are not always more competitive than individuals 
A further issue that merits comment relates to whether groups are 'always' 
more competitive than individuals. Clearly in many instances groups are 
cooperative and individuals are competitive and it would be foolish to 
maintain otherwise. Even though the above point has been repeatedly made 
(for example, Insko et aI., 1993; Schopler & Insko, 1992), it is sometimes 
asserted that we believe otherwise. For example, Rabbie and Lodewijkx 
(1994) claim that 'Schopler and Insko (1992) have concluded, based on their 
research with the PDG, that groups are invariably more competitive and 
less cooperative than individuals' (p. 1 58). Actually, Schopler and Insko 
(1992) and Insko et al. (1993) argued that Kelley and Thibaut's (1978) index 
of correspondence provides a basis for predicting when groups will be more 
competitive than individuals. Kelley and Thibaut's index of correspondence, 
in a situation in which the range of outcomes for the two players are 
the same, is the correlation between the outcomes across the cells. With the 
PDG and with other matrices that we have investigated, this correlation 
is moderately negative. Clearly, one would not expect discontinuity with 
matrices involving totally correspondent outcomes (+ 1 correlation). With 
such matrices both groups and individuals should coordinate their responses 
and there thus would be no discontinuity effect. Likewise in a zero-sum 
situation with totally non-correspondent outcomes (-1 correlation) both 
individuals and groups have no alternative to competing, and there would 
again be no basis for a discontinuity effect. Given this theoretical argument, 
the empirical challenge is to determine whether the circumstances in which 
groups are more competitive than individuals are, indeed, situations with 
somewhat, but not totally, non-correspondent outcomes. This is an issue 
that is currently being investigated. 
A further consideration that is relevant to the issue as to whether groups 
are always more competitive than individuals is the fact that our investi­
gations of the discontinuity effect have almost always been in situations in 
which individuals and groups communicate with each other prior to making 
simultaneous and irrevocable choices. Typically the communication has 
been face-to-face communication, but sometimes the communication has 
been via an intercom. Furthermore, Insko et al. (1993) obtained evidence 
indicating that in one-trial situations the absence of communication via an 
intercom produced a larger decrease in cooperation between individuals 
than between groups. Further evidence indicated that the presence of 


Mere categorization and competition 
193 
communication produced a greater decrease in expected competItIOn 
between individuals than between groups. Such evidence is consistent with 
the possibility, as discussed below, that there is an outgroup schema leading 
groups not to trust each other and thus for communication between groups 
to be of relatively low credibility. All of which is to suggest that without 
some form of communication the discontinuity effect should be markedly 
smaller and may, under some circumstances, be non-existent. Typically, of 
course, relations between individuals and relations between groups do occur 
in the context of some form of communication. 
The theoretical basis f
or discontinuity: ingroup (greed) and outgroup 
(f
ear) schemata 
As mentioned earlier, research investigating the discontinuity effect allows 
for careful consideration of the roots and nature of competition between 
groups. Previously mentioned research has provided evidence for the link 
between the belief in the appropriateness of competition (group schema) 
and competitive behaviour. Actually there is reason to believe that the 
schema concerning the believed appropriateness of competition can be 
subdivided into two schemata, the ingroup, or greed, schema and the 
outgroup, or fear, schema. At this point we would like to ruminate more 
extensively on these two theoretical bases for the discontinuity effect, 
The ingroup (greed) schema 
If a set of beliefs regarding the appropriate­
ness of competition exist, it is reasonable to suppose that some of those 
beliefs relate to what is appropriate and expected within the ingroup. But 
what might those beliefs be? One possibility relates to the beliefs that are 
postulated by the altruistic rationalization hypothesis. According to this 
hypothesis, group members assume that they are able to rationalize their 
competitiveness in intergroup interactions by believing that such behaviour 
is for the benefit of in 
group members. Thus, such competition is altruistic in 
regards to other members of the ingroup. The hypothesis bears some 
resemblance to Snyder's concept of attributional ambiguity (Snyder, Kleck, 
Strenta & Mentzer, 1979; Snyder & Wicklund, 198 1). Snyder maintains that 
self-interested behaviour is more likely to occur when there is attributional 
ambiguity as to whether the behaviour is self-interested in nature. Unfor­
tunately, there is evidence that when subjects are explicitly given the 
opportunity to share outcomes, and thus make altruistic rationalizations, 
they do not become more competitive. Insko, Pinkley, Hoyle, et al. (1987) 
conducted two studies employing a condition in which individuals were 
bound together by outcome interdependence. Within this condition, there 
were three simultaneous dyad interactions, with all subjects having a separ­
ate room, three on each side of the centre hall. However, subjects were told 
that, while they would interact alone with an individual on the other side 
of the hall, their money would be divided equally with the other two people 
on the same side of the hall. According to the altruistic rationalization 


1 94 
Social identity: international pers
pectives 
hypothesis, this outcome interdependence condition should have created 
more competition than the individual interaction condition. However, the 
means for the two conditions were virtually identical, thus making the 
altruistic rationalization hypothesis implausible. 
A more compelling argument for the creation of a competition-allowing 
ingroup schema revolves around the social support f
or shared self-interest 
that is available for groups but not for individuals. A belief in the avail­
ability of within-group support is quite likely based on a wealth of within­
group experience. We thus postulate that a history of mutual support for 
short-term self-interest has produced an ingroup schema that supports 
greedy behaviour. Individuals, by definition, are not afforded this support 
and therefore, apart from fantasised social support, are less likely to develop 
this schema. In characterizing social support for shared self-interest as a 
schema we are explicitly postulating that individuals in groups expect that 
they will be supported when they suggest a course of action consistent with 
the group's self-interest. Beyond this, however, we also postulate that greed­
based competitiveness is most likely in a situation in which the other group 
is expected to be cooperative and is therefore vulnerable. (Combativeness 
that occurs in a context in which the other group is expected to be com­
petitive and is therefore dangerous is more reasonably attributed to fear 
than to greed.) Schopler et al.'s (1993) study 2 was designed to demonstrate 
the resultant competitive behaviour when there is social support for self­
interest, and an expectation that the other group will be cooperative. 
Confederates were placed in each group and told to give a single suggestion 
as to what the group should choose in a standard PDG. The confederates 
were told to suggest to the other two subjects either that they should choose 
X (the cooperative choice) or Z (the competitive choice). The use of 
confederates also allowed for manipulation of the feedback concerning what 
the other group actually chose. By selecting the confederates as represen­
tatives it was possible to give the subjects false feedback that the other 
group was either choosing all X, all Z, or a 50-50 mixture of the two. The 
feedback manipulation was conceived as a manipulation of the other 
group's expected behaviour. The prediction was of an interaction between 
the feedback and suggestion factors such that only with cooperative 
feedback would the cooperative suggestion succeed in reducing intergroup 
competitiveness. Within the context of the repeated cooperative feedback 
where the other group is perceived as vulnerable, the cooperative suggestion 
theoretically should have reduced the within-group social support for 
combativeness, and thus reduced competitiveness in this condition relative 
to all the other conditions. The predicted pattern of results occurred. 
The outgroup (f
ear) schema While social support may foster greed through 
the bolstering of a competition-allowing ingroup schema, belief that the 
other group will be competitive is also expected to create competition within 
the PDG paradigm. Competition as the result of a competitive outgroup 
schema has been termed the 'fear' hypothesis. 


Mere categorization and competition 
Cooperation 
Withdrawal 
Competition 
Cooperation 
Withdrawal 
Competition 
Figure 1 3.3 
A Prisoner's Dilemma Game alternative matrix (PDG-Alt. ) 
195 
There is some evidence that individuals do maintain beliefs describing 
other groups as competitive, aggressive and dishonest in nature. Doise 
(1969) reported that before any interaction subjects attribute fewer coopera­
tive motives to the opposing group than to themselves or their own group 
members. Hoyle, Pinkley and Insko (1989) found that the source of this 
negative expectancy was not whether one was alone or in a group, but 
whether the opponent was a group or an individual. The expectations for 
the outgroup were significantly more competitive and aggressive regardless 
of whether one was alone or in a group. We do not have definitive evidence 
that such expectations, or beliefs, constitute a schema, but we are explicitly 
postulating that they do. 
There is evidence relating the postulated outgroup schema to the discon­
tinuity effect itself. This evidence comes from research using a new three­
choice matrix referred to as a 
Dilemma Game (PDG-Alt.) matrix (see 
Figure 1 3.3). The PDG-Alt. matrix includes a third choice in addition to the 
cooperative and competitive choices of the PDG matrix. This new choice 
guarantees intermediate outcomes regardless of the opponent's choice. This 
new choice, labelled 'withdrawal', is thus a safe choice that should be 
rationally preferred when there is fear of the opponent's competitive intent. 
Insko et al.'s (1990) second study used the PDG-Alt. matrix. The results 
indicated a discontinuity effect for both competition and withdrawal. Not 
only did groups compete and withdraw more than individuals, they com­
peted and withdrew to approximately the same extent. Thus, the discon­
tinuity effect cannot be explained by greed alone. The negativity of the 
outgroup schema (that is, the expectation that another group IS more 
dangerous than another individual) also affects group behaviour. 


196 
Social identity: international perspectives 
What is not clear, however, is the basis for this negative outgroup schema. 
One could suggest a number of hypotheses regarding its development. The 
first could simply be experience. We may develop general negative outgroup 
schemata based upon negative experiences with specific outgroups. Such 
experiences may cloud interactions with any group, just as the competitive 
PDG experience made competition more likely in the mere categorization 
paradigm, and just as max. reI. did not emerge until after a few trials using 
the PDR and PDC matrices. Consistent with this speculation, Pemberton, 
Insko and Schopler (1997) obtained evidence indicating that intergroup 
relations are recalled as more competitive than interindividual relations, and 
also that intergroup relations are experienced as more competitive than 
interindividual relations. The experiential data were obtained from 'diaries' 
that subjects carried with them for seven days (Reis & Wheeler, 1991). 
A second hypothesis concerns the role of false consensus. This negative 
view of outgroups in general may stem from our knowledge of our own 
competitive tendencies within groups. Thus, we may only need to look at 
ourselves and by projection develop a prediction of how the other group 
may behave. Further research should expand our understanding of these 
matters. 
Regardless of how the negative outgroup schema develops, its effects on 
behaviour are predictable. When the only choice is between cooperation 
and competition, as in the standard PDG, groups lash out in fear. It may be 
that the root of many group behaviours that are intensely competitive in 
nature are simply the result of fear of the other group and not a reflection of 
ingroup greed. This is certainly apparent when there is the chance to 
withdraw from the situation with certainty of outcomes, as in the PDG-Alt. 
In such situations groups tend to withdraw as often as they attack. 
Reducing discontinuity through recategorization: preliminary results 
There is an 'older' literature and a 'newer' literature on the overcoming of 
ingroup ethnocentricism. The 'older' literature tended to specify remedial 
interventions in terms of interactive behaviour between two groups. There 
were such global recommendations as joint interactions to attain super­
ordinate goals (Sherif, Harvey, White, Hood & Sherif, 1961), and there 
were the specific behavioural rules contained in Osgood's (1962) Graduated 
and Reciprocated Initiatives in Tension-reduction (GRIT). The 'newer' 
literature focuses on the cognitive mechanisms presumed to be antecedents 
to intergroup bias, and seems to vary with respect to the directness with 
which the intervention affects relevant cognitions. The most direct methods 
are those that degrade the categorized representation by such methods as 
recategorization (Gaertner, Dovidio, Anastasio, Bachman & Rust, 1993; 
Gaertner, Mann, Murrell & Dovidio, 1989) or criss-crossing category 
membership (Brewer, Ho, Lee & Miller, 1987). With the exception of 
modelling GRIT (Lindskold, 1983), both traditions tested for a reduction in 
ethnocentrism with measures of individual perceptions of ingroup and 


Mere categorization and competition 
197 
outgroup members. To our knowledge no one has attempted to assess 
whether the intervention was successful in increasing such behaviours as 
intergroup cooperation. We believe the ultimate effectiveness of any tech­
nique resides in its ability to create intergroup cooperation between intact 
groups. 
We have pilot tested some recategorization procedures that, in terms of 
reduced intergroup conflict, were spectacular failures. The pilot studies 
began with the formation of four three-person groups, who were sub­
sequently combined into two rivalrous six-person groups or, in a different 
version, two, independent, six-person groups each working cooperatively on 
a task. The recategorization worked well, in the sense of increasing liking 
for former outgroup members and creating satisfactory working relations 
with them. But, when we regrouped participants into their initial three­
person groups and asked them to interact with previous group members on 
a PDG, their rate of competition was just as large as the rate for groups 
whose members were never recategorized. All of which makes us suspect 
that if at the end of camp Sherif and his associates (Sherif et aI., 1961) had 
put the Eagles and Rattlers into a mixed-motive situation, they would have 
been just as competitive as they were in the tournament. Quite possibly 
Sherif and others have underestimated the power of the social setting. 
Conclusions 
That individuals gain self-definition and bolster self-esteem by belonging to 
groups is certainly a reasonable working hypothesis. Furthermore, we find it 
plausible that the need to support high self-esteem through the demon­
strated superiority of one's own group may be at least a partial explanation 
for between-group competitiveness. This main tenet of social identity theory 
has not been questioned. However, the belief that mere categorization into 
groups produces such intergroup competitiveness, a second tenet of the 
theory, has been questioned. An abundance of research has demonstrated 
that when more differentiated social orientations are used, mere 
categorization does not produce the relative competitiveness predicted by 
social identity theory. 
Instead, research detailing the discontinuity effect has suggested that what 
is need to produce the intergroup competitiveness beyond mere categoriza­
tion is a belief that competition is appropriate. There is evidence that 
competitive behaviour increases following competitive experience. This is 
true even when the opponent is a group other than the one with whom the 
prior competitive experience occurred. Such evidence is consistent with the 
possibility that beliefs in the appropriateness of competition can produce 
competition between categories. We have speculated that such beliefs flow 
both from an ingroup schema relating to expected social support for self­
interested, or greedy, behaviour, and from an outgroup schema relating to 
the expected competitiveness, and dangerousness, of an outgroup. 


198 
Social identity: international pers
pectives 
Notes 
Work on this chapter was supported by a National Institute of Mental Health Grant 
(5ROI MH53258) to the last two authors. 
I .  Given the existence of difference in the interaction of two groups and the interaction of 
two individuals, we are inclined to define aggregates of people as groups whenever behaviour 
between these aggregates is different from behaviour between individuals. However, in order to 
avoid circularity it is necessary to use the term 'group' in two different senses. Groups in the 
sense of aggregates, or sets of individuals, is assumed in the definition of discontinuity. The 
existence of discontinuity, however, provides evidence for the existence of groups in the more 
psychological meaning of the term. 
2. In seeming conflict with the above results, Platow, McClintock and Liebrand (1990) 
found that categorized subjects would select max. reI. on the Bornstein et al. (1983a, 1983b) 
matrices. Two differences between this experiment and all, or much, of the above research are 
that Platow et al. used points and not money for outcomes, and did not have a cover story. The 
combination of points and the lack of a non-game cover story may have led some subjects to 
interpret the situation as game-like, and thus to attribute relative value to the points. We 
believe that researchers should be careful not to create a game-like atmosphere and should use 
outcomes that have value outside of the experimental context. 


PART IV 
THE IMPACT OF 
CULTURE ON IDENTITY 
AND CATEGORIZATION 
1 4  
Discrimination and Beliefs on 
Discrimination in Individualists and 
Collectivists 
J. Francisco Morales, Mercedes L6pez-Saez and Laura Vega 
The concept of individualism 
Individualism refers not to the individual as an empirical subject or 
'individual specimen of the human species', but rather to the individual as a 
'moral, independent, autonomous being, . . .  
bearer of our supreme values'. 
In this second sense, individualism is the prevailing contemporaneous 
ideology of person and society (see Dumont, 1987, pp. 37ff.). 
As opposed to individualism, which ascribes to the individual the main 
value in society, for holism the supreme value lies in society as a whole. 
Historically, individualism appears in holistic societies as a source of 
opposition to them and as a sort of supplement. Here, the Indian 'renouncer' 
is an important landmark. In order to achieve the utmost truth, the 
renouncer relinquishes social life and its limitations and devotes himself to 
his own destiny. Similarly, in the Hellenistic period, against the self-sufficient 
polis of 
Plato and Aristotle, the idea was postulated that in order to become a 
wise person it was necessary to renounce the world. At that time, non­
learned people were considered as victims of the worldly life. 
From the beginning of the modem era, society is usually conceived as an 
'association', which suggests that the individuals who belong to it have 
associated with each other 'voluntarily', and thus they are prior to the groups 
and/or the relations they establish among themselves. This conception entails 


200 
Social identity: international perspectives 
two characteristic features: equality and freedom. There is a belief that one's 
own selfish interest is the best guarantee of social order and of general 
satisfaction for all members of society. 
Nevertheless, quite often, the way many people have used freedom has 
prevented others from achieving their own. Here is the central 'paradox' of 
individualism, according to Camps. Ethics must necessarily be individua­
listic, because it has to preserve the individual. But it must be achieved in 
such a way that it protects everyone at the same time, not some at the 
expense of others. When this is not achieved, individualism is equivalent to 
'anti-ideology', insolidarity, insensitivity to unequalities, lack of interest in 
public affairs. And that is what happens to most developed societies, 'whose 
only expectation is the perpetuation of group' (see Camps, 1 993, pp. 14f.). 
Cultural aspects of individualism 
Hofstede (1980) reports the results of the HERMES project, which involved 
the study of employees from 40 different countries, all of them belonging to 
the same multinational business corporation. From the responses given by 
this sample of subjects to 14 work goals, this author identified a factor 
representing a dimension of individualism-collectivism. 
The individualism-collectivism factor loaded positively on the following 
goals, from highest to lowest loading: (a) enough free time for oneself and 
one's family; (b) freedom to follow one's own way at work; (c) having a job 
which provides a personal sense of performance; and (d) living in a place 
desirable to oneself and one's family. The goals with negative loadings, also 
in a decreasing order, were: (a) opportunities to learn new skills; (b) good 
physical working conditions for work; (c) use of skills and abilities at work; 
and (d) good marginal profits. According to Hofstede, goals with positive 
loadings have in common that they emphasize personal independence vis-a­
vis the organization. On the other hand, goals with negative loadings refer 
to things the organization can do for the individual. 
The factor analysis conducted by Hofstede has been called 'ecological', 
because the loading of each goal is the correlation coefficient between each 
country's factor score and its mean score in that goal. This correlation is 
calculated across all 40 countries. From factor scores an Individualism 
Index is calculated for each country, ranging from 0 to 100 due to a 
mathematical transformation. 
Psychological aspects of individualism 
Hui (1988) postulates that individualism is also an individual tendency or 
predisposition, since there is, within any culture, subcultural variation 
among individuals regarding this dimension. This author's work has two 
main characteristics. First, the development of a scale to measure the 


Discrimination and belie
f
s
 on discrimination 
201 
individualism-collectivism construct, named INDCOL, consisting of eight 
subscales. Thus, the construct is considered as multifaceted and is measured 
in relation to eight specific targets: spouse or partner, relatives, parents, 
neighbours, close friends, co-workers, acquaintances and siblings. 
Second, INDCOL measures beliefs, attitudes, behavioural intentions and 
behaviours. As Hui himself notes, this has both theoretical and methodo­
logical consequences. From a methodological standpoint, generality of 
measurement increases, although it is at the expense of a part of the scale's 
internal consistency. From a theoretical point of view, individualism­
collectivism is not considered as a unitary disposition, but rather as a 
syndrome. In fact, by adding up the scores of the eight subscales, the 
General Index of Collectivism is obtained. Hui defines this index as the 
degree to which the person believes the group, and not the individual, is the 
basic unit of survival. 
Based on INDCOL and other scales developed by himself, Triandis, 
Bontempo, Villareal, Asai and Lucca (1988) produced the Three Factor 
Questionnaire. They applied this questionnaire to samples from the United 
States and found, as did Hui, that individualism is a multifaceted construct, 
since it appears characterized by the following features: (a) priority given to 
one's own goals over those of the ingroup; (b) less attention paid to the 
views of the ingroup; (c) self-reliance with competition; (d) distance from 
ingroup; (e) making decisions on one's own before asking others' point of 
view; (f) less concern for ingroup. Therefore, in the event of a conflict 
between ingroup and individual, individualism will lead the person to do 
what he or she, and not the ingroup, considers appropriate. 
Correlates of individualism: self, relationship with in 
group and 
outgroup, emotional experience 
Besides the above-mentioned work by Hui (1988) and Triandis et al. (1988), 
many others have studied the differences in social behaviour between 
individualists and collectivists. Among the best known the following can be 
cited: Hui and Triandis (1986), Markus and Kitayama (J991a), Triandis 
(1 993), Triandis, Leung, Villareal and Clack (1985), Triandis et al. (1 986) 
and Triandis, McCusker and Hui (1990). Table 14.1 summarizes the main 
differences regarding self and relationship with ingroups and outgroups. 
In order to better understand the content of this table, two points must be 
noted. First, not all ingroups are equally important for collectivists. On the 
contrary, there is usually a preference or priority ranking among them. 
According to Triandis (1993), in many societies the order is (a) family, (b) 
nation, (c) religious group and (d) work group. However, in Japan the 
religious group does not exist, and the work group ranks higher. Together 
with the priority order of ingroups, we must take into account the existence 
of vertical collectivism and horizontal collectivism. The former implies 


202 
Social identity: international pers
pectives 
Table 14. 1 
Summary of some dif
f
erences between individualists and 
collectivists 
Diff
erence target 
Individualists 
Collectivists 
Content of self 
Individual diff
erences 
Social categories 
Way of achieving self-
'I can do whatever I want' 
'I am not a burden to my 
actualization 
group' 
Basic unit of survival (belief) 
Individual 
Group 
Regulation of behaviour 
Personal attitudes and cost-
Ingroup norms 
benefit analysis 
Personal goals vs ingroup 
Personal > ingroup 
Ingroup > personal 
goals 
Diff
erence between ingroup 
Weak 
Strong 
and outgroup 
Ingroup and outgroup 
Outgroup more 
Ingroup more homogeneous 
homogeneity 
homogeneous 
Kind of relations 
Horizontal 
Vertical 
submission to ingroup authority. The latter means a fusion with in 
group 
identity. Both kinds of collectivism may coexist, but the prevalence of one 
or the other depends on the type of society concerned. 
Markus and Kitayama (199Ia) designate individualist cultures as 'inde­
pendent self' cultures, and collectivist ones as 'interdependent self' cultures. 
According to these authors, there are important differences between both 
types of culture as far as emotional experience is concerned, since the 
fundamental emotions, their sustaining basis and the functions they fulfil 
do not coincide. Thus, in 'independent self' cultures, the most important 
emotions are ego-focused, specifically anger, frustration, pride, guilt and 
haughtiness, their basis consists on the subject's internal attributes, and their 
function is the expression of inner feelings. On the other hand, emotions in 
'interdependent self' cultures are focused on others. Among them shame 
and sympathy stand out, the basis of these being a special sensitivity to 
others, an ability to take the perspective of the other and an active search 
for interdependence. Their function is to express the outcome of 
interpersonal relationships. 
Individualism and social identity: Hinkle and Brown's proposal 
As the various chapters in this book have shown, social identity theory 
(SIT) deals with processes of group comparison and identification, rela­
tionships between groups, the formation of psychological group, and the 
relative weight of personal and social identity, among many other processes. 
Results obtained by those who have studied the repercussions of the 
individualism-collectivism orientation, summarized in Table 14.2, lead us 
to expect a convergence between studies on individualism and those carried 
out within the framework of SIT. 


Discrimination and belie
f
s
 on discrimination 
Table 14.2 Relation between identification with the 
group and intergroup dif
f
erentiation as a f
unction of 
individualism/collectivism and autonomous/relational 
orientation 
Orientation 
Orientation 
Autonomous 
Relational 
Individualist 
.03 
.33* 
n = 17 
n = 24 
Collectivist 
.47** 
.79*** 
n = 25 
n = 1 8  
* p < . 10; ** p < .05; * * *  
p < .01 
203 
Hinkle and Brown (1990) were the first to pay attention to this con­
vergence. From a theoretical point of view, they state that processes of 
social identity studied by SIT will take place mainly in 'collectivist kind of 
groups' (p. 66), since it seems reasonable to expect only collectivist groups 
or individuals to worry about the outcome of comparisons between ingroup 
and outgroups. The general underlying idea is that a collectivist orientation 
seems important for the manifestation of social identity processes (Hinkle & 
Brown, p. 67). 
From an empirical point of view, Hinkle, Brown and Ely (1992) found 
that the relation between identification with the group and intergroup 
differentiation is significantly greater in collectivists than in individualists. 
This is a proof, they argue, that only collectivists are really concerned with 
the outcome of comparisons between ingroup and outgroups (See Table 
14.2). 
Individualism and identity theory: other perspectives 
SIT is also a theory of the formation of a psychological group. Since the 
earlier experiments using the minimal group paradigm, both Tajfel and 
Turner interpreted the discrimination taking place in that situation as 
evidence of group formation. Insko and Schopler (1987, p. 2 1 5) note that 
Tajfel considers the effect of mere categorization as a proof of entitativity, 
meaning the feeling of shared group membership. As early as 1 98 1 ,  Brown 
and Turner proposed using an interpersonal-group continuum instead of 
the famous interpersonal-intergroup continuum. The reason was, according 
to these authors, that 'the same processes may tend to distinguish both 
intragroup and intergroup processes of interpersonal relationships' (Brown 
& Turner, 198 1 ,  p. 40). In other words, intergroup differentiation implies 
group formation and vice versa. 
Nevertheless, research by Triandis and his co-workers suggests that one 
of the main differences between individualists and collectivists is the 


204 
Social identity: international perspectives 
relationship they maintain with their own groups. In principle, the idea 
must be rejected that individualists find it difficult to join a group. Even 
accepting the schematic picture of Table 14. 1, the very fact of having an 
ingroup and an outgroup implies group identification, as minimal as this 
may be. Even more, it would not be incorrect to predict a greater ease for 
individualists to form a group, having in mind that their differentiation vis­
a-vis outgroup people is weaker. For the same reason, it seems difficult for 
collectivists to be able to form a group with people not belonging formerly 
to any of their few ingroups. 
At this point we must mention an interesting result reported by Hinkle, 
Brown and Ely (1992, p. 107-108). For subjects showing an autonomous 
as well as individualist orientation, the correlation between identification 
and differentiation was .03, as can be seen in Table 14.2. However, these 
subjects did show positive discrimination in favour of their group. There­
fore, we tried to find out: (a) whether individualists show discrimination in 
the minimal group paradigm situation; (b) whether there are differences 
between them and collectivists in their showing intergroup discrimination. 
First experiment 
Subjects were 52 students from adult education courses, aged between 20 
and 60, and being of a low socio-economic level. All of them had previously 
completed the Spanish version of Triandis et al.'s (1988) Three Factor 
Questionnaire, validated in this country. Two groups of 26 subjects were 
formed, each consisting of 1 3  subjects above and 1 3  subjects below the 
median of the questionnaire. Thus, both groups were homogeneous in this 
variable. One of the groups was put to the standard categorization pro­
cedure. This was the experimental or categorized group. The other one, not 
categorized, was the control group. Therefore, there were two variables -
categorization and individualism - each having two values. Subjects 
answered only once to the six Tajfel matrices and to the 10 Insko matrices. 
The MANOV 
A carried out on Tajfel's strategies showed that the 
experimental group (categorized) was more discriminatory than the control 
group (not categorized). Thus the experimental group was significantly 
higher in maximum ingroup profit (MIP) over maximum joint profit (M1P): 
6.17 vs I, F(I, 39) = 1 2.93, p < .01 .  It was also higher in maximum 
(ingroup) difference (MD): 5.29 vs -.58, F(I, 39) = 1 8.61, P < .01 and in 
MIP over fairness (F): 5.96 vs -.26, F(I, 39) = 16.96, P < .01 .  On the other 
hand, the control group was higher in fairness or equity: 7.84 vs 2.38, F(I, 
39) = 7.78, P < .01 .  
The MANOV 
A carried out on Insko's strategies yielded similar results. 
The experimental group was higher in max. reI. own: .86 vs - 1 .0 
I, F(I, 39) 
= 10.83, P = .06. The control group was higher in max. joint other: .25 vs 
-.93, F(I, 39) = 9. 1 5, p < .06, and also in min. diff.: 1 .83 vs .61, F(I, 39) 
-3.76, p < .06. 


Discrimination and belie
f
s
 on discrimination 
205 
No effects of individualism were found, nor interaction between cat­
egorization and individualism. This means that the classic effect of ingroup 
favouritism, first found by Tajfel and afterwards by many other researchers, 
appears again in our study, both for individualists and for collectivists in the 
two types of matrices. 
In order to find out whether there were any differences in the kind of 
discrimination showed by individualists and collectivists, within-cell analysis 
of the preferences in the use of different strategies were carried out, both in 
experimental and in control group. No significant differences appeared 
either. 
Second experiment 
Within-cell comparisons in the first experiment were based on a small 
number of subjects (13). We carried out a second experiment to increase 
that number. We decided not to include the control group, since the effect 
of categorization seemed beyond doubt, and focused only on the individual­
ism variable. Ninety-two subjects participated, of whom 47 were indi­
vidualists and 45 were collectivists. 
The procedure was similar: subjects completed the six Tajfel matrices and 
the 10 Insko ones, which provided the first measure of the dependent 
variable, named first choice. However, a change was introduced. After the 
booklet where they had marked their preferences had been taken away, 
subjects were given a second one where they were asked: (a) to complete the 
16  matrices again, recording their previous choice in each, but explaining 
why they had made it - since there were 1 6  different matrices, subjects had 
to provide 1 6  reasons; this second measure was called justified choice; (b) to 
record the choice they thought people of their group had made, ingroup 
choice; and (c) to record the choice they thought people of the other group 
had made, outgroup choice. 
Next, we will present the results obtained for each of these measures. But 
before we do so, we must stress that the effect of categorization, in the 
absence of a control group, is evident for every measure in the prevalence of 
discriminating over non-discriminating strategies. Thus, in the first choice 
only the fairness/equity strategy in Tajfel's matrices is higher than discrimi­
nating ones: 6.87. But both MIP over MJP (3.89), and MD (3.47), as well as 
MIP over fairness (3.01) are higher than MJP (.30) and MJP + MIP (.47). 
Similar results are found in the other three measures of the dependent 
variable. Insko's matrices show a similar pattern, with the exception of, 
only, the strategy max. joint own, which is as high as the discriminating 
strategies max. own and max. reI. own. 
Results regarding first choice 
The MANOV 
A carried out on Tajfel's matrices yielded significant values. 
The univariate Fs show that individualists are lower than collectivists in 


206 
Social identity: international pers
pectives 
MIP over fairness, 1 .60 vs 4.44, F(I ,  89) = 6.98, p < .01 ,  which is evidence 
of their lesser degree of discrimination. At the same time, they are higher in 
fairness, 7.69 vs 6.04, F (1, 89) 
= 2.00, p < . 1 6. MANOVA was not 
significant for Insko's matrices. 
However, in these latter matrices differences do appear when relative 
preferences of individualists and collectivists are compared by means of the 
Wilcoxon test. Thus, we find that individualists are different from collec­
tivists in that they prefer the strategy of fairness (min. diff.) over max. reI. 
own and max. joint own; also in their preference for max. joint own over 
max. own. Collectivists, on the other hand, choose max. reI. own more 
often than max. joint own. All this converges in a description of indi­
vidualists as more prone than collectivists to fairness and to joint and non­
discriminating strategies, bearing in mind that this differential tendency is 
found within a discrimination framework shared by both kinds of subjects, 
which explains the lack of differences in the MANOV 
A. 
Results regarding justified choice 
The MANOV 
A did not yield any significant differences, either in Tajfel's 
matrices or in Insko's ones. 
As we have mentioned, after subjects had completed the matrices, they 
were asked a number of questions. The first one was about the reason why 
they had chosen a particular strategy in allocating points between two 
persons. More specifically, the question was: 'Why did you choose that cell?' 
Each subject had to record the reason corresponding to each cell. Since 
these amounted to 1 6, each subject provided 1 6 reasons to explain his/her 
choice-allocating strategies. 
After analysing the reasons given by all subjects, we found they could be 
grouped into 10 main categories: 
1 
Fairness or equality 
34% 
2 
Equality to balance 
4.2% 
3 
Equality when it is the only choice not detrimental to one's 
group 
0.58% 
4 
Maximum ingroup profit 
24.8% 
5 
Maximum ingroup difference 
8.53% 
6 
Maximum ingroup joint profit or maximum joint own 
1 5.31% 
7 
Maximum joint other 
3.2% 
8 
Maximum outgroup difference 
0.77% 
9 
Others 
0.84% 
10 Chance 
7% 
Six of these reasons correspond exactly to the strategies of matrices. These 
are f
airness or equality (F) (min. diff.), maximum ingroup pro
fit (MIP) (max. 
own), maximum ingroup dif
f
erence (MD) (max. reI. own), maximum ingroup 
joint pro
fit (MJP) (max. joint own), maximum joint other (max. joint other) 
and maximum outgroup diff
erence (max. reI. other). 


Discrimination and belie
f
s
 on discrimination 
207 
These six strategies tend to be the most often used by subjects. In fact, a 
high percentage of them repeat the reason given for the first cell in the 
remaining 1 5, that is, does not show 'alternation' in its reasons. In these 
cases, which we have called 'Use of Single Strategy', or USS, subjects tend 
to use, almost always, some of these six reasons. 
On the other hand, reasons 2, 3 and 9, respectively categorized as 
'equality to balance', 'equality when it is the only choice not detrimental to 
one's group' and 'others', are subsidiary to the former ones. This means that 
they are used only in cases of 'alternation' of strategies, or SA, where 
subjects do not want to limit themselves to only one reason and resort to 
other complementary ones. 
Let us see an example. Subject 34 uses a perfect alternation of reasons. 
On eight occasions he employs reason 4 (MIP) (max. own), which is dis­
criminatory. In the remaining eight, he uses reason 2 (equality to balance), 
which is egalitarian. This subject's words are worth being cited. When 
he chooses reason 4, he justifies his choice as follows: 'To give points to the 
people having the same code as myself'. When he chooses reason 2, he 
states: 'In order not to increase too much the difference in favour of my 
group.' 
In the case of this subject, as well as in other similar ones, reason 2 is 
clearly subsidiary to reason 4. The subject wants to benefit his/her group 
more, but does not want to do so in an extreme or radical way. Therefore, 
this justice or equality pursued arises from a wish to compensate a previous 
discrimination. 
Reason 3 is even more subsidiary than reason 2. It is usually employed by 
subjects with a discriminatory orientation when they meet a Tajfel matrix in 
which they cannot benefit their group more than the opposite group. This 
is the case of subject 39, who shows a discriminatory pattern of reasons, 
choosing 1 5  times reason 4 (MIP) (max. own), and only once reason 3. In 
justifying his discriminating strategy, this subject states: 'To give more 
points to my group.' However, when choosing the equality strategy in 
Tajfe\'s last matrix, he says: 'To give as few points as possible to the other 
group.' Effectively, in this latter matrix it is not possible to discriminate in 
favour of one's own group. 
It is clear, then, that reason 3 justifies equality in a subsidiary way, as a 
sort of lesser evil to prevent one's own group from getting less points than 
the other. 
The last subsidiary reason is 9, which we have called 'others'. Sometimes, 
subjects look for justice allocating one group more points in one cell and 
less in the next one, since they think that a strict equality strategy will be 
detrimental to both groups in rendering their whole absolute profit lower. 
Thus, subject 1 8  states: 'I have tried to balance the two groups. Since 
compensation is sometimes difficult, it is rather a random criterion.' 
Reason 1 0, 'chance', embodies those reasons not related to the strategy 
chosen. Five subjects used only this reason and were removed from the 
analyses, as was another who chose it most of the times. 


208 
Social identity: international pers
pectives 
Nevertheless, the great majority of reasons (82.64%) were distributed 
among the four main categories, which closely coincided with the main 
allocating strategies. These four reasons are fairness (34%), maximum 
ingroup profit (24.8%), maximum difference (8.53%) and maximum joint 
profit (1 5.3%). 
Although each subject had to justify 1 6 different choices, a thorough 
analysis of the pattern of justifications showed that most subjects used only 
one reason for all 1 6  occasions or alternated between two. There were no 
differences between individualists and collectivists in the regularity of use of 
justification patterns: 74.4% of the former and 73.3% of the latter used one 
or two reasons to justify their respective choices. 
When only the justifications of these 'regular' subjects are considered, the 
four above-mentioned categories explained 96.5% of the reasons used 
(instead of 82.64% which appeared when all subjects were computed). Dis­
criminatory strategies, on the one hand (maximum ingroup profit and 
maximum difference), and non-discriminatory ones, on the other (fairness 
and maximum joint profit), were added up. This calculation led to the 
discovery that in individualists the use of discriminatory reasons (28.1%) is 
less than the use of non-discriminatory reasons (69.2%), a difference which 
is significant. However, in collectivists, though the use of discriminatory 
reasons is also less (42% and 54%, respectively), the difference is far from 
being significant. 
Ingroup choice 
By means of the Wilcoxon test we tried to find out whether there were 
differences between individualists and collectivists in their preferences for 
any strategy in the various choices. In other words, when individualists 
proceed from first choice and justified choice to ingroup choice, do they 
change their preferences as compared with collectivists? If this is true, the 
conclusion would be that individualists see ingroup members in a different 
way from how collectivists see members of their own ingroup. 
Accepting p < .05 as a criterion of statistical significance, we found that, 
in effect, there were strong differences between collectivists and individual­
ists. They certainly coincide in thinking that members of their own group use 
the fairness/equity strategy of Tajfel's matrices less often than themselves, 
which means they consider themselves as fairer than their ingroup's mem­
bers. But individualists, as opposed to collectivists, think that ingroup mates 
are higher than themselves in the three Tajfel discriminatory strategies, 
namely MIP over MJP, MD, and MIP over fairness. 
In Insko's matrices something similar happens. Collectivists consider 
themselves as more just than in 
group members. And so do individualists. 
Moreover, collectivists see themselves as less discriminatory than ingroup 
members in max. reI. own. And so do individualists. But these think they 
are less discriminatory than ingroup members in max. own, and here they 
are opposed to collectivists. 


Discrimination and belie
f
s
 on discrimination 
209 
Outgroup choice 
As in the previous choice, the Wilcoxon test was employed to look for 
differences between individualists and collectivists in the use of strategies in 
this choice as compared to the two fonner ones. 
In Tajfel's matrices, results are exactly the same as in the previous choice. 
Collectivists consider themselves as more prone to fairness (F) than 
outgroup members. But collectivists do not think outgroup members to be 
more discriminatory than themselves. Individualists, however, do think so 
in the following strategies: MIP over MJP, MD, and MIP over F. 
Results are also very similar in Insko's matrices. Both collectivists and 
individualists consider themselves as more fair than outgroup members and 
less discriminatory in max. reI. own. But individualists, as opposed to 
collectivists, think they are less discriminatory than outgroup members in 
max. own. 
First choice-justified choice and ingroup choice-outgroup choice 
comparisons 
Considering the six Tajfel strategies and the seven Insko ones, as well as the 
two groups, individualists and collectivists, a total of 26 first choice-justified 
choice comparisons and 26 ingroup choice-outgroup choice comparisons 
were carried out. None of them yielded significant differences, which, in 
view of the previous results, suggests that subjects are very homogeneous 
when answering for themselves (whether having to justify their answer or 
not) and when they answer in the place of others (be these from in 
group or 
from outgroup). 
Conclusion 
Two conclusions can be drawn from the results of our two experiments. 
First, both individualists and collectivists discriminate in favour of their 
own group in a minimal group situation. As we have seen, there is a slight 
tendency for individualists to be less discriminatory. However, on the 
whole, both kinds of subjects provide evidence of group fonnation, along 
the lines suggested by Insko and Schopler (1987). 
Second, as far as beliefs on discrimination are concerned, it seems clear 
that individualists consider their point allocation to be less discriminatory 
than that of collectivists. The reasons they provide to justify their second 
choice seem to imply that individualists interpret their allocating behaviour 
as a search for justice and a wish to share with the other group. On a much 
lower number of occasions they admit to being motivated by a wish to 
dominate the opposite group. 
Moreover, the divergence between individualists and collectivists appears 
again in the clear-cut distinction the fonner establish between themselves 


210 
Social identity: international perspectives 
and people of ingroup and outgroup. Collectivists tend to consider them­
selves as more just in their comparisons to ingroup and outgroup, but this 
result can be easily explained as due to social desirability. Individualists also 
see themselves as more fair and, at the same time, as less discriminatory. 
This explicit and public acceptance of their non-following ingroup norm 
seems to speak for a claim of independence vis-a.-vis the group, of a certain 
critical positioning against it, despite the fact that actual allocating 
behaviour is discriminatory and, as such, is a proof that the person feels as 
a member of the group to all purposes. 
All the aforementioned leads us to the conclusion already outlined in 
previous pages. It is not a matter of individualists experiencing insurmount­
able difficulties to form groups or identify themselves with them. Rather, 
what happens is that they do so while trying to preserve their individual 
initiative, in this particular case restricted to beliefs. Of course, other inter­
pretations of our data are possible, this allowing and, at the same time, 
encouraging future research on a subject which is of interest to understand 
processes mediating behaviour of people in groups and in society as a 
whole. 
Note 
This chapter was translated from the Spanish by Elena Gaviria Stewart. 


1 5  
Constructing Social Identity: The Role of 
Status, Collective Values, Collective Self­
Esteem, Perception and Social Behaviour 
Daria Paez, Cristina Martinez-Taboada, Juan Jose Arr6spide, 
Patricia Insua and Sabino Ayestaran 
Social identity theory (SIT) and Turner's self-categorization theory (SeT), a 
more cognitive version of the former, have stimulated research on group 
processes and are currently some of the most heuristic theories in this 
domain. Nevertheless, some psychological aspects of SIT have not received 
systematic empirical support. The most important ones are the following: 
(a) the belief in the relationship between salience of the categorization and 
intergroup discrimination; (b) the relationship between strong identification 
and perceptual and behavioural biases which favour the ingroup; (c) the 
relationship between discrimination and an increase in self-esteem (Messick 
& Mackie, 1 989). These results suggest that the relationship put forward by 
SIT, and also partly by SeT, between salience and favourable behaviour 
towards the ingroup, coupled with an increase in self-esteem, is not so 
general as was once thought. However, we must state that these criticisms 
are valid only for a more individualistic approach to SIT and SeT. 
Following Turner, Hogg, Oakes, Reicher and Wetherell (1987), psycho­
logical processes of identification, self-esteem, discrimination, and so on, are 
only one aspect of intergroup behaviour. Socio-structural position and 
socially shared belie
f
s
 on the legitimization of power/status differences and 
individual/collective strategies to cope with these social differences are two 
main explanatory processes of intergroup behaviour. From this perspective 
it is a mistake to focus only on the intra 
psychological relationships of 
identification, self-esteem and discrimination. 
This chapter will analyse certain themes of debate which have arisen with 
respect to the concept of social identity: When does it regulate a person's 
behaviour and perception? When does it produce a favourable perception of 
the in 
group and discriminatory behaviour towards the outgroup? Hinkle, 
Brown and Ely's (1992) revision of 
the literature has shown that there is not 
always an association between group identification and ingroup favouritism, 
and that there is no consistent association between discrimination in favour 
of one's group and increasing one's self-esteem. Hinkle et al. confirmed that 


212 
Social identity: international perspectives 
subjects who share collectivistic values are those who show more association 
between identification and ingroup favouritism. Crocker and Luhtanen 
(1990) have found that the variable linked to intergroup discrimination is 
collective and not personal self-esteem: it is those subjects with a high 
collective self-esteem who show a discriminatory bias in situations which 
threaten one's social identity. Abrams (1990, 1 994) has shown the moder­
ating role of private self-consciousness and strong identification: those 
subjects with a strong group identification and high private self­
consciousness are those who show a stronger behaviour regulation on the 
basis of one's social identity. We will present Spanish data which confirm 
and specify the role of these moderating variables in the relationship 
between social identity, perception and social behaviour. Finally low-status 
groups' socio-cognitive functioning does not always respond to the logic 
derived from SIT's statement of the need to defend one's identity: instead of 
showing ethnocentrism subjects present centrism or outgroup favouritism. 
Turner's perspective emphasizes the role of socially shared belie
f
s
 and socio­
structural position in order to explain these results. In our case we will focus 
on the role played by individualistic versus collectivistic values and in the 
role of collective self-esteem. This is why we will analyse social status and 
the predominance of the collectivistic cultural values as factors which will 
moderate the regulation of one's behaviour by social identity. 
We believe it is specially important to contrast the macro 
social and 
transcultural validity of the above-mentioned theories. This is one of the 
reasons why first of all we will devote special attention to the studies 
conducted in Spanish and Latin American contexts on the issues of 
collective identity, stereotypes and discrimination (Morales & Paez, 1 996). 
Individualism-collectivism, status and the importance of personal and 
social identity 
SIT states that there is a personal-intergroup continuum. On the intergroup 
pole, subjects see themselves as interchangeable members of the category or 
group. On the personal pole, subjects will be defined by their singularity and 
individual specificity. It is thought that the coping strategies which the 
members of low- or high-status groups rely on will be located as opposites 
in this dimension: on one extreme we will find individual mobility, on the 
other side social competition. Nevertheless, research carried out in Latin 
America has shown that Chilean workers simultaneously agreed with social 
conflict strategies and using strategies aimed towards individual improve­
ment of their siblings' position by means of education (Quevedo, 1 972). 
This study did not analyse the correlation between the agreement with both 
strategies. Even though this correlation could be negative, and so confirm 
the hypothesis which states the opposite nature of both these strategies, the 
fact that the average agreement with both strategies was high confirms that 
they may coexist on a group level. The opposition between the salience of 


Constructing social identity 
213 
identification levels which have different inclusiveness (that is, regional and 
supra-national) has not been confinned as a general phenomenon. We 
must state that some subjects show strong regional identification and a low 
identification on a higher level. But other subjects simultaneously show a 
strong identification with different levels of inclusiveness (as was the case 
of a simultaneously strong identification with the Polish and European 
national identities, Mlicki & Ellemers, 1996). Finally it is possible that in 
collectivistic cultures the difference between social and personal identity is 
much less than in individualistic cultures. SIT posits that an important 
dimension of identity is the social dimension. Nevertheless, the identity or 
individual self-concept may be more or less shaped by social identity. 
There is both a trait and situational variability with regard to the 
integration of social identity in personal identity or the fonner's relevance. 
The importance of social identity and its integration within personal 
identity will be mediated by the subjects and the historical context which 
contextualizes them. The importance of social identity is stronger in 
collectivistic and sociocentric cultures in comparison to individualistic and 
egocentric ones. 
Individualism is a set of beliefs, values and cultural practices in which 
people manifest an emotional independence from groups and organizations. 
On the other hand, collectivism as a cultural syndrome is associated with an 
emotional dependence towards a group. A collectivistic culture is that in 
which there is a stable relationship between the individual and certain 
infonnal groups to which one belongs, a subordination of individual objec­
tives to those of the collective, and an adjustment between the person's and 
the group's objectives. There is a tendency towards not leaving the group 
even though the requirements may be high. An individualistic culture is that 
in which a person has negotiable relationships with various groups. S/he 
may abandon these groups when they are very demanding (or potentially 
may be so), and the individual's goals are more important than the group's 
goals (Morales, Lopez & Vega, 1992; Triandis, 1992, 1994; Triandis, 
Bontempo, Villareal, Asai & Lucca, 1988). Cooperation with the group, 
sharing resources and spaces with members of certain stable ingroups, col­
lective achievements and a strong affective link all characterize collectivistic 
subjects. Interpersonal competition, having important private spaces and 
resources, individual achievements and distancing oneself from the group 
are all characteristics of an individualistic subject (Hinkle et aI., 1992). 
From the point of view of the importance of social or collective identity 
for one's self-image or for the social representation of the individual (see 
Doise, Chapter 2 in this volume), individualistic cultures stress a concept of 
the individual as a separate, autonomous, self-contained and independent 
being. Collectivistic cultures emphasize a representation of the person as an 
entity centred in his or her relationships with others, connected to, and 
dependent on, the social context (Markus & Kitayama, 1991 
b). 
The group's social status also affects the degree to which individuals 
define themselves in tenns of their social or personal identity. High-status 


214 
Social identity: international perspectives 
groups perceive themselves as a set of individuals. Low-status groups define 
themselves as an aggregate of subjects who share a series of similar 
characteristics (as Lorenzi-Cioldi et al.'s studies have shown). Members of 
dominant groups will define their identity mainly on the basis of personal 
factors, while members of dominated groups will define their self-concept on 
the basis of mainly social factors (Lorenzi-Cioldi & Doise, 1 994). As a 
summary we may state that dominated groups and collectivistic cultures will 
reinforce the relevance of social identity over personal identity. On the other 
hand, individualistic cultures and dominant groups will pay more attention 
to, and have a better evaluation of, an individual who possesses resources 
while also stressing the relevance of personal identity over social identity. 
Low status, salience of the intergroup comparison, outgroup 
favouritism and group perception 
Culture and status have an influence not only on the importance of the 
social categories in the representation of one's self, but also in assessing 
the representation of group perception. Collectivistic cultures, which posi­
tively value the group, reinforce the perception of intragroup homogeneity, 
while individualistic cultures, which positively value an autonomous subject, 
reinforce the perception of ingroup variability (Devos, Comby & 
Deschamps, 1 995). 
From the point of view of intra- and intergroup social perception, we 
must remember that there is a tendency towards perceiving the ingroup as 
better in certain attributes. For instance Rodriguez (1992) found that 
subjects from Venezuela evaluated themselves better than subjects from 
Colombia in four out of six positive attributes, whilst the Colombians did 
the same in five out of six attributes. 
In general, subjects differ positively in those dimensions which define the 
ingroup, although they may accept outgroup superiority in another set of 
dimensions. This social perception which favours the in 
group is modelled 
by social status: high-status groups tend towards a stronger perceptive 
differentiation in favour of the ingroup than those subjects who hold a 
lower status, as Mullen, Brown and Smith's (1992) meta-analytic review 
showed. 
In relation to the variability of stereotypical attributes, the ingroup is seen 
as more complex, variable and heterogeneous than the outgroup. There is a 
tendency towards perceiving the outgroup as more homogeneous and the 
ingroup as more heterogeneous. As Mullen and Hu's (1989) meta-analytic 
review has confirmed, this effect is further stressed in the case of natural 
groups. Status also has an influence in the perception of variability: minority 
or threatened groups will perceive their group as less variable. This is a way 
of reinforcing its distinctiveness and cohesion. Results have also shown that 
groups perceive themselves as more homogeneous on those attributes or 
dimensions which define them (Leyens, yzerbyt & Schadron, 1 994). 


Constructing social identity 
215 
'Cold' socio-cognitive theory (Leyens et ai., 1994) suggests that having 
had a direct experience with the group will lead to a more variable, 
heterogeneous or individualized representation of the ingroup. 
From an SIT perspective, cognitive categorization as a member of a 
group, or the identification with this group, will stress intergroup differences 
and intragroup similarities. Individualization or aggregation would be a 
function of the defence of one's social identity, on the basis of the existing 
power and status relationships. Social comparison will influence not only 
the perception of differences but also homogeneity. It will not always be the 
outgroup who will be perceived as less variable, simple or homogeneous. 
The ingroup will be perceived as more homogeneous in those attributes 
which define their social identity and in situations in which this identity is 
under threat (Leyens et ai., 1994; Marques & Paez, 1 994). 
In a study conducted by us we confirmed the predominance of this 
normative logic of identity defence, instead of a cognitive logic of experience 
(see Marques, Paez and Abrams, Chapter 9, this volume). We asked 
psychology students (N = 109) to evaluate the ingroup's (psychologists) and 
outgroup's (engineers) variability in stereotypical attributes (Table 15. 1). 
These attributes had been generated in previous studies and suggested that 
being friendly and sociable are typical traits awarded to psychologists, whilst 
being competitive, intellectual and having a strong leadership capacity are 
typical of engineers. On the basis of these answers we calculated the Pd 
(perceived difference) and Var (variability). The estimated variabilities are 
significantly different, although not regularly higher in the ingroup. 
Variability was higher in the outgroup on those attributes stereotypical of 
the ingroup and vice versa. In other words, psychologists perceived them­
selves as more homogeneously friendly and sociable in comparison to 
engineers, while this last group was seen as more homogeneously competitive 
and intellectual, although in this last attribute differences were smaller and 
other studies have found that it is associated with the psychologists' 
stereotype. In those variables in which we did not find significant differences, 
means were in the hypothesized direction. Variability was smaller for 
engineers in leadership and in being normative. 
The tendency towards accentuating ingroup favouritism on important or 
consensual dimensions seems an increase in competition or salience of the 
comparison conditions, as Doise and Sinclair's study (reanalysed by van 
Knippenberg & Wilke, mentioned in Morales & Huici, 1994, p. 737) has 
shown. Nevertheless in a study conducted in Brazil by Nascimento-Schulze 
(1993) on bank clerks, a social comparison situation led to a worse image of 
the group than a situation in which there was no social comparison, and the 
same happened when the context made group membership more salient. As 
we can see, this is contrary to the fact that group salience should activate a 
process of favouritism or defence of the group's image. The author 
interprets these results as showing that these groups of bank clerks in social 
comparison and salient conditions are aware of their problematic work 
conditions and of their unstable status, which is reflected in a diminution of 


216 
Social identity: international pers
pectives 
Table 15.1 Perceived ingroup and outgroup heterogeneity 
Variability means of 
Engineers 
Psychologists 
Attributes 
Mean 
SD 
Mean 
SD 
Friendly 
0.88 
0.56 
0.68 
0.54 
Competitive 
0.60 
0.61 
0.82 
0.63 
Sociable 
0.8 1  
0.72 
0.56 
0.52 
Intellectual 
0.64 
0.56 
0.76 
0.53 
** p < .01 ;  *** p < .001 
t (1 1 3) 
3.86*** 
-3.17** 
4.03** 
-2.72** 
their self-image. This socio-cognitive functioning of groups with unstable 
and relatively insecure status does not seem to follow an ethnocentric logic. 
In a conceptual replication of this research, we have confirmed that in 
conditions of salience of the comparison subjects increase the superiority of 
the outgroup in consensual or stereotypical attributes. In order to manipu­
late the salience of the comparison, we asked subjects enrolled in a group 
dynamics course (N = 1 1 8) to perform a cognitive creativity task: to list the 
attributes which characterize a good leadership. Subjects were randomly 
assigned to a condition of activation of the comparison with the engineer 
outgroup, or of ingroup comparison. All the subjects had to read a first set 
of attributes. Depending on the condition, subjects were told that this list 
had been written by psychologists (ingroup) or engineers (outgroup). Apart 
from performing the task, subjects also had to estimate ingroup and out­
group variability on a series of six attributes: leadership, friendly, competi­
tive, normative or followers of the norm, sociable and intellectual. On a 
five-point scale subjects had to decide, for instance, what percentage of 
engineers and psychologists had no, some, regular, a high or an extremely 
high intellectual capacity. This cognitive task allowed us to obtain vari­
ability indices such as the Pd and the Var, whilst also allowing us to infer 
the group's mean on that attribute. All the subjects evaluated both psy­
chologists and engineers. The F of the between-group interaction with 
regard to the group that was being evaluated (ingroup or outgroup) and the 
salience manipulation was significant (F( l ,  1 1 3) = 4.89; p < .05). Specifically 
psychologists were attributed a much lower leadership average when the 
social comparison was made salient. This reflects the realistic perception of 
the inferior nature, in terms of status, of the psychologists compared to 
engineers (see Table 1 5.2). 
In order to confirm that previous results (psychologists perceived them­
selves as more homogeneously friendly and sociable in comparison to 
engineers; the engineers group was seen as more homogeneously com­
petitive in comparison to psychologists) followed a normative or social 
desirability logic, we asked a sample of psychology students (N = 9), 
psychologists and mental health workers (N = 9), and qualified workers, 
including physicians and engineers, and non-qualified workers (N = 1 2) to 


Constructing social identity 
217 
Table 1 5.2 
Perception of ingroup and outgroup by salience versus non­
salience o
f
 social comparison 
Attributes 
Leaders 
Salience of social comparison 
('between' variable) 
Engineers 
Psychologists 
('within' variable) 
Inferred mean 
Inferred mean 
3.31 
2.71 
Non-salience 
('between' variable) 
Engineers 
Psychologists 
('within' variable) 
Inferred mean 
Inferred mean 
3.47 
3.34 
'evaluate to which degree the following characteristics are desirable or 
positive in each jobs' performance'. As we can see in Table 15.3 both the 
engineers' higher status and the normative nature for psychologists on those 
attributes in which there was a smaller variability (friendly, sociable) was 
confirmed. Exactly the opposite took place for the attribute 'competitive' 
(less variability and it was attributed more to engineers). The only exception 
was found in the attribute 'intellectual', in which psychologists have a 
higher score and that in the first study showed more variability, although 
the difference in variability means was the smallest of all. In order to 
confirm that the attribution of the traits followed both a professional social 
desirability logic as well as a group identification logic, we correlated the 
subjects' level of identification with psychologists ('we ask you to indicate 
your identification with the professional category termed "psychologists"') 
with a composite rating of attributes (friendly, sociable, intellectual) of the 
group. This composite score showed a satisfactory reliability (alpha = .67) 
and a significant correlation of 0.35; p < .05 with one's identification with 
the group of psychologists. This identification was individually associated 
with believing that it was desirable that the psychologists should be friendly 
(r = .36; p < .05), with sociable (r = .27; P < .10) and, although of marginal 
significance, with intellectual (r = . 1 5; P < . 1 0). 
On the other hand, identification with psychologists was not associated 
with the agreement that psychologists were competitive and normative (r = 
< .09). However, those features were typical of the engineers. Identification 
with psychologists was associated with the leadership attribute (r = .42; P < 
.05). Bearing in mind one's professional experience, in comparison to 
students, psychologists and mental health experts stated that it was more 
socially desirable for a psychologist to have leadership capacity (mean = 
5.1 1), and both these groups believed in this fact more than did qualified 
and non-qualified workers (mean = 3.75; F(2, 27) = 4.07; P < .03). Group 
identification was linked with a stronger attribution to this group of the 
positive attributes, especially when these define their professional identity. 
Moreover, even when the task was a comparison with a group of engineers, 
those psychologists and mental health experts more professionally involved 
attributed positive traits (e.g. leadership) to the ingroup, though it has been 
shown that these traits tend to be associated with high status groups. In 


218 
Social identity: international pers
pectives 
Table 1 5.3 Perceived social desirability o
f
 attributes f
or engineers and 
psychologists 
Means of social desirability for 
Engineers 
Psychologists 
Attributes 
Mean 
SD 
Mean 
SD 
Friendly 
3.70 
1 .61 
5.90 
1 .26 
Competitive 
5.70 
0.95 
4.17 
1 .58 
Sociable 
4.67 
1 .40 
5.97 
1 .22 
Intellectual 
5.00 
1 .19 
5.87 
0.90 
Leadership 
5.80 
1 .3 1  
4.30 
1 .31 
Normative 
5.50 
1 .07 
4.70 
1 .58 
Means of group status for 
Engineers 
Psychologists 
Status 
6.27 
0.69 
4.00 
1 .08 
I 
= not at all; 7 = very much 
* p < .05; ** p < .01 ;  *** = p < .001 
t (29) 
-7.05*** 
4.63*** 
-4.33*** 
-3.98** 
3.84** 
2.68** 
1 1 .84*** 
summary we may state that the lower variability traits associated with 
the psychologists are in general those which are seen as socially desirable 
for the job of being a psychologist. These traits are also associated with the 
identification with this group and involvement in one's professional activity 
(see Table 1 5.3). 
We may conclude that a certain relative ingroup bias is present in social 
perception and that it follows a normative logic - of defence of a positive 
social identity by means of stressing the positive aspects of the ingroup in 
those attributes used to define the group. Moreover, in subordinated groups 
this tendency is not shown in social comparison situations. But this does not 
imply that subordinated groups do not show in general outgroup favour­
itism. Bourhis, Gagnon and MOise (1994) have shown how members of 
dominated groups choose and evaluate more positively dominant outgroup 
symbols. In studies conducted in Latin America, different authors have 
found that the Venezuelan ingroup was evaluated worse than the European 
(Italian) and North American outgroups (Montero, 1987, 1990). Even 
though the subjects identified themselves with the ingroup (Venezuela), 
subjects had a better evaluation of the European groups (England and 
Spain) and of certain American nations such as Argentina, although they 
had a bad image of other American countries such as Colombia. Due to the 
results obtained from similar psychosocial, anthropological and historical 
studies, which show that in Latin America, Asia and Africa there is a 
devaluation of the national ingroup and favouritism towards First World 
outgroups, the concept of Altercentrism has been coined. Altercentrism is 
an important theoretical concept as it allows us to pay attention to a basic 
problematic found in SIT: how do we explain the socio-cognitive func-


Constructing social identity 
219 
tioning of 
the dominated groups who do not mobilize themselves, or cannot 
be individually included in higher status groups? 
Various explanatory processes put forward by SIT suggest how domi­
nated groups manage their ambivalent image, showing at the same time 
favouritism towards dominant groups. 
Intragroup social comparison with more disadvantaged subsectors, or, 
alternatively, an advantaged social comparison with a group of similar and 
pertinent attributes, would be another mechanism which would allow the 
showing of favouritism towards the dominant outgroups while at the same 
time retaining a partly positive image of the ingroup. Members of sub­
ordinated ingroups (that is, Venezuelans) accept the superiority of dominant 
outgroups (that is, North Americans), although at the same time they stress 
their superiority over other groups who hold a similar position to theirs and 
whom they undervalue (that is, Colombians) (Montero, 1 987; Salazar, 
1989). 
Another phenomenon that characterizes dominated groups is a type of 
symbolic individual mobility - of which Bovarism and the cosmopolitanism 
of the dominant classes of Latin America is a good example. Bovarism 
(permanently believing one is living in Paris, London or New York, 
although one is really living in Puerto Hundido) is a form of symbolic and 
ideal identification with the dominant outgroup. Latin Americans, 
especially, although not only, those from dominant classes, view themselves 
as second-class or 'transplanted' Europeans or North Americans. The use of 
surnames (for instance, the large amount of people called Johnny, William 
or Peter Gonzalez), adopting customs and identifying with the values of the 
dominant cultures constitute a coping strategy of the accepted negative 
national identity. Studies carried out in developed countries have shown 
that this symbolic individual mobility strategy takes place even though there 
are no real possibilities of integrating oneself into the dominant group 
(Wright, Taylor & Moghaddam, 1990, mentioned in Bourhis et a!., 1 994). 
Finally, low-status or dominated groups use cognitive creativity, redefin­
ing as important a series of new dimensions in which the ingroup is superior 
(see Smith & Bond, 1 993). Research conducted by Salazar et a!. on national 
stereotypes in Latin America has confirmed the existence of a status/ 
competence dimension, a sociability dimension and a third, less important, 
education/instruction factor. There is evidence to suggest that the instru­
mental or status/power dimensions and the expressiveness or sociability 
ones are transculturally valid dimensions of social perception (Zebrowitz, 
1990). Various studies have shown that subordinated groups accept the 
superiority of dominant groups on the competence/ability dimensions, or on 
status, power and resources, but they view themselves as superior in the 
expressive or sociability dimensions. For instance, nurses with a high status 
viewed themselves as superior in technical knowledge, while lower status 
nurses perceived themselves as being more friendly (Smith & Bond, 1993, p. 
62). Studies conducted in Latin America confirm that the self-stereotypes 
reaffirm a positive evaluation of sociability attributes: attributes stereo-


220 
Social identity: international pers
pectives 
Table 1 5.4 Attributes f
or engineers and psychologists 
Engineers 
Psychologists 
Attributes 
Inferred mean 
SD 
Inferred mean 
SD 
Leadership 
3.40 
0.83 
3.07 
0.80 
3.13* 
Friendly 
2.98 
0.66 
3.83 
0.77 
-9.08** 
Competitive 
4. 1 1  
0.70 
2.78 
0.80 
14.40** 
Normative 
3.56 
0.90 
2.69 
0.84 
7.57** 
Sociable 
3.00 
0.66 
3.96 
0.56 
-1 1 .75** 
Intellectual 
3.42 
0.83 
3.27 
0.67 
6.49** 
* = p < .01; ** = p < .001 
typically favourable to the Latin American national groups were of the 
sociable type (that is, happy), while the stereotypical positive attributes 
towards European and North American groups were of a more instru­
mental nature (that is, practical, workers and so on) (Rodriguez, 1992). 
Our own research conducted in Spain on psychologists and engineers 
confirmed that the perceptive superiority of a lower status ingroup (psycho­
logists) was found on the expressive dimension (they viewed themselves as 
more friendly and sociable), whilst they also accepted the superiority of the 
higher status outgroup on more instrumental attributes. They perceived the 
engineers as more competitive, intellectual, and so on (see Table 1 5.4). 
The above-mentioned results may be interpreted as the effect of a 
normative process, in the sense that subjects try to defend their group by 
perceiving it as an entity which adjusts itself to the norms and values which 
relevantly define its subculture more than the outgroup does (Marques & 
Paez, 1994; also see Devos et al., 1995). Nevertheless, although low-status 
group members may use symbolic individual mobility, comparison with 
disadvantaged groups and cognitive creativeness in order to explain one's 
ambivalent social image, according to the psychological version of SIT, low 
status must have an influence on one's self-esteem. 
Status, self-esteem, collectivistic values and social identity 
Status is an important variable for SIT due to the fact that it is associated 
with the prestige and self-esteem one may infer from one's group 
membership. According to SIT, belonging to a social category or group is a 
source of self-esteem. Belonging to low-status groups, which possess nega­
tive attributes, would motivate a search for individual mobility, cognitive 
creativity and social competition. What is implicit in this approach is the 
idea that subjects from disadvantaged classes and groups should suffer a 
deficit in self-esteem. This idea is included within the hypothesis stating that 
a deficit or threat to self-esteem would be an antecedent or cause of the 
ingroup's favourable social perception, of ethnocentric attitudinal prejudice 
and behavioural discrimination. Personal self-esteem as an evaluative 


Constructing social identity 
221 
component of personal identity may be differentiated from self-esteem 
referring to group belonging or being a part of a certain social category, in 
other words collective self-esteem. 
One answer to the fact that the salience of the categorization does not 
directly lead to differentiation, or that it leads to an increase in self-esteem, 
is methodological. The relationship between discrimination or intergroup 
differentiation and self-esteem should take place and be proven not on the 
basis of personal self-esteem, but on the basis of collective self-esteem, in 
other words on the basis of the favourable evaluation of the category or 
social group to which one belongs. Nascimento-Schulze's (1993) results with 
Brazilian workers confirm that there is no relationship between intergroup 
discrimination and an increase in general personal self-esteem. Intergroup 
discrimination produced an increase in the group's evaluation or collective 
self-esteem. In this sense, the development of collective self-esteem scales is 
a step forward in order to better understand the relationship between social 
identity and social behaviour. Crocker and Luhtanen, 1990; Luhtanen & 
Crocker, 1992) have developed a scale which measures stable individual 
differences in the level of a person's positive social identity (high collective 
self-esteem) and also a positive personal identity. These authors developed a 
scale with a private collective self-esteem dimension, a public collective self­
esteem dimension, the feeling of being a member of the group, and the 
importance of this identity. These scales had structural validity and alpha 
coefficients which scored higher than . 
73 (in fact the public self-esteem scale 
reached a score of .80). The private collective self-esteem scale and the one 
referring to the importance of being a member of the group, which meas­
ures the most individual aspects of social identity, also correlated signifi­
cantly with personal self-esteem scales. The collective self-esteem scores 
were positively and coherently associated with the agreement with collective 
values - measured using Hui's individualism-collectivism scale (Luhtanen 
& Crocker, 1992). 
Our results with the Spanish translation of this scale also show satis­
factory results in relation to internal consistency. Alpha coefficients were 
satisfactory for the private and public collective self-esteem scales, identi­
fication with the ingroup and belonging to this same group (Cronbach 
alphas of .83, .78, .66 and .71, respectively). Moreover, they also showed 
convergent validity with Hui's collectivistic values scale (see below). 
Experimental research found that, using the above-mentioned scale, 
subjects high in collective self-esteem, contrary to those with low collective 
self-esteem, showed more positive biases in relation to their group in 
situations in which there was a threat to their group identity. In situations of 
failure, subjects with a high collective identity changed their scores in such a 
way that they mitigated the threat to the group's worth. Personal self-esteem 
did not play a regulating role in the answers (Crocker & Luhtanen, 1990). 
These results show that we have to differentiate between the need for 
individual and collective esteem. Moreover, they also suggest that the SIT 
statement referring to the tendency for discriminatory biases which favour 


222 
Social identity: international pers
pectives 
the ingroup with the aim of obtaining a positive social identity is valid only 
for those subjects with a stable tendency towards positively evaluating social 
identity or socially sharing a positive definition of collective self. 
Studies conducted on relative deprivation and political participation 
reinforce the idea that it is the threat to collective self-esteem, and not the 
lack of personal self-esteem, which is associated with the regulation of 
perception and social conduct by social identity. Let us remember that 
deprivation or personal dissatisfaction is not associated with socio-political 
participation behaviours, which we may view as examples of social conflict, 
but that in fact it is associated with anomie and individual affective 
alterations. It is sociocentric or fraternal relative deprivation (in relation to 
the group) which is linked to political participation (Guimond & Tougas, 
1 994). On the other hand, research on political participation shows that it is 
those subjects who are socially integrated, with a perception of political self­
efficiency, but who are also dissidents or critical of the social system which 
they view as the cause of the group's disadvantaged position, who mobilize 
themselves - and not those subjects who are most isolated, deprived or 
disadvantaged (Kinder & Sears, 1985). This is why Crocker, Luhtanen, 
Blaine and Broadnax (1994) have found that it is those subjects with high 
collective (socially shared beliefs or collective affective evaluation with 
regard to the group and not the person) self-esteem (those who are integrated 
and efficient) who show more ingroup bias in threatening conditions (in 
other words, of relative deprivation in which the group does not receive what 
it expected). As we can see, those subjects who regulate their perception and 
behaviour on the basis of their social identity and those who show ingroup 
favouritism and outgroup discrimination are the ones who have high col­
lective self-esteem in threatening situations. 
The differentiation between social and personal identity as well as the 
relative independence of public, private and personal collective self-esteem 
are two mechanisms which may explain why members of dominated and 
stigmatized classes do not show a deficit in personal self-esteem (Phinney, 
1 990). 
An explanatory process for the lack of deficit in personal and collective 
self-esteem in a dominated group included in an individualistic culture is the 
frequently found phenomenon of the acceptance of the disadvantaged 
nature of the group, while at the same time rejecting its individual dis­
advantage (Crosby, Cordova & Jaskar, 1994). For instance, Puerto Rican 
subjects accepted that Puerto Ricans are dependent, but they themselves are 
not like that. Personal self-valuation was mainly positive, while collective 
self-valuation was mainly negative (Rivera, 1991). As Rivera (1991) stated: 
The results of this first study were the pronounced and consistent divergence 
between the collective and personal valuation; the latter was invariably higher than 
the collective valuation. These results suggest that participants may have inter­
nalized, to some degree, the Puerto Ricans' negative stereotype, depersonalizing 
these categories in order to show that the rest of 
the Puerto Ricans are that way, but 
not themselves, who are attributed positive personal valuations. (pp. 10 1-102) 


Constructing social identity 
223 
Another explanatory process for the lack of low personal self-esteem in 
members of dominated groups is that subjects will not link other people's 
opinions about their group with their own personal self-esteem. Crocker et 
al. (1994) have found that the relationship between public (how the subject 
believes that other people evaluate hislher group) and private collective self­
esteem (how the subject evaluates hislher own group) is medium in 
individualistic cultures (Anglo-Americans), positive and strong in collecti­
vistic cultures (Asian-Americans) and null in dominated groups (African­
Americans). Let us add that private collective self-esteem is strongly linked 
in general with personal self-esteem. Nevertheless, in the sample of black 
subjects there was no relationship between public and private collective self­
esteem. These black subjects strongly differentiate how they believe other 
people value their group and how they themselves do so - and probably 
how they value themselves as individuals. In this way it is possible to cope 
with discrimination and prejudice separating collective negative judgements 
from one's image of oneself. On the other hand, for Asian subjects the 
vision the other person has is strongly associated with one's own view of the 
ingroup. This suggests that the relationship between personal self-esteem 
and group self-esteem is more underlined in collectivistic cultures. 
Self-consciousness, individualistic-collectivistic cultural values, social 
identity and behaviour regulation 
As Abrams (1990) has stated, sometimes behaviour is not regulated by 
social identity. Categorization in a group and salience of group membership 
will not be sufficient conditions for ingroup discrimination and its function 
of augmenting self-esteem. Moreover, a person must strongly identify with 
the group. Abrams (1990, 1 994) has shown in different studies that group 
identification and self-consciousness or focalizing attention on the private 
and public aspects of one's self converge in order to regulate behaviour. 
Subjects with strong identification and high private self-consciousness 
regulated behaviour by means of social identity. 
There is correlational evidence which confirms that high identification with 
a group is associated with a good evaluation of this same group. Garza and 
Herringer (1986) developed an instrument which measured important 
dimensions in group identity, the evaluation of the group, its emotional 
valuation and the stability of group identity. In a replication of this study, we 
applied the Spanish version of this scale (Paez, Arrospide, Martinez-Taboada 
& Ayestaran, 1 992). Our sample was composed of 120 subjects, 46% were 
men and 54% women with a mean age of 24.7 years. Forty-seven subjects 
were students who were starting their participation in a group dynamics 
course, and the rest of the sample were relatives of fifth-year psychology 
students enrolled in the Psychology Faculty of the Basque Country 
University. Students answered the questionnaires in class while the relatives 
were interviewed at home by the students as part of their class curricula. 


224 
Social identity: international perspectives 
Table 1 5.5 Factorial analysis o
f
 the 'most important social identity scale' 
(Garza & Herringer, 1986) 
Factor 
2 
3 
EV Good-bad 
.88 
EV Likes-dislikes 
.86 
EV Positive-negative 
.87 
EM Pleasant-unpleasant 
.83 
EM Comfortable-uncomfortable 
.80 
1M Important-unimportant 
.67 
1M Valid-invalid 
.80 
ST 
Permanent-impermanent 
.69 
ST 
Stable-unstable 
.75 
ST 
Secure-insecure 
.81 
ST 
Cooperative-competitive 
.64 
ST 
Dominant-dominated 
.81 
ST 
With power-without power 
.84 
EM Amusing-boring 
.70 
ST Abstract 
-concrete 
.77 
ST 
Private-public 
.67 
EV = Evaluative dimension; EM = Emotional dimension; 1M = Importance; ST = Stability 
We found three clearly differentiated factors in this scale (see Table 1 5.5). 
The first factor explaining 47% of the variance was saturated by the four 
dimensions mentioned by Garza and Herringer (1986): emotion, evaluation, 
importance and stability, and also by item number 1 3, which refers to 
competition-cooperation. The second factor was composed of items related 
to power and dominance, and they explained 9.4% of the variance. The 
third factor included those items referring to the abstract versus the concrete 
and the private versus the public. This final factor explained 8% of the 
variance. 
These results reaffirm that the evaluation of the group, the emotional 
response to the group, the importance of the group for identity and the 
stability of the identification with a group are congruently associated. With 
regard to the relationship between identification and stereotypes, we 
have found that in Latin America those subjects who strongly identified with 
their national groups (samples from Venezuela, Chile and Colombia) 
rejected sloth, while stressing happiness as a description of a Latin American 
person when answering closed items. With regard to open answers, these 
samples underlined characteristics associated with the Latin American's 
positive sociability: happy, hospitable. Moreover, they mentioned history 
and socio-political processes such as oppression and dependency as identity 
elements. Those subjects who shared a low identification mentioned more 
frequently deficitary instrumental attributes, such as sloth and irresponsi­
bility, and they also made more mention of underdevelopment, poverty and 
language as identity factors (Salazar, 1 989). In other words, those people 
with a strong identification show perceptive favouritism stressing the 
stereotype's positive attributes, relativizing the negative ones, while those 


Constructing social identity 
225 
subjects who did not identify share the stereotypical negative attributes of 
their social category. 
As a summary, we may state that group identification and focalizing 
attention will interact in order to regulate behaviour. It is easy to believe 
that it will be those individuals with a high private self-consciousness and 
collective identity who will be more stable in their beliefs and behaviours 
linked to obtaining group goals. They will also be more emotionally 
reactive to stimuli such as group threats. 
Of course not only situational factors or internal tendencies of group 
identity salience and attention directed towards aspects of collective identity 
influence behaviour regulation. Macrosocial phenomena such as collecti­
vistic orientation cultures reinforce the importance of collective or social 
identity, while individualistic cultures reinforce personal or individual 
identity, as we have already seen (Gudykunst, 1988; Smith & Bond, 1 993). 
Collectivistic cultures emphasize an intensive relationship with ingroup 
members, while in individualistic cultures people behave in a more extensive 
than intensive fashion - even with the ingroup members. Consequently, the 
difference between behaviour towards the outgroup and behaviour towards 
the ingroup is not so important (Triandis, 1994). There is partial evidence 
which confirms that individuals from collectivistic cultures are more sensible 
to outgroup and ingroup membership in the regulation of their behaviour 
(Triandis, 1994). People from collectivistic countries show a stronger ten­
dency towards cooperation, and people from individualistic-type countries 
tend more towards competition. Nevertheless, when they must interact with 
outgroup members, collectivistic persons become more competitive. For 
example, Argentinean and Indian children who had to share game rewards 
with subjects who did not belong to their group were more competitive than 
children from the United States (who in general are competitive) (Smith & 
Bond, 1 993). 
In the following experiment we empirically confirmed the importance of 
collectivistic values in behaviour and perception regulation. We asked 
students (N = 48) enrolled in a group dynamics class to perform a cognitive 
creativeness task: to list the attributes which characterize good leadership. 
All the subjects had to read a first set of attributes supposedly previously 
generated by a group. Depending on the condition, subjects were told that 
this list had been generated by psychologists (ingroup) or engineers 
(outgroup). This first list and the one they created after 10 minutes of 
individual reflection and another 10 minutes of group discussion were 
evaluated as: good-bad, complete-incomplete, unreal-realistic, and 
typical-not at all typical. The response scale ranged from I = the first 
attribute of the pair to 7 = the second attribute of the set. For instance I = 
good, 7 = bad. 
Differences in scores between their own list and the previous one were 
taken as an index of favouritism. The difference was established between the 
first score (the score given to the supposedly other group) and the second 
score (given to one's own group). The subtraction was performed in such a 


226 
Social identity: international perspectives 
way that a positive difference meant a better evaluation of one's own group. 
We subtracted from the scale with highest negative scores (bad 
= 7; 
incomplete = 7; not at all typical = 7) the score given to the first set (other 
group) from the score given to the second list (one's own group). For 
example, if the first list ( 
other group) is considered bad = 7 and the second 
list (own group) is good = 1 ,  the positive difference of +6 was a sign of this 
favouritism. When the scale had a higher score and also a positive 
evaluation (realistic = 7), the second score was subtracted (own group list) 
from the first list. If the ingroup list was awarded a 7 (very realistic) and the 
first list (other group) was awarded a 1 (not at all realistic), the subtraction 
(7 - 1) would yield a positive score which would reflect ingroup favouritism. 
In other words, a positive score always meant ingroup favouritism. 
The scores in Crocker and Luhtanen's collective self-esteem scale, Hui's 
collectivistic values and Carver, Antoni and Scheier's ( 1 985) self­
consciousness scale were used as covariables and their influence was 
controlled. The variance and covariance multivariate analysis showed a 
significant effect if the first comparison group was that of engineers or 
psychologists. On a univariate level the comparison with an ingroup or 
outgroup had a significant effect for the positive evaluation of the list (F(5, 
46) = 1 3. 1 ; p < .001) and near significance for the complete list (F(5, 46) = 
2.74; p < . 10). The other group's list was evaluated as better and more 
complete when it was attributed to psychologists (ingroup), and was 
evaluated as worse and less complete when it was attributed to the engineers 
(outgroup). 
The private and public self-consciousness variables, as well as the group 
identification variables, did not interact with the experimental manipula­
tion. Nevertheless the individualistic/collectivistic values did. Subjects with a 
high score in collectivism behaved in a different way depending on whether 
the first list was attributed to the outgroup or to the ingroup. The multi­
variate interaction effect was significant (F( l l ,  38) = 2.33; P < .026). The 
univariate contrasts indicate that those subjects who share more collec­
tivistic values view the first list as more believable (F(1, 48) = 4.88; P < .03), 
realistic (F(1, 48) = 6.66; p < .02), complete (F( I ,  48) = 5.75; p < .02), and 
less typical (F( l ,  48) = 1 1 .04; p < .002) when this list was attributed to 
psychologists. In other words, when the stimulus was associated with the 
general ingroup and subjects were collectivistic (with regard to their values) 
it was evaluated more positively and as more original. Exactly the opposite 
happened with subjects low on collectivism; these subjects viewed the list 
that was attributed to the outgroup (engineers) as less typical, more 
believable, realistic and complete (see Table 1 5.6). 
The lack of self-consciousness and identity effects may be explained in the 
first case because the scale measures the tendency to focus attention on the 
public and private aspects of the self - and not necessarily on the collective 
dimension of the self. In the second case, the scale measures the general 
tendency to identify with groups in general - and not with the group which 
was used in the manipulation (psychologists). Finally the low number of 


Constructing social identity 
Table 1 5.6 Dif
f
erences between engineers and psychologists as a 
f
unction o
f
 their degree of collectivism 
First list 
Engineers 
Psychologists 
Differences in 
Low 
High 
Low 
High 
evaluation 
collectivism 
collectivism 
collectivism 
collectivism 
Believable 
-1 .30 
-0. 12 
-0. 17 
-\ .00 
Incomplete 
\ .50 
0.25 
-0.80 
0.73 
Realistic 
-1 .50 
-0.63 
-0. 18 
-\ .80 
Not typical 
- 1 .30 
0.60 
0.08 
-1.\0 
227 
subjects does not allow us to perform a contrast simultaneously choosing 
subjects with high group identification and high private self-consciousness. 
The above-mentioned research suggests that those subjects who show 
more ingroup bias in conditions in which the intergroup comparison is 
salient are those who share more the collectivistic values. Moreover, an index 
of collectivistic values is associated with social identification, collective self­
esteem and public and private collective self-consciousness, confirming the 
relationship between these processes. 
Applying the private and public self-consciousness items to the attention 
on social identity's external and internal aspects, we have developed a 
collective self-consciousness scale which we validated by comparing it to 
Garza and Herringer's scale (Paez et aI., 1992). This scale was developed on 
the basis of the items extracted from the original private and public self­
consciousness scale - referring to the most important group a person 
belongs to. The items and instructions were as follows: Think of the most 
relevant group to which you belong (family, friends, and so on). (I) I always 
try to analyse myself in relation to my group's objectives; (2) I am interested 
in how I do things when I do them in my group's name; (3) I reflect a lot 
upon myself in relation to my group's values and beliefs; (4) In relation to 
my group's objectives I am very careful with the image I give to other 
people; (5) I am well aware of the image I give in relation to my group; (6) I 
constantly think of the reasons for my behaviour in relation to my group's 
objectives; (7) I normally think of myself as a representative of my group; 
(8) I am worried about the image my group gives; (9) When my group gets 
what it wants I am happy; (10) I dream about myself many times in relation 
to my group; (I I) I do not feel responsible or guilty if a member of my 
closest group (family) fails; (12) I do not feel responsible or guilty if a friend 
fails; (13) When my group does not achieve what it wants I get very 
annoyed; (14) When a member of my group achieves what s/he wants I feel 
well. Questions number 1 1  and 12 are to be inversely scored. The scale was 
passed with a score of I = totally disagree to 7 = total agreement. In our 
sample the reliability of our collective self-consciousness scale was .73. This 
collective self-consciousness scale had a satisfactory convergent validity - a 
correlation of .39, p < .01 with Hui's individualism-collectivism scale; .22, 


228 
Social identity: international pers
pectives 
p < .05 with Garza and Herringer's scale of group identification; and .21 , 
p < .05 with a scale of importance of group membership for one's self (Paez 
et ai., 1992). 
These results suggest that the chronic accessibility of the collective 
dimension of the self, or, in other words, the stable focalization of internal 
and external attention on the collective or group dimensions of identity, is 
coherently associated with a positive evaluation of collectivism, with a 
positive identification with the groups and with a structure of one's self 
based on one's group membership. 
Conclusions 
This chapter has examined the conditions and processes by which social 
identity regulates social perception and behaviour. We have emphasized the 
role of socio-structural variables and of socially shared beliefs in order to 
explain this process. This is why we underlined the importance of social 
status and the predominance of collectivistic cultural values as factors which 
will moderate the regulation of one's behaviour by social identity. 
We stressed the stronger influence which social identity has in collectivistic 
and sociocentric cultures in comparison to individualistic or egocentric 
cultures. We also showed the same pattern of results for dominated or low­
status groups. On the other hand, dominant classes and individualistic 
cultures reflect more the attention and valuation of a subject who possesses 
resources, augmenting the salience of personal over social identity. 
High-status groups will stress the personal aspects of identity, the per­
ception of ingroup individuation and dif
f
erentiation. Dominated groups will 
have a self-concept which will place more emphasis on group membership, 
stressing the perception of intragroup homogeneity - the same as subjects 
from collectivistic cultures. 
Groups in general view themselves as superior in comparison to other 
groups on those attributes which define them or which are typical of them, 
and tend to perceive outgroups as less variable or more homogeneous. 
Nevertheless, in threatening situations and in relation to defining or typical 
attributes of one's collective identity, exactly the opposite takes place: they 
are perceived as more homogeneous. 
Social status was another variable which moderated the relationship 
between social identity and ingroup favouritism. Subjects perceive their 
ingroup as superior to the outgroup on those attributes which define it, 
although they do not necessarily show general favouritism. This favouritism 
will be most clearly found among high-status groups. Although in general 
subjects represent their ingroup as more heterogeneous than the outgroup, 
at the same time they state that it is more homogeneous, or less variable, on 
those attributes which define it. Low-status or -prestige groups show out­
group f
avouritism or altercentrism, while at the same time they view them­
selves as superior on specific dimensions, especially sociability and 


Constructing social identity 
229 
expressiveness. Our own study on psychologists and engineers in Spain 
confirmed that the perceptive superiority o
f
 a lower status ingroup (psy­
chologists) was found in an expressive dimension (they viewed themselves as 
more friendly and sociable), accepting the fact that the high-status outgroup 
(engineers) was superior to them in more instrumental attributes. 
SIT states that group membership is a source of self-esteem, and that 
belonging to a disadvantaged group will motivate favourable social com­
parison and social competition. In a sense, dominated or low-status groups 
do not show a deficit in self-esteem. The differentiation between personal 
and social identity, as well as the relative independence of private, public 
and personal collective self-esteem, are two mechanisms which explain why 
the members of dominated and stigmatized classes do not show a deficit in 
personal self-esteem. 
Also, as we have already stated, the relationship between discrimination 
and the increase in self-esteem is not consistent. The difference between 
individual and collective self-esteem partly explains the lack of association 
between identification, discrimination and self-esteem. Studies conducted on 
collective self-esteem and on fratemalistic relative deprivation and political 
participation suggest that a threatened high collective self-esteem is one of 
the antecedents of a biased social perception and of discriminatory 
behaviour. 
Tendencies towards ingroup favouritism are clearer among those indi­
viduals who share a strong group identification, high collective self-esteem 
and collectivistic values. Those individuals who value belongingness to 
collectivities also think that their group membership is important and they 
positively value this membership for their identity. These are the people 
who have stronger reactions when their social identity is under threat; they 
increase intergroup differentiation, positively compare their ingroup to 
other groups, and increase the identification and attraction of their group. 
Our correlational studies confirmed that individuals with a strong group 
identification show high private and public group self-consciousness, more 
stability, a better valuation, a positive emotional reaction, and attach more 
importance to belonging to a group for their identity. 
Note 
We wish to thank John C. Turner for his comments on an early version of this chapter. 


References 
Abrams, D. (1990). How do group members regulate their behaviour? An integration of social 
identity and self-awareness theories. In D. Abrams & M.A. Hogg (Eds.), Social identity 
theory: Constructive and critical advances. Hemel Hempstead: Harvester Wheatsheaf. 
Abrams, D. (1994). Social self-regulation. Personality and Social Psychology Bulletin, 20, 473-
483. 
Abrams, D., & Hogg, M.A. (1988). Comments on the motivational status of self-esteem in 
social identity and intergroup discrimination. European Journal o
f
 Social Psychology, IS, 
3 17-334. 
Abrams, D., & Hogg, M.A. (1990). Social identification, self-categorization and social 
influence. In W. Stroebe & M. Hewstone (Eds.), European review o
f
 social psychology, Vol. 
1. Chichester: Wiley. 
Abrams, D., Wetherell, M.S., Cochrane, S., Hogg, M.A., & Turner, J.e. (1990). Knowing 
what to think by knowing who you are: Self-categorization and the nature of norm 
formation, conformity and group polarization. British Journal o
f
 Social Psychology, 29, 97-
1 19. 
Abric, J.e. (1994). Pratiques sociales et representations. Paris: Presses Universitaires de France. 
Adamopoulos, J. (1982). Analysis of interpersonal structures in literary works of three 
historical periods. Journal o
f
 Cross-Cultural Psychology, 13, 1 57-168. 
Adamopoulos, J., & Bontempo, R.N. (1986). Diachronic universals in interpersonal structures. 
Journal o
f
 Cross-Cultural Psychology, 1 7, 169-189. 
Aebischer, V., Hewstone, M., & Henderson, M. (1984). Minority influence and musical 
preference: Innovation by conversion not coercion. European Journal o
f
 Social Psychology, 
14, 23-33. 
Allen, V.L., & Wilder, D.A. (1975). Categorization, belief similarity, and intergroup 
discrimination. Journal o
f
 Personality and Social Psychology, 32, 971-977. 
Allen, V.L., & Wilder, D.A. (1979). Group categorization and attribution of belief similarity. 
Small Group Behaviour, 10, 73-80. 
Allport, F.H. (1924). Social psychology. Boston: Houghton Mifflin. 
Allport, G.W. (1961). The ego in contemporary psychology. Psychological Review, 50, 471-
478. 
Allport, G.W., & Odbert, H.S. (1936). Trait-names: A psycho-lexical study. Psychological 
Monographs, 47. 
Altman, I., & Low, S.M. (Eds.). (1992). Place attachment. New York: Plenum. 
Amancio, L. (1989). Social differentiation between 'dominant' and 'dominated' groups: Toward 
an integration of social stereotypes and social identity. European Journal o
f
 Social 
Psychology, 19, 1-10. 
Anderson, B. (1983). Imagined communities: Re
flections on the origin and spread o
f
 nationalism. 
London: Verso. 
Apter, D.E. (1964). Ideology and discontent. In D.E. Apter (Ed.), Ideology and discontent. 
London Collier Macmillan. 
Ardener, S. (1983). Arson, nudity and bombs among the Canadian Doukhobors: A question of 
identity. In G.M. Breakwell (Ed.), Threatened identities. Chichester: Wiley. 
Ardrey, R. (1967). T
he territorial imperative. London: Collins. 
Argyle, M. (1992). The social psychology o
f
 everyday lif
e. London: Routledge. 
Asch, S.E. (1952). Social psychology. Englewood Cliffs, NJ: Prentice Hall. 


Re
f
erences 
231 
Asch, S.E. (1956). Studies on independence and conformity: A minority of one against a 
unanimous majority. Psychological Monographs, 70, 456. 
Asch, S.E., & Zuckier, H. (1984). Thinking about persons. Journal o
f
 Personality and Social 
Psychology, 46, 1230- 1240. 
Ayesteran, S. (Ed.). (1985). Psicologia y enf
ermedad mental: ldeologia y representacion social de 
la enf
ermedad mental. Bilbao: Universidad del Pais Vasco. 
Bakan, D. (1966). The duality o
f
 human existence. Boston: Beacon. 
Ball, P., Giles, H., & Hewstone, M. (1984). Secondary language acquisition: The intergroup 
theory with catastrophic dimensions. In H. Tajfel (Ed.), The social dimension: European 
developments in social psychology, Vol. 2. Cambridge: Cambridge University Press. 
Barsalou, L.W. (1983). Ad hoc categories. Memory and Cognition, 11, 2 1 1 -227. 
Barsalou, L. 
W. (1987). The instability of graded structure: Implications for the nature of 
concepts. In U. Neisser (Ed.), Concepts and conceptual development: Ecological and 
intellectual f
actors in categorization. Cambridge: Cambridge University Press. 
Barsalou, L.W. (1989). Intraconcept similarity and its implications for interconcept similarity. 
In S. Vosniadou & A. Ortony (Eds.), Similarity and analogical reasoning. Cambridge: 
Cambridge University Press. 
Barsalou, L.W., & Sewell, D.R. (1984). Constructing representations o
f
 categories f
rom diff
erent 
points o
f
 view (Emory Cognition Projects Report No. 2). Atlanta: Emory University. 
Barsalou, L.W., Sewell, D.R., & Ballato, S.M. (1986). Assessing the stability o
f
 category 
representations with graded structure. Unpublished manuscript. 
Bar-Tal, D. (1986). The Masada Syndrome: A case of central belief. In N. Milgram (Ed.), 
Stress and coping in time o
f
 war. New York: Brunner/Maze!' 
Bar-Tal, D. (1990). Group belie
f
s: A conce
ption f
or analyzing group structure, processes, and 
behavior. New York: Springer Verlag. 
Bar-Tal, D. (1993). American convictions about conflictive USA-USSR relations: A case of 
group beliefs. In S. Worchel & 1. Simpson (Eds.), Conflict between people and peoples. 
Chicago: Nelson-Hall. 
Bar-Tal, D. (in press). Societal belie
f
s
 o
f
 ethos: A social psychological analysis o
f
 society. 
London: Sage. 
Bar-Tal, D., Raviv, A., & Brosh, M. (1991). Perception of epistemic authority and attribution 
for the choice as a function of knowledge area and age. European Journal o
f
 Social 
Psychology, 21, 477-492. 
Bar-Tal, Y. (1989). Can leaders change followers' stereotypes? In D. Bar-Tal, C. Grauman, 
A.W. Kruglanski & W. Stroebe (Eds.), Stereotyping and pre
judice: Changing conceptions. 
New York: Springer Verlag. 
Bellelli, G. (Ed.). (1 987). La representation sociale de la maladie mentale. Naples: Liguori. 
Bellelli, G. (Ed.). (1994). L'altra malattia: Come la societa pensa la malattia mentale. Naples: 
Liguori. 
Bern, D.J. (1967) Self perception: An alternative interpretation of cognitive dissonance 
phenomena. Psychological Review, 74, 1 83-200. 
Bern, D.l. (1970). Belie
f
s, attitudes, and human a
f
f
airs. Belmont, CA: Brooks/Cole. 
Bern, S.L. (1981). Gender schema theory: A cognitive account of sex typing. Psychological 
Review, 88, 354-364. 
Bern, S.L. (1985). Androgyny and gender schema theory: A conceptual and empirical 
integration. In T.B. Soderberegger (Ed.), Psychology and gender: Nebraska Symposium on 
Motivation, Vol. 32. Lincoln: University of Nebraska Press. 
Billig, M. (1985). Prejudice, categorization and particularization: From a perceptual to a 
rhetorical approach. European Journal o
f
 Social Psychology, 15, 79-103. 
Billig, M., & Tajfel, H. (1973). Social categorization and similarity in intergroup behaviour. 
European Journal o
f
 Social Psychology, 3, 27-52. 
Boltanksi, 1.L. (1982). Les cadres: La f
ormation d'un groupe social. Paris: Minuit. 
Bomstein, G., Crum, L., Wittenbraker, 1., Harring, K., Insko, C.A., & Thibaut, 1. (l983a). On 
the measurement of social orientation in the minimal group paradigm. European Journal o
f
 
Social Psychology, 13, 321 -350. 


232 
Social identity: international perspectives 
Bomstein, G., Crum, L., Wittenbraker, 1., Harring, K., Insko, C.A., & Thibaut, 1. (1983b). 
Reply to Turner's comments. European Journal o
f
 Social Psychology, 13, 369-381 .  
Bourhis, R.Y., Gagnon, A., & MOise, L.C. (1 994). Discrimination et relation intergroupe. In 
R.Y. Bourhis & l.-P. Leyens (Eds.), Stereotypes, discrimination et relations intergroupes. 
Brussels: Mardaga. 
Brewer, M.B. (1979). Ingroup bias in the minimal intergroup situation: A cognitive­
motivational analysis. Psychological Bulletin, 86, 307-324. 
Brewer, M.B. (1988). A dual process model of impression formation. In T.K. Srull & R.S. 
Wyer (Eds.), Advances in social cognition, Vol. 1. Hillsdale, Nl: Erlbaum. 
Brewer, M.B. (1991). The social self: On being the same and different at the same time. 
Personality and Social Psychology Bulletin, 1 7, 475-482. 
Brewer, M.B., Dull, Y., & Lui, L. (1981). Perceptions of the elderly: Stereotypes as prototypes. 
Journal o
f
 Personality and Social Psychology, 41, 656-670. 
Brewer, M.B., Ho, H., Lee, 1., & Miller, N. (1987). Social identity and social distance among 
Hong Kong school children. Personality and Social Psychology Bulletin, 13, 1 56-165. 
Brewer, M.B., & Lui, L. (1984). Categorization of the elderly by the elderly: Effects of 
perceiver's category membership. Personality and Social Psychology Bulletin, 10, 585-595. 
Brewer, M.B., & Silver, M. (1978). Ingroup bias as a function of task characteristics. European 
Journal o
f
 Social Psychology, 8, 393-400. 
Brown, R. (1954). Mass phenomena. In G. Lindzey (Ed.), Handbook o
f
 social psychology, Vol. 
2. Cambridge, MA: Addison-Wesley. 
Brown, R.l. (1988). Group processes: Dynamics within and between groups. Oxford: Blackwell. 
Brown, R.l., Hinkle, S., Ely, P.G., Fox-Cardamone, L., Maras, P., & Taylor, L.A. (1992). 
Recognising group diversity: Individualist-collectivist and autonomous-relational social 
orientations and their implications for intergroup processes. British Journal o
f
 Social 
Psychology, 31, 327-342. 
Brown, R.1., & Ross, G.F. (1982). The battle for acceptance: An investigation into the 
dynamics of intergroup behaviour. In H. Tajfel (Ed.), Social identity and intergrou
p 
relations. Cambridge: Cambridge University Press. 
Brown, R.l., & Turner, 1.C. (1979). The criss-cross categorization effect in intergroup 
discrimination. British Journal o
f
 Social and Clinical Psychology, 18, 371 -383. 
Brown, R.l., & Turner, 1.C. (1981). Interpersonal and intergroup behaviour. In 1.C. Turner & 
H. Giles (Eds.), I
ntergroup behaviour. Oxford: Blackwell. 
Bruner, 1.S. (1957). On perceptual readiness. Psychological Review, 64, 123-151. 
Burnstein, E. (1982). Persuasion as argument processing. In H. Brandstatter, 1. Davis & G. 
Stocker-Kreichgauer (Eds.), Group decision-making. London: Academic Press. 
Butera, F., Huguet, P., Mugny, G., & Perez, 1.A. (1994). Socio-epistemic conflict and 
constructivism. Swiss Journal o
f
 Psychology, 53, 229-239. 
Butera, F., Legrenzi, P., Mugny, G., & Perez, 1.A. (1990-1991). Influence sociale et 
raisonnement. Bulletin de Psycho 
logie, 45, 144-154. 
Byrne, D. (1969). Attitudes and attraction. In L. Berkowitz (Ed.), Advances in ex
perimental 
social psychology, Vol. 4. New York: Academic Press. 
Byrne, D. (1971). The attraction problem. New York: Academic Press. 
Caddick, B. (1982). Perceived illegitimacy and intergroup relations. In H. Tajfel (Ed.), Social 
identity and intergroup relations. Cambridge: Cambridge University Press. 
Campbell, D.T. (1958). Common fate, similarity, and other indices of the status of aggregates 
of persons as social entities. Behavioral Science, 3, 14-25. 
Campbell, D.T. (1965). Ethnocentric and other altruistic motives. In D. Levine (Ed.), Nebraska 
Symposium on Motivation, Vol. 13. Lincoln: University of Nebraska Press. 
Camps, Y. (1993). Parado
jas del individualismo. Barcelona: Critica. 
Cannon, T. (1973). Black Panther party program: What we want, and what we believe. In R.R. 
Evans (Ed.), Social Movements. Chicago: Rand McNally. 
Cantor, N., & Mischel, W. (1979). Prototypes in person perception. In L. Berkowitz (Ed.), 
Advances in experimental social psychology, Vol 12. New York: Academic Press. 
Capozza, D., Bonaldo, E., & Di Maggio, A. (1982). Problems of identity and social conflict: 


Re
f
erences 
233 
Research on ethnic groups in Italy. In H. Tajfel (Ed.), Social identity and intergroup 
relations. Cambridge: Cambridge University Press. 
Carroll, 0.1., & Chang, J.J. (1970). Analysis of individual differences in multidimensional 
scaling via N-way generalisation of 'Eckart-Young' decomposition. Psychometrika, 35, 
283-319. 
Cartwright, D. (1968). The nature of group cohesiveness. In D. Cartwright & A. Zander (Eds.), 
Group dynamics (3rd ed.). London: Tavistock; New York: Harper & Row. 
Cartwright, D., & Zander, A. (Eds.). (1968). Group dynamics (3rd ed.). London: Tavistock; 
New York: Harper & Row. 
Carver, c., Antoni, M., & Scheier, M. (1985). Self-consciousness and self-assessment. Journal 
o
f
 Personality and Social Psychology, 48, 1 1 7-124. 
Clark, R.D., III, & Maass, A. (1988a). Social categorization in minority influence: The case of 
homosexuality. European Journal o
f
 Social Psychology, 18, 347-364. 
Clark, R.D., HI, & Maass, A. (I 
988b). The role of social categorization and perceived source 
credibility in minority influence. European Journal o
f
 Social Psychology, 18, 381 -394. 
Codol, J.-P. (1975). On the so-called 'superior conformity of the self' behaviour: Twenty 
experimental investigations. European Journal o
f
 Social Psychology, 5, 457-501 .  
Codol, J.-P. (1979). Semblables et difJerents: Recherches sur la quete de la similitude et de la 
difJ
erenciation sociale. Thesis, Universite de Provence, Aix-en-Provence. 
Codol, J.-P. (1982). Diff
erentiating and non-differentiating behavior: A cognitive approach to 
the sense of identity. In J.-P. Codol & J.-P. Leyens (Eds.), Cognitive analysis o
f
 social 
behavior. The Hague: Martinus Nijhoff. 
Codol, J.-P. (1984a). On the system of representations in an artificial social situation. In R.M. 
Farr & S. Moscovici (Eds.), Social representations. Cambridge: Cambridge University Press. 
Codol, J.-P. (1 984b). La perception de la similitude interpersonnelle: Influence de 
I'appartenance categorielle et du point de ref
erence de la comparaison. L'Annee 
Psychologique, 84, 43-56. 
Codol, J.-P. (1987). Comparability and incomparability between oneself and others: Means of 
differentiation and comparison ref
erence point. CaMers de Psychologie Cognitive, 7, 1- 18. 
Codol, J.-P. (1990). Studies on self-centered assimilation processes. In J.P. Caverni, J.M. Fabre 
& M. Gonzalez (Eds.), Cognitive biases. Amsterdam: North Holland. 
Condor, S. (1989). 'Biting into the future': Social change and the social identity of women. In S. 
Skevington & D. Baker (Eds.), The social identity o
f
 women. London: Sage. 
Condor, S. (1 990). Social stereotypes and social identity. In D. Abrams & M.A. Hogg (Eds.), 
Social identity theory: Constructive and critical advances. Hemel Hempstead: Harvester 
Wheatsheaf. 
Condor, S., Brown, R.1., & Williams, J. (1987). Social identification and intergroup behaviour. 
Quarterly Journal o
f
 Social Science Af
f
airs, 3, 299-317. 
Converse, P.E. (1964). The nature of the belief systems in mass pUblics. In D.E. Apter (Ed.), 
Ideology and discontent. New York: Free Press. 
Coutant, D., Worchel, S., & Grossman, M. (in preparation). Minority influence over time. 
Coutant-Sassic, D. (1991). DifJ
erential cognitive processing o
f
 a ma
jority vs. a minority in
fluence 
on the same or difJ
erent task. Unpublished master's thesis, Texas A&M University. 
Cowan, G., & Hoffman, C.D. (1986). Gender stereotyping in young children: Evidence to 
support a concept-learning approach. Sex Roles, 14, 2 1 1 -224. 
Crano, W.O. (1989). Judgmental sub
jectivity, objectivity, and minority influence. Paper 
presented to the third workshop on minority influence, Perugia, Italy, June. 
Crocker, J., & Luhtanen, R. (1990). Collective self-esteem and ingroup bias. Journal o
f
 
Personality and Social Psychology, 58, 323-338. 
Crocker, J., Luhtanen, R., Blaine, B., & Broadnax, S. (1 994). Collective self-esteem and 
psychological well-being among white, Black and Asian college students. Personality and 
Social Psychology Bulletin, 20, 503-51 3. 
Crosby, F., Cordova, D., & Jaskar, K. (1994). On the failure to see oneself as disadvantaged: 
Cognitive and emotional components. In M.A. Hogg & D. Abrams (Eds.), Motivation: 
Social psychological perspectives. Hemel Hempstead: Harvester Wheatsheaf. 


234 
Social identity: international pers
pectives 
Cylwik, A. (1992). Przychylnosc wobec innych narodowosci i mniejszosci narodowych a schematy 
la/My. Unpublished master's thesis, Wydzial Psychologii, Uniwersytet Warzawski, Warsaw. 
Damon, W., & Hart, D. (1982). The development of self-understanding from infancy through 
adolescence. Child Development, 53, 841-864. 
Deaux, K. (1992). Personalizing identity and socializing self. In G. Breakwell (Ed.), Social 
psychology o
f
 identity and the self
-conce
pt. London: Academic Press. 
Deaux, K. (1993). Reconstructing social identity. Personality and Social Psychology Bulletin, 
19, 4-12. 
Deconchy, J.P. (1984). Rationality and social control in orthodox systems. In H. Tajfel (Ed.), 
The social dimension: European developments in social psychology, Vol. 1. Cambridge: 
Cambridge University Press. 
Delumeau, J.J. (1993). La peur en Occident. Paris: Hachette: Pluriel. 
Deschamps, J.-C. (1972-1973). Imputation de la responsabilite de I'echec (ou de la reussite) et 
categorisation sociale. Bulletin de Psychologie, 26, 794-806. 
Deschamps, J.-C. (1977). L'attribution et la categorisation sociale. Berne: Peter Lang. 
Deschamps, J.-C. (1979). Diff
erenciation categorielle et differenciation de soi par rapport a 
autrui. Recherches de Psychologie Sociale, 1, 29-38. 
Deschamps, J.-c. (1980). L'identite sociale et les rapports de domination. Revue Suisse de 
Sociologie, 1, 1 1 1-122. 
Deschamps, J.-c. ( l982a). Differenciation entre soi et autrui et entre groupes: Recherches sur la 
'covariation' entre les differenciations inter-individuelles et inter-groupes. In J.-P. Codol & 
J.-P. Leyens (Eds.), Cognitive analysis o
f
 social behavior. The Hague: Martinus Nijhoff. 
Deschamps, J.-C. (1982b). Social identity and relations of power between groups. In H. Tajfe\ 
(Ed.), Social identity and intergroup relations. Cambridge: Cambridge University Press. 
Deschamps, J .-c. (1983-1984). Categorisation sociale et differenciations interindividuelles. 
Bulletin de Psycho 
logie, 37, 489-500. 
Deschamps, J.-c. (1984). Identite socia 
Ie et differenciation categorielle. Cahiers de Psychologie 
Cognitive, 4, 449-474. 
Deschamps, J.-C. (1991). Identites, appartenances sociales et differenciations individuelles. 
Cahiers Internationaux de Psychologie Sociale, 9-10, 49-6\. 
Deschamps, J.C., Clemence, A., & Roux, P. (1983). Differ
enciations inter-groupes et inter­
individuelles. Paper presented at Congreso della Divisione di Psicologia della SIPS, Padua. 
Deschamps, J.-C., & Doise, W. (1978). Crossed category memberships in intergroup relations. 
In H. Tajfel (Ed.), Diff
erentiation between social groups: Studies in the social psychology o
f
 
intergroup relations. London: Academic Press. 
Deschamps, J.-c., & Lorenzi-Cioldi, F. (1981). 'Egocentrisme' et 'sociocentrisme' dans les 
relations entre groupes: Recherche experimentale. Revue Suisse de Psychologie Pure et 
Applique, 40, 108-131. 
Deschamps, J.-c., & Serino, C. (1996). Mondo al singolare: Pros
pettive socio-cognitive 
sull'identita. Naples: Liguori. 
Deschamps, J.-c., & Volpato, C. (1984). Identita individuale nelle relazioni tra gruppi. Giornale 
Italiano di Psicologia, 11, 275-301 .  
Deutsch, M. (1968). The eff
ects of cooperation and competition upon group process. I n  D. 
Cartwright & A. Zander (Eds.), Group d
ynamics (3rd ed.). London: Tavistock; New York: 
Harper & Row. 
Deutsch, M., & Gerard, H.B. (1955). A study of normative and informational social influence 
upon individual judgment. lournal o
f
 Abnormal and Social Psychology, 51, 629-636. 
Devos, T., Comby, L., & Deschamps, J.-c. (\995). Appartenances sociales et jugement de 
variabilite intra-groupe. Cahiers Internationaux de Psychologie Sociale, 27, 9-43. 
Diamond, M. (1976). The American idea of man: The view from the founding. In T. Kristol & 
P. Weaver (Eds.), The Americans: 1976, Vol. 2. Lexington, MA: Lexington Books. 
Diehl, M. (\990). The minimal group paradigm: Theoretical explanations and empirical 
findings. In W. Stroebe & M. Hewstone (Eds.), European review o
f
 social 
psychology, Vol. 1. 
Chichester: Wiley. 


Re
f
erences 
235 
Doise, W. (1969). Strategies de 
jeu a I'interieur et entre des groupes de nationalites diff
erentes. 
Bulletin du Centre d'Etudes et Recherches Psycho 
techniques, 18, 1 3-26. 
Doise, W. (1976). L'articulation psychosociologique et les relations entre groupes. Brussels: De 
Boeck. 
Doise, W. (1986). Les representations sociales: Definition d'un concept. In W. Doise & A. 
Palmonari (Eds.), L'erude des representations sociales. Paris: Delachaux et NiestIe. 
Doise, W. (1988). Individual and social identities in intergroup relations. European Journal o
f
 
Social Psychology, /8, 99-1 1 \ .  
Doise, W., Clemence, A., & Lorenzi-Cioldi, F .  (1993). The quantitative anal
ysis o
f
 social 
representations. Hemel Hempstead: Harvester Wheatsheaf. 
Doise, W., Csepeli, G., Dann, H.D., Gouge, C., Larsen, K., & Ostell, A. (1972). An experi­
mental investigation into the formation of intergroup representations. European Journal o
f
 
Social Psychology, 2, 202-204. 
Doise, W., Deschamps, J.-c., & Meyer, G. (1 978). The accentuation of intracategory 
similarities. In H. Tajfel (Ed.), Diff
erentiation between social groups: Studies in the social 
psychology o
f
 intergroup relations. London: Academic Press. 
Doise, W., Gachoud, J.P., & Mugny, G. (1986). Influence directe et indirecte entre groupes 
dans des choix esthetiques. Cahiers de Psychologie Cognitive, 6, 283-30\ .  
Doise, W., & Lorenzi-Cioldi, F .  (1990). Levels of analysis and social identity. In D .  Abrams & 
M.A. Hogg (Eds.), Social identity theory: Constructive and critical advances. Hemel 
Hempstead: Harvester Wheatsheaf. 
Doise, W., & Lorenzi-Cioldi, F. (1991). L'identite comme representation sociale. In V. 
Aebischer, J.P. Deconchy & E.P. Lipiansky (Eds.), Ideologie et representations sociales. 
Cousset: Delval. 
Doise, W., & Palmonari, A. (1986). L'etude des representations sociales. Paris: Delachaux et 
Niestle. 
Doob, L.W. (1964). Patriotism and nationalism. New Haven, CT: Yale University Press. 
Doosje, B., Ellemers, N. & Spears, R. (1996). Perceived intragroup variability as a function of 
group status and identification. Journal o
f
 Experimental Social Psychology, 3/, 410-436. 
Dumont, L. (1987). Ensayos sobre el individualismo: Una pers
pectiva antropologica sobre la 
ideologia mode 
rna. Madrid: Alianza. 
Dunbar, R. (1992). Why gossip is good for you. New Scientist, 21 November, 28-3\ .  
Dunning, D., Perie, M., & Story, A.L. (1991). Self-serving prototypes of social categories. 
Journal o
f
 Personality and Social Psychology, 6/, 957-968. 
Durand Delvigne, A. (\ 992). Identite et modides sexues des personnes. These de Doctorat d'Etat 
de Lettres et Sciences H 
umaines, Paris. 
Durkheim, E. (1898). Representations individuelles et representations collectives. Revue de 
Metaphysique, 6, 274-302. 
Duveen, G., & Lloyd, B. (1990). Social representations and the development o
f
 knowledge. 
Cambridge: Cambridge University Press. 
Eagly, A.H., & Chaiken, S. (1993). The psychology o
f
 attitudes. New York: Harcourt Brace 
Jovanovich. 
Echebarria, A., Elajabarrieta, F., Valencia, J., & Villareal, M. (1992). Representations sociales 
de l'Europe et identites sociales. Bulletin de Psychologie, 45, 280-288. 
Eiser, J.R. (1983). Psicologia sociale cognitivista. Bologne: II Mulino. (Original English 
publication: Maidenhead: McGraw-Hill, 1980.) 
Ellemers, N., Wilke, H., & van Knippenberg, A. (1993). Effects of the legitimacy of low group 
or individual status on individual and collective status-enhancement strategies. Journal o
f
 
Personality and Social Psychology, 64, 766-778. 
Emler, N., & Hopkins, N. (1990). Reputation, social identity and the self. In D. Abrams & 
M.A. Hogg (Eds.), Social identity theory: Constructive and critical advances. Hemel 
Hempstead: Harvester Wheatsheaf. 
Emler, N., & Reicher, S. (1995). Adolescence and delinquency. Oxford: Blackwell. 
Epstein, A.L. (1978). Ethos and identity. London: Tavistock. 
Epstein, S. (\ 980). The self-concept: A review and the proposal of an integrated theory of 


236 
Social identity: international pers
pectives 
personality. In E. Staub (Ed.), Personality: Basic issues and current research. Englewood 
Cliffs, NJ: Prentice Hall. 
Erikson, E. (1956). The problem of ego identity. Journal o
f
 the American Psychiatric 
Association, 1, 58-121. 
Erikson, E. (1977). Childhood and society. London: Paladin. 
Etaugh, C, & Riley, S. (1983). Evaluating competence of women and men: Effects of marital 
and parental status and occupational sex-typing. Sex Roles, 9, 943-952. 
Etzioni, A. (1975). Complex organizations (rev. ed.). New York: Free Press. 
Falomir, J.M., Mugny, G., & Perez, I.A. (1996). Le paradoxe du denio In J.-C Deschamps & 
J.-L. Beauvois (Eds.), Des attitudes aux attributions. Grenoble: Presses Universitaires de 
Grenoble. 
Farr, R.M., & Moscovici, S. (1984). Social representations. Cambridge: Cambridge University 
Press. 
Fells, J., & Niznik, J. (1992). Conclusion: What is Europe? International Journal o
f
 Sociology, 
26, 201 -206. 
Festinger, L. (1950). Informal social communication. Psychological Review, 57, 271-282. 
Festinger, L. (1954). A theory of social comparison processes. Human Relations, 7, 1 17-140. 
Festinger, L., Riecken, H.W., & Schachter, S. (1956). W
hen prophecy f
ails. Minneapolis: 
University of Minnesota Press. 
Festinger, L., & Thibaut, J. (1951). Interpersonal communication in small groups. Journal o
f
 
Abnormal and Social Psychology, 46, 92-99. 
Festinger, L., Schachter, S., & Back, K. (1950). Social pressures in inf
ormal groups: A stud
y o
f
 
human f
actors i
n
 housing. New York: Harper. 
Fiske, S.T., & Neuberg, S.L. (1990). A continuum of impression formation, from category­
based to individuating processes: Influences of information and motivation on attention and 
interpretation. In M.P. Zanna (Ed.), Advances in ex
perimental social psychology, Vol. 23. 
New York: Academic Press. 
Fiske, S.T., & Pavelchak, M.A. (1986). Category-based versus piecemeal-based aff
ective 
responses: Developments in schema-triggered affect. In R.M. Sorrentino & E.T. Higgins 
(Eds.), Handbook o
f
 motivation and cognition: Foundations o
f
 social behavior, Vol. 1. New 
York: Guilford. 
Forgas, J.P. (1979). Social episodes: T
he stud
y o
f
 interaction routines. New York: Academic 
Press. 
Forsyth, D.R. (\995). Our social world. Pacific Grove, CA: Brooks/Cole. 
Fraser, S.L., & Lord, R.G. (l988). Stimulus prototypicality and general leadership impressions: 
Their role in leadership and behavioural ratings. Journal o
f
 Psychology, 122, 291-303. 
Gaertner, S.L., Dovidio, J.F., Anastasio, P.A., Bachman, B.A., & Rust, M.C (1993). The 
common ingroup identity model: Recategorization and the reduction of intergroup bias. In 
W. Stroebe & M. Hewstone (Eds.), European review o
f
 social psychology, Vol. 4. Chichester: 
Wiley. 
Gaertner, S.L., Mann, J., Murrell, A., & Dovidio, J.F. (\989). Reducing intergroup bias: The 
benefits of recategorization. Journal o
f
 Personality and Social Psychology, 57, 239-249. 
Garza, R.T., & Herringer, L.G. (\986). Social identity: A multidimensional approach. Journal 
o
f
 Social Psychology, 127, 299-308. 
Gergen, K.J. (1965). Interaction goals and personalistic feedback as factors affecting the 
presentation of the self. Journal o
f
 Personality and Social Psychology, 1, 413-424. 
Gergen, KJ. (1982). From self to science: What is there to know? In I.M. Suls (Ed.), 
Psychological pers
pectives on the self. Hillsdale, Nl: Erlbaum. 
Gillin, J. (\ 955). National and regional cultural values in the United States. Social Forces, 34, 
107- 1 1 3. 
Goethals, G.R. (1972). Consensus and modality in the attribution process: The role of 
similarity and information. Journal o
f
 Personality and Social Psychology, 21, 84-92. 
Goffman, E. (I 956}. The presentation o
f
 self in everyday lif
e. Edinburgh: Social Science 
Research Centre - University of Edinburgh. 
Goffman, E. (\ 968). Stigma. Englewood, Cliffs: Prentice Hall. 


Re
f
erences 
237 
Gol̗b, A., & Reykowski, 1. (1985). Studia nad rozwojem standardow ewaluatywnych. Wroclaw: 
Ossolineum. 
Goldberg, L.R. (1990). An alternative 'Description of Personality': The big-five factor 
structure. Journal o
f
 Personality and Social Psychology, 59, 1216-1229. 
Gottman, 1.M. (1981). Time series analysis: A comprehensive introduction f
or social scientists. 
Cambridge: Cambridge University Press. 
Greenwald, A.G, & Banaji, M. (1995). Implicit social cognition: Attitudes, self-esteem and 
stereotypes. Psychological Review, 12, 1-19. 
Greenwald, A.G., & Pratkanis, A.R. (1984). The self. In R.S. Wyer & T.K. Srull (Eds.), 
Handbook o
f
 social cognition, Vol. 1. Hillsdale, NJ: Erlbaum. 
Gudkyunst, W. (1988). Culture and intergroup processes. In M. Bond (Ed.), T
he cross-cultural 
challenge to social psychology. Newbury Park, CA: Sage. 
Guimond, S., & Tougas, F. (1994). Sentiment d'injustice et action collective: La privation 
relative. In R.Y. Bourhis & J.-P. Leyens (Eds.), Stert!otypes, discrimination et relations 
intergroupes. Brussels: Mardaga. 
Hamilton, V.L., & Rauma, D. (1995). Social psychology of deviance and law. In K.S. Cook, 
G.A. Fine & 1.S. House (Eds.), Sociological pers
pectives on social psychology. Boston: Allyn 
& Bacon. 
Hardoin, M., & Codol, J.-P. (1984). Description de soi et d'autrui: Influence de l'ordre de 
description sur les categories de reponse utilisees. Cahiers de Psychologie Cognitive, 4, 295-
302. 
Harkins, S., & Petty, R.E. (1981). The multiple source effect in persuasion: The effects of 
distraction. Personality and Social Psychology Bulletin, 21, 84-92. 
Harkins, S., & Petty, R.E. (1983). Social context effects in persuasion: The effects of multiple 
source and multiple targets. In P.B. Paulus (Ed.), Basic group process. New York: Springer 
Verlag. 
Harkins, S., & Petty, R.E. (1987). Inf
ormational utility and the multiple-source effect. Journal 
o
f
 Personality and Social Psychology, 52, 260-268. 
Harrison, A.A., Moore, M., & Rucker, M.H. (1985). Further evidence on career and family 
compatibility among eminent women and men. Archivio di Psicologia, Neurologia e 
Psichiatria, 46, 140-1 55. 
Harter, S. (1985). The self-perception pro
file f
or children: Revision o
f
 the Perceived Competence 
Scale f
or children. Manual, University of Denver. 
Haslam, S.A., McGarty, C, Oakes, P.J., Turner, J.C, & Onorato, R. (1995). Contextual shifts 
in the prototypicality of extreme and moderate outgroup members. European Journal o
f
 
Social Psychology, 25, 509-530. 
Haslam, S.A., Oakes, P.J., Turner, J.C, & McGarty, C (1996). Social identity, self­
categorization and the perceived homogeneity of in 
groups and outgroups: The interaction 
between social motivation and cognition. In R.M. Sorrentino & E.T. Higgins (Eds.), 
Handbook o
f
 motivation and cognition, Vol. 3. New York: Guilf
ord. 
Haslam, S.A., & Turner, 1.C (1992). Context-dependent variation in social stereotyping 2: The 
relationship between frame of ref
erence, self-categorization and accentuation. European 
Journal of Social Psychology, 22, 251-277. 
Haslam, S.A., Turner, J.C, Oakes, PJ., McGarty, C., & Hayes, B.K. (1992). Context­
dependent variation in stereotyping I :  The effects of intergroup relations as mediated by 
social change and frame of reference. European Journal of Social Psychology, 22, 3-20. 
Hayes, CH. (1926). Nationalism as a religion. In CH. Hayes, Essays on nationalism. New 
York: Macmillan, 1962. 
Hertz, F. (1944). Nationalism in history and politics. London: Routledge & Kegan Paul. 
Hewstone, M., Hantzi, A., & lohnston, L. (1991). Social categorization and person memory: 
The pervasiveness of race as an organizing principle. European Journal o
f
 Social Psychology, 
21, 517-528. 
Higgins, E.T., & King, G. (1981). Accessibility of social constructs: Information-processing 
consequences of individual and contextual variability. In N. Cantor & J.F. Kihlstrom (Eds.), 
Personality, cognition, and social interaction. Hillsdale, NJ: Erlbaum. 


238 
Social identity: international pers
pectives 
Hinkle, S., & Brown, R.J. (1990). Intergroup comparisons and social identity: Some links and 
lacunae. In D. Abrams & M.A. Hogg (Eds.), Social identity theory: Constructive and critical 
advances. Hemel Hempstead: Harvester Wheatsheaf. 
Hinkle, S., Brown, R.J., & Ely, P.G. (1992). Social identity theory processes: Some limitations 
and limiting conditions. Revista de Psicologia Social, Monogra
jico, 99- 1 1 1 .  
Hinkle, S., & Crook, K.F. (1987). Identification, importance o
f
 comparison dimensions and 
intergroup diff
erentiation. Paper presented at the International Conference on Social Identity, 
Exeter, UK, July. 
Hinkle, S., Fox-Cardamone, D.L., Taylor, L.A., & Crook, K.F. (1988). Identification, outgroup 
threat and dimensions o
f
 intergrou
p comparison. Paper presented at the Invitational 
Conference on Aggression, Conflict and Negotiation, Kill Devil Hills, NC. June. 
Hoffman, M.L. (1989). Empathy and pro 
social activities. In N. Eisenberg, J. Rekowski & E. 
Staub (Eds.), Social and moral values. Hillsdale, NJ: Erlbaum. 
Hofstede, G. (1980). Culture's consequences: International dif
f
erences in work-related values. 
Beverly Hills, CA: Sage. 
Hogg, M.A. (1987). Social identity and group cohesiveness. In l.e. Turner, M.A. Hogg, P.l. 
Oakes, S.D. Reicher & M.S. Wetherell, Rediscovering the social group: A self-categorization 
theory. Oxford: Blackwell. 
Hogg, M.A. (1992). The social psychology o
f
 group cohesiveness. Hemel Hempstead: Harvester 
Wheatsheaf. 
Hogg, M.A. (1993). Group cohesiveness: A critical review and some new directions. In W. 
Stroebe & M. Hewstone (Eds.), European review o
f
 social psychology, Vol. 4. Chichester: 
Wiley. 
Hogg, M.A. (1996). Intragroup processes: group structure, and social identity. In P. Robinson 
(Ed.), Social groups and identity: Developing the legacy o
f
 Henri T
a
j
f
el. Oxford: Butterworth 
Heinemann. 
Hogg, M.A., & Abrams, D. (1988). Social identifications: A social psychology o
f
 intergroup 
relations. London and New York: Routledge & Kegan Paul. 
Hogg, M.A., & Abrams, D. (1990). Social motivation, self-esteem and social identity. In D. 
Abrams & M. Hogg (Eds.), Social identity theory: Constructive and critical advances. Hemel 
Hempstead: Harvester Wheatsheaf. 
Hogg, M.A., Cooper-Shaw, L., & Holzworth, D.W. (in press). Studies of group prototypicality 
and depersonalized attraction in small interactive groups. Personality and Social Psychology 
Bulletin. 
Hogg, M.A., & Hardie, E.A. (1991). Social attraction, personal attraction, and self­
categorization: A field study. Personality and Social Psychology Bulletin, 1 7, 175-180. 
Hogg, M.A., & Hardie, E.A. (1992). Prototypicality, conformity and depersonalized attraction: 
A self-categorization analysis of group cohesiveness. British Journal o
f
 Social Psychology, 
31, 45-56. 
Hogg, M.A., & McGarty, C. (1990). Self-categorization and social identity. In D. Abrams & 
M.A. Hogg (Eds.), Social identity theory: Constructive and critical advances. Hemel 
Hempstead: Harvester Wheatsheaf. 
Hogg, M.A., & Turner, l.e. (l985a). When liking begets solidarity: An experiment on the role 
of interpersonal attraction in psychological group formation. British Journal o
f
 Social 
Psychology, 24, 267-281 .  
Hogg, M.A., & Turner, l.C. (1985b). Interpersonal attraction, social identification, and 
psychological group formation. European Journal o
f
 Social Psychology, 15, 51-66. 
Hogg, M.A., & Turner, J.e. (1987). Social identity and conformity: A theory of referent 
information influence. In W. Doise & S. Moscovici (Eds.), Current issues in European social 
psychology, Vol. 2. Cambridge: Cambridge University Press; Paris: Maison des Sciences de 
I'Homme. 
Hogg, M.A., Turner, J.C., & Davidson, B. (1990). Polarized norms and social frames of 
reference: A test of the self-categorization theory of group polarization. Basic and Applied 
Social Psychology, 11, 77-100. 


Re
f
erences 
239 
Holyoak, K.J., & Gordon, P.e. (1983). Social reference points. Journal o
f
 Personality and 
Social Psychology, 44, 881-887. 
Holyoak, K.J., & Gordon, P.e. (1984). Information processing and social cognition. In R.S. 
Wyer & T.K. Srull (Eds.), Handbook o
f
 social cognition, Vol. 1. Hillsdale, NJ: Erlbaum. 
Homans, G. (1950). The human group. New York: Harcourt Brace. 
Horwitz, M., & Rabbie, J.M. (1989). Stereotypes of groups, group members. and individuals in 
categories: A differential analysis. In D. Bar-Tal, e.F. Grumann, A. Kruglanski & W. 
Stroebe (Eds.), Stereotyping and 
pre
judice. New York: Springer Verlag. 
Hostetler, J.A. (1968). Amish society (rev. ed.). Baltimore: Johns Hopkins University Press. 
Hoyle, R.H., Pinkley, R.L., & Insko, e.A. (1989). Perceptions of behavior: Evidence of 
differing expectations for interpersonal and intergroup interactions. Personality and Social 
Psychology Bulletin, 15, 365-376. 
Huguet, P., Mugny, G., & Perez, J.A. (1990-1991). Influence sociale et processus de 
decentration. Bulletin de Psychologie, 45, 155-163. 
Hui, e.H. (1988). Measurement of individualism-collectivism. Journal o
f
 Research in 
Personality, 28, 17-36. 
Hui, e.H., & Triandis, H.e. (1986). Individualism-collectivism: A study of cross-cultural 
researchers. Journal o
f
 Cross-Cultural Psychology, 1 7, 225-248. 
Huici, e. (1 984). The individual and social functions of sex role stereotypes. In H. Tajfel (Ed.), 
The social dimension: European developments in social psychology, Vol. 2. Cambridge: 
Cambridge University Press. 
Huici, C. (1985). Relaciones intergrupales II: EI paradigma del Grupo Minimo. In e. Huici 
(Ed.), Estructura y procesos de Gru
po, Vol. 2. Madrid: UNED. 
Hurtig, M.e., & Pichevin, M.F. (1986). La dif
ference des sexes. Paris: Tierce. 
Hurtig, M.e., Pichevin, M.F., & Piolat, M. (1992). Sex and age as factors of asymmetry in the 
perceived similarity between two persons. In L. Arcuri & e. Serino (Eds.), Asymmetry 
phenomena in interpersonal comparison: Cognitive and social issues. Naples: Liguori. 
Inoff, G., Halverson, e.F., & Pizzigati, K.A.L. (1983). The influence of sex-role standards on 
children's self
- and peer-attributions. Sex Roles, 9, 1205-1222. 
Insko, e.A., Hoyle, R.H., Pinkley, R.L., Hong, G., Slim, R., Dalton, G., Lin, Y., Ruffin, P.P., 
Dardis, G.J., Bernthal, P.R., & Schopler, J. (1988). Individual-group discontinuity: The role 
of a consensus rule. Journal o
f
 Ex
perimental Social Psychology, 24, 505-5\9. 
Insko, e.A., Pinkley, R.L., Harring, K., Holton, B., Hong, G., Krams, D.S., Hoyle, R.H., & 
Thibaut, J. (1987). Beyond categorization to competition: Expectations of appropriate 
behavior. Representative Research in Social Psychology, 1 7, 5-36. 
Insko, C.A., Pinkley, R.L., Hoyle, R.H., Dalton, G., Hong, G., Slim, R., Landry, P., Holton, 
B., Ruffin, P.P., & Thibaut, J. (1987). Individual-group discontinuity: The role of 
intergroup contact. Journal o
f
 Ex
perimental Social Psychology, 23, 250-267. 
Insko, e.A., & Schopler, J. (1987). Categorization, competition, and collectivity. In e. 
Hendrick (Ed.), Review o
f
 personality and 
social psychology: Group processes, Vol. 8. Beverly 
Hills, CA: Sage. 
Insko, e.A., Schopler, J., Drigotas, S.M., Graetz, K.A., Kennedy, J., Cox, e., & Bomstein, G. 
(1993). The role of communication in interindividual-intergroup discontinuity. Journal o
f
 
Conflict Resolution, 37, 108-1 38. 
Insko, e.A., Schopler, J., Hoyle, R.H., Dardis, G.J., & Graetz, K.A. (1990). Individual-group 
discontinuity as a function of fear and greed. Journal o
f
 Personality and Social Psychology, 
58, 68-79. 
Insko, e.A., Schopler, J., Kennedy, J.F., Dahl, K.R., Gratez, K.A., & Drigotas, S.M. (1992). 
Individual-group discontinuity from the differing perspectives of Campbell's realistic group 
conflict theory and Tajfel and Turner's social identity theory. Social Psychological Quarterly, 
55, 272-291 . 
Jackson, L.A. (1993). Achieving positive social identity. Paper presented at the Invitational 
Conference on Self and Social Identity. Highland Park, FL, June. 
James, W. (\890). The princi
ples o
f
 psychology. New York: Holt. 
James, W. (1892). Psychology: A brie
f
er course. Greenwich, CT: Fawcett. 


240 
Social identity: international perspectives 
Janis, I.L. (1968). Group identification under conditions of external danger. In D. Cartwright & 
A. Zander (Eds.), Group d
ynamics (3rd ed.). London: Tavistock; New York: Harper & Row. 
Jarymowicz, M. (1982). Perception of self-others dissimilarity and signs of attractiveness or 
aversiveness of diversity from others. Polish Psychological Bulletin, 13, 39-44. 
Jarymowicz, M. (1987). Perceiving one's own individuality: Estimation and attractiveness o
f
 
one's own distinctiveness f
rom others. Warsaw: Wydawnictwa Uniwersytetu Warszawskiego. 
Jarymowicz, M. (l99Ia). The self-distinctive traits as personal identity attributes. CaMers de 
Psychologie Cognitive, 11, 679-695. 
Jarymowicz, M. (199Ib). La distinctive Soi-Nous-Autre comme base d'identite individuelle et 
socia 
Ie. Cahiers Internationaux de Psychologie Sociale, 9-10, 103-1 16. 
Jarymowicz, M. (I 
992a). Self, We and Other(s): Schemata distinctiveness and altruism. In P. 
Oliner, S. Oliner, L. Baron, D. Krebs & Z. Smoleska (Eds.), Embracing the other: 
Philosophical, psychological, and historical perspectives on altruism. New York: University of 
New York Press. 
Jarymowicz, M. ( l992b). Introduction to the Social Self 
Symposium: Searching 
f
or the social self 
concept. Paper presented to the EAESP/SESP Joint Meeting, LeuveniLouvain-la-Neuve, 
July. 
Jarymowicz, M. (l993a). Categorial vs cross-categorial social identity and outgroup involvement. 
Paper presented to the EAESP General Meeting, Lisbon, September. 
Jarymowicz, M. (I 993b). An attempt to operationalize 'social' versus 'personal identity' 
constructs. In M. Jarymowicz (Ed.), To know self - to understand others. Delf
t: Eburon. 
Jarymowicz, M. (1994a). W stron̘ indywidualnej podmiotowoSci i zblizen z innymi: 
Podmiotowe podstawy spolecznych identyfikacji. In M. Jarymowicz (Ed.), Poza 
egocentrycznt;l perspektywt;l widzenia siebie i swiata. Warsaw: Wydawnictwo Instytutu 
Psychologii PAN. 
Jarymowicz, M. (1994b). 0 formach umyslowego ujmowania MY i ich zwilizku ze 
spostrzeganiem Innych. In M. Jarymowicz (Ed.), Poza egocentrycznt;l perpsktywt;l widzenia 
siebie i swiata. Warsaw: Wydawnictwo Instytutu Psychologii PAN. 
Jarymowicz, M. (1994c). Poznawcza indiwiduacja a spoleczne identyfikacje. In M. Jarymowicz 
(Ed.), Poza egocentrycznt;l perspektywt;l widzenia siebie i swiata. Warsaw: Wydawnictwo 
Instytutu Psychologii PAN. 
Jarymowicz, M., & Greenwald, A.G. (1993). The anchoring function of the self in social 
judgments: Introduction. In M.C. Hurtig, M.F. Pichevin & M. Piolat (Eds.), Studies on the 
self and social cognition. London: World Scientific Publishing. 
Jodelet, D. (Ed.) (1989). Les representations sociales. Paris: Presses Universitaires de France. 
Jones, E.E., & DeCharms, R. (1967). Changes in social perception as a function of the personal 
relevance of behavior. Sociometry, 20, 175-185. 
Jones, E.E., & Gerard, H.B. (1967). Foundations o
f
 social psychology. New York: Wiley. 
Jones, E. E., Wood, G., & Quattrone, G. (1981 ). Perceived variability of personal characteristics 
in in-groups and out-groups: The role of knowledge and evaluation. Personality and Social 
Psychology Bulletin, 7, 523-528. 
Jung, C.G. (1943). The psychology of the unconscious. In Collected Works. Princeton, NJ: 
Princeton University Press. 
Karylowski, J. (1982). On the two types of altruistic behavior: doing good to feel good versus to 
make others feel good. In V. Derlega and J. Grzelak (Eds.), Cooperation and Helping 
Behavior. New York: Academic Press. 
Kedourie, E. (1960). Nationalism. London: Hutchinson. 
Kelley, H.H., & Stahelski, A. (1970). The social interaction basis of cooperators' and 
competitors' beliefs about others. Journal o
f
 Personality and Social Psychology, 16, 66-9\ .  
Kelley, H.H., & Thibaut, J.W. (1978). Interpersonal relations: A theory o
f
 interdependence. New 
York: Wiley. 
Kelly, J. (1988). Entertainment in individual and group behavior. In J.E. McGrath (Ed.), The 
social psychology o
f
 time. Newbury Park, CA: Sage. 
Kelman, H.C. (1969). Patterns of personal involvement in the national system: A social 


Re
f
erences 
241 
psychological analysis of political legitimacy. In 1.N. Rosenau (Ed.), International politics 
and f
oreign policy: A reader in research and theory. Glencoe, IL: Free Press. 
Kenny, D.A., & La Voie, L. (1985). Separating individual and group effects. Journal o
f
 
Personality and Social Psychology, 48, 339-348. 
Kiesler, CA., & Kiesler, S.B. (1969). Con
f
ormity. Reading, MA: Addison-Wesley. 
Killian, L.M. (1964). Social movements. In R.E.L. Faris (Ed.), Handbook o
f
 modern sociology. 
Chicago: Rand McNally. 
Kinder, D.R., & Sears, D.O. (1995). Public opinion and political action. In G. Lindzey & E. 
Aronson (Eds.), Handbook o
f
 social psychology. Vol. 2. New York: Random House. 
Klein, M. (1948). Contributions to psychoanalysis 1921-1945. London: Hogarth. 
Kluckhohn, C (1951). Values and value-orientations in the theory of action: An exploration in 
definition and classification. In T. Parsons & E. Shils (Ed.), Toward a general theory o
f
 
action. Cambridge, MA: Harvard University Press. 
Kobuszewska, E. (1989). Mechanizmy altruizmu w warunkach deindywiduacji. Unpublished 
master's thesis, Wydzial Psychologii, Uniwersytet Warszawski, Warsaw. 
Kohlberg, L. (1976). Moral stages and moralization: The cognitive-developmental approach. In 
T. Lickona (Ed.), Moral development and behavior. New York: Holt, Rinehart & Winston. 
Kohn, H. (1944). The idea of 
nationalism: A stud
y o
f
 its origins and background. New York: 
Macmillan. 
Kohn, H. (1962). The age o
f
 nationalism: The first era o
f
 global history. New York: Harper. 
Kotkin, M. (1983). Sex roles among married and unmarried couples. Sex Roles, 9, 975-985. 
Krech, D., Crutchfield, R.S., & Ballachey, E.L. (1962). Individual in society. New York: 
McGraw-Hill. 
Kruglanski, A.W. (1989). Lay epistemics and human knowledge: Cognitive and motivational 
bases. New York: Plenum. 
Kuhn, M.H., & McPartland, T.S. (1954). An empirical investigation of self-attitudes. American 
Sociological Review, 19, 68-76. 
Lalonde, R.N. (1992). The dynamics of group differentiation in the face of defeat. Personality 
and Social Psychology Bulletin, 18, 336-342. 
Lane, R.E. (1962). Political ideology. New York: Free Press. 
Lassaline, M.E., Wisniewski, E.l., & Medin, D.L. (1992). Basic levels in natural categories: Are 
all basic levels created equal? In B. Bums (Ed.), Percepts, concepts and categories. New 
York: Elsevier Science Publishers. 
Latani:, 8., & Darley, 1.M. (1970). T
he unres
ponsive bystander: Why doesn't he hel
p? New 
York: Appleton-Century-Crofts. 
Latane, B., Williams, K., & Harkins, S. (1979). Many hands make light the work: The causes 
and consequences of social loafing. Journal o
f
 Personality and Social Psychology, 37, 822-
832. 
LeBon, G. (1 896). The crowd. London: Unwin. 
L'Ecuyer, R. (1981). The development of the self-concept through the life span. In M.D. 
Lynch, A.A. Norem-Hebeisen & K. Gergen (Eds.), Self concept: Advances in research. 
Cambridge, MA: Ballinger. 
Lemaine, G. (1974). Social differentiation and social originality. European Journal o
f
 Social 
Psychology, 4, 17-52. 
Lemaine, G., Kastersztein, 1., & Personnaz, B. (1978). Social differentiation. In H. Tajfel (Ed.), 
Dif
f
erentiation between social groups: Studies in the social psychology o
f
 intergroup relations. 
London: Academic Press. 
Levine, I.M. (1980). Reactions to opinion deviance in small groups. In P.B. Paulus (Ed.), 
Psychology o
f
 group in
fluence. Hillsdale, Nl: Erlbaum. 
Levine, 1.M., & Moreland, R.L. (1994). Group socialization: Theory and research. In W. 
Stroebe and M. Hewstone (Eds.), European review o
f
 social psychology, Vol. 5. Chichester: 
Wiley. pp. 305-336. 
Le Vine, R.A., & Campbell, D.T. (1972). Ethnocentrism: T
heories o
f
 con
flict, ethnic attitudes 
and group con
flict. New York: Wiley. 


242 
Social identity: international pers
pectives 
Levy, S.B. (1975). Shifting patterns of ethnic identification among the Hassidim. In J.W. 
Bennet (Ed.), The new ethnicity: Pers
pectives 
f
rom ethnology. St Paul, MN: West Publishing. 
Lewin, K. (1948). Resolving social con
flicts. New York: Harper & Row. 
Lewin, K., Lippitt, R., & White, R. (1939). Patterns of aggressive behavior in experimentally 
created social climates. Journal o
f
 Psychology, 10, 271-299. 
Leyens, J.-P. (1983). Sommes nous tous des psychologues? Brussels: Mardaga. 
Leyens, J.-P., Yzerbyt, V.Y., & Schadron, G. (1994). Stereotypes and social cognition. London: 
Sage. 
Lindskold, S. (1983). Conflict and conciliation with groups and individuals. In A.P. Hare, V. 
Kent & M. Davies (Eds.), Small groups and social interaction, Vol. 2. New York: Wiley. 
Linville, P.W., & Jones, E.E. (1980). Polarized appraisals of out-group members. Journal o
f
 
Personality and Social Psychology, 38, 689-703. 
Linz, J.J. (1986). El con
flicto en Euskadi. Madrid: Espasa-Calpe. 
Lord, R.G., Foti, R.J., & De Vader, c.L. (1984). A test of leadership categorization theory: 
Internal structure, information processing, and leadership perceptions. Organizational 
Behaviour and Human Per
f
ormance, 34, 343-378. 
Lorenzi-Cioldi, F. (1988). Individus dominants et groupes domines: Images masculines et 
jeminines. Grenoble: Presses Universitaires de Grenoble. 
Lorenzo-Cioldi, F. (1993). They all look alike, but so do we . . .  
sometimes. Journal o
f
 Social 
Psychology, 32, 1 1 1-24. 
Lorenzi-Cioldi, F. (1994). Les androgynes. Paris: Presses Universitaires de France. 
Lorenzi-Cioldi, F., & Doise, W. (1990). Levels of analysis and social identity. In D. Abrams 
and M.A. Hogg (Eds.), Social identity theory, constructive and critical advances. Hemel 
Hempstead: Harvester Wheatsheaf. 
Lorenzi-Cioldi, F., & Doise, W. (1994). Identite personelle et identitee sociale. In R.Y. Bourhis 
& J.-P. Leyens (Eds.), Stereotypes, discrimination et relations intergroupes. Brussels: 
Mardaga. 
Lorenzi-Cioldi, F., Eagley, A.M., & Stewart, T. (1995). Homogeneity of gender groups in 
memory. Unpublished manuscript, University of Geneva. 
Lott, A.J., & Lott, B.E. (1965). Group cohesiveness as interpersonal attraction: A review of 
relationships with antecedent and consequent variables. Psychological Bulletin, 64, 259-309. 
Luthanen, R., & Crocker, J. (1 992). A collective self-esteem scale: Self-evaluation of one's 
social identity. Personality and Social Psychology Bulletin, 3, 302-318. 
Luria, A.R. (1 976). La storia sociale dei processi cognitivi. Florence: Giunti-Barbera. 
Maass, A., & Clark, R.D., III. (1984). Hidden impact of minorities: Fifteen years of minority 
influence research. Psychological Bulletin, 95, 428-450. 
Maass, A., Clark, R.D., III, & Haberkorn, G. (1982). The effects of differential ascribed 
category membership and norms on minority influence. European Journal o
f
 Social 
Psychology, 12, 89-104. 
Mackie, D.M., & Cooper, J. (1984). Attitude polarization: The effects of group membership. 
Journal o
f
 Personality and Social Psychology, 46, 575-585. 
Mackie, D.M., Worth, L.T., & Asuncion, A.G. (1990). Processing of persuasive in-group 
messages. Journal o
f
 Personality and Social Psychology, 58, 812-822. 
Makris-Botsaris, E., & Robinson, P.W. (1991). Harter's self-perception profile for children: A 
cross-cultural validation in Greece. Evaluation and Research in Evaluation, 5, 135-143. 
March, J.G., & Simon, H.A. (1958). Organizations. New York: Wiley. 
Marcia, J.E. (1966). Development and validation of ego identity status. Journal o
f
 Personality 
and Social Psychology, 3, 551 -558. 
Mariategui, J.C. (1925). EI hombre y el mito. In J.C. Mariategui, El alma matinal. Lima: 
Amatuam, 1 950. 
Markus, H. (1977). Self-schemas and processing information about the self. Journal o
f
 
Personality and Social Psychology, 35, 63-78. 
Markus, H. (1984). Possible selves. Paper presented at the Boston University Symposium for 
the Interdisciplinary Study of Personality. 


Re
f
erences 
243 
Markus, H., & Kitayama, S. (l99Ia). Culture and the self: Implications for cognition, emotion 
and motivation. Psychological Review, 98, 224-253. 
Markus, H., & Kitayama. S. (l99Ib). Cultural variations in the self-concept. In J. Strauss & 
G.R. Goethals (Eds.), The sel
f: An interdisc i
plinary approach. New York: Springer Verlag. 
Markus, H., & Smith, J. (1981). The influence of self-schemata on the perception of others. In 
N. Cantor & J. Kihlstrom (Eds.), Personality, cognition and social interaction. Hillsdale, NJ: 
Erlbaum. 
Marques, J.M. (1990). The black sheep effect: Outgroup homogeneity in social comparison 
settings. In D. Abrams & M.A. Hogg (Eds.), Social identity theory: Constructive and critical 
advances. Hemel Hempstead: Harvester Wheatsheaf. 
Marques, J.M., Abrams, D., & Paez, D. (1996). Unpublished manuscript. 
Marques, J.M., Brito, P.N., Correia, M.G., & Serodio, R.G. (1996). Evaluation of attractive 
and unattractive ingroup and outgroup members as a function of group uniformity. 
Unpublished manuscript. 
Marques, J.M., & Paez, D. (1994). The black sheep eff
ect: Social categorization, rejection of 
ingroup deviates, and perception of group variability. In M. Hewstone & W. Stroebe (Eds.), 
European review o
f
 social psychology, Vol. 5. Chichester: Wiley. 
Marques, I.M., Robalo, E.M., & Rocha, S.A. (1992). Ingroup bias and the black sheep effect: 
Assessing the impact of social identification and perceived variability on group judgments. 
European Journal o
f
 Social Psychology, 22, 331-352. 
Marques, J.M., & Yzerbyt, V.Y. (1988). The black sheep effect: Judgmental extremity towards 
ingroup members in inter- and intragroup situations. European Journal o
f
 Social Psychology, 
18, 287-292. 
Marques, J.M., Yzerbyt, V.Y., & Leyens, J.-P. (1988). The black sheep effect: Judgmental 
extremity towards ingroup members as a function of group identification. European Journal 
o
f
 Social Psychology, 18, 1-16. 
Martin, R. (I 988a). Ingroup and outgroup minorities: Diff
erential impact upon public and 
private response. European Journal o
f
 Social Psychology, 18, 39-52. 
Martin, R. (1988b). Minority influence and social categorization: A replication. European 
Journal of Social Psychology, 18, 369-373. 
Martin, R. (1988c). Minority influence and 'trivial' social categorization. European Journal o
f
 
Social Psychology, 18, 465-470. 
Maslow, A.H. (1954). Motivation and personality. New York: Harper & Row. 
McAdams, D.P. (1980). A thematic coding system for the intimacy motive. Journal o
f
 Research 
in Personality, 14, 412-432. 
McAdams, D.P. (1985). The 'Imago': A key narrative component of identity. In P. Shaver 
(Ed.), Self, situations and social behavior. Beverly Hills, CA: Sage. 
McCallum, D.M., Harring, K., Gilmore, R., Drenan, S., Chase, J., Insko, CA., & Thibaut, 1. 
(1985). Competition between groups and between individuals. Journal o
f
 Ex
perimental 
Social Psychology, 21, 301-320. 
McDougall, W. (1920). The group mind. London: Putnam. 
McGarty, C, & Turner, J.C (1992). The effects of categorization on social judgement. British 
Journal o
f
 Social Psychology, 31, 253-268. 
McGarty, C, Turner, J.C., Hogg, M.A., David, 8., & Wetherell, M.S. (1992) Group 
polarization as conformity to the prototypical group member. British Journal o
f
 Social 
Psychology, 31, 1-20. 
McGrath, J.E. (1988). Time and social psychology. In J.E. McGrath (Ed.), The social 
psychology o
f
 time. Newbury Park, CA: Sage. 
McGuire, W.J., & McGuire, C.V. (1988). Content and process in the experience of the self. In 
L. Berkowitz (Ed.), Advances in experimental social psychology, Vol. 21. New York: 
Academic Press. 
McPhail, C. (1991). The myth o
f
 the madding crowd. New York: Aldine de Gruyter. 
Mead, G.H. (1934). Mind, self and society. Chicago: University of Chicago Press. 
Medin, D.L. (1989). Concepts and conceptual structure. American Psychologist, 44, 1469-1481. 


244 
Social identity: international pers
pectives 
Medin, D.L., & Ortony, A. (1989). Psychological essentialism. In S. Vosniadou & A. Ortony 
(Eds.), Similarity and analogical reasoning. Cambridge: Cambridge University Press. 
Medin, D.L., & Wattenmaker, W.D. (1987). Category cohesiveness, theories, and cognitive 
archaeology. In U. Neisser (Ed.), Concepts and conceptual development: Ecological and 
intellectual f
actors in categorization. Cambridge: Cambridge University Press. 
Medin, D.L., Wattenmaker, W.D., & Hampson, S.E. (1987). Family resemblance, conceptual 
cohesiveness and category construction. Cognitive Psychology, 19, 242-279. 
Mervis, CB., & Rosch, E. (1981). Categorization of natural objects. In M.R. Rozenzweig & 
L.W. Porter (Eds.), Annual review o
f
 psychology, Vol. 32. Palo Alto, CA: Annual Reviews. 
Messick, D., & Mackie, D. (1989). Intergroup relations. Annual Review o
f
 Psychology, 40, 45-
81. 
Mlicki, P.P., & Ellemers, N. (1996). Being different or being better? National stereotypes and 
identifications of Polish and Dutch students. European Journal o
f
 Social Psychology, 26, 97-
1 14. 
Monteil, J.M. (1992). Intergroup differentiation and individuation: The effect of social 
deprivation. European Bulletin o
f
 Cognitive Psychology, 12, 189-203. 
Montero, M. (1987). A traves del espejo: Una aproximacion teorica al estudio de la conciencia 
social en America Latina. In M. Montero (Ed.), Psicologia Politica Latinoamerica. Caracas: 
Panapo. 
Montero, M. ( 1990). Autoimagen de los venezolanos: Lo positivo y 10 negativo. In H. 
Riquelme (Ed.), Buscando America Latina. Caracas: Nueva Sociedad. 
Morales, J 
.F. (1989). Identidad social y personal. In J. Mayor & J.L. Pinillos (Eds.), Tratado de 
psicologia general. 7: Creencias, actitudes y valores. Madrid: Alhambra. 
Morales, J.F., & Huici, C. (1994). Procesos grupales. In J.F. Morales, M. Moya, J.M. 
Fernandez-Dols, C Huici, J. Marques, D. Paez & J.A. Perez (Eds.), Psicologia social. 
Madrid: McGraw-Hill. 
Morales, J.F., Lopez, M., & Vega, L. (1992). Individualismo, collectivismo e identidad social. 
Revista de Psicologia Social, Monogra
fico, 49-72. 
Morales, J.F. & Paez, D. (1996). Estereotipos, discriminacion y relaciones intergrupos en 
Espana y Latinoamerica. In R.Y. Bourhis & J.-P. Leyens (Eds.), Estereotipos discriminacion 
y relaciones entre grupos. Madrid: McGraw-Hill. 
Moreland, R.L. (1987). The formation of small groups. In C. Hendrick (Ed.), Review o
f
 
personality and social psychology: Group processes, Vol. 8. Beverly Hills, CA: Sage. 
Moreland, R.L., & Levine, J.M. (1982). Socialization in small groups: Temporal changes in 
individual-group relations. In L. Berkowitz (Ed.), Advances in experimental social 
psychology, Vol. 15. New York: Academic Press. 
Moreland, R.L., & Levine, J.M. (1988). Group dynamics over time: Development and 
socialization in small groups. In J.E. McGrath (Ed.), The social psychology o
f
 time. 
Newbury Park, CA: Sage. 
Morris, W.N., Worchel, S., Bios, J., Pearson, J., Roundtree, C, Samaha, G., Wachtler, J., & 
Wright, S. (1976). Collective coping with stress: Group reactions to fear, anxiety, and 
ambiguity. Journal o
f
 Personality and Social Psychology, 33, 674-679. 
Moscovici, S. (1976). Social in
fluence and social changes. London: Academic Press. 
Moscovici, S. (1980). Toward a theory of conversion behavior. In L. Berkowitz (Ed.), Advances 
in experimental social psychology, Vol. 13. New York: Academic Press. 
Moscovici, S., & Mugny, G. (1983). Minority influence. In P. Paulus (Ed.), Basic group process. 
New York: Springer Verlag. 
Moscovici, S., & Nemeth, CJ. (1974). Minority influence. In CJ. Nemeth (Ed.), Social 
psychology: Classic and contemporary integrations. Chicago: Rand McNally. 
Moya, M.C (1987). Percepcion de es
pisodios y estereotipia sexual. Tesis Doctoral, Servicio de 
Publicaciones, Universidad de Granada. 
Moya, M.C. (1989). Genero, estereotipia y relaciones intergrupales. In P. Ballarin & T. Ortiz 
(Eds.), La mu
jer en Andalucia. Granada: Universidad de Granada. 
Moya, M.C, & Peplau, L.A. (in preparation). 1dentidad de genero y relaciones de pare
ja. 
Mugny, G. (1982). The power o
f
 minorities. London: Academic Press. 


Re
f
erences 
245 
Mugny, G., Kaiser, e., & Papastamou, S. (1983). Influence minontaue, identification et 
relations entre groupes: Etude experimentale autour d'une votation. Cahiers de Psychologie 
Sociale, 19, 1-30. 
Mugny, G., Kaiser, e., Papastamou, S., & Perez, J.A. (1984). Intergroup relations, 
identification and social influence. British Journal o
f
 Social Psychology, 23, 31 7-322. 
Mugny, G., & Papastamou, S. (1982). Minority influence and psycho-social identity. European 
Journal o
f
 Social Psychology, 12, 379-394. 
Mugny, G., & Perez, J.A. (1985). Influence sociale, conflit et identification: Etude 
experimentale autour d'une persuasion 'manquee' lors d'une votation. Cahiers de 
Psychologie Sociale, 6, 1-13. 
Mugny, G., & Perez, J.A. (199Ia). T
he social psychology o
f
 minority influence. Cambridge: 
Cambridge University Press. 
Mugny, G., & Perez, J.A. (199Ib). EI efecto de criptomnesia social. Anthropos: Revista de 
Documentacian Cientifica de la Cultura, 27, 1 10-1 12. 
Mullen, B., Brown, R.J., & Smith, C. (1992). Ingroup bias as a function of salience, relevance 
and status: An integration. European Journal o
f
 Social Psychology, 22, 103-1 12. 
Mullen, B., & Hu, L. (1989). Perceptions of in 
group and outgroup variabilty: A meta-analytic 
integration. Basic and Applied Social Psychology, 10, 233-252. 
Murphy, G.L., & Medin, D.L. (1985). The role of theories in conceptual coherence. 
Psychological Review, 92, 289-3 16. 
Mussolini, B. (1935). Fascism: Doctrine and institutions. Rome: Ardita. 
Nakbi, J.L. (1990). Evolution de I'identite psychosociale feminine et masculine de l'enfance a 
l'age adulte. Cahiers de Psychologie Sociale, 34, 1 -9. 
Nakbi, J.L., & Arnal-Duchemin, M.J. (1987). Dynamique des representations chez les femmes 
et les hommes. Cahiers de Psychologie Sociale, 34, 1-9. 
Nascimento-Schulze, C.M. (1993). Social comparison, group identity and prof
essional identity: 
A study with bank clerks. Revista de Psicologia Social, 8, 69-82. 
Neisser, U. (Ed.). (1987). Concepts and conceptual development: Ecological and intellectual 
f
actors in categorization. Cambridge: Cambridge University Press. 
Nemeth, C. (1986). Differential contributions of majority and minority influence. Psychological 
Review, 93, 1-10, 23-32. 
Nemeth, e., & Staw, B.M. (1989). The tradeoffs of social control and innovation in groups and 
organizations. In L. Berkowitz (Ed.), Advances in experimental social psychology, Vol. 23. 
Orlando, FL: Academic Press. 
Oakes, P.J. (1987). The salience of social categories. In J.e. Turner, M.A. Hogg, P.l. Oakes, 
S.D. Reicher & M.S. Wetherell, Rediscovering the social group: A self
-categorization theory. 
Oxford: Blackwell. 
Oakes, P.J. (1996). The categorization process: Cognition and the group in the social 
psychology of stereotyping. In P. Robinson (Ed.), Social groups and identity: Developing the 
legacy o
f
 Henri T
a
j
f
el. Oxford: Butterworth Heinemann. 
Oakes, P.J., Haslam, S.A., & Reynolds, K.J. (in press). Social categorization and social 
context: An example from stereotype change. In D. Abrams & M.A. Hogg (Eds.), Social 
identity and social cognition. Oxford: Blackwell. 
Oakes, P.J., Haslam, A.S., & Turner, J.e. (1994). Stereotyping and social reality. Oxford: 
Blackwell. 
Oakes, P.J., & Turner, J.C. (\986). Distinctiveness and the salience of social category 
memberships: Is there an automatic perceptual bias towards novelty? European Journal o
f
 
Social Psychology, 16, 325-344. 
Oakes, P.J., & Turner, J.C. (\990). Is limited infonnation processing capacity the cause of 
social stereotyping? In W. Stroebe and M. Hewstone (Eds.), European review o
f
 social 
psychology, Vol. I. Chichester: Wiley. pp. 1 1 1-136. 
Oakes, P.J., Turner, J.e., & Haslam, S.A. (1991). Perceiving people as group members: The 
role of fit in the salience of social categorizations. British Journal o
f
 Social Psychology, 30, 
125-144. 
O'Dea, T.F. (1957). The Mormons. Chicago: University of Chicago Press. 


246 
Social identity: international pers
pectives 
Olden 
quist, A. (1992). Autonomy, social identities, and alienation. In F. Geyer & W.R. Heinz 
(Eds.), Alienation, society, and the individual. New Brunswick, NJ: Transaction Publishers. 
Osgood, C.E. (1962). An alternative to war or surrender. Urbana: University of Illinois Press. 
Ostrom, T.M., & Sedikides, C. (1992). Out-group homogeneity effect in natural and minimal 
groups. Psychological Bulletin, 112, 536-552. 
Paez, D., Arrospide, l., Martinez-Taboada, C., & Ayestaran, S. (1992). Identidad social, 
autoesquematizacion y autoconciencia collectiva: Investigaciones correlacionales en la 
perspectiva de la teona de la identidad social. Revista de Psicologia Social, Monogrti
fico, 3-
20. 
Palmonari, A., Carugati, F., Ricci-Bitti, P., & Sarchielli, G. (1979). Identita imper
f
ette. 
Bologne: II Mulino. 
Palmonari, A., Pombeni, M.L., & Kirchler, E. (1989). Peergroups and evolution of the self­
system in adolescence. European Journal o
f
 Psychology o
f
 Education, 4, 3-15. 
Parelman, A. (1983). Emotional intimacy in marriage: A sex-roles perspective. Ann Arbor, MI: 
UMI Research Press. 
Park, B., & Rothbart, M. (1982). Perception of out-group homogeneity and levels of social 
categorization: Memory for subordinate attributes of in-group and out-group members. 
Journal o
f
 Personality and Social Psychology, 42, 1051-1068. 
Parsons, T. (1968). The position of identity in the general theory of action. In G. Gordon & 
KJ. Gergen (Eds.), The self in social interaction. New York: Wiley. 
Pavelchak, M.A. (1989). Piecemeal and category-based evaluation: An idiographic analysis. 
Journal o
f
 Personality and Social Psychology, 56, 354-363. 
Pawluczuk, B. (1992). Indywiduac
ja systemu norm a toleranc
ja wobec obcych. Unpublished 
research report, Wydzial Psychologii, Uniwersytet Warszawski, Warsaw. 
Peabody, D., & Goldberg, L.R. (1989). Some determinants of factor structures from 
personality-descriptors. Journal o
f
 Personality and Social Psychology, 57, 552-567. 
Pemberton, MJ., Insko, C.A., & Schopler, l. (1996). Memory for and experience of differential 
competitive behavior of individuals and groups. Journal o
f
 Personality and Social 
Psychology, 71, 5, 953-966. 
Peplau, L.A. (1983). Roles and gender. In H.H. Kelley, E. Berscheid, A. Christensen, l. 
Harvey, T.L. Huston, G. Levinger, E. McClintock, A. Peplau & D.R. Peterson (Eds.), Close 
relationshi
ps. New York: W.H. Freeman & Co. 
Peplau, L.A., & Gordon, S.L. (1985). Women and men in love: Gender differences in close 
heterosexual relationships. In V.E. O'Leary, R.K. Under & B.S. Wallston (Eds.), Women, 
gender, and social psychology. Hillsdale, Nl: Erlbaum. 
Perez, l.A., & Mugny, G. (1985a). Categorizacion e influencia minoritaria. Anuario de 
Psicologia. 32, 100-1 16. 
Perez, l.A., & Mugny, G. (1985b). Influencia minoritaria sobre las opiniones frente al aborto y 
los anticonceptivos. Estudios de Psicologia, 23124, 29-54. 
Perez, l.A., & Mugny, G. (I 986a). Efectos paradojicos de la categorizaci6n en la influencia 
minoritaria. Boletin de Psicologia, 12, 65-89. 
Perez, l.A., & Mugny, G. (1986b). Induction experimentale d'une influence minoritaire 
indirecte. Cahiers de Psychologie Sociale, 32, 15-24. 
Perez, l.A., & Mugny, G. (1987). Paradoxical effects of categorization in minority influence: 
When being an out-group is an advantage. European Journal o
f
 Social Psychology, 17, 1 57-
169. 
Perez, l.A., & Mugny, G. (1990). Minority influence, manifest discrimination and latent 
influence. In D. Abrams & M.A. Hogg (Eds.), Social identity theory: Constructive and 
critical advances. Heme! Hempstead: Harvester Wheatsheaf. 
Perez, l.A., Mugny, G., Llavata, E., & Fierres, R. (1993). Paradoxe de la discrimination et 
conflit culture!: Etudes sur Ie racisme. In l.A. Perez, G. Mugny, et aI., I
n
fluences sociales: La 
theorie de l'elaboration du con
flit. Paris: Delachaux et Niestle. 
Perez, l.A., Mugny, G., & Moscovici, S. (1986). Les eff
ets paradoxaux du deni dans l'influence 
sociale. Cahiers de Psychologie Sociale, 32, 1-14. 
Perez, l.A., Mugny, G., & Navarro, E. (1991). EI efecto de la 'paralisis intragrupal': Niveles de 


Re
f
erences 
247 
categorizaci6n del yo y niveles de influencia social. Anthropos: Revista de Documentacion 
Cientifica de la Cultura, 27, 121-132. 
Perez, J.A., Mugny, G., Roux, P., & Butera, F. (1991). Influence via la comparaison et 
influence via la validation. In J.L. Beauvois, R.V. Joule & J.M. Monteil (Eds.), Perspectives 
cognitives et conduites sociales, Vol. 3. Cousset: Delval. 
Pfost, K.S., & Fiore, M. (1990). Pursuit of nontraditional occupations: Fear of success or fear 
of not being chosen? Sex Roles, 23, 1 5-24. 
Phinney, J.S. (1990). Ethnic identity in adolescents and adults: Review of research. 
Psychological Bulletin, 108, 499-514. 
Piaget, J. (1975). L'equilibration de structures cognitives: Probleme central du developpement. 
Paris: Presses Universitaires de France. 
Platow, MJ., McClintock, e.G., & Liebrand, W. (1990). Predicting intergroup fairness and 
ingroup bias in the minimal group paradigm, European Journal o
f
 Social Psychology, 20, 
221-239. 
Pratto, F., & Bargh, J.A. (1991). Stereotyping based on apparently individuating information: 
Trait and global components of sex stereotypes under attention overload. Journal o
f
 
Experimental Social Psychology, 27, 26-47. 
Proshansky, H.M., Fabian, A.K., & Kaminoff, R. (1983). Place identity: Physical world 
socialization of the self. Journal o
f
 Environmental Psychology, 3, 57-83. 
Pryor, J.B., & Ostrom, T.M. (1987). Social cognition theory of group processes. In B. Mullen & 
G.R. Goethals (Eds.), Theories o
f
 group behavior. New York: Springer Verlag. 
Quattrone, G.A. (1986). On the perception of a group's variability. In S.G. Worchel & W.G. 
Austin (Eds.), Psychology o
f
 intergroup relations. Chicago: Nelson-Hall. 
Quevedo, S. (1972). Revindicaci6n urbana y lucha politica. Revista Latinoamericana de 
Estudios Urbanos Regionales, 2, 12-43. 
Rabbie, J.M., & Lodewijkx, H.G.M. (1994). Conflict and interaction: An individual-group 
continuum. In B. Markovsky, K. Heimer & J. O'Brien (Eds.), Advances in group processes, 
Vol. 11. Greenwich, CT: JAI Press. 
Ravaud, J.F., Beaufils, B., & Paicheler, H. (1986). Handicap et integration scolaire. Sciences 
Sociales et Sante, 3-4, 167-194. 
Reicher, S., & Emler, N. (1986). The management of delinquent reputations. In H. Beloff (Ed.), 
Getting into lif
e. London: Methuen. 
Reis, H.T., Nezlek, J., & Wheeler, L. (1980). Physical attractiveness in social interaction. 
Journal o
f
 Personality and Social Psychology, 38, 604-617. 
Reis, H.T., & Wheeler, L. (1991). StUdying social interaction with the Rochester Interaction 
Record. In M.P. Zanna (Ed.), Advances in experimental social psychology, Vol. 24. New 
York: Academic Press. 
Reykowski, J. (1979). Motywac
ja, postawy prospoleczne a osobowoSi:. Warsaw: PWN. 
Reykowski, J. (1989). Dimensions of the development in moral values. In N. Eisenberg, J. 
Reykowski & E. Staub (Eds.), Social and moral values. Hillsdale, NJ: Erlbaum. 
Reynolds, V., Falger, V., & Vine, I. (1987). The sociobiology o
f
 ethnocentrism. London: Croom 
Helm. 
Rips, L.J. (1989). Similarity, typicality, and categorization. In S. Vosniadou & A. Ortony 
(Eds.), Similarity and analogical reasoning. Cambridge: Cambridge University Press. 
Rivera, A.N. (1991). Psicologia y colonizaci6n. In M. Montero (Ed.), Accion y discurso: 
Problemas de psicologia politica en America Latina. Caracas: Eduven. 
Robinson, J.P. (1988). Time diary evidence about the social psychology of 
everyday life. In J 
.E. 
McGrath (Ed.), The social psychology o
f
 time. Newbury Park, CA: Sage. 
Rodriguez, C.P. (1992). Estereotipos de nacionalidad en estudiantes colombianos y 
venezolanaos. Boletin A vepso, 15, 65-74. 
Rogers, C. (1951). Client-centered therapy. Boston: Houghton-Mifflin. 
Rokeach, M. (1973). The nature o
f
 human values. New York: Free Press. 
Rook, D.S. (1987). Social support versus companionship: Effects of life stress, loneliness, and 
evaluations by others. Journal o
f
 Personality and Social Psychology, 52, 1 1 32-1 147. 
Rosch, E. (1973). On the internal structure of perceptual and semantic categories. In T.E. 


248 
Social identity: international perspectives 
Moore (Eds.), Cognitive development and the acquisition o
f
 language. New York: Academic 
Press. 
Rosch, E. (1975). Cognitive reference points. Cognitive Psychology, 7, 532-547. 
Rosch, E. (1978). Principles of categorization. In E. Rosch & B.B. Lloyd (Eds.), Cognition and 
categorization. Hillsdale, NJ: Erlbaum. 
Rosch, E., Mervis, CB., Gray, W.D., Johnson, D.M., & Boyes-Braem, P. (1976). Basic objects 
in natural categorization. Cognitive Psychology, 15, 346-378. 
Rosenberg, S., & Gara, M. (1985). The multiplicity of personal identity. Review o
f
 Personality 
and Social Psychology, 6, 87-1 13. 
Roth, E.M., & Schoben, E.1. (1983). The effect of context on the structure of categories. 
Cognitive Psychology, 15, 346-378. 
Rubin, Z., Peplau, L.A., & Hill, CT. (1981). Loving and leaving: Sex differences in romantic 
attachments. Sex Roles, 7, 821-835. 
Rush, M.C, & Russell, J.E.A. (1988). Leader prototypes and prototype-contingent consensus 
in leader behaviour descriptions. Journal o
f
 Experimental Social Psychology, 24, 88-104. 
Salazar, J.M. (1983a). Bases psicologicas del nacionalismo. Mexico: Trillas. 
Salazar, J.M. (1983b). On the psychological viability of 'Latin-Americanism'. International 
Social Science Journal, 35, 295-308. 
Salazar, J.M. (1989). Niveles de identificacion y estructura cognoscitiva en relacion con el 
latinoamericano. Revista de Psicologia Social, 4, 13-22. 
Salisbury, R.H. (1975). Interest groups. In F.I. Greenstein & N.W. Polsby (Eds.), Handbook o
f
 
political science, Vol. 4. Reading, MA: Addison-Wesley. 
Sanchez-Mazas, M., Perez, J.A., Navarro, E., Mugny, G., & Jovanovic, J. (1993). De la 
paralysie intragroupe au confiit normatif: Etudes dur I'avortement, la contraception et la 
xenophobie. In J.A. Perez & G. Mugny (Eds.), Influences sociales: La tMorie de l'eboration 
du confiit. Paris: Delachaux et Niestle. 
Sande, G.N. (1990). The multifaceted self. In J.M. Olson & M.P. Zanna (Eds.), Self
-in
f
erence 
processes: The Ontario Symposium, Vol. 6. Hillsdale, NJ: Erlbaum. 
Sanders, G.S., & Baron, R.S. (1977). Is social comparison irrelevant for producing choice 
shifts? Journal o
f
 Experimental Social Psychology, 13, 303-314. 
Sarbin, T.R., & Allen, V.L. (1968). Role theory. In G. Lindzey & E. Aronson (Eds.), Handbook 
o
f
 social psychology, Vol. 1 (2nd ed.). Reading, Mass: Addison-Wesley. 
Schachter, S. (1951). Deviation, rejection, and communication. Journal o
f
 Abnormal and Social 
Psychology, 46, 190-207. 
Schopler, J., & Insko, CA. (1992). The discontinuity effect in interpersonal and intergroup 
relations: Generality and mediation. In W. Stroebe & M. Hewstone (eds.), European review 
o
f
 social psychology, Vo!' 1. Chichester: Wiley. 
Schopler, J., Insko, CA., Graetz, K.A., Drigotas, S.M., & Smith, V.A. (1991). The generality 
of the individual-group discontinuity eff
ect: Variations in positivity-negativity of outcomes, 
players' relative power, and magnitude of outcomes. Personality and Social Psychology 
Bulletin, 1 7, 612-624. 
Schopler, J., Insko, CA., Graetz, K.A., Drigotas, S.M., Smith, VA, & Dahl, K. (1993). 
Individual-group discontinuity: Further evidence for mediation by fear and greed. 
Personality and Social Psychology Bulletin, 19, 419-43 1. 
Semin, G.R. (1987). On the relationship between representation of 
theories in psychology and 
ordinary language. In W. Doise & S. Moscovici (Eds.), Current issues in European social 
psychology, Vol. 2. Cambridge: Cambridge University Press. 
Serino, C (1988a). Biases egocentrici e fenomeni di asimmetria nei confronti Se/Altro. Giornale 
Italiano di Psicologia, 15, 213-233. 
Serino, C. (I 
988b). Strategies et structure de la comparaison sociale: Quelques aspects de 
l'asymmetrie soi/autrui dans les relations entre groupes. Cahiers de Psychologie Cognitive, 8, 
627-648. 
Serino, C (1989). Processi di confronto e argomentazione: In G. Bellelli (Ed.), II metodo del 
discorso. Naples: Liguouri. 
Serino, C (1990). La comparaison socia 
Ie et Ie structure des systemes categoriels: Quelques 


Re
f
erences 
249 
fI!flexions sur la prototypicalite du soi. Cahiers Internationaux de Psychologie Socia 
Ie, J, 73-
96. 
Serino, C (1992). From the asymmetry phenomena to the prototypicality of self: Studies on the 
structure of social categories. In L. Arcuri & C Serino (Eds.), Asymmetry phenomena in 
interpersonal comparison: Cognitive and social issues. Naples: Liguori. 
Serino, C (1993). Information processing and self/others comparison: Social-cognitive 
extensions of a non-metrical approach to similarity. In M.C. Hurtig, Pichevin, M.F., & 
M. Piolat (Eds.), Studies on the self and social cognition. London: World Scientific 
Publishing. 
Serino, C (1996). Similarity: Inter
personal and intergroup comparisons in an asymmetrical 
f
rame. Paper presented at the Small Group Meeting on 'Intergroup Relations: Current Work 
and Future Perspectives', Catania, Italy, 26-9 September. 
Serino, C, & Mazzara, B. (1995). Scripts e processi di confronto nella rappresentazione sociale 
della persona handicappata: Una ricerca sperimentale. Giornale Italiano di Psicologia, 22, 
265-292. 
Seton-Watson, H. (1965). Nationalism. old and new. Sydney: Sydney University Press. 
Shaw, M.E. (1976). Group d
ynamics: The psychology o
f
 small group behavior. New York: 
McGraw-Hill. 
Sherif, M. (1936). The psychology o
f
 social norms. New York: Harper & Brothers. 
Sherif, M. (1951). A preliminary experimental study of inter-group relations. In 1.H. Rohrer & 
M. Sherif (Eds.), Social psychology at the crossroads. New York: Harper & Brothers. 
Sherif, M. (1966). In common predicament: Social psychology o
f
 intergroup conflict and 
cooperation. Boston: Houghton Mifflin. 
Sherif, M. (1967). Group conflict and cooperation: T
heir social psychology. London: Routledge 
& Kegan Paul. 
Sherif, M., Harvey, 0.1., White, B.1., Hood, W.R., & Sherif, CW. (1961). I
ntergroup conflict 
and cooperation: The Robbers' Cave experiment. Norman, OK: University Book Exchange. 
Shils, E. (1968). Ideology. In D.L. Sills (Ed.), International encylopedia o
f
 the social sciences. 
Vol. 7. New York: Macmillan and Free Press. 
Sills, D.L. (1958). The volunteers. New York: Free Press. 
Simmel, G. (1955). Conflict: T
he web o
f
 group a
ffiliations. Chicago: Chicago Free Press. 
Simon, B. (1994). The perception of ingroup and outgroup homogeneity: Reintroducing the 
intergroup context. In W. Stroebe and M. Hewstone (Eds.), European review o
f
 social 
psychology. Vol. 3. Chichester: Wiley. pp. 1-3. 
Simon, B., & Brown, R. (1987). Perceived intragroup homogeneity in minority-majority 
contexts. Journal o
f
 Psychology and Social Psychology, 53, 703-71 1 .  
Simon, B., & Hamilton, D.L. (1994). Self stereotyping and social context: The effects of relative 
in-group size and in-group status. Journal o
f
 Personality and Social Psychology, 66, 699-
7 1 1 .  
Singh, A .  (1979). Reliability and validity of self-reported delinquency studies. Psychological 
Reports, 44, 987-993. 
Singleton, R., 1r. (1979). Another look at the conformity explanation of group-induced shift in 
choice. Human Relations, 32, 37-56. 
Smith, A.D. (I 984a). National identity and myths of ethnic descent. Research in Social 
Movements. Con
flict and Change, 7, 95-230. 
Smith, A.D. (1984b). Ethnic myths and ethnic revivals. European Journal o
f
 Sociology, 25, 
283-305. 
Smith, A.D. (199 1). National identity. London: Penguin. 
Smith, P.B., & Bond, M.H. (1993). Social psychology across cultures. New York: Harvester 
Wheatsheaf. 
Snyder, CR., & Fromkin, H.L. (1980). Uniqueness: the human pursuit o
f
 d
iff
erence. New York: 
Plenum Press. 
Snyder, M.L., Kleck, B.E., Strentna, A., & Mentzer, S.l. (1 979). A comparison of egotism, 
negativity, and learned helplessness as explanations for poor performance after unsolvable 
problems. Journal o
f
 Personality and Social Psychology, 40, 24-30. 


250 
Social identity: international pers
pectives 
Snyder, M.L., & Wicklund, R. (1981). Attribute ambiguity. In J.H. Harvey, W. Ickes & R.R. 
Kidd (Eds.), New directions in attribution research. New York: Halsted. 
Sobczyk, K. (1992). Schematy Ja/My/I
nni a spostrzegany d
ystans i ustosunkowania wobec 
narodowosci blizsze
j i dalsze
j. Unpublished master's thesis, Wydzial Psychologii, Universytet 
Warszawski, Warsaw. 
Sokolowska, G. (1992). Tozsamosciowe przes/anki ustosunkowim wobec obcych. Unpublished 
research report, Wydzial Psychologii, Uniwersytet Warszawski, Warsaw. 
Spears, R., Lea, M., & Lee, S. (1990). De-individuation and group polarization in computer­
mediated communication. British Journal o
f
 Social Psychology, 29, 121-134. 
Srull, T.K., & Gaelik, L. (1983). General principles and individual diff
erences in the self as a 
habitual ref
erence point: An examination of self-other judgments of similarity. Social 
Cognition, 2, 108-121. 
Stephenson, G.M. (1984). Interpersonal and intergroup dimensions of bargaining and 
negotiation. In H. Tajfel (Ed.), The social dimension: European developments in social 
psychology, Vol. 2. Cambridge: Cambridge University Press. 
Stroup, H.H. (1945). T
he Jehovah's Witnesses. New York: Russell & Russell. 
Suls, J.M., & Miller, R.L. (Eds.). (1977). Social comparison processes: Theoretical and empirical 
pers
pectives. Washington, DC: Hemisphere. 
Sumner, W.G. (1906). Folkways. New York: Ginn. 
Szuster-Zbrojewicz, A. (1988). Z badan nad zwiazkami pobodienstwalodn;:nosci Ja-Inni z 
gotowoSci̕ do egzocentrycznego angaiowania sil;: w sprawy pozaosobiste. In M. Jarymowicz 
(Ed.), Studia nad spostrzeganiem relac
ji Ja-I
nni: Tozsamosc, ind
ywiduac
ja, przynaleznosc. 
Wroclaw: Ossolineum. 
Szuster-Zbrojewicz, A. (1 990). Formy reprezentacji MY a prze
jawy pozaosobistego 
zaangazowania. Unpublished report G-MEN-9/90, Wydzial Psychologii, Uniwersytet 
Warszawski, Warsaw. 
Tajfel, H. (1959). Quantitative judgment in social perception. British Journal o
f
 Psychology, 50, 
16-29. 
Tajfel, H. (1969a). Cognitive aspects of prejudice. Journal o
f
 Social Issues, 25, 79-97. 
Tajfel, H. (1969b). The formation of national attitudes: A social psychological perspective. In 
M. Sherif & c. Sherif (Eds.), Interdisci
plinary relationshi
ps in the social sciences. Chicago: 
Aldine. 
Tajfel, H. (1970a). Aspects of national and ethnic loyalty. Social Science I
n
f
ormation, 9, 1 19-
144. 
Tajfel, H. (1970b). Experiments in intergroup discrimination. Scientific American, 223, 96-102. 
Tajfel, H. (I 
972a). La categorisation sociale. In S. Moscovici (Ed.), Introduction a la 
psychologie sociale, Vol. 1 .  Paris: Larousse. 
Tajfel, H. (1972b). Experiments in a vacuum. In J. Israel & H. Tajfel (Eds.), T
he context o
f
 
social psychology: A critical assessment (European Monographs in Social Psychology No. 2). 
London: Academic Press. 
Tajf
el, H. (1974). Social identity and intergroup behavior. Social Science Inf
ormation, 13, 65-
93. 
Tajfel, H. (1978a). Interindividual behaviour and intergroup behaviour. In H. Tajfel (Ed.), 
Diff
erentiation between social groups: Studies in the social psychology o
f
 intergroup behaviour. 
London: Academic Press. 
Tajfel, H. (1978b). Social categorization, social identity and social comparison. In H. Tajfel 
(Ed.), Dif
f
erentiation between social groups: Studies in the social psychology o
f
 intergroup 
relations. London: Academic Press. 
Tajfel, H. (I978c). The achievement of group diff
erentiation. In H. Tajfel (Ed.), Diff
erentiation 
between social groups: Studies in the social psychology o
f
 intergroup relations. London: 
Academic Press. 
Tajfel, H. (Ed.). (1978d). Dif
f
erentiation between social groups: Studies in the social psychology 
of intergroup relations. London: Academic Press. 
Tajfel, H. (1978e). The social psychology o
f
 minorities. London: Minority Rights Group. 


Re
f
erences 
251 
Tajfel, H. (1979). Individuals and groups in social psychology. British Journal o
f
 Social 
Psychology, 18, 183-190. 
Tajfel, H. (\981). Human groups and social categories: Studies in social psychology. Cambridge: 
Cambridge University Press. 
Tajfel, H. (1982a). Instrumentality, identity, and social comparison. In H. Tajfel (Ed.), Social 
identity and intergroup relations. Cambridge: Cambridge University Press. 
Tajfel, H. (Ed.). (\982b). Social identity and intergroup relations. Cambridge: Cambridge 
University Press. 
Tajfel, H. (1982c). Social psychology of intergroup relations. In M.R. Rosenzweig & L.R. 
Porter (Eds.), Annual review o
f
 psychology, Vol. 33. Palo Alto, CA: Annual Reviews. 
Tajfel, H., & Billig, M. (\974). Familiarity and categorization in intergroup behavior. Journal 
o
f
 Experimental Social Psychology, 10, 159-170. 
Tajfel, H., Flament, e., Billig, M., & Bundy, R.P. (1971). Social categorization and intergroup 
behaviour. European Journal o
f
 Social Psychology, 1, 149-178. 
Tajfel, H., & Turner, J.C. (\979). An integrative theory of intergroup conflict. In W.G. Austin 
& S. Worchel (Eds.), The social psychology o
f
 intergroup relations. Monterey, CA: Brooks/ 
Cole. 
Tajfel, H., & Turner, J.e. (1986). The social identity theory of intergroup behavior. In S. 
Worchel & W.G. Austin (Eds.), The psychology o
f
 intergroup relations. Chicago: Nelson­
Hall. 
Tajfel, H., & Wilkes, A.L. (1963). Classification and quantitative judgement. British Journal o
f
 
Psychology, 54, 101-1 14. 
Tap, P. (1979). L'identification: Est-elle une alienation de l'identite? In P. Tap (Ed.), Jdentite 
individuelle et personalisation. Toulouse: Privat. 
Tap, P. (1985). Masculin et f
eminin chez I'en
f
ant. Toulouse: Privat. 
Taylor, D.M., & Moghaddam, F.M. (1987). T
heories o
f
 intergroup relations. New York: 
Praeger. 
Toch, H. (\965). The social psychology o
f
 social movements. Indianapolis: Bobbs-Merrill. 
Torregrosa, J.R., & Ramirez, S. (1991). Identidades nacionales y estados plurinacionales: Una 
perspectiva psicosociol6gica. lnteraccion Social, 1, 123-1 36. 
Triandis, H.e. (\ 
988). Collectivism vs. individualism. In G. Verma & e. Bagley (Eds.), Cross­
cultural studies of 
personality, attitudes, and cognition. London: Macmillan. 
Triandis, H.e. (1992). Comment on S. Hinkle, R.J. Brown & P. Ely, Social identity processes: 
some limitations and limiting conditions. Revista de Psicologia Social, Monogra
fico, 1 1 3-
123. 
Triandis, H.e. (1993). Paper for Psychology in the Americas: North-South Exchange, 
Communicaci6n Presentada al Congreso Interamericano, Santiago de Chile. 
Triandis, H.e. (\994). Recherches recentes sur I'individualisme et collectivisme. Cahiers 
1nternationaux de Psychologie Sociale, 23, 14-27. 
Triandis, H.C., Bontempo, R., Betancourt, H., Bond, M., Leung, D., Brenes, A., Georgas, G., 
Hui, e.H., Marin, G., Setiadi, B., Sinha, J.B.P., Verma, J., Spangenberg, J., Touzard, H., & 
Montmollin, G. (1986). The measurement of etic aspects of individualism and collectivism 
across cultures. Australian Journal o
f
 Psychology, 38, 257-267. 
Triandis, H.C., Bontempo, R., Villareal, M.J., Asai, M., & Lucca, N. (\988). Individualism 
and collectivism: Cross-cultural perspectives on self-ingroup relationships. Journal o
f
 
Personality and Social Psychology, 54, 323-338. 
Triandis, H.e., Leung, K., Villareal, M.J., & Clack, F.L. (1985). Allocentric versus idiocentric 
tendencies: Convergent and discriminant validation. Journal o
f
 Research in Personality, 19, 
395-415. 
Triandis, H.C., McCusker, e., & Hui, e.H. (1990). Multimethod probes of individualism and 
collectivism. Journal o
f
 Personality and Social Psychology, 59, 1006-1020. 
Triska, J.F. (Ed.). (\962). Soviet communism: Programs and rules. San Francisco: Chandler. 
Truman, D.B. (\951). The governmental process. New York: Knopf. 
Tuckman, B.W. (1965). Developmental sequence in small groups. Psychological Bulletin, 63, 
384-399. 


252 
Social identity: international pers
pectives 
Tuckman, B.W., & Jensen, M.A.e. (1977). Stages of small group development revisited. Group 
and Organization Studies, 2, 419-427. 
Turner, J.e. (l975a). Social comparison and social identity: Some prospects for intergroup 
behaviour. European Journal o
f
 Social Psychology, 5, 5-34. 
Turner, J.e. (1975b). Social categorization and social comparison in intergroup relations. PhD, 
University of Bristol. 
Turner, J.C. (1978). Social categorization and social discrimination in the minimal group 
paradigm. In H. Tajfel (Ed.), Diff
erentiation between social groups: Studies in the social 
psychology o
f
 intergroup relations. London: Academic Press. 
Turner, J.e. (l98Ia). Towards a cognitive redefinition of the social group. Cahiers de 
Psychologie Cognitive, 1, 93-1 18. 
Turner, J.e. (198Ib). The experimental social psychology of intergroup behavior. In J.C. 
Turner & H. Giles (Eds.), Intergroup behavior. Chicago: University of Chicago Press. 
Turner, J.C. (1982). Towards a cognitive redefinition of the social group. In H. Tajfel (Ed.), 
Social identity and intergroup relations. Cambridge: Cambridge University Press. 
Turner, J.C. ( l983a). Some comments on the measurement of social orientations in the minimal 
group paradigm. European Journal o
f
 Social Psychology, 13, 351-368. 
Turner, J.e. (I 
983b). A second reply to Bornstein, Crum, Wittenbraker, Harring, Insko, and 
Thibaut on the measurement of social orientations. European Journal o
f
 Social Psychology, 
13, 383-387. 
Turner, J.e. (1985). Social categorization and the self-concept: A social cognitive theory of 
group behaviour. In E.J. Lawler (Ed.), Advances in group processes, Vol. 2. Greenwich, CT: 
JAI Press. 
Turner, J.C. (1987). In J.e. Turner, M.A. Hogg, P.J. Oakes, S.D. Reicher & M.S. Wetherell, 
Rediscovering the social group: A self-categorization theory. Oxford: Blackwell. 
Turner, J.e. (1988). Comments on Doise's 'Individual and social identities in intergroup 
relations'. European Journal o
f
 Social Psychology, 18, 1 1 3-1 16. 
Turner, J.C. (1991). Social influence. Milton Keynes: Open University Press. 
Turner, J.e., & Brown, R.I. (1978). Social status, cognitive alternatives and intergroup 
relations. In H. Tajfel (Ed.), Diff
erentiation between social groups: Studies in the social 
psychology o
f
 intergroup relations. London: Academic Press. 
Turner, J.e., Brown, R.J., & Tajfel, H. (1979). Social comparison and group interest in 
intergroup favoritism. European Journal o
f
 Social Psychology, 9, 187-204. 
Turner, J.C., & Giles, H. (Eds.). (1981). Intergroup behaviour. Oxford: Blackwell; Chicago: 
University of Chicago Press. 
Turner, J.e., Hogg, M.A., Oakes, P.I., Reicher, S.D., & Wetherell, M.S. (1987). Rediscovering 
the social group: A self-categorization theory. Oxford: Blackwell. 
Turner, J.e., Hogg, M.A., Oakes, P.I., & Smith, P.M. (1984). Failure and defeat as 
determinants of group cohesiveness. British Journal o
f
 Social Psychology, 23, 97-1 1 1 .  
Turner, J.e., & Oakes, P.J. (1986). The significance of the social identity concept for social 
psychology with ref
erence to individualism, interactionism, and social influence. British 
Journal o
f
 Social Psychology, 25, 237-252. 
Turner, J.e., & Oakes, P.J. (1989). Self-categorization theory and social influence. In P.B. 
Paulus (Ed.), The psychology o
f
 group influence. Hillsdale, NJ: Erlbaum. 
Turner, J.C., Oakes, P.I., Haslam, S.A., & McGarty, e.A. (1994). Self and collective: 
Cognition and social context. Personality and Social Psychology Bulletin, 20, 454-63. 
Turner, J.e., Sachdev, I., & Hogg, M.A. (1983). Social categorization, interpersonal attraction 
and group formation. British Journal o
f
 Social Psychology, 22, 227-239. 
Turner, J.e., Wetherell, M.S., & Hogg, M.A. (1989). Referent informational influence and 
group polarization. British Journal o
f
 Social Psychology, 28, 135- 147. 
Tversky, A. (1977). Features of similarity. Psychological Review, 84, 327-352. 
Vander Zanden, J.W. (1960). The Klan revival. American Journal o
f
 Sociology, 65, 426-459. 
van Knippenberg, A. (1978). Status differences, comparative relevance and intergroup 
differentiation. In H. Tajfel (Ed.), Diff
erentiation between social groups: Studies in the social 
psychology o
f
 intergroup relations. London: Academic Press. 


Re
f
erences 
253 
van Knippenberg, A. (1984). Intergroup diff
erences in group perception. In H. Tajfel (Ed.), The 
social dimension: Developments in European social psychology, 
Vol. 
2. Cambridge: 
Cambridge University Press. 
van Knippenberg, A., & Wilke, H. (1988). Social categorization and attitude change. European 
Journal of Social Psychology, 18, 395-406. 
Verkuyten, M. (1990). Self-esteem and the evaluation of ethnic identity among Turkish and 
Dutch adolescents in the Netherlands. Journal o
f
 Social Psychology, 130, 285-297. 
Volpato, C., Maass, A., Mucchi-Faina, A., & Vitti, E. (1990). Minority influence and social 
categorization. European Journal o
f
 Social Psychology, 20, 1 19-132. 
Wetherell, M.S. (1987). Social identity and group polarization. In J.c. Turner, M.A. Hogg, P.J. 
Oakes, S.D. Reicher & M.S. Wetherell, Rediscovering the social group: A self
-categorization 
theory. Oxford: Blackwell. 
Wheeler, L., Reis, H., & Nezlek, J. (1983). Loneliness, social interaction, and sex roles. Journal 
o
f
 Personality and Social Psychology, 45, 943-953. 
Wilder, D.A. (1977). Perception of groups, size of oppositions, and social influence. Journal o
f
 
Experimental Social Psychology, 13, 253-268. 
Wilder, D.A. (1978). Homogeneity of jurors: The majority's influence depends upon their 
perceived independence. Law and Human Behavior, 2, 363-376. 
Wilder, D.A. (1990). Some determinants of the persuasive power of in-groups and out-groups: 
Organization of information and attribution of independence. Journal o
f
 Personality and 
Social Psychology, 59, 1 202-1213. 
Wilkes, A., & Tajfel, H. (1966). Types de classification et importance du contraste relatif. 
Bulletin du CERP, 15, 77-81 .  
Williams, J.A., & Giles, H .  (1978). The changing status of women in society: An intergroup 
perspective. In H. Tajfel (Ed.), Diff
erentiation between social groups: Studies in the social 
psychology o
f
 intergroup Behaviour. London: Academic Press. 
Williams, R.M. (1970). American society (3rd ed.). New York: Knopf. 
Winter, D.G. (1973). The power motive. New York: Free Press. 
Wish, M. (1976). Comparisons among multidimensional structures of interpersonal relations. 
Multivariate Behavioral Research, 11, 297-324. 
Wish, M., Deutsch, M., & Kaplan, S.J. (1976). Perceived dimensions of interpersonal relations. 
Journal o
f
 Personality and Social Psychology, 33, 409-420. 
Wittenbraker, J. (1983). The etiology o
f
 intergroup discrimination: A developmental approach. 
Unpublished doctoral dissertation, University of North Carolina, Chapel Hill. 
Worchel, S., Coutant-Sassic, D., & Grossman, M. (1992). A developmental approach to group 
dynamics: A model and illustrative research. In S. Worchel, W. Wood & J.A. Simpson 
(Eds.), Group process and productivity. Newbury Park, CA: Sage. 
Worchel,. S., Coutant-Sassic, D., & Wong, F. (1993). Toward a more balanced view of conflict: 
There is a positive side. In S. Worchel & J.A. Simpson (Eds.), Con
flict between people and 
groups. Chicago: Nelson-Hall. 
Worchel, S., Grossman, M., & Coutant-Sassic, D. (1993). Minority influence in the group 
context: How group factors aff
ect when the minority will be influential. In A. Mucchi Faina 
& S. Moscovici (Eds.), Minority influence. Chicago: Nelson-Hall. 
Wright, S.c., Taylor, D.M., & Moghaddam, F.M. (1990). Responding to membership in a 
disadvantaged group: From acceptance to collective protest. Journal o
f
 Personality and 
Social Psychology, 58, 994-1 003. 
Wyer, R.S., & Podeschi, D.M. (1978). The acceptance of generalizations about persons, objects 
and events. In R. Revlin & R. Mayer (Eds.), Human reasoning. Washington, DC: Winston. 
Yamagishi, T. (1995). Social dilemmas. In K.S. Cook, G.A. Fine & J.S. House (Eds.), 
Sociological perspectives on social psychology. Boston: Allyn & Bacon. 
Zander, A. (1968). Group aspirations. In D. Cartwright & A. Zander (Eds.), Group d
ynamics 
(3rd ed.). London: Tavistock; New York: Harper & Row. 
Zander, A., Stotland, E., & Wolfe, D. (1960). Unity of group, identification with group, and 
self-esteem of members. Journal o
f
 Personality, 28, 463-478. 


254 
Social identity: international perspectives 
Zavalloni, M. (1973). Social identity: Perspectives and prospects. Social Science I
n
f
ormation, 
12, 65-9J. 
Zavalloni, M. (1983). Ego-ecology: The study of the interaction between social and personal 
identities. In A. Jacobson-Widdings (Ed.), Identity: Personal and socio-cultural. Stockholm: 
Almqvist & Wicksell International. 
Zavalloni, M., & Louis-Guerin, C. (1984). Identite sociale et conscience: Introduction a /'ego­
ecologique. Toulouse: Privat. 
Zebrowitz, L. (1990). Social perception. Buckingham: Open University Press. 
Zeegers, W. (1988). Andere tijden, andere mensen: De sociale re
presentatie van identiteit. 
Amsterdam: Uitgeveri
j Bert Bakker. 
Zeigarnik, B. (1927). fIber das Behalten von erledigten und unerledigten Handlungen. 
Psychologische Forschung, 9, 1-85. 
Ziller, R. (1964). Individuation and socialization: A theory of assimilation in large 
organizations. Human Relations, 1 7, 341-360. 
Zimbardo, P.G. (1969). The human choice: Individuation, reason, and order versus 
deindividuation, impulse, and chaos. In W.J. Arnold & D. Levine (Eds.), Nebraska 
Symposium on Motivation, Vol. 1 7. Lincoln: University of Nebraska Press. 
Zimbardo, P.G., & Leippe, M.R. (1991). T
he psychology o
f
 attitude change and social influence. 
New York: McGraw-HilI. 


Index 
Abrams, D., 82, 124-41, 154, 1 55, 166, 167, 
169, 177-8, 212, 223 
Abric, I.e., 13 
Adamopoulos, I., 17 
Aebischer, V., 144 
agency, 18, 19 
aggregate groups, 20-1 ,  132 
Allen, V.L., 2, 95, 108 
Allport, F.H., 1 80 
Allport, G.W., 16, 44, 81 
altercentrism, 218-19, 228 
Altman, 1., 1 17 
altruistic rationalization hypothesis, 193-4 
Amancio, L., 22 
Amish society, 96, 102 
Amnesty International, 97-8 
Anastasio, P.A., 1 96 
anchored identities, 20-3 
Anderson, B., 1 20 
Antoni, M., 226 
Apter, D.E., 100 
Ardener, S., 96 
Ardrey, R., 1 17 
Argyle, M., 55 
Arnal-Duchemin, M.J., 15 
Arr6spide, 1.1., 21 1-29 
Asai, M., 201, 213 
Asch, S.E., 36, 81, 104, 151 
assimilation effect, 4 
Asuncion, A.G., 143 
asymmetry phenomena in similarity 
judgments, 37-41, 42 
attraction see interpersonal attraction; social 
attraction 
auto-stereotypy, 7 
Ayestanin, S., 21 1-29 
Bachman, B.A., 196 
Back, K., 126 
Bakan, D., 18 
Ball, P., 169 
Ballachey, E.L., 1 12 
Ballato, S.M., 88 
Banaji, M., 42 
Bargh, I.A., 36 
Baron, R.S., 81 
Barsalou, L.W., 88, 90, 92 
Bar-Tal, D., 93-1 13 
Bar-Tal, Y., 109 
Basques, 1 23 
Beaufils, B., 13 
beliefs, group, 93-1 13 
Bern, D.I., 101, 141 
Bern, S.L., 20 
Billig, M., 24, 32, 105, 1 10, 1 19, 145, 1 52, 
181 
Black Panthers, 107-8 
'black sheep eff
ect', 1 29-31, 150 
Blaine, B., 222 
Boltanski, 1.L., 37 
Bonaldo, E., 170 
Bond, M.H., 219, 225 
Bontempo, R.N., 17, 201 , 213 
Bornstein, G., 183, 184 
Bourhis, R.Y., 218 
Bovarism, 219 
Boyes-Braem, P., 75 
Brewer, M.B., 36, 41, 45, 47, 76, 80, 83, 87-8, 
90, 104, 130, 181, 1 96 
Brito, P.N., 133 
Broadnax, S., 222 
Brosch, M., 109 
Brown, R., 1 32, 1 80 
Brown, R.I., 46, 83, 125, 1 27, 1 55, 167, 168, 
170, 171, 172, 175, 176, 177, 182, 202-3, 
204, 2 1 1 ,  214 
Bruner, I.S., 31, 128 
Bundy, R.P., 24, 105, 145 
Burnstein, E., 81 
Butera, F., 149, 151 
Byrne, D., 45, 47, 154 
Caddick, B., 168, 171-2, 175 
Campbell, D.T., 77, 104, 1 15, 1 19, 180, 
188 
Camps, V., 200 
Cannon, T., 107-8 
Cantor, N., 31-2, 41 
Capozza, D., 170 
Carroll, D.I., 159 


256 
Social identity: international pers
pectives 
Cartwright, D., 98, 125, 126 
Carugati, F., 13 
Carver, C., 226 
categorial diff
erentiation see differentiation, 
intergroup 
categorization, 4-5, 32 
and intergroup competition, 180-97 
recategorization and the discontinuity 
eff
ect, 196-7 
see also self-categorization; self­
categorization theory; social 
categorization 
Cavanni, L., 158 
centrality of group beliefs, 101-2 
Chaiken, S., 188 
Chang, 1.1., 1 59 
CIC theory see social identity theory 
Clack, F.L., 201 
Clark, R.D., 70, 142, 143 
Clemence, A., 13, 25 
Cochrane, S., 82 
Codol, I.-P., I, 9, 13, 14, 27, 28, 34, 35, 47, 
48 
cognitive centrism, 9 
cognitive creativity, 219 
cognitive differentiation eff
ect, 4 
cognitive stereotype effect, 4 
collection groups, 20-1 ,  1 32 
collectivism, 201-2, 203, 204, 205-6, 212-14, 
214, 225, 228 
Comby, L., 214 
common knowledge, in personal identity, 
14-17, 19-20 
communion, 18, 19 
communist ideology, \00 
'comparative fit' principle, 79, 127, 128 
competition and conflict 
intergroup, 5, 66, 67, 68-9, \06-7, 176, 
180-97 
interpersonal, 181-2, 192-3, 213 
concentricity thesis, 121, 122-3 
Condor, S., 41, 163, 172 
confidence in group beliefs, 101, l lO-l l  
Consumer Federation of America, 99 
context, dependence of categorization on, 
75-92 
contrast or cognitive diff
erentiation effect, 
4 
Converse, P.E., I I I  
Cooper, 1., 82 
Cooper-Shaw, L., 84 
cooperation between groups, 68-9, 106-7, 
168, 176 
Cordova, D., 222 
Correia, M.G., 133 
couples, relationships between see 
interpersonal relationships 
Coutant, D., 71 
Coutant-Sassic, D., 57, 69, 71, 72 
covariation hypothesis, xviii, 8-l l ,  26, 27, 30 
Cowan, G., 16 
Crano, W.o., 71 
Crocker, 1., 212, 221 ,  222, 223, 226 
Crook, K.F., 172-3 
Crosby, F., 222 
Crutchfield, R.S., 1 12 
Csepeli, G., 83 
culture, 56, 1 1 7-18, 120-1, 199-229 
Cylwik, A., 51 
Damon, W., 14 
Dann, H.D., 83 
Dardis, G.I., 181 
Darley, I.M., 180 
David, B., 79 
Davidson, B., 82 
Deaux, K., 160 
decay of groups, 61-2, 63, 73 
DeCharms, R., 126 
Deconchy, I.P., 101 
delinquency, adolescent, 22-3 
Delumeau, 1.1., 124 
depersonalization, 7, 77, 84, 91, 120 
Deschamps, I.-C., xviii, I -l l ,  25, 26, 27, 30, 
55, 127, 214 
Deutsch, M., 17, 131, 148 
De Vader, C.L., 88 
deviates, derogation of, 124-5, 132-4, 
1 35-41 
Devos, T., l - l l ,  214, 220 
Diamond, M., 97 
Diehl, M., 181 
diff
erentiation, 27-30 
and groups beliefs, 95, \08 
intergroup, 3, 4, 5-7, 8-l l ,  14, 24, 25-6, 
77, 121-2, 126-41 passim, 199-210, 
203 
intragroup, 14, 25-6, 124-41 
and the meta-contrast principle, 77-80 
and national identity, 120-1 
see also covariation hypothesis; 
prototypicality 
DiMaggio, A., 170 
discontinuity effect, 180-2, 187-97 
dissociation theory, 148-51, 1 52, 153 
Doise, W., 4, 9, 1 3-23, 26, 27, 55, 83, 
127, 132, 142, 143, 147, 195, 213, 
214, 215 
dominant groups, 20, 26, 27, 132, 214, 
228 


dominated groups, 20, 26, 27, 132, 214-20, 
222, 223, 228 
Doob, L.W., 1 17 
doomsday cult, 59 
Doosje, B., 132 
Doukhobors, 96 
Dovidio, J.F., 80, 196 
Drigotas, S.M., 180-97 
Dull, V., 80 
Dumont, L., 199 
Dunbar, R., 120 
Dunning, D., 85 
Durand Delvigne, A., 1 3, 21-2 
Durkheim, E., 180 
Duveen, G., 13 
Eagly, A.H., 132, 188 
Echebarria, A., 122 
ego and the self, 44 
egocentrism, 9, 52 
Eiser, J.R., 25 
Elejabarrieta, F., 122 
Ellemers, N., 1 32, 178, 213 
Ely, P.G., 166-79, 203, 204, 21 1 
Emler, N., 6, 22-3 
emotional experience and individualist! 
collectivist culture, 202 
'entitativity', 180, 183, 203 
Epstein, A.L., 103 
Epstein, S., 44 
Erikson, E., 14, 18, 44 
Etaugh, C., 1 56 
ethnicity, 103, 1 18, 120-1 
ethnocentrism, 1 1 5, 121 
ethnocentrism theory, 1 19 
Etzioni, A., 98 
Fabian, A. 
K., 1 17 
failure of groups, 64, 65 
Falger, V., 1 1 5  
Falomir, J.M., 152 
Farr, R.M., 13 
fascist ideology, 100 
fear-based competition, 194-6 
Fells, J., 122 
Festinger, L., 59, 106, 1 10, 126 
Fierres, R., 1 52 
Fiore, M., 1 57 
Fiske, S.T., 23, 36, 76 
Flament, C., 24, 105, 145 
Forgas, J.P., 158-9 
Forsyth, D.R., 129, 141 
Foti, R.J., 88 
Fox-Cardamone, L., 166-79 
Fraser, S.L., 88 
Index 
Fromkin, H.L., 45, 47 
fundamental group belief, 94, 103-6 
Gachoud, J.P., 142 
Gaelik, L., 28 
Gaertner, S.L., 80, 196 
Gagnon, A., 218 
Gara, M., 164 
Garza, R.T., 223, 224, 227 
gender, 19, 20, 21-2, 26, 143 
Gerard, H.B., 1 10, 126, 148 
Gergen, K.J., 3 
Giles, H., 76, 155, 169 
Gillin, J., 97 
globalization, 1 14, 123 
goals, group, 98-9 
Goethals, G.R., 1 52 
Goffman, E., 165 
Goldberg, L.R., 16, 17 
Gordon, P.c., 28, 32, 48 
Gordon, S.L., 1 56 
Gotab, A., 46 
Gouge, c., 83 
257 
Graduated and Reciprocated Initiatives in 
Tension-reduction (GRIT), 196 
Graetz, K.A., 181 
Gray, W.D., 75 
greed-based competition, 193-4 
Greenwald, A.G., 42, 44 
Grossman, M., 57, 69, 71 
group beliefs, 93- 1 1 3  
group concept, xviii 
group identity, 53-123 
developmental view of, 56-73 
group polarization, 81-3, 91 
group productivity, 60, 63, 67, 68 
Gudkyunst, W., 225 
Guimond, S., 222 
Haberkorn, G., 143 
Halverson, C.F., 16 
Hamilton, D.L., 132 
Hamilton, V.L., 124, 141 
Hampson, S.E., 89 
Hantzi, A., 24 
Hardie, E.A., 84-5 
Hardoin, M., 34, 48 
Harkins, S., 61, 180 
Harring, K., 181, 185, 187, 191 
Harrison, A.A., 156 
Hart, D., 14 
Harter, S., 19 
Harvey, O.J., 54, 107, 196 
Haslam, S.A., 31, 75-92, 127, 128 
Hayes, C.H., 92, 120 


258 
Social identity: international pers
pectives 
Henderson, M., 144 
HERMES project, 200 
Herringer, L.G., 223, 224, 227 
Hertz, F., 1 1 5  
Hewstone, M., 24, 144, 169 
Higgins, E.T., 102 
Hill, C.T., 156 
Hinkle, S., 166-79, 202-3, 204, 21 1-12 
Ho, H., 196 
HolTman, C.D., 16 
HolT
man, M.L., 46 
Hofstede, G., 56, 200 
Hogg, M.A., 76, 79, 82, 83, 84-5, 86, 88, 105, 
1 19, 125-31 passim, 139, 146, 1 54, 1 55, 
162, 166, 167, 1 77-8, 2 1 1  
Holyoak, K.J., 28, 32, 48 
Holzworth, D.W., 84 
Homans, G., 96 
Hood, W.R., 54, 107, 196 
Hopkins, N., 6 
Horwitz, M., 180 
Hostetler, J.A., 96, 102 
Hoyle, R.H., 181, 1 93, 1 95 
Hu, L., 214 
Huguet, P., 149, 1 5 1  
Hui, C.H., 200-1 
Huici, C., 154, 155, 215 
human identity, 31, 77 
Hurtig, M.C., 21, 37-8 
identification stage of group development, 
58-60, 62, 68, 73 
ideology, 99-101 
Imago theory, 18 
INDCOL scale, 201 
independence, 54-6 
individualism, 199-210, 212-14, 214, 225, 
228 
individualization, 35, 215 
individuation stage of 
group development, 
60-1, 63, 73 
influence, 126 
informational, 1 26, 1 34, 1 39-41 
of majorities, 142 
of minorities, 70-2, 73, 147-8 
normative, 1 26, 1 34 
and social categorization, 142-53 
Inoff, G., 16 
Insko, C.A., 180-97, 203, 209 
InstIa, P., 2 1 1 -29 
interest groups, 99 
intergroup relationships, 66-9, 73 
competitive, S, 66, 67, 68-9, 106-7, 176, 
180-97 
cooperative, 68-9, 106-7, 168, 176 
interpersonal relationships as, 1 54, 
1 55-65 
use of new comparison dimensions in, 
166-79 
interpersonal attraction, 1 54 
and group behaviour, 83-4, 85, 91 
interpersonal relationships, S, 6, 55-6 
competition and conflict in, 181-2, 192-3, 
213 
as intergroup relationships, 154, 1 55-65 
Jackson, L.A., 178 
James, W., 2, 18, 44 
Janis, I.L., 126 
Jarymowicz, M., 44-52 
Jaskar, K., 222 
Jehovah's Witness group, 107 
Jensen, M.A.C., 57 
Jodelet, D., I 3  
Johnson, D.M., 75 
Johnston, L., 24 
Jones, E.E., 69, 1 10, 126, 134 
Jovanovic, M., 146 
Jung, C.G., 18 
Kaiser, c., 143, 144 
KaminolT, R., 1 17 
Kaplan, S.J., 17 
Karytowski, J., 45 
Kastersztein, J., 168 
Kedourie, E., 1 15 
Kelley, H.H., 1 54, 183, 187, 192 
Kelly, J., 54 
Kelman, H.C., 120 
Kenny, D.A., 1 29 
Kiesler, C.A. and Kiesler, S.B., 81 
Killian, L.M., 103 
Kinder, D.R., 222 
King, G., 102 
Kirchler, E., 1 5  
Kitayama, S., 201 ,  202 
Kleck, B.E., 1 93 
Klein, M., 1 8  
Kluckholn, c., 97 
Kobuszewska, E., 52 
Kohlberg, L., 46 
Kohn, H., 1 14, 1 1 5 
Kotkin, M., 1 56 
Krech, D., 1 12 
Kruglanski, A.W., 101, 109 
Ku Klux Klan, 98 
Kuhn, M.H., 14 
Lalonde, R.N., 171 
Lane, R.E., 100 


Index 
259 
language, 1 17-18 
Larsen, K., 83 
Lassaline, M.E., 75 
Latane, B., 61, 180 
La Voie, L., 129 
Lea, M., 82 
leadership, 88, 178 
LeBon, G., 180 
L'Ecuyer, R., 14 
Lee, J., 196 
Lee, S., 82 
legitimacy/illegitimacy, 167, 170-2, 175-6 
Legrenzi, P., 149 
Leippe, M.R., 141 
Lemaine, G., 151, 168, 169, 171, 174, 176, 
177 
Leung, K., 201 
Levine, J.M., 57, 125, 126 
Le Vine, R.A., 1 19 
Levy, S.B., 102 
Lewin, K., 54 
Leyens, J.-P., 17, 129, 214, 215 
Lindskold, S., 196 
Linville, P.W., 69 
Linz, J.J., 1 23 
Lippitt, R., 54 
Llavata, E., 152 
Lloyd, B., 13 
Lodewijkx, H.G.M., 1 92 
L6pez-Saez, M., 1 99-210 
Lord, R.G., 88, 90 
Lorenzi-Cioldi, F., 9, 10, 1 3, 1 5, 1 9, 20-1 , 22, 
26, 1 32, 214 
Lott, A.J.and Lott, B.E, 83, 131 
Louis-Gueerin, c., 1 3  
Low, S.M., 1 17 
Lubovitcher Hassidim, 102 
Lucca, N., 201 ,  213 
Luhtanen, R., 212, 221 ,  222, 226 
Lui, L., 80, 87 
Luria, A.R., 27 
Maass, A., 70, 142, 143 
Mackie, D., 82, 143, 144, 146, 2 1 1  
majorities, influence of, 142 
Makris-Botsaris, E., 1 9  
Mann, J., 80, 196 
Maoris, 59 
Maras, P., 176 
March, J.G., 98 
Marcia, J.E., 14 
Mariategui, J.c., 1 1 8  
Markus, H., 1 8, 47, 48, 201 , 202 
Marques, J.M., 124-41, 1 50, 215, 220 
Martin, R., 143, 144 
Martinez-Taboda, M., 21 1-29 
Maslow, A.H., 54 
Mazzara, B., 36 
McAdams, D.P., 18, 19 
McCallum, D.M., 181, 1 87 
McCusker, C., 201 
McDougall, W., 180 
McGarty, C., 77, 79, 82, 86, 90, 91, 92, 127, 
139 
McGrath, J.E., 54 
McGuire, W.J. and McGuire, C.V., 14, 23 
McPartland, T.S., 14 
McPhail, C., 180 
Mead, G.H., 2, 47 
Medin, D.L., 75, 77, 88-9, 89-91, 92 
Mentzer, S.J., 193 
Mervis, C.B., 75 
Messick, D., 21 1 
meta-contrast principle, 77-80, 91, 120-1, 
127, 128, 1 36-7, 138 
Meyer, G., 55, 127 
Miller, N., 196 
Miller, R.L., 47 
minimal group paradigrn, 1 10, 1 82-3 
minorities, 142, 166 
influence of, 70-2, 73, 147-8 
Mischel, W., 31-2, 41 
Mlicki, P.P., 213 
Moghaddam, F.M., 178, 219 
Moise, L.C., 218 
Montei1, J.M., 35 
Montero, M., 218 
Moore, M., 156 
morale, 178 
Morales, J.F., 164, 199-210, 215 
Moreland, R.L., 57, 104, 105, 125, 126 
Mormons, 102 
Morris, W.N., 56 
Moscovici, S., 13, 70, 142, 149 
motivation, 24-5, 28, 178 
Moya, M.C., 1 54-65 
Mucchi-Faina, A., 142 
Mugny, G., 70, 125, 142-53 
Mullen, B., 214 
Multiple Alternative Matrices (MAMs) 
experiments, 183-6 
Murphy, G.L., 89 
Murrell, A., 80, 196 
Mussolini, B., 100 
'myths of descent', 1 18 
Nakbi, J.L., 1 5  
Nascimento-Schulze, C.M., 215, 221 
nation-state, 1 14-15, 1 16, 1 1 8-19, 121 
national identity, 1 14-23 


260 
Social identity: international pers
pectives 
National Rifle Association, 99 
National Right to Life Committee, 99 
Navarro, E., 146 
needs, satisfaction of, 104 
Neisser, D., 75, 92 
Nemeth, c., 70, 72, 149 
Neuberg, S.L., 23, 76 
Nezlek, J., 56 
T
he Night Porter (film), 1 58 
Niznik, J., 1 22 
'normative fit' principle, 79, 91, 127-8 
normative influence, 126, 1 34 
norms, group, 96, 106, 1 33-4 
denotative, 127-9, 1 34, 1 35, 1 36-7 
prescriptive, 129, 134, 1 35, 1 36-7, 138 
Oakes, P.J., 31, 32, 33, 75-92, 1 19, 1 27, 1 28, 
146, 162, 2 1 1  
Odbert, H.S., 16 
O'Dea, T.F., 102 
Oldenquist, A., 55 
Onorato, R., 86 
Ortony, A., 90 
Osgood, C.E., 1 96 
Ostell, A., 83 
Ostrom, T.M., 8, 104 
'outgroup homogeneity effect', 8 
Paez, D., 1 24-41, 2 1 1 -29 
Pages, R., 2 
Paicheler, H., 1 3  
Palmonari, A., 1 3, 1 5  
Papastamou, S., 143, 144 
paralysis, ingroup, 146, 148 
Parelman, A., 1 56 
Park, B., 35, 69 
Parsons, T., 97 
particularization, 32, 35 
Pavelchak, M.A., 23, 36 
Pawluczuk, B., 5 1  
Peabody, D., 1 6  
Pemberton, M.J., 196 
Peplau, L.A., 1 55, 156, 162 
Perez, J.A., 1 25, 142-53 
Perie, M., 85 
persecution, collective, 1 24-5 
personal attraction see interpersonal 
attraction 
personal identity, 24-5, 31, 47-9, 54-6, 76, 
77, 160, 228 
common knowledge in, 14-17, 19-20 
and social identity, 1-1 1 , 213 
as a social representation, 1 3-23 
personal relationships see interpersonal 
relationships 
Personnaz, B., 168 
Pfost, K.S., 157 
Phinney, J.S., 222 
Piaget, J., 46 
Pichevin, M.F., 21, 37-8 
Pinkley, R.L., 181, 1 85, 187, 191, 193, 1 95 
Piolat, M., 37-8 
Pizzigati, K.A.L., 16 
Podeschi, D.M., 32 
polarization, group, 81-3, 9 1  
political participation, 222, 229 
Pombeni, M.L., 15 
positive self-identity/evaluation, 6, 24-5, 
181 
see also self-esteem, personal 
positive social identity, 93, 121-2, 221 
and manifest influence, 148 
use of new dimensions of intergroup 
comparison to attain, 166-79 
see also self
-esteem, collective 
Pratkanis, A.R., 44 
Pratto, F., 36 
primus inter pares (PIP) effect, 27 
see also superior conformity of the self 
productivity, group, 60, 63, 67, 68 
Proshansky, H.M., 1 17 
prototypicality, 30-1, 32, 33-5, 37, 38, 41-2, 
75-92, 127-9, 134 
Pryor, J.B., 104 
Quattrone, G., 69 
Questionnaire o
f
 Social Perception, 48 
Quevedo, S., 212 
Rabbie, J.M., 1 80, 1 82, 1 90 
Ramirez, S., 123 
Rauma, D., 124, 141 
Ravaud, J.F., 13, 14-15 
Raviv, A., 109 
realistic conflict theory, 1 88 
recategorization, and discontinuity eff
ect, 
1 96-7 
referent informational influence, 1 39 
regionalism, 1 16, 1 17, 121 
Reicher, S., 22-3, 76, 1 19, 1 27, 146, 162, 
21 1 
Reis, H.R., 56, 1 96 
relationships see intergroup relationships; 
interpersonal relationships 
Reykowski, J., 45, 46 
Reynolds, Y., 92, 1 1 5  
Ricci Bitti, P., 1 3  
Richard III Society, 99 
Riecken, H.W., 59 
Riley, S., 156 


Index 
261 
Rips, L.J., 90 
Rivera, A.N., 222 
Robalo, E.M ., 130 
Robbers' Cave experiments, 106-7 
Robinson, P.W., 19 
Rocha, S.A., 130 
Rodriguez, C.P., 214 
Rogers, C., 54 
Rokeach, M., 96 
Rook, D.S., 56 
Rosch, E., 30-1, 41, 75, 76, 77, 80, 87 
Rosenberg, S., 164 
Ross, G.F., 168, 170, 171, 175 
Roth, E .M., 89 
Rothbart, M., 35, 69 
Roux, P., 25 
Rubin, Z., 1 56 
Rucker, M.H., 156 
Rush, M.C., 88 
Russell, CT., 107 
Russell, J.E.A., 88 
Rust, M.C., 196 
Rutherford, J.F., 107 
Sachdev, I., 83, 105 
Salazar, J .M., 1 14-23, 219, 224 
Salisbury, R.H., 99 
Sanchez-Mazas, M., 146, 148, 150 
Sande, G.N., 16 
Sanders, G.S., 81 
Sarbin, T.R., 2 
Sarchielli, G., 13 
Schachter, S., 54, 59, 1 26 
Schadron, G., 214 
Schei̖r, M., 226 
schematic traits, 47 
Schoben, E.J., 89 
Schopler, J., 180-97, 203, 209 
Sears, D.O., 222 
Sedikides, C., 8 
self 
and the ego, 44 
fluidity of the, 3 
as object, 44 
and prosocial attachment, 44-6 
social, 44-6 
as subject, 44 
superior conformity of the, 9, 16, 27 
self-categorization, levels of, 14, 31, 36, 
146 
self-categorization theory, 7-8, 30-3, 34, 35, 
65, 93-4, 105 
and idea of depersonalized social 
attraction, 84 
and group polarization, 8 1-3 
prototypicality in, 30-1, 33, 34, 76-81 ,  
85-6, 91, 127-9 
and social influence, 147-8 
self-consciousness, 212, 226-8, 229 
self-definition, 7-8 
self-esteem 
collective, 212, 220-3, 229 
personal, 6, 19, 20, 21 1-12, 220-3, 229 
self-other comparison, 24-42 
see also differentiation; similarity 
relationships 
self-permanence, 3 
self-presentation, 177-8 
self-representation, duality of, 2 
self-stereotyping, 126-7, 1 28, 131 
self structure, 45 
self-we-others schemata, 44-52 
Semin, G.R., 17 
Serino, C., 24-42 
Serooio, R.G., 133 
Seton-Watson, H., 1 1 5  
Sewell, D.R., 88 
Shaw, M.E., 126 
Sherif, CW., 54, 107, 196 
Sherif, M., 5, 54, 69, 81, 83, 106-7, 155, 196, 
197 
Shils, E., 100 
Sills, D.L., 98 
Silver, M., 83 
similarity relationships, 3, 4, 6-8, 14, 27-30, 
33-5, 44-5, 47, 89-90 
asymmetries in, 37-41, 42 
and group beliefs, 95, 108 
intergroup, 14, 14-17, 20-2 
and the meta-contrast principle, 77-80 
metrical approach to, 28-9 
and national identity, 120-1 
see also covariation hypothesis; 
prototypicality 
Simmel, G., 125 
Simon, B., 132 
Simon, H.A., 98 
Sinclair, A., 215 
Singh, A., 22 
Singleton, R., 81 
Smith, A.D., 1 1 5, 1 1 8  
Smith, C., 214 
Smith, J., 47 
Smith, P.B., 219, 225 
Smith, P.M., 83 
Smith, V.A., 181 
Snyder, CR., 45, 47, 193 
Sobczyk, K., 52 
social attraction, and prototypicality, 83-5, 
91 


262 
Social identity: international pers
pectives 
social categorization, 4-8, 105 
context dependence of, 75-92 
deductive effect of, 1 5  
graded structure of, 75-92 
productive character of, 31-2 
and social influence, 142-53 
social comparison see diff
erentiation; self­
other comparison; similarity 
relationships 
social constructivism, and latent influence, 
149 
social control, and intragroup differentiation, 
1 24-41 
social creativity, 166-79 
social identity theory, 4-8, 55, 65, 93, 105, 
1 19, 167-9, 181-6 
social influence see influence 
social loafing, 61 
social representations, in personal identity, 
1 3-23, 213 
social self, 44-6 
sociocentrism, 9 
Sokolowska, G., 52 
Spears, R., 82, 132 
Srull, T.K., 28 
stability/instability, 167, 170, 171, 175-6 
Stahelski, A., 1 87 
state, 1 14-1 5, 1 16, 1 18-19, 1 20 
stateless nationalism, 1 16 
status, 167-8, 170-2, 175-6, 212-23, 228-9 
Staw, B.M., 70 
Stephenson, G.M., 1 57 
stereotypes, 4, 7, 1 5, 2 14, 215 
see also self-stereotyping 
Stewart" 1 32 
Story, A.L., 85 
Stotland, E., 103 
Strenta, A., 193 
Stroup, H.H., 107 
subjective group dynamics, 131-2, 1 39-41 
success of groups, 64-5 
Suls, 1.M., 47 
superior conformity of the self, 9, 16, 27 
supra-nationalism, 1 16, 1 18, 121, 122 
symbolic individual mobility strategy, 219, 
220 
Szuster-Zbrokewicz, A., 51, 52 
t-group movement, 57 
Tajfel, H., xviii, 4, 5, 6, 26, 41, 55, 77, 105, 
1 19, 1 25, 126, 127, 1 3 1 , 145, 1 55, 166, 
168, 169, 181 
on legitimacy/stability of 
groups, 167, 171, 
175-6 
minimal group paradigm, 1 10, 1 82-3 
on minority groups, 103 
on national and ethnic identity, 1 14 
on similarity/difference judgements, 24, 25, 
29 
on temporal aspects of social identity, 163 
Tap, P., 1 3, 19, 46 
Taylor, D.M., 178, 219 
Taylor, L.A., 166-79 
temporal aspects of social identity, 163, 176-7 
territoriality, 1 17, 120-1 
Thibaut, 1.W., 106, 1 54, 183, 192 
Toch, H., 100, 109 
Torregrosa, 1.R., 123 
Tougas, F., 222 
Trans-Species Unlimited, 98-9 
transcendence, and national identity, 120, 121 
Triandis, H.C., 56, 201, 203, 213 
Triska, J.F., 100 
Truman, D.B., 99 
Tuckman, B.W., 57 
Turner, J.e., xviii, 5, 14, 24, 34, 35, 36, 41, 55, 
65, 75-92, 93, 105, 1 19, 1 26, 128, 1 32, 
1 39, 146, 163, 166, 167, 169, 181, 1 84, 
1 85, 203, 21 1 
on comparative fit, 127 
on group polarization, 81, 82, 83 
on influence of groups, 147-8 
on meta-contrast principle, 77, 79, 80, 1 27, 
162 
and the minimal group paradigm, 9, 182 
on normative fit, 1 27-8 
on positive self-identity, 6 
and prototypicality theory, 30, 75-92, 127 
on referent informational influence, 139 
on self-definition, 7-8 
see also self-categorization theory 
Tversky, A., 27-8, 29, 34, 41, 77 
Valencia, J., 122 
values, 96-8, 212, 220-3, 225, 226 
Vander Zanden, J.W., 98 
van Knippenberg, A., 83, 168, 171, 174, 176, 
177, 178, 215 
Vega, L., 199-210 
Verkuyten, M., 1 9  
Villareal, M., 1 22, 201, 2 1 3  
Vine, I., 1 1 5 
Vitti, E., 142 
Volpato, C., 9, 10, 142, 144 
Wattenmaker, W.O., 89 
Wetherell, M.S., 76, 79, 81, 82, 83, 1 19, 127, 
146, 162, 2 1 1  
Wheeler, L., 56, 1 96 
White, B.J., 54, 107 


White, R., 54 
Wicklund, R., 193 
Wilder, D.A., 95, 108, 143, 144, 145-6 
Wilke, H., 83, 178, 215 
Wilkes, A.L., 25, 29, 127 
Williams, J., 1 55, 172 
Williams, K., 61, 180 
Williams, R.M., 97 
Winter, D.G., 18 
Wish, M., 17 
Wisniewski, E.J., 75 
witch-hunting, 124 
Wittenbraker, J., 184 
Wolfe, D., 103 
Wood, G., 69 
Worchel, S., 53-73 
Index 
Worth, L.T., 143 
Wright, S.c., 178, 219 
Wyer, R.S., 32 
Yamagishi, T., 124 
Yzerbyt, V.Y., 129, 214 
Zander, A., 98, 103, 125 
Zavalloni, M., 13, 25, 26 
Zebrowitz, L., 219 
Zeegers, W., 13, 16 
Zeigamik, B., 64 
Ziller, R., 46 
Zimbardo, P.G., 141, 180 
Zuckier, H., 36 
263 


Social Identity, 
Intergroup Conflict, and 
Conflict Reduction,
Volume 3
Richard D. Ashmore
Lee Jussim
David Wilder,
Editors
OXFORD UNIVERSITY PRESS


SOCIAL IDENTITY,
INTERGROUP CONFLICT, and
CONFLICT REDUCTION


RUTGERS SERIES ON SELF AND SOCIAL IDENTITY
Series Editors
Richard D. Ashmore and Lee Jussim
Rutgers, The State University of New Jersey
Editorial Board
Kay Deaux, City University of New York
Kenneth J. Gergen, Swarthmore College
Dorothy C. Holland, University of North Carolina at
Chapel Hill
Daniel M. Ogilvie, Rutgers, The State University of
New Jersey
Suzanne Ouellette, City University of New York
Seymour Rosenberg, Rutgers, The State University of
New Jersey
Diane N. Ruble, New York University
Ronald T
aylor, T
emple University
Peggy A. Thoits, Vanderbilt University
David Wilder, Rutgers, The State University of New
Jersey
Volume 1
Self and Identity: Fundamental Issues
Volume 2
Self, Social Identity, and Physical Health: Interdisciplinary 
Explorations
Volume 3
Social Identity, Intergroup Conflict, and Conflict Reduction


RUTGERS SERIES ON SELF AND SOCIAL IDENTITY
Volume 3
SOCIAL IDENTITY,
INTERGROUP CONFLICT,
and CONFLICT REDUCTION
Edited by
Richard D. Ashmore
Lee Jussim
David Wilder
1
2001


3
Oxford
New York
Athens
Auckland
Bangkok
Bogotá
Buenos Aires
Cape T
own
Chennai Dar es Salaam Delhi Florence Hong Kong Istanbul
Karachi
Kolkata
Kuala Lumpur Madrid Melbourne Mexico City Mumbai
Nairobi
Paris São Paulo Shanghai Singapore T
aipei T
okyo T
oronto Warsaw
and associated companies in
Berlin Ibadan
Copyright © 2001 by Oxford University Press
Published by Oxford University Press, Inc.
198 Madison Avenue, New York, New York 10016
Oxford is a registered trademark of Oxford University Press.
All rights reserved. No part of this publication may be reproduced, 
stored in a retrieval system, or transmitted, in any form or by any means, 
electronic, mechanical, photocopying, recording, or otherwise, 
without the prior permission of Oxford University Press.
Library of Congress Cataloging-in-Publication Data
Social identity, intergroup conflict, and conflict reduction / [edited by] Richard D. Ashmore,
Lee Jussim, and David Wilder.
p. cm.—(Rutgers series on self and social identity ; 3)
Includes bibliographical references and index.
ISBN 0-19-513742-6; 0-19-513743-4 (pbk.)
1. Group identity. 2. Conflict management. 3. Ethnic relations. I. Ashmore, Richard D. II.
Jussim, Lee J. III. Wilder, David, 1949–
IV. Series
HM753.S6197
2001
302.4—dc21
00-053761
9 8 7 6 5 4 3 2 1
Printed in the United States of America
on acid-free paper


Preface to Volume 3
This volume is important to me both professionally and personally: it is a
significant part of my evolving life story. At the same, my co-editors and I
believe that this book provides a start toward an important new perspective
on intergroup violence, a problem that brings extensive material destruc-
tion and widespread human suffering around the globe.
In the spring of 1964, while still an undeclared and uncertain history ma-
jor, I read Gordon Allport’s The Nature of Prejudice for a course entitled,
“Medieval Anti-Semitism.” It was an epiphany—I would become a social
psychologist, and I would seek to understand the causes and cures of inter-
group conflict. That summer I enrolled in my first psychology classes, fin-
ished the psychology major during my senior year, and went on to get my
Ph.D. in social psychology with a specialization in intergroup relations.
When I read Allport and chose my occupation and specialty area, I did
not know Erikson’s seminal work on identity development. As I look back,
though, his ideas about adolescent identity search fit my experience very
well. Becoming a social psychologist and seeking to understand prejudice
and discrimination allowed me to bring together my childhood identifica-
tions (with two parents who were simply good people, with a succession of
outstanding teachers [from Mrs. Cirino in the fourth grade to Professor
Lewis Spitz at Stanford who, to this day, provides the model for how I teach
a large undergraduate class]), my then current concerns and values (I be-
lieved in 1964 that conflict among groups was the greatest problem facing
humans; I felt that we should and could do something about this), and my
future work (becoming a scientist and uncovering why people engage in
group-based harm-doing).


Now it is 2000. With many twists and turns, I today still see myself as an
intergroup relations researcher. And I continue to believe that intergroup
conflict is the major problem facing the world. What is different, one of
those twists and turns, is that I now consider myself also an identity re-
searcher. This volume allows me to bring together, then, my personal life
story, especially my commitment to understanding intergroup conflict, with
my evolving professional self-definition as a social scientist who studies self
and identity.
Although the world today is, in many ways, quite different from that of
almost 40 years ago, group-based violence, aggression, and suppression are
no less in evidence. Although the global cold war has ended, we now ex-
perience a seemingly never-ending succession of small and local “wars.”
These are global in a new way—group-based strife occurs on all continents
and every corner of the world and yet, because of powerful new communi-
cations technologies, the whole world often is involved if only as CNN
viewers.
Since I entered the field in the mid-1960s, the social and behavioral 
sciences have made great strides in seeking to understand and ameliorate
intergroup conflict. At the same time, there are still many gaps in our
knowledge.
My co-editors and I believe that social identity may provide a powerful
perspective for filling in some of these gaps. Social identity has certainly
been implicated in past work on group strife, but these efforts have been
piecemeal and isolated from one another across disciplinary boundaries.
Our goal in this volume was to bring together scientists from diverse social
and behavioral sciences to address the multiple ways that social identity and
intergroup conflict are interrelated. We believe that our contributors have
done an excellent job, on the one hand, of highlighting major current theo-
ries, concepts, and methods and, on the other, of identifying the crucial 
issues that need to be addressed if we are to develop a more complete and
coordinated social identity perspective on intergroup conflict. We hope that
this volume contributes to the development of such a paradigm and that it,
along with other approaches, facilitates social technologies that significantly
reduce the incidence and intensity of people harming others because of
group membership.
Piscataway, New Jersey
Richard D. Ashmore
May 2000
vi
Preface to Volume 3


Acknowledgments
Many individuals and organizations helped us as we planned and carried out
the Third Rutgers Symposium on Self and Social Identity and produced
this volume. We thank you all. We received so much help from so many
sources that we are sure to have forgotten to mention some here, and for
that we apologize.
We first thank the symposium speakers/chapter authors. They were 
chosen because they each have distinguished scientific careers, and that past
behavior predicted their performance in this endeavor. They produced
high-quality symposium presentations and excellent chapters appearing in
this book. In addition, they had to deal with fairly picky editors. We thank
each of you for sticking with us and producing the present polished contri-
butions.
Next, we wish to thank our publisher, Oxford University Press, and our
editors there, Joan Bossert, Philip Laughlin, and Catharine A. Carlin. Joan
has been an invaluable source of guidance and support since she first re-
ceived the proposal for the Rutgers Series, and Philip and Catharine ably 
assisted us as we moved the present volume from an idea, to a conference,
to a book.
There are many people to thank for sharing their ideas about the topic of
social identity and intergroup conflict, for suggesting speakers/authors, and
for providing comments on chapter drafts. We especially wish to thank the
members of our editorial board, who provided valuable feedback on initial
and revised proposals. In addition, the following people provided input
concerning one or more facets of the conference/book: Myron Aronoff,
Mahzarin R. Banaji, James Jones, Louis Kreisberg, B. Ann Bettencourt,


Yael Zerubavel, David Sears, Daniel M. Ogilvie, Jessica Heppen, and Rick
Lau.
This volume, as with its two predecessors, deserves to be part of the Rut-
gers Series on Self and Social Identity. We have been helped at every step of
the way by the faculty, staff, students, and administrators of Rutgers–The
State University of New Jersey. The following members of the Rutgers
University community were particularly helpful: Steven Lione, Thelma
Collins, Deana Guden-Pagnozzi, and the staff of the University Inn and
Conference Center; Stacey Hersh-Ehling of the Rutgers News Service;
Joanne Aguglia of the Douglass College Center; Rosemary Manero and her
colleagues in the Rutgers Dining Service; Marc Beebe, Justine Cardone,
Kathy Maloney, Rae Frank, Maria Brewer, Anne Sokolowski, and Alison
Smith, all of the Department of Psychology, who exhibited patience and
good humor in their extremely efficient handling of day-to-day tasks and
deadline pressures.
Other members of the Rutgers community deserve special thanks. First,
Charles Flaherty, chair of the Department of Psychology, has supported the
Rutgers Series since its inception, and he made departmental staff and other
resources available for our use in putting together the third symposium and
present associated book. Second, we express our gratitude to the faculty and
graduate students of the doctoral program in social psychology. Members
of the Social Psychology Area served as hosts for speakers, helped with
symposium registration, transported speakers to and from airports and train
stations, directed traffic between symposium sessions, and helped conduct
myriad other essential tasks. Third, we thank the following members of the
Rutgers community for their generous financial assistance: Joseph J. Seneca
(Vice President for Academic Affairs), and Richard F. Foley (Dean, Faculty
of Arts and Sciences).
We are also grateful for grants from the National Institutes of Health
(R13MH60129-01; Carolyn Morf, Chief, Personality and Social Cognition
Program) and from the National Science Foundation (SBR-9811485;
Steven J. Breckler, Program Director, Social Psychology) that supported
both the Third Rutgers Symposium and the production of this volume.
viii
Acknowledgments


Contents
Contributors
xi
1
introduction: Social Identity and Intergroup Conflict
3
Lee Jussim, Richard D. Ashmore, and David Wilder
PART I / The Contribution of Individuals’ Identities and the Collective
Identities of Social Groups to Intergroup Conflict
2
Ingroup Identification and Intergroup Conflict: When Does Ingroup Love
Become Outgroup Hate?
17
Marilynn B. Brewer
3
Ethnic Identity, National Identity, and Intergroup Conflict: The 
Significance of Personal Experiences
42
Thomas Hylland Eriksen
PART II / The Contribution of Ethnic and National Identities to 
Political Conflict in the United States
4
The Meaning of American National Identity: Patterns of Ethnic Conflict
and Consensus
71
Jack Citrin, Cara Wong, and Brian Duff


5
Communal and National Identity in a Multiethnic State: A Comparison of
Three Perspectives
101
Jim Sidanius and John R. Petrocik
PART III / The Contribution of Social Identity to Violent Intergroup
Conflict
6
Social and Role Identities and Political Violence: Identity as a Window on
Violence in Northern Ireland
133
Robert W. White
7
Individual and Group Identities in Genocide and Mass Killing
159
Ervin Staub
PART IV / The Role of Social Identity in Reducing Intergroup Conflict
8
The Role of National Identity in Conflict Resolution: Experiences from 
Israeli-Palestinian Problem-Solving Workshops
187
Herbert C. Kelman
9
conclusion: T
oward a Social Identity Framework for 
Intergroup Conflict
213
Richard D. Ashmore, Lee Jussim, David Wilder, and Jessica Heppen
Index of Subjects
251
Index of Names
265
x
Contents


xi
Contributors
Richard D. Ashmore, Department of Psychology, Tillett Hall, Rutgers, The State
University of New Jersey, 53 Avenue E, Piscataway, NJ 08854-8040. E-mail: 
ashmore@rci.rutgers.edu.
Marilynn B. Brewer, Department of Psychology, Ohio State University, 1885 Neil
Avenue, Columbus, OH 43210. E-mail: brewer.64@osu.edu.
Jack Citrin, Department of Political Science, 210 Barrows Hall #1950, University 
of California, Berkeley, Berkeley, CA 94720-1950. E-mail: gojack@uclink.
berkeley.edu.
Brian Duff, Department of Political Science, 210 Barrows Hall #1950, University 
of California, Berkeley, Berkeley, CA 94720-1950. E-mail: duff@socrates.
berkeley.edu.
Thomas Hylland Eriksen, Department of Social Anthropology, University of Oslo,
Moltke Moesvei 31, Eilert Sundts hus, 6th floor, P. O. B. 1091 Blindern, N-0317
OSLO, Norway. E-mail: t.h.eriksen@ima.uio.no.
Jessica Heppen, Department of Psychology, Tillett Hall, Rutgers, The State Uni-
versity of New Jersey, 53 Avenue E, Piscataway, NJ 08854-8040. E-mail: 
jheppen@rci.rutgers.edu.
Lee Jussim, Department of Psychology, Tillett Hall, Rutgers, The State University
of New Jersey, 53 Avenue E, Piscataway, NJ 08854-8040. E-mail: jussim@rci.
rutgers.edu.
Herbert C. Kelman, Psychology Department, Harvard University, Cambridge, MA
02138. E-mail: hck@wjh.harvard.edu.
John R. Petrocik, 210 McAlester Hall, Department of Psychology, University of 
Missouri-Columbia, Columbia, MO 65211. E-mail: PetrocikJ@missouri.edu.


Jim Sidanius, Department of Psychology, 1285 Franz Hall, P. O. Box 951563, 
405 Hilgard Avenue, Los Angeles, CA 90095-1563. E-mail: Sidanius@psych.
ucla.edu.
Ervin Staub, Department of Psychology, University of Massachusetts, Amherst, MA
01003. E-mail: estaub@psych.umass.edu.
Robert W. White, Department of Sociology, Indiana University at Indianapolis, 
Cavanaugh Hall 303, 425 University Boulevard, Indianapolis, IN 46202. E-mail:
rwwhite@iupui.edu.
David Wilder, Department of Psychology, Tillett Hall, Rutgers, The State Uni-
versity of New Jersey, 53 Avenue E, Piscataway, NJ 08854-8040. E-mail:
wilder@psych-b.rutgers.edu.
Cara Wong, Department of Political Science, 210 Barrows Hall #1950, University 
of California, Berkeley, Berkeley, CA 94720-1950. E-mail: cwong@bravo.
berkeley.edu.
xii
Contributors


SOCIAL IDENTITY,
INTERGROUP CONFLICT, and
CONFLICT REDUCTION


This page intentionally left blank 


Lee Jussim
1
Richard D. Ashmore
David Wilder
Introduction
Social Identity and Intergroup Conflict
Intense conflict between nations, ethnic groups, and political groups has
created some of the most enduring and lethal social problems humans have
ever faced. Slavery, conquest, exploitation, war, mass murder, and genocide
are among the most extreme manifestations of intergroup conflict. Such
conflicts remain as much a part of the human condition today as ever. 
Over the last two decades, social and behavioral scientists have increas-
ingly emphasized the role of self and identity in the causes and conse-
quences of intergroup hostility and harm-doing. This is, in part, because
issues of self and identity occur at multiple levels of analysis: individual
within a social context, groups within a society, and nations of the world.
At the level of individual within a social context, ethnic identity may con-
tribute to both in-group bonds and hostility toward other groups (Jones,
1997). At the level of groups within a society, social identity theory and
self-categorization theory emphasize the potential for group-based identi-
ties to foster support for the status quo among higher power and status
groups, and to foster intergroup competition and movements for political
change among lower power and status groups (T
ajfel, 1981; Thoits & Vir-
shup, 1997; Turner, 1987). Finally, at the level of modern nations, identity
is implicated in a variety of conflicts—state violence against internal mi-
norities (e.g., Goldhagen 1996; Staub, 1989), armed uprisings by ethnic
groups against central governments (e.g., Gurr, 1993), and international
conflicts (e.g., Greenfeld & Chirot, 1994; Keegan, 1993). Furthermore,
addressing issues of identity may be necessary to reduce certain conflicts
between groups (e.g., Kelman, 1997; Kriesberg, Northrup, & Thorson,
1989).
3


Although self and identity play a critical role in some (perhaps most) in-
tergroup conflicts, social scientific analyses of that role have thus far been
largely isolated from each other. This is unfortunate because, just as the sci-
entific analysis of conflict has blossomed in recent years, so too has there
been a resurgence in work on self and identity (Ashmore & Jussim, 1997).
Multiple and diverse lines of contemporary social science research and
theory demonstrate that self and identity are central to the understanding
of human thought, feeling, and action. They are also crucial in accounting
for interrelationships between the individual and larger sociocultural insti-
tutions and systems (Holland, 1997; T
ajfel, 1981; Thoits & Virshup, 1997).
Thus, there is much in the areas of self and social identity that might inform
our understanding of intergroup conflict.
The main purposes of this book are to elucidate ways in which: 
• Social identities create and exacerbate intergroup conflicts.
• Intergroup conflict influences social identity.
• Addressing social identity issues might help reduce some inter-
group conflicts. 
T
o achieve these goals, we invited a group of talented social and behav-
ioral scientists to address relationships between intergroup conflict and
identity. First, we asked them to address the locus of social identity. Are so-
cial identities primarily phenomena of individual psychology or are social
identities primarily phenomena that reside in the larger social structure?
That is, do social, group-based identities occur primarily in the minds of in-
dividual people? Or are they embedded in the culture, language, norms, and
social practices of a group of people? Or do social identities emerge from
some combination of individual psychology and socio-cultural discourses,
beliefs, and practices? 
Second, we sought contributors whose work addresses different types of
intergroup conflict. Although developing a taxonomy of the dimensions on
which intergroup conflicts may vary is beyond the scope of this book, it is
clear that social identities may be involved in a great variety of intergroup
conflicts. For example, whether it is “identity politics” in American elec-
tions, ethnic cleansing in Yugoslavia, or riots targeting Muslims in India,
group-based identities often play a critical role in maintaining, and some-
times exacerbating, intergroup conflicts. Although addressing all types of
intergroup conflicts is beyond the scope of any single book, we sought con-
tributions that would address a wide range of conflicts: conflicts played out
within normal legal structures; conflicts involving illegal acts of aggression;
conflicts over symbolic and ideological issues; and conflicts over concrete
resources and power.
Although each contributor did not address the role of identity in this full
range of conflicts, the contributions addressed many different types of con-
flicts (rather than focus on a single type of conflict—e.g., genocide). There-
4
Introduction


fore, the first pair of contributions (by Brewer and Eriksen) focuses on 
identifying broad principles or common themes underlying a range of in-
tergroup conflicts. The second pair (by Citrin, Wong, and Duff, and Sida-
nius and Petrocik) focuses on the extent to which ethnic identity in the
United States remains strong and predicts strength of attachment to U.S.
political institutions and culture. The third pair of contributions (by White
and Staub) addresses the role of social identity in lethal violence by and
against the state. The seventh chapter (by Kelman) addresses the role of 
social identity in reducing intergroup conflict. 
In setting the stage for the chapters to follow, we address the following
questions: What do we mean by social identity? What do we mean by con-
flict? What is the organizational scheme for this volume? What specific
topics do the chapters address?
What Do We Mean by Social Identity?
Self and identity are not simple concepts with widely agreed upon mean-
ings. As we pointed out in the first volume in this series (Ashmore & Jussim,
1997), the terms have been used in a bewildering diversity of ways. Some-
times, the same word or phrase is used by different writers to refer to differ-
ent concepts; sometimes different words or terms are used to refer to what
appears to be the same construct. T
o make matters worse, self and identity
are both evolving concepts, which have meant different things in different
historical epochs (Baumeister, 1997; Danziger, 1997). Self and identity are
also doubly contested concepts today. First, the nature and meaning of group-
based identities, such as national, ethnic, or gender-based identities, are 
often contested in the context of sociocultural discourses (for example, con-
sider the potentially strong reactions, both pro and con, that the statement
“My country right or wrong” sometimes evokes). Second, the meaning and
nature of self and identity as social scientific constructs are contested among
researchers from different traditions (Danziger, 1997; Holland, 1997). So-
cial constructivists, for example, often argue against the possibility of even
studying self and identity objectively (Holland, 1997), whereas many psy-
chologists and sociologists implicitly assume that self and identity can be
studied objectively (Jussim & Ashmore, 1997). 
This volume, however, does not attempt to address the multiplicity of
controversies surrounding self and identity (though see Ashmore & Jussim,
1997, which does just this). This volume focuses on social identity (and 
its role in intergroup conflict). Social identity is a narrower and more spe-
cific subset of the phenomena included under the general rubric of “self 
and identity.” As such, it may be at least somewhat easier to introduce the
term in a way that will meet with some agreement across disciplines and
perspectives.
Social Identity and Intergroup Conflict
5


Within psychology and sociology, one of the clearest and most common
definitions of social identity was provided by T
ajfel (1981, p. 255): “that part
of an individual’s self-concept which derives from his knowledge of his
membership in a social group (or groups) together with the value and emo-
tional significance attached to that membership.” Thus, according to T
ajfel,
social identity has two components: belief that one belongs to a group (e.g.,
“I am an American”) and the importance of that group membership to one’s
self (e.g., “and I am damn proud to be a citizen of the greatest country on
earth”). 
We have not insisted that the contributors use this definition; they were
free to define social identity in the manner they saw as most appropriate to
their work. Nonetheless, five of the seven contributors cite and use T
ajfel’s
definition. Limitations to this definition, however, were implicitly high-
lighted by the two authors (Eriksen, Kelman) who did not use it. T
ajfel’s 
definition is almost purely individualistic, focusing exclusively on how the
individual thinks and feels about group memberships. Both Eriksen and
Kelman suggest that ethnic and national identities are not purely individu-
alistic—that such identities emerge within specific sociocultural contexts.
Sociocultural discourses, national myths, and intergroup relations are all es-
sential to the creation of any particular ethnic or national identity, so that,
according to Eriksen and Kelman, social identity resides at least partly
within one’s national or cultural community, rather than exclusively within
the individual.
What Do We Mean by Intergroup Conflict?
Conflict, of course, can occur between almost any type of groups (e.g.,
sports teams, corporations, university departments, nations, political par-
ties, ethnic groups). The chapters in this book, however, primarily (but not
exclusively) emphasize conflict between ethnic, racial, religious, and na-
tional groups. This emphasis reflects some of the dominant concerns within
the social sciences. Those concerns, we suspect, have largely been moti-
vated by a longstanding desire to understand some of humanity’s deepest
and most tragic social ills—war, genocide, mass murder, and institutional-
ized oppression.
Within social psychology, conflict is often defined as some sort of incom-
patibility of goals, beliefs, attitudes, and/or behavior (e.g., Myers, 1999).
The nature of the conflicts addressed in this volume are many and varied.
Some address economic and political discrimination within a nation; others
address how such discrimination can flare up into murderous intergroup 
violence. Five chapters focus on specific, real-world conflicts. Eriksen ad-
dresses conflicts between Serbs and Bosnians; Fijians and ethnic Indians in
Fiji; Hindus and Muslims in India. Citrin et al. and Sidanius and Petrocik
address conflicts between white Americans and ethnic minorities in
6
Introduction


America. White discusses the multi-sided conflict in Northern Ireland, and
Kelman describes his own workshops focusing on reducing the conflict be-
tween Israelis and Palestinians. 
Three chapters (Brewer, Eriksen, Staub) seek to provide broad, general
insights into the intersection of social identity and intergroup conflict.
Brewer develops a theoretical perspective on the steps leading to progres-
sively greater intergroup hostility. Eriksen derives general insights into in-
tergroup hostility on the basis of his analysis of three real-world conflicts.
Staub focuses specifically on the origins of genocide and mass murder. 
The starting point for understanding intergroup conflict often varies
across the different social science disciplines. Within social psychology,
conflict of some type or another is often addressed, implicitly or explicitly,
in theories and approaches to intergroup relations (e.g., Allport, 1954;
Oakes, Haslam, & Turner, 1994; Sherif, 1958; T
ajfel, 1981). The goal of
such research is typically to provide broadly applicable general explanations
of phenomena such as bias, prejudice, and discrimination. Social psycho-
logical theories customarily provide individual-level explanations for conflict
and the method of choice is, in general, laboratory experimentation. The
starting point for many such perspectives, therefore, is the accumulated ex-
perimental evidence addressing issues such as stereotypes and prejudice.
This is in stark contrast to many of the other social sciences, such as an-
thropology, political science, and sociology. Work in these disciplines often
focuses on understanding the history, sources, and nature of particular in-
tergroup conflicts. When general principles are sought, however, the start-
ing point is usually real-world conflicts between nations, ethnic groups, or
religious groups. There is often more emphasis on organizational, institu-
tional, and sociocultural levels of explanation. The methods of choice tend
to be surveys, interviews, ethnographies, archives, and other sources for his-
torical analysis. Consequently, the literatures addressing intergroup conflict
in the different social sciences have developed largely independently, and
often may appear to be talking past one another. One purpose of this book,
therefore, was to begin to redress this tendency by drawing contributors
from several social sciences (anthropology, political science, psychology,
and sociology). 
Themes Underlying This Book
Several themes provide the framework for the specific topics addressed by
the volume: (1) identity issues are important to diverse intergroup conflicts;
(2) conflict influences social identity; (3) social identity can contribute to in-
tergroup conflict; (4) addressing issues of social identity may be useful or
even necessary for reducing some conflicts; and (5) to understand how iden-
tity figures in intergroup strife, it is necessary to consider multiple levels of
analysis.
Social Identity and Intergroup Conflict
7


Identity Issues Are Important to Diverse Intergroup Conflicts
Regardless of whether intergroup conflicts are played out peacefully in cul-
tural and political arenas, or become violent and murderous, social identity
is usually involved. On the peaceful side, consider the various conflicts
within the United States over issues such as affirmative action, immigration,
and official use of the English language. These conflicts often seem to in-
volve social identities—with policy preferences and attitudes often varying
among different ethnic groups in such a manner as to enhance the power,
status, or privileges of one’s own group.
Or consider the more brutal side: the various incidents of mass murder
and genocide that have occurred throughout human history. In many cases,
extreme nationalist movements link a deep commitment to a national or
ethnic identity to hatred of a minority group—a pattern that can be found
in the Turkish genocide of the Armenians, the German genocide of the
Jews, and in Serbian mass murder of Bosnians.
Conflict May Influence Social Identity
People are usually affected, sometimes deeply, by intergroup conflicts. Cul-
tural stereotypes and prejudices can lead members of stigmatized groups to
disidentify with and disengage from tasks in which they are most heavily
stigmatized (Steele, 1997). Some African-American youths, for example,
disengage from academic achievement because negative intellectual stereo-
types about African Americans are too threatening and because high
achievement may be seen as “acting white” (Fordham & Ogbu, 1986;
Steele, 1997).
Furthermore, groups that have once been victimized by conquest, en-
slavement, or genocide may, in turn, undergo identity changes that render
them more vulnerable to committing atrocities against other groups in the
future. At least sometimes, a “we better get them before they get us” atti-
tude becomes a critical component of a national or ethnic identity. When
such an attitude takes hold on a large scale, the society may become less
likely to consider making peace with enemies, more likely to seek violent
solutions to conflicts, and more likely to commit genocide (du Preez, 1994;
Goldhagen, 1996; Staub, 1989).
Social Identity May Contribute to Intergroup Conflict
Intergroup violence can be fostered by social identities as well as stem from
group-based self-interest (the notion of realistic group conflict [Ashmore &
Del Boca, 1976]). For example, the recent Serbian-Kosovar conflict was
started over a material issue: who, Serbs or Kosovars, would have political
authority over Kosovo? But one of the reasons this was important to the
Serbian government was the historical linkage of Kosovo with Serbian
8
Introduction


identity (it was the location of a 600-year-old stand by Serbs against invad-
ing Turks).
Addressing Issues of Social Identity May Be Useful for Reducing Some Conflicts
At the same time, social identity can assist in alleviating intergroup conflict.
For example, Sherif’s (1958) classic work on “superordinate goals” demon-
strated that members of two groups in conflict can work toward a shared
goal. Doing so can involve a new sense of “us,” which likely contributes 
to the lessened hostility that “superordinate goals” produce (cf. Gaertner,
Dovidio, Anastasio, Bachman, & Rust, 1993).
To Understand How Identity Figures in Intergroup Strife, It Is Necessary 
to Consider Multiple Levels of Analysis
The incorporation of the final organizing variable—level of analysis—
underscores that the links of identity to intergroup conflict cannot be re-
duced either to the characteristics of individuals or to the qualities of
groups. Fortunately, however, different social and behavioral sciences tend
to emphasize different levels of analysis. Thus, it was necessary that con-
tributors to this volume come from diverse academic disciplines that cover
the full continuum—from those that focus primarily on the individual (as
most psychologists do), to others that emphasize understanding the rela-
tionship of individuals to institutions and social structure within a society
(as do many sociologists and political scientists), to still others that investi-
gate multiple societies (as do many anthropologists). 
The Chapters in This Volume
Three sections, each including two chapters, address different aspects of re-
lations between identity and conflict: (1) general analyses of how social
identities contribute to conflicts between groups and how conflict can influ-
ence social identity; (2) analyses of the extent to which ethnic identities 
influence degree of patriotism and attachment to mainstream American po-
litical culture and institutions; and (3) perspectives on how social identity
contributes to lethal intergroup violence. The fourth topic is conflict reso-
lution, which is addressed in a chapter that describes a specific conflict re-
duction intervention and the role of social identity in this intervention.
How Social Identities Contribute to Intergroup Conflict
The first section analyzes the role of identity, as both a group- and indi-
vidual-level variable, in contributing to intergroup conflict. Ethnic and na-
tional rivalries have a long history of leading to intergroup conflicts with
Social Identity and Intergroup Conflict
9


horrendous consequences. In the extreme, ethnic identity has led to mass
murder (e.g., in the former Yugoslavia). National identities, especially in the
form of aggressive nationalism, have contributed to genocides and global
war (e.g., the Nazi Holocaust and World War II). Ethnic identities have
also contributed to slavery, discrimination, and exploitation. 
In chapter 2, Brewer reviews the psychological research on social 
identity and intergroup strife. Social psychologists have long studied how
individuals’ social identities and group memberships influence their reac-
tions to, and perceptions of, others. Much work has addressed the role of
individual-level stereotypes, prejudice, and discrimination in intergroup
conflict. In addition, two related theories—social identity theory and self-
categorization theory (T
ajfel, 1981; Turner, 1987) have explicitly identified
ways in which individuals’ group memberships and social identities may
lead to intergroup strife. Brewer integrates existing theoretical perspectives
on social identity with social cognition research on stereotypes and preju-
dice and with her own work on optimal distinctiveness into a new perspec-
tive. Her analysis suggests that intergroup harm-doing is the end result of 
a series of progressively more hostile steps that include the psychological
categorization of people into groups, ingroup trust and preference, per-
ceived moral superiority of the ingroup over outgroups, and antagonism to-
ward outgroups. She argues that although progression from each step to the
next is not inevitable, the occurrence of each step increases the likelihood
that the next will occur.
In chapter 3, Eriksen addresses this topic by focusing on three case stud-
ies of the role of identity in intergroup conflict: the Serbian-Bosnian con-
flict; the Hindutva movement in India; and relations between ethnic Indians
and Fijians on Fiji. In all three cases, Eriksen describes how conceptions of
“us” and “them” (social identities and intergroup relations) are influenced
by societal/cultural phenomena such as the existing power structure, na-
tionalistic leaders, national myths, and modernization. He then discusses
how these socially and historically constructed views of “us” and “them” in
turn contribute to intergroup conflict. 
ToWhat Extent Do Ethnic Identities Influence Attachment 
to the Mainstream Political Culture of the United States?
The second section of the book focuses specifically on relations between
ethnicity and attachment to mainstream American political culture. T
o what
extent have different ethnic groups adopted mainstream American values
and become strongly attached to American political institutions? T
o what
extent does a strong ethnic identity undermine attachment to American 
political culture? T
wo chapters address these issues.
In chapter 4, Citrin, Wong, and Duff assess ethnic similarities and dif-
ferences regarding key, symbolic aspects of American identity. Levels of
chauvinism, patriotism, and pride in America are strikingly similar among
10
Introduction


blacks, whites, Asians, and Hispanics. Personal commitment to an Ameri-
can identity serves to bind Americans to the United States and to bring di-
verse groups together in providing a common identity, and, to some extent,
a common set of interests. 
However, alongside this common sense of attachment to America, many
people, especially members of minority groups, express a strong sense of 
attachment to their ethnic groups. Among whites, furthermore, a strong at-
tachment to a white racial identity seems to foster intergroup conflict.
Specifically, such whites tend to be most resistant to policies aimed at im-
proving the social and political conditions facing Blacks. In contrast, strong
ethnic attachment among minority groups did not seem to foster inter-
group hostility.
In chapter 5, Sidanius and Petrocik compare three general perspectives—
melting pot, pluralist, and social dominance—regarding relations between
ethnic identity and national identity among African-American, white, La-
tino, Asian, and Jewish United States citizens, and among resident aliens liv-
ing in the United States They find that: (1) ethnic identities remain strong
for many people; (2) ethnocentrism exists among all groups they studied; 
(3) differences in patriotism between whites and the non-black ethnic groups
is small or nonexistent; but (4) differences in patriotism between African
Americans and whites are larger. Sidanius and Petrocik interpret these results
as most consistent with the social dominance model, which predicts that the
strongest support for a society’s culture and institutions will be found among
those groups that benefit most from that culture and those institutions.
Social Identity and Lethal Intergroup Conflict
The third section of the book is devoted to the role of national and ethnic
identity in the instigation of deadly violence. Intergroup hostility, even
when intense, does not always erupt into violence. And even when violent,
conflicts do not always erupt into mass killing, ethnic cleansing, or geno-
cide. Nonetheless, systematic and widespread state-sponsored violence is
clearly one of the ultimate manifestations of conflict between groups. From
biblical stories of conquest, to the Romans’ treatment of foes and rebellious
peoples (e.g., Carthage, Israel), the Crusades, the Turkish slaughter of Ar-
menians, the German Holocaust, the Cambodian genocide, and widespread
killings in Rwanda and Bosnia, mass murder and genocide have emerged as
gruesome recurrent themes of intergroup relations. Thus, two chapters ad-
dress the role of social identity in lethal intergroup conflicts.
In chapter 6, White presents an analysis of the ongoing hostilities in
Northern Ireland. In tracing the historical roots of the existing conflict, he
identifies three primary perpetrators of violence: the Irish Republican para-
militaries, Protestant paramilitaries, and the state’s security forces. He ana-
lyzes the role of religious, political, and national identities in perpetuating
the conflict and the violence. He concludes that: (1) social identities play a
Social Identity and Intergroup Conflict
11


crucial role in maintaining the conflict; and (2) very different types of iden-
tities underlie each group’s involvement. IRA members often view them-
selves as working for freedom and to overcome the oppression of the Irish
by the British; Protestant paramilitary members view themselves as fighting
for the defense of their religious/ethnic group; and British political leaders
view themselves, at least according to their public statements, as unbiased
peacekeepers (a view that White argues is not completely consistent with
their actions there). 
In chapter 7, Staub presents a social psychological analysis of factors
leading to, and preventing, genocide. He describes the social and historical
conditions that often predispose a society toward genocidal violence. In so
doing, he highlights the various ways in which self and social identity are in-
volved in the processes of mass killing. For example, he concludes that “in-
jured” social identities—the belief that one’s group has been unfairly vic-
timized in the past and, therefore, is at greater risk of being victimized in
the future—is one major precursor of genocide (“we have to get them be-
fore they get us”). Staub also discusses the types of interventions that seem
to avert genocide. 
Social Identity Approaches to Reducing Intergroup Conflict
In chapter 8, Kelman describes his extensive experience running conflict-
resolution workshops for Palestinians and Israelis. Through these work-
shops, it has become clear that each side tends to have components of a na-
tional identity that seem to reduce the conflict to a zero-sum game. Indeed,
even acknowledging the identity and existence of the other group has often
been viewed as a threat to one’s own identity and existence as a nation.
However, Kelman also points out that, although the core identity of both
peoples may be non-negotiable, there are many peripheral aspects of those
identities that have emerged from these workshops as negotiable. Kelman’s
chapter, therefore, describes the role of identity in both contributing to the
Israeli-Palestinian conflict and in seeking solutions to that conflict.
Conclusion
Research on identity and on conflict have long histories within the social
and behavioral sciences. Until relatively recently, however, research on rela-
tions between identity and conflict has been less common. In the last 20
years, researchers in anthropology, political science, psychology, and soci-
ology have begun to study how group identifications contribute to conflict,
how intergroup conflict can influence identity, and, most recently, how
identities might be harnessed to reduce intergroup hostility. This volume
does not summarize all such existing work. The following chapters do pre-
sent a sophisticated and detailed analysis of some of the most fundamental
12
Introduction


issues involved in understanding identity and conflict. Chapters address
conflict between ethnic, racial, religious, and national groups; they address
ethnic differences in national attachment; political conflicts over territory;
and genocidal conflicts. As the chapters take stock of existing work on iden-
tity and conflict within several social scientific disciplines, we suspect that
both their diversity and some similarities in underlying themes will stand
out.
References
Allport, G. (1954). The nature of prejudice. Cambridge, MA: Addison-Wesley.
Ashmore, R. D., & Del Boca, F. K. (1976). Psychological approaches to under-
standing intergroup conflict. In P. Katz (Ed.), Towards the elimination of racism
(pp. 73–123). Elmsford, NY: Pergamon Press.
Ashmore, R. D., & Jussim, L. (Eds.) (1997). Self and identity: Fundamental issues
(Rutgers Series on Self and Social Identity, Volume 1). New York: Oxford Uni-
versity Press.
Baumeister, R. F. (1997). The self and society: Changes, problems, and opportuni-
ties. In R. D. Ashmore, & L. Jussim, (Eds.), Self and identity: Fundamental issues
(Rutgers Series on Self and Social Identity, Volume 1; pp. 191–217). New York:
Oxford University Press.
Danziger, K. (1997). The historical formation of selves. In R. D. Ashmore, & L. Jus-
sim, (Eds.), Self and identity: Fundamental issues (Rutgers Series on Self and Social
Identity, Volume 1; pp. 137–159). New York: Oxford University Press.
du Preez, P. (1994). Genocide: The psychology of mass murder. Southampton, UK: 
Bowerdean Publishing Co.
Fordham, S., & Ogbu, J. U. (1986). Black students’ school success: Coping with the
“burden of acting White.” Urban Review, 18, 176–206.
Gaertner, S. L., Dovidio, J. F., Anastasio, P. A., Bachman, B. A., & Rust, M. C.
(1993). The common ingroup identity model: Recategorization and the reduc-
tion of intergroup bias. In W. Stroebe & M. Hewstone (Eds.), European review of
social psychology, 4, pp. 1–26, London: Wiley.
Goldhagen, D. (1996). Hitler’s willing executioners: Ordinary Germans and the holo-
caust. New York: Alfred A. Knopf.
Greenfeld, L., & Chirot, D. (1994). Nationalism and aggression. Theory and Society,
23, 79–130.
Gurr, T. R. (1993). Why minorities rebel: A global analysis of communal mobiliza-
tion and conflict since 1945. International Political Science Review, 14, 161–201.
Holland, D. (1997). Selves as cultured: As told by an anthropologist who lacks a
soul. In R. D. Ashmore, & L. Jussim, (Eds.), Self and identity: Fundamental Issues
(Rutgers Series on Self and Social Identity, Volume 1, pp. 160–190). New York:
Oxford University Press.
Jones, J. (1997). Prejudice and Racism (second edition). New York: McGraw-Hill.
Jussim, L., & Ashmore, R. D. (1997). Fundamental issues in the study of self and
identity—contrasts, contexts, and conflicts. In R. D. Ashmore, & L. Jussim,
(Eds.), Self and identity: Fundamental issues (Rutgers Series on Self and Social
Identity, Volume 1, pp. 218–230). New York: Oxford University Press.
Social Identity and Intergroup Conflict
13


Keegan, J. (1993). A history of warfare. New York: Alfred A. Knopf.
Kelman, H. C. (1997). Some determinants of the Oslo breakthrough. Negotiaion
Journal, 13, 327–340.
Kreisburg, L., Northrup, T. A. & Thorson, S. J. (1989). Intractable conflicts and
their transformation. Dynamics of Identity, 4, 55–82. 
Myers, D. G. (1999). Social psychology (fifth edition). New York: McGraw-Hill.
Oakes, P. J., Haslam, S. A., & Turner, J. C. (1994). Stereotyping and social reality.
Cambridge, MA: Blackwell.
Reich, W. A. (1997). An expanded role for identity in the analysis of intergroup con-
flict. Unpublished manuscript.
Sherif, M. (1958). Superordinate goals in the reduction of intergroup conflict.
American Journal of Sociology, 63, 349–363.
Staub, E. (1989). The roots of evil: The origins of genocide and other group violence. Cam-
bridge: Cambridge University Press.
Steele, C. (1997). A threat in the air: How stereotypes shape intellectual identity and
performance. American Psychologist, 52, 613–629. 
T
ajfel, H. (1981). Social identity and intergroup relations. London: Cambridge Univer-
sity Press.
Thoits, P. A., & Virshup, L. K. (1997). Me’s and we’s: Forms and functions of social
identities. In R. D. Ashmore, & L. Jussim, (Eds.), Self and identity: Fundamental
issues (Rutgers Series on Self and Social Identity, Volume 1; pp 106–133). New
York: Oxford University Press.
Turner, J. C. (1987). Rediscovering the social group: A self-categorization theory. New
York: Basil Blackwell.
14
Introduction


PART I
THE CONTRIBUTION OF 
INDIVIDUALS’ IDENTITIES
AND THE COLLECTIVE 
IDENTITIES OF SOCIAL
GROUPS TO INTERGROUP
CONFLICT


This page intentionally left blank 


Marilynn B. Brewer
2
Ingroup Identification and 
Intergroup Conflict
When Does Ingroup Love Become Outgroup Hate?
Social psychologists’ interest in social identity and ingroup bias arose in the
context of research on intergroup relations (T
ajfel &Turner, 1986; Brewer &
Brown, 1998). As a consequence, there has been a coupling of the study of in-
group favoritism and outgroup prejudice, the implicit assumption being that
the more we know about why individuals become attached to their ingroups,
the more we will understand why they come to dislike or derogate outgroups.
The present chapter seeks to challenge that assumption by examining more
carefully the nature of the relationship between ingroup love and outgroup
hate. The basic thesis is that the formation of ingroups and ingroup identifi-
cation arises independently of attitudes toward outgroups. Attachment to in-
groups is presumed here to be the primary process, fundamental to indi-
vidual survival and well-being. Ingroup identification thus precedes
outgroup hostility and intergroup conflict but is not by itself a sufficient ex-
planation for such conflict. Instead, as this chapter will argue, we need to
trace the steps that lead from ingroup formation to outgroup hostility, speci-
fying the conditions under which the two become entangled.
Sumner’s Theory of Ethnocentrism
An early theory of the interrelationship between ingroup attachment and
identification and outgroup antagonism was implicit in the definition of
“ethnocentrism” when the term was introduced by Sumner (1906). Ethno-
centrism was described by Sumner as a universal characteristic of human
social groups whereby
17


a differentiation arises between ourselves, the we-group, or in-group, and
everybody else, or the others-group, out-groups. The insiders in a we-group
are in a relation of peace, order, law, government, and industry, to each other.
Their relation to all outsiders, or others-groups, is one of war and plunder.
. . . Ethnocentrism is the technical name for this view of things in which
one’s own group is the center of everything, and all others are scaled and rated
with reference to it. . . . Each group nourishes its own pride and vanity,
boasts itself superior, exalts its own divinities, and looks with contempt on
outsiders. . . . (pp. 12–13)
The features characteristic of Sumner’s ethnocentrism syndrome include
negative attitudes toward outgroups as well as positive feelings and evalua-
tions of the ingroup. Further, an explicit negative correlation between in-
group and outgroup attitudes is postulated: the greater the attachment and
solidarity within the ingroup, the greater the hostility and contempt di-
rected toward outgroups. This hypothesized relationship between ingroup
love and outgroup hate derived from Sumner’s functional theory of the ori-
gins of social groups and intergroup conflict:
The relation of comradeship and peace in the we-group and that of hostility
and war towards others-groups are correlative to each other. The exigencies
of war with outsiders are what make peace inside, lest internal discord should
weaken the we-group for war. . . . Thus war and peace have reacted on each
other and developed each other, one within the group, the other in the inter-
group situation. . . . Sentiments are produced to correspond. Loyalty to the
ingroup, sacrifice for it, hatred and contempt for outsiders, brotherhood
within, warlikeness without—all group together, common products of the same
situation. (1906, pp. 12–13, emphasis added)
The last 20 years of social psychological research on ingroup bias and in-
tergroup relations have cast doubt on the validity of the functional relation-
ship between ingroup and outgroup attitudes and behavior presumed by
Sumner. Results of the laboratory experiments using the “minimal group
paradigm” designed by T
ajfel and his colleagues (T
ajfel, Billig, Bundy, & Fla-
ment, 1971) explicitly challenged the assumption that differential treatment
of ingroup and outgroup members derives from cooperative interdepen-
dence within groups and competitive relations between groups. In the mini-
mal group experiments, mere categorization of persons into differentiated
social groupings has proved to be sufficient to give rise to preferential dis-
crimination in favor of ingroup over outgroup members, in the absence of
any differential interdependence (T
ajfel et al., 1971; Turner, 1978; Brewer,
1979; Diehl, 1990). Field studies assessing the relationship between identifi-
cation and attachment to ingroups and discrimination against outgroups
have also failed to find any straightforward negative correlation between in-
group positivity and intergroup attitudes (Hinkle & Brown, 1990). For ex-
ample, in a study of the reciprocal attitudes among 30 ethnic groups in East
Africa, Brewer and Campbell (1976) found that almost all of the groups ex-
hibited systematic differential positive evaluation of the ingroup over all out-
18
The Contribution of Identities to Intergroup Conflict


groups on dimensions such as trustworthiness, obedience, friendliness, and
honesty, but the correlation between this measure of ingroup positivity and
distancing from outgroups was essentially .00 across the 30 groups (p. 85).
Thus, the question of when and how ingroup identification is related to out-
group hostility and intergroup conflict requires further investigation.
Decomposing Ethnocentrism: Ingroup, Intergroup, and 
Outgroup Orientations
Sumner’s definition of ethnocentrism cited at the beginning of this chapter
contains four separate propositions. The four distinguishable elements can
be characterized as follows:
1. Human social groups are organized into discrete ingroup-out-
group categories (the social categorization principle).
2. Individuals value their ingroups positively and maintain positive, 
cooperative relationships with members of the ingroup (the in-
group positivity principle).
3. Ingroup positivity is enhanced by social comparison with out-
groups in which ingroup attributes and outcomes are evaluated as
better than or superior to those of outgroups (the intergroup compar-
ison principle).
4. Relationships between ingroup and outgroups are characterized by 
antagonism, conflict, and mutual contempt (the outgroup hostility
principle).
In contrast to Sumner’s (1906) expectation that these four elements nec-
essarily cohere into a pattern that is universally characteristic of intergroup
relations, a review of relevant social psychological research suggests that the
components can be distinguished both empirically and conceptually. How-
ever, the components are not entirely independent features of ethnocentric
attitudes, either. Instead, I propose that the elements represent a systematic
progression along a continuum of possible relationships between ingroup
formation and intergroup behavior in which each element in the progres-
sion provides a necessary but not sufficient condition for occurrence of the
subsequent elements. Within this system, the first two elements are proba-
bly universal characteristics of human social groups (as Sumner postulated),
but the third and fourth elements require additional social-structural and
motivational conditions that are not inherent in the processes of group for-
mation itself.
Social Categorization:The Basic Substrate
Virtually all current theories of social identity and intergroup relations rec-
ognize social categorization as the basic cognitive process underlying all in-
Ingroup Identification and Intergroup Conflict
19


tergroup phenomena (e.g, Allport, 1954; T
ajfel, 1969). Categorization (a)
partitions the multidimensional variability among human beings into dis-
crete subsets, accompanied by (b) accentuation of perceived intracategory
similarities and intercategory differences (Doise, Deschamps, & Meyer,
1978; T
ajfel, 1969; T
ajfel & Wilkes, 1963). The basic processes of catego-
rization and category accentuation are presumed to be the same whether we
are talking about individuals’ partitioning of the world of physical objects
and events or of the social world. Individuals learn to classify objects as
functionally interchangeable and develop concepts that distinguish mem-
bers of one category from those of another as a fundamental tool for nego-
tiating the physical and social environment. 
The assimilation and contrast processes underlying category formation
and accentuation are independent of any self-involvement in the category
system. Just as one can acquire a concept of “apple,” which may include a
representation of the prototypic member of the category, a theory of the
“essence” of appleness, and an understanding of the variability among ex-
emplars of the category, so one can acquire similar conceptualizations of so-
cial objects such as “librarian” or “African American.” Stereotypic represen-
tations (prototypes) of social categories are a natural consequence of
categorization and concept formation processes, motivated by the need for
cognitive simplification and social structure. As with any other category
concept, whether social category stereotypes are positive or negative (or
mixed) is a function of what has been learned about the content of the cate-
gory representation.
What makes social categorization different from nonsocial categoriza-
tion systems, however, is the potential for self-involvement in that social
perceivers are themselves members of some social categories (ingroups) and
not others. Thus, an additional categorical distinction between ingroup and
outgroups may be superimposed onto particular social classifications. In the
absence of the engagement of this ingroup-outgroup distinction, social 
categories are merely cognitive concepts that help individuals to make sense
of their social world and to direct their behavior toward other individuals.
When social categories are also recognized as social ingroups, however, ad-
ditional processes associated with self-categorization (Turner, Hogg, Oakes,
Reicher, & Wetherell, 1987) are engaged that lay the groundwork for fur-
ther social differentiation. Categorization of the self as an ingroup member
entails assimilation of the self to the ingroup category prototype and en-
hanced similarity to other ingroup members. Self-categorization provides
the cognitive substrate for attachment to ingroups and differentiation from
outgroups—the first step toward ingroup bias and discrimination. 
Ingroup Positivity:The Causes and Consequences of Social Identification
Differentiation between ingroups and outgroups does not necessarily carry
evaluative implications, and acknowledging one’s ingroup membership is
20
The Contribution of Identities to Intergroup Conflict


not equivalent to identification with that social group. This is clear from
T
ajfel’s (1981) definition of social identity as “that part of an individual’s
self-concept which derives from his knowledge of his membership of a so-
cial group . . . together with the value and emotional significance attached
to that membership” (p. 255; emphasis added). Social identification, then,
entails affective and evaluative processes that are above and beyond mere
cognitive classification of the self and others into a shared social category.
The affective significance of social identification arises from the felt attach-
ment between the self and the ingroup as a whole. 
Social identification represents the extent to which the ingroup has been
incorporated into the sense of self, and at the same time, that the self is ex-
perienced as an integral part of the ingroup. With high levels of social iden-
tification, the group’s outcomes and welfare become closely connected to
one’s own sense of well-being (Brewer, 1991). According to social identity
theory, it is this engagement of the self that accounts for the positive valua-
tion of the ingroup and positive orientations toward fellow ingroup mem-
bers. Self-worth is both projected onto and derived from positive ingroup
evaluation. The distinction between ingroup membership and ingroup
identification implies that social identification is not an automatic by-
product of categorization. Identification implies some motivation to define
oneself in terms of that group membership and to achieve and maintain 
inclusion in the ingroup category. Since identification entails some sacrifice
of an autonomous self-concept, the question arises as to why individuals
would attach some measure of their self-worth and well-being to the fate of
a collective. 
One explanation that has been offered is that social identity and ingroup
positivity are engaged in the interests of enhanced self-esteem (Turner,
1975; T
ajfel, 1978). But group outcomes are not always positive or self-
aggrandizing, and identification often entails assimilation of group failures
as well as successes (Brewer & Weber, 1994; Crocker & Major, 1989; Hirt,
Zillmann, Erickson, & Kennedy, 1992; Turner, Hogg, Turner, & Smith,
1984). Overall, results of research on the causal role of self-esteem in social
identification are far from convincing (Abrams & Hogg, 1988; Rubin &
Hewstone, 1998). In fact, there is good reason to believe that attachment of
the self to the group (or other objects) leads to positive valuation of the
group, rather than vice versa (Greenwald & Banaji, 1995). 
If strength of ingroup identification is largely independent of ingroup
status, then motives other than self-enhancement must be implicated in se-
lecting and maintaining social identities. An alternative motivational expla-
nation is provided by optimal distinctiveness theory (Brewer, 1991). The
theory postulates that human beings have two powerful social motives: a
need for inclusion that motivates assimilation of the self into large, imper-
sonal social collectives, and an opposing need for differentiation that is satis-
fied by distinguishing the self from others. As opposing motives, the two
needs hold each other in check. When a person feels isolated or detached
Ingroup Identification and Intergroup Conflict
21


from any larger social collective, the drive for inclusion is aroused; on the
other hand, immersion in an excessively large or undefined social collective
activates the search for differentiation and distinctiveness. 
Identification with distinctive social groups emerges from the processes
involved in satisfying both needs simultaneously. Groups that are exclusive
rather than inclusive, groups that have clear categorical boundaries—satisfy
the need for inclusion (intragroup assimilation) at the same time that they
provide a basis for satisfying the need for differentiation (intergroup con-
trast). Clear ingroup boundaries serve to secure both inclusion and exclu-
sion. According to optimal distinctiveness theory, ingroups that meet these
simultaneous conditions become an integral part of the individual’s sense of
self and the basis for a secure and stable self-concept. Group identification
in this sense involves “transformations in the definition of self and the basis
for self-evaluation. When the definition of self changes, the meaning of
self-interest and self-serving motivation also changes accordingly” (Brewer,
1991, p. 476). 
Optimal distinctiveness theory proposes that social identification is the
product of the search for inclusion and differentiation, rather than a conse-
quence of the search for self-esteem. Nonetheless, although the desire for
self-enhancement may not be a primary cause of ingroup identification,
identification may well lead to a motivation to view the ingroup in the most
favorable possible terms. Drawing on Heider’s (1958) principles of cogni-
tive balance, the connection between self and ingroup constitutes a unit re-
lationship. Assuming that affect toward the self is primarily one of postive
regard, positive attitudes toward the group creates balance between unit
and sentiment relations. There is empirical evidence that positivity biases
associated with ingroup identification arise automatically and without
awareness. Perdue, Dovidio, Gurtman, and Tyler (1990), for instance,
demonstrated that ingroup signifiers (terms such as “we,” “us,” and “our”)
automatically activate positive evaluative responses, as indicated by differ-
ential reaction time to recognize subsequent words as “good” (vs. “bad”).
Consistency motives are also implicated in relationships between atti-
tudes and behavior; positive evaluations and expectations give rise to trust
and cooperative behavior that in turn justify positive feelings and future ex-
pectancies. Thus, the process of assimilation between self and others that
underlies ingroup formation instigates a “benign cycle” of positive affect
and behavior.
The emphasis at this stage on positive affect and evaluations directed to-
ward the ingroup is not devoid of implications for intergroup relations.
Positive attitudes and expectations about ingroup members, perceived simi-
larity, and cooperative orientation are all limited to those who share mem-
bership in the common ingroup. The extension of positive self-regard to
others stops at the boundary between ingroup and outgroups; attitudes to-
ward those outside the boundary are, at best, characterized by indifference.
Differential indifference is not without its consequences. When positive re-
22
The Contribution of Identities to Intergroup Conflict


gard, empathy, and cooperation are extended to some individuals but not
others on the basis of category membership, an initial form of intergroup
discrimination is evident (Brewer, 1979, 1996). Compared to ingroupers,
outgroupers are less likely to be helped in ambiguous circumstances (Frey
& Gaertner, 1986; Gaertner & Dovidio, 1977; Gaertner, Dovidio, & John-
son, 1982), more likely to be seen as provoking aggression (Baron, 1979;
Rogers & Prentice-Dunn, 1981), less likely to receive the benefit of the
doubt in attributions for negative behaviors (Weber, 1994), and likely to be
seen as less deserving of public welfare (Pettigrew & Meertens, 1995).
Studies of ethnic and racial prejudice in the United States and Europe
confirm that the essence of “subtle racism” is not the presence of strong
negative attitudes toward minority outgroups but the absence of positive sen-
timents toward those groups (e.g., Dovidio & Gaertner, 1993; Pettigrew &
Meertens, 1995; Stangor, Sullivan, & Ford, 1991, Study 1). As Gaertner,
Dovidio, Anastasio, Bachmann, & Rust. (1993) put it, “Initially it was 
assumed that in situations in which blacks were treated differently than
whites, whites discriminated against blacks, and thus reflected subtle, indi-
rectly expressed negative racial attitudes. Alternatively, it is possible that, 
at some fundamental level, aversive racism reflects a pro-white (i.e., pro-
ingroup) rather than the solely anti-black bias that was originally pro-
posed” (p. 4).
If social groups differ in access to resources and power, the benefits of in-
group favoritism accrue to members of some groups more than others and
contribute to substantial differences in outcomes between the groups as a
whole. Ultimately, many forms of discrimination may develop not because
outgroups are hated, but because positive affects such as admiration, sympa-
thy, and trust are reserved for the ingroup and withheld from outgroups
(Pettigrew & Meertens, 1995). Differential positivity toward ingroups over
outgroups also sets the stage for more pernicious forms of intergroup 
discrimination.
Intergroup Comparison: Social Competition for Positive Distinctiveness
Ingroup positivity implies a sort of self-absorption whereby group members
attend to the concerns and welfare of their own group with disregard for
the status or outcomes of those outside the ingroup boundary. The orienta-
tion is simply “us” versus “not us,” with little need to differentiate among
outgroups or to take outgroup outcomes into account while pursuing in-
group goals. Satisfaction with the ingroup’s well-being is evaluated against
internal standards and expectations, independent of the status of those out-
side the group.
The idea that assessments of goodness and satisfaction can be made in an
absolute sense against clear internal standards may hold true for some di-
mensions of evaluation or some conditions of group living. In studies of the
individual self-concept, Neidenthal and Beike (1997) demonstrated that
Ingroup Identification and Intergroup Conflict
23


some self-evaluations are affected by comparative context, but others re-
main stable and independent of comparative information. A parallel distinc-
tion may hold for ingroup self-evaluations. Groups that are physically or
psychologically isolated from outsiders, for instance, may measure success
in terms of basic sustenance or survival in the face of environmental hard-
ships, or in terms of adherence to dogma or beliefs and practices that are
not shared by those outside the group. 
In most cases, however, standards of evaluation are not so clear-cut, and
knowledge of the status of outgroups becomes relevant to assessing the state
of welfare of the ingroup. The conditions of social comparison at the inter-
group level are parallel to those for interpersonal comparison in the pursuit
of self-evaluation as outlined by Festinger (1954). The need for social com-
parison is aroused when there is uncertainty about one’s standing on some
dimension of self-evaluation, uncertainty that can be resolved by comparing
one’s own position to that of relevant others. The relevance requirement is
particularly interesting because it implies that the self and other (or ingroup
and outgroup) must be perceived as similar in some sense (i.e, as subject to
the same values or sharing a common frame of reference for evaluation) in
order to be useful targets for comparative appraisal. Thus, when ingroup
evaluation is dependent upon intergroup comparison, it is necessary to
identify specific outgroups for comparison and to distinguish between out-
groups that are relevant to ingroup self-regard and those that are not. 
Comparison (whether interpersonal or intergroup) is not inherently
competitive. This depends on what motives social comparison is serving
(Wood, 1989). Comparison can be undertaken for purposes of objective
self-appraisal or to motivate self-correction and improvement. When as-
sessments are relative and evaluative, however, the better the other is judged
to be, the worse the self-evaluation. In this case, ingroup favoritism takes a
subtle shift from motivated perceptions that “we are good” to perceptions
that “we are better.” When all of the groups in a social milieu are jockeying
for this type of ingroup advantage, ingroup-outgroup relations are charac-
terized by what Turner (1975) referred to as “social competition” for posi-
tive identity. Combined with motivations for self-enhancement, then, social
comparison becomes social competition and the pursuit of positive self-
regard can be achieved only at the expense of the other. 
In this case, competition is not necessarily inherent in the availability of
positive outcomes but is created by the evaluation that is attached to out-
comes for the ingroup and outgroup. If resources (economic and social 
benefits) are valued in an absolute sense, increases in ingroup welfare can be
achieved even if the outgroup is also benefitting. However, when valuation
(e.g., subjective states such as status) is relative, the welfare of the ingroup
and the outgroup is psychologically a zero-sum distribution. We cannot 
improve our position or sense of well-being unless the outgroup is doing 
less well than we are. When resources are also attached to intergroup differ-
ences in status and power, this zero-sum mentality is exacerbated. Ingroup
24
The Contribution of Identities to Intergroup Conflict


bias motivated by a desire to establish or maintain positive distinctiveness 
of the ingroup relative to the outgroup is a second form of intergroup dis-
crimination that can be differentiated, at least conceptually, from simple 
ingroup favoritism (Brewer, 1996). 
Interestingly, the more similar groups are in their values and aspirations,
the more acute the intergroup social competition. When the ingroup and
outgroup are arrayed along a single dimension of evaluation that all groups
regard as important and relevant to collective self-esteem, comparative ap-
praisals are inevitably competitive. In a context of self-enhancement mo-
tives, outgroup achievements are threatening to ingroup position and the
closer the outgroup, the greater the threat (T
esser, 1988). Social competi-
tion and intergroup threat are diluted when the outgroup is very remote or
distant from the ingroup on the dimension of evaluation. Ingroups that en-
joy very high status often display largess or magnanimity, rather than in-
group discrimination, toward outgroups that are much lower down on the
status ladder. Conversely, low-status groups often exhibit outgroup posi-
tivity in evaluations that are related to the status distinction, conceding out-
group superiority on dimensions where the status difference is clear and
large (Hinkle & Brown, 1990; Mullen, Brown, & Smith, 1992). 
Either of these forms of outgroup favoritism, however, is qualified by
strategies that serve ingroup positivity overall. Whereas low-status groups
exhibit outgroup positivity on status-relevant evaluations, they generally
display greater ingroup bias (compared to their high-status counterparts)
on dimensions unrelated to the status differential (Brewer & Campbell,
1976; Brewer, Manzi, & Shaw, 1993; Mullen, Brown, & Smith, 1992). Simi-
larly, outgroup largess on the part of high-status groups quickly reverts to
ingroup favoritism when the status gap between ingroup and outgroup is
perceived to be closing (Sachdev & Bourhis, 1991; Sidanius, 1993).
In a pluralistic context where multiple dimensions of evaluation are possi-
ble, social competition for positive distinctiveness may be more diffused and
less intense. In research on individual self-evaluation maintenance it has
been demonstrated that individuals feel threatened by a friend’s achieve-
ments when the domain of performance is highly important and relevant to
their own self-evaluation, but a friend’s accomplishments are celebrated
when the domain is not self-important (T
esser, 1988). Similar effects have
been observed in research on ingroup bias and intergroup discrimination.
Mummendey and Schreiber (1983) found that when individuals are given the
opportunity to choose dimensions of evaluation, they evaluate ingroup prod-
ucts more favorably than outgroups on dimensions chosen by the ingroup
but rate the outgroup positively on other dimensions. In general, in-group
favoritism is highest on dimensions of evaluation that are important to the
ingroup, but is mitigated for dimensions that are important to the outgroup
but relatively unimportant to the ingroup (Mummendey & Schreiber, 1984;
Mummendey & Simon, 1989). 
In interpreting these findings, Mummendey and Schreiber (1983) sug-
Ingroup Identification and Intergroup Conflict
25


gest that social competition occurs only when there is no alternative for es-
tablishing one’s own positive identity than at the expense of the outgroup.
When positive identity is possible for both groups at the same time, judg-
ments are more influenced by fairness than by ingroup bias. Thus, positive
distinctiveness does not necessitate across-the-board favoritism of ingroup
over outgroups. Ingroup distinctiveness at the expense of outgroups obtains
only when valuation is relative and on dimensions that are particularly im-
portant to the group’s collective self-worth.
Outgroup Antagonism and Intergroup Aggression
Intergroup comparison and social competition provide an initial connection
between valuing the ingroup and devaluing relevant outgroups. In the lit-
erature on social motives (McClintock, 1972), however, a clear distinction is
made between competition—the motivation to seek relative gain for the in-
group over others—and aggression—the motivation to harm the other as an
end in itself. Following this distinction, we can distinguish between inter-
group discrimination that is based solely on ingroup favoritism, and dis-
crimination or prejudice that entails an active component of outgroup dero-
gation and aggression (Levin & Sidanius, 1999; Struch & Schwartz, 1989).
Aggression directed against an outgroup may originate in the service of
protection or enhancement of the ingroup, but it may also become an end
in itself—outgroup hate that has become dissociated from ingroup en-
hancement or desire to benefit the ingroup.1
When outcomes are objectively (or perceived to be) zero-sum, it be-
comes difficult to pin down the locus of bias (Brewer, 1979). If ingroup out-
comes improve only as outgroup benefits decrease, it is not clear whether
the motivation for discriminatory behavior is to benefit the ingroup (i.e.,
discrimination for us) or to disadvantage the outgroup (discrimination
against them). Still, the experimental literature on intergroup discrimina-
tion provides evidence that the primary motivation is to benefit the ingroup
rather than harm the outgroup. In studies where evaluations of the ingroup
can be assessed independently of evaluations of the outgroup, conditions
that increase ingroup identification and favoritism generally involve en-
hancement of ingroup evaluations rather than decreased evaluation or dero-
gation of outgroups (e.g., Bettencourt & Dorr, 1998; Brewer, 1979; Brewer,
Manzi, & Shaw, 1993). 
More direct evidence is provided by research on the asymmetry of posi-
tive and negative ingroup bias (Mummendey & Otten, 1998). Although the
initial experiments by T
ajfel et al. (1971) claimed that ingroup bias in the
minimal intergroup situation occurred for allocations of both positive and
negative outcomes, subsequent research suggests that when the outcomes
to be distributed are negative or harmful, the usual intergroup discrimina-
tion may be lessened or may disappear all together. Hewstone, Fincham,
and Jaspars (1981) modified the normal paradigm by asking group members
26
The Contribution of Identities to Intergroup Conflict


to subtract money from ingroup and outgroup recipients. Although some
ingroup bias was observed in this context, the levels were lower than those
obtained with the standard allocation matrices. Mummendey et al. (1992)
extended this finding by asking participants to distribute different durations
of an unpleasantly high-pitched tone to ingroup and outgroup members.
With harm as an outcome, ingroup bias was generally eliminated and
strategies to equalize or minimize the amount of aversive stimulation were
observed.
Mummendey and Otten (1998) have suggested that the positive-negative
asymmetry in intergroup discrimination is explained by normative con-
staints that make it more difficult to justify differential allocations that harm
others directly than just benefit some more than others. As long as everyone
obtains some benefit, it is relatively easy to find justification for why the in-
group might be entitled to benefit more than the outgroup. There is even
evidence that such ingroup benefiting is considered normative in its own
right (Blanz, Mummendey, & Otten, 1997; Platow, O’Connell, Shave, &
Hanning, 1995). Thus, allocations that favor the ingroup over the outgroup
may be motivated primarily by intentions to be a “good” ingroup member,
rather than by any sentiments against the outgroup. 
Inflicting harm, however, is not sanctioned so directly. It is more difficult
to justify protecting the ingroup if it comes at the expense of inflicting harm
on others (Mummendey & Otten, 1998). Thus, even when the ingroup may
be disadvantaged, intergroup discrimination is constrained when there 
is no direct motivation to harm or derogate the outgroup. These are the
constraints that are lifted when outgroups are viewed with hatred or con-
tempt—emotions that justify outgroup harm above and beyond ingroup
benefit. Yet outside of the laboratory, the world is full of evidence of the 
existence of such self-justified harm toward outgroups, even at great costs
to the ingroup. A desire for positive distinctiveness can account for prefer-
ential hiring of ingroup members over outgroup members and for selective
investment in projects that benefit one’s own group over others. But wars 
of conquest, pograms, and ethnic cleansing require explanation that goes
beyond that of achieving positive distinctiveness for the ingroup.
Social identity theory provides an adequate framework for understand-
ing ingroup bias and intergroup discrimination of the positive types de-
scribed earlier. Assimilation of the self to the ingroup accounts for ingroup
positivity, and from there, the extension to relative enhancement of the in-
group over outgroups under conditions of intergroup comparison is not a
big motivational leap. The story becomes more complicated, however,
when one tries to use the same framework to account for more virulent out-
group hate and intergroup hostility. T
o justify aggression against outgroups
in the interest of the ingroup, the very existence of the outgroup, or its
goals and values, must be seen as a threat to the maintenance of the ingroup
and to one’s own social identity. Thus, understanding the relationship be-
tween ingroup identification and outgroup hostility requires understanding
Ingroup Identification and Intergroup Conflict
27


how the interests of the ingroup and those of the outgroup come to be 
perceived as in conflict. For that purpose, we need to reconsider the funda-
mental motivations underlying ingroup attachment and the path from in-
group identity to outgroup threat.
From Ingroup Formation to Intergroup Conflict: Another Look
Realistic competition over scarce vital resources provides one explanation for
perceived conflict of interests. When resources for survival or flourishing are
scarce and groups are locked into negative interdependence (competing for
the same resources), the link between ingroup survival and destruction of the
outgroup is direct. But analyses of warfare and intergroup violence call into
question the necessary role of realistic, objective conflicts of interest in many
cases of intergroup conflict (e.g., Sears & Funk, 1991). Perceived threat is of-
ten subjective and symbolic, and overt conflict often appears to be the prod-
uct of prior intergroup antagonism rather than its cause.
The thesis that is being developed here gives prior status to ingroup 
categorization and identification for understanding the origins of intergroup
behavior. This view assumes that ingroups are formed for reasons indepen-
dent of intergroup conflict and competition over resources. The social and
psychological functions served by ingroup formation and identification pro-
vide the necessary conditions for collective action—either intragroup or 
intergroup. Hence, ingroup attachment and loyalty lay the groundwork for
intergroup conflict, but outgroup antagonism is not a necessary extension of
ingroup positivity and enhancement. Explicating the relationship between
social identity and intergroup conflict requires a fuller understanding, first,
of the functions of ingroup identification and, then, of the additional condi-
tions that lead to a connection between ingroup attachment and outgroup
hostility.
Ingroup Formation: The Boundaries of Cooperation and Trust
If ingroup-outgroup distinctions do not always involve intense (or even
mild) competition or conflict over scarce resources, there is need for a
theory of the evolution of social groups that does not depend on intergroup
conflict per se. Such a theory starts from the recognition that group living
represents the fundamental survival strategy that characterizes the human
species. In the course of our evolutionary history, humans abandoned most
of the physical characteristics and instincts that make possible survival and
reproduction as isolated individuals or pairs of individuals in favor of other
advantages that require cooperative interdependence with others in order
to survive in a broad range of physical environments. In other words, as a
species we have evolved cooperation rather than strength and social learn-
ing rather than instinct as basic adaptations.
28
The Contribution of Identities to Intergroup Conflict


The result is that, as a species, human beings are characterized by obliga-
tory interdependence (Brewer, 1997; Caporael, 1997). For long-term survival,
we must be willing to rely on others for information, aid, and shared re-
sources, and we must be willing to give information and aid and to share 
resources with others. At the individual level, the potential benefits (receiv-
ing resources from others) and costs (giving resources to others) of mutual
cooperation go hand-in-hand and set natural limits on cooperative interde-
pendence. The decision to cooperate (to expend resources to another’s 
benefit) is a dilemma of trust since the ultimate benefits depend on every-
one else’s willingness to do the same (Brewer, 1981). A cooperative system
requires that trust dominate over distrust. But indiscriminate trust (or indis-
criminate altruism) is not an effective individual strategy; cooperation must
be contingent on the probability that others will cooperate as well.
Social categorization and clear group boundaries provide one mechanism
for achieving the benefits of cooperative interdependence without the risk of
excessive costs. Ingroup membership is a form of contingent altruism. By
limiting aid to mutually acknowledged ingroup members, total costs and
risks of nonreciprocation can be contained (see T
akagi, 1996, for a related ar-
gument). Thus, ingroups can be defined as bounded communities of mutual
trust and obligation that delimit mutual interdependence and cooperation. 
An important aspect of this mutual trust is that it is depersonalized
(Brewer, 1981), extended to any member of the ingroup whether personally
related or not. Psychologically, expectations of cooperation and security
promote positive attraction toward other ingroup members and motivate
adherance to ingroup norms of appearance and behavior that ensure one
will be recognized as a good or legitimate ingroup member. Symbols and
behaviors that differentiate the ingroup from local outgroups become par-
ticularly important here, to reduce the risk that ingroup benefits will be 
inadvertently extended to outgroup members, and to assure that oneself 
is recognized as a member of the ingroup and entitled to those benefits. As-
similation within and differentiation between groups is thus mutually rein-
forcing, along with ethnocentric preference for ingroup interactions and 
institutions.
The optimal distinctiveness model of social identity (Brewer, 1991) was
based on this evolutionary perspective on the functions of ingroup for-
mation and differentiation. The theory assumes that psychological me-
chanisms at the individual level (opposing needs for differentiation and 
inclusion) co-evolved with structural requirements at the group level for
successful group functioning and coordination (size and boundedness). The
relative strength of inclusion and differentiation needs at the individual
level interacts with group properties such as size and permeability to de-
termine social identification and ingroup attachment and loyalty. The psy-
chology of assimilation and differentiation limits the extent to which strong
social identification can be indefinitely extended to highly inclusive, super-
ordinate social groups or categories. 
Ingroup Identification and Intergroup Conflict
29


Thus, one implication of optimal distinctiveness theory is that ingroup
loyalty, and its concomitant depersonalized trust and cooperation, is most
effectively engaged by relatively small, distinctive groups or social cate-
gories. Groups that satisfy these needs are those that provide the necessary
boundaries on mutual obligation and trust. Optimal group identities be-
come essential to a secure self-concept. Once group identification has been
achieved, maintaining a secure sense of inclusion and boundedness becomes
tantamount to protecting one’s own existence. When optimal inclusion
within a distinctive ingroup is not achieved, the sense of self is threatened
and vulnerable.
Intergroup Differentiation and Outgroup Stereotypes
A consequence of ingroup identification and intergroup boundaries is that
individuals modify their social behavior depending on whether they are in-
teracting with ingroup or outgroup members. Ingroup behavior is governed
by norms and sanctions that reinforce expectations of mutual cooperation
and trustworthiness. Depersonalized trust is supported by implicit under-
standings that ingroup members will monitor the behavior and interactions
of other group members, sanctioning deviations from group expectations
about appropriate ingroup attitudes and behavior. Thus, shared ingroup
membership may be taken as prima facie evidence that other members of the
group will live by the codes of conduct that bind them as a group (Kramer,
Brewer, & Hanna, 1996).
Just as there is a realistic basis for ethnocentric trust of ingroups, differ-
ences in norms and sanctions applied to ingroup behavior compared to be-
havior in interactions with outgroup members provide a realistic basis for
outgroup distrust and negative stereotypes. At the same time that groups
promote trust and coopration within, they caution wariness and constraint
in intergroup interactions. Thus, even in the absence of overt conflict be-
tween groups, the differentiation between ingroup and outgroup behavior
creates a kind of self-fulfilling prophecy in the realm of intergroup percep-
tions. As LeVine and Campbell (1972) put it, “if most or all groups are, in
fact, ethnocentric, then it becomes an ‘accurate’ stereotype to accuse an
outgroup of some aspect of ethnocentrism” (p. 173). 
Combined with the accentuation principle that exaggerates perceived
differences between social categories, this leads to a set of “universal stereo-
types” to characterize ingroup-outgroup differences. Whereas “we” are
trustworthy, peaceful, moral, loyal, and reliable, “they” are clannish, exclu-
sive, and potentially treacherous. What is particularly interesting about this
pattern of stereotypes is that the same behaviors that are interpreted as rea-
sonable caution on the part of the ingroup in dealings with outgroup mem-
bers become interpreted as “clannishness” and indicators of potential
treachery when exhibited by outgroupers toward the ingroup. 
Evidence for this hypothesized universal difference in ingroup versus out-
30
The Contribution of Identities to Intergroup Conflict


group stereotypes was provided by data from our survey of intergroup per-
ceptions in East Africa (Brewer & Campbell, 1976). Respondents in the 
survey were presented with a lengthy list of character traits (both positive 
and negative) and asked to indicate which groups were most likely to exhibit
each trait. When responses to this trait list were factor analyzed, the first 
factor that emerged was a bipolar evaluative factor interpretable as trustwor-
thiness versus untrustworthiness. Positive loadings on this factor were asso-
ciated with traits such as “peaceful,” “obedient,” “honest,” “gentle,” and
“friendly.” Negative loadings were obtained for traits such as “quarrelsome,”
“disobedient,” “dishonest with others,” “cruel,” and “hot-tempered.” An 
index of positive regard was constructed by taking the frequency with which
a group was mentioned in connection with positive traits on this factor and
subtracting the frequency of mentions in connection with the negative traits.
Comparing ingroup and outgroup ratings on this index revealed that all 30
groups in the survey rated their own group more positively than the average
rating given to outgroups. Further, 27 of the groups rated the ingroup higher
than any outgroup in the sample (Brewer & Campbell, 1976, p. 79). Although
the assignment of negative traits such as “quarrelsome,” “dishonest,” and
“cruel” varied considerably across outgroups, attributions of positive traits
such as “peaceful,” “honest,” and “friendly” were almost universally reserved
for the ingroup.
The socialization experiences that lead to differential expectations and
comfort in ingroup dealings compared to dealings with outgroups may be
generalized in the form of a generic schema that is brought to all intergroup
contexts. Whenever ingroup-outgroup distinctions become salient, the in-
tergroup schema is activated. Just such a generic schema was suggested by
T
ajfel in his initial explanations for the minimal intergroup discrimination
findings (T
ajfel et al., 1971). At that time, T
ajfel (1970) speculated about a
“generic norm” of intergroup behavior that prescribed outgroup competi-
tion. More recently, Insko and Schopler (1987) also postulated such a group
schema as one explanation for their robust finding that in interdepen-
dent situations, people are consistently more competitive in intergroup ex-
changes than when facing the same situation in an interpersonal context—a
finding that is known as the “individual-group discontinuity effect” (Insko
et al., 1987, 1990). A primary component of the group schema, as devel-
oped by Insko and his colleagues, is schema-based distrust—the learned be-
lief or expectation that intergroup relations are inherently competitive and
therefore the outgroup cannot be trusted and the ingroup’s welfare must be
protected (see also Fiske & Ruscher, 1993). In an experimental test of the
schema-based mistrust hypothesis, Insko et al. (1990) found that group dis-
cussion preceding an intergroup prisoner’s dilemma game was characterized
by frequent references to the potential untrustworthiness of the other
group, and there was a strong negative correlation between the amount of
recorded distrust during discussion and later cooperation in the PDG play.
Further, groups more than individuals chose to withdraw from play when
Ingroup Identification and Intergroup Conflict
31


that option was made available, where withdrawal serves as an indicator of
fear that the other player will prefer competition to cooperation.
The Role of Emotions in Intergroup Contexts
When distrust leads groups to compete in an effort to protect themselves
from anticipated outgroup defection, a self-fulfilling cycle of competitive-
ness is set in motion (Kelley & Stahelski, 1970). What is of importance here
is the idea that this cycle is the product of a misattribution process whereby
universal ethnocentric behavior (intragroup trust and lack of trust in inter-
group encounters) is used to infer malevolent intent on the part of out-
groups. Similar misattribution processes may be at work in interpreting and
evaluating intergroup differences more generally. If individuals feel discom-
fort and unease in interactions with outgroups, those feelings may well be
attributed to intergroup differences in values, attitudes, and cultural prac-
tices. Hence, negative affect becomes associated with intergroup differences
and, through a process of misattribution of arousal, intergroup anxiety is
transformed into more virulent intergroup emotions of fear, hatred, or dis-
gust (Smith, 1993; Stephan & Stephan, 2000). It is this emotional compo-
nent that is postulated here to be the critical ingredient that turns inter-
group comparison into intergroup antagonism.
One perspective on the etiology of intergroup emotions can be derived
from the theory of optimal distinctiveness (Brewer, 1991) described earlier.
The theory holds that strong ingroup identification is associated with meet-
ing needs for secure inclusion (though similarity to the ingroup) and secure
differentiation (through intergroup distinctions). From this perspective,
there are two fundamental sources of potential threat to a person’s social
identity. One threat comes from feelings of marginalization from the 
desired ingroup—loss of a secure sense of inclusion and fit. One way of cop-
ing with this type of threat is to reaffirm support for ingroup positivity 
and moral superiority. However, hypervaluation of the ingroup alone may
actually increase the distance between the self and the ingroup ideal. Dero-
gation of outgroups, however, increases the contrast between ingroup 
and outgroup and, indirectly, the relatively similarity between the self 
and the ingroup by comparison. Thus, threats to inclusion are predicted 
to heighten feelings of moral superiority, intolerance of difference, and con-
comitant emotions of contempt and disgust toward relevant outgroups.
The emotions associated with moral superiority may justify some nega-
tive discrimination against outgroups, but do not necessarily lead directly to
hostility or conflict. In various contexts, groups have managed to live in a
state of mutual contempt over long periods without going to war over their
differences. The emotions of contempt and disgust are associated with
avoidance rather than attack, so intergroup peace may be maintained
through segregation and mutual avoidance. However, social changes that
give rise to the prospect of close contact, integration, or influence from the
32
The Contribution of Identities to Intergroup Conflict


outgroup engage the second form of threat to social identity—the threat of
loss of distinctiveness, accompanied by feelings of invasion. The emotions
engendered by this form of threat are fear and anger—emotions that propel
action against the outgroup rather than mere avoidance. This analysis sug-
gests that the combined emotions of contempt (engendered by moral supe-
riority in the service of insecure identity) and anger (engendered by fear of
invasion and loss of distinctiveness) provide the potent ingredients that are
sufficient to kindle hatred, expulsion, and even ethnic cleansing.
Some Moderating Conditions
The preceding analysis indicates that there are a number of steps by which
the effects of social categorization and ingroup-outgroup differentiation
may provide a fertile ground for conflict and hate. As the steps progress, the
line is crossed between the absence of trust and the presence of active dis-
trust, or between noncooperation and overt conflict. In this final section I
consider a few of the conditions that may exacerbate or moderate the
process whereby maintaining ingroup integrity and loyalty paves the way to
outgroup hate and hostility.
Common Goals.
The presence of superordinate goals or common threat is
widely believed to provide the conditions necessary for intergroup coopera-
tion and reduction of conflict (e.g., Sherif , 1966). This belief is an extrapo-
lation of the general finding that intragroup solidarity is increased in the face
of shared threat or common challenge. It may be true that loosely knit in-
groups become more cohesive and less subject to internal factioning when
they can be rallied to the demands of achieving a common goal. The dy-
namics of interdependence are quite different, however, in the case of
highly differentiated social groups. Among members of the same ingroup,
engaging the sense of trust necessary for cooperative collective action is es-
sentially nonproblematic. In an intergroup context, however, perceived in-
terdependence and the need for cooperative interaction make salient the
absence of mutual trust. Without the mechanism of depersonalized trust
based on common identity, the risk of exploited cooperation looms large
and distrust dominates over trust in the decision structure. It is for this rea-
son that I have argued elsewhere (Brewer, 2000) that the anticipation of posi-
tive interdependence with an outgroup, brought on by perceptions of com-
mon goals or common threat, actually promotes intergroup conflict and
hostility. When negative evaluations of the outgroup such as contempt are
also already present, common threat in particular may promote scapegoat-
ing and blame rather than mutual cooperation.
Perceived positive interdependence with the outgroup also threatens 
intergroup differentiation. T
o the extent that feelings of secure inclusion,
ingroup loyalty, and optimal identity are dependent upon the clarity of in-
group boundaries and intergroup distinctions, shared experiences and co-
Ingroup Identification and Intergroup Conflict
33


operation with the outgroup threaten the basis for social identification. Par-
ticularly for individuals who are exclusively vested in a single group identity,
the threat of lost distinctiveness may override the pursuit of superordinate
goals and lead to resistance to cooperation (collaboration) even at the cost
of ingroup self-interest.
Power Politics. Moral superiority, distrust of outgroups, and social compari-
son are all processes that emerge from ingroup maintenance and favoritism
and can lead to hostility and conflict between groups even in the absence of
realistic conflict over material resources or power. When groups are politi-
cal entities, however, these processes may be exacerbated through deliber-
ate manipulation by group leaders in the interests of mobilizing collective
action to secure or maintain political power. Social category differentiation
provides the fault lines in any social system that can be exploited for politi-
cal purposes. When trust is ingroup-based, it is easy to fear control by out-
siders; perceived common threat from outgroups increases ingroup cohe-
sion and loyalty; appeals to ingroup interests have greater legitimacy than
appeals to personal self-interest. Thus politicization—an important mecha-
nism of social change—can be added to the factors that may contribute to a
correlation between ingroup love and outgroup hate.
Culture. In addition to psychological needs, there are social structural and
cultural factors that may promote the intensity of ingroup attachment and
loyalty and enhance the distance between ingroup and outgroups. When
the salience and strength of intragroup interdependence and mutual obliga-
tion is increased, the importance of maintaining group boundaries is also
increased. Hence we might expect that in collectivist societies ingroup-
outgroup distinctions and distrust of outgroups would be higher than in 
individualistic societies where social interdependence is less emphasized.
Indeed, findings from cross-cultural studies of ingroup bias in collectivist
and individualist societies supports this prediction (Triandis, 1995).
Social Structural Complexity. A strong relationship between intense ingroup
favoritism and outgroup antagonism might also be expected in highly seg-
mented societies that are differentiated along a single primary categoriza-
tion, such as ethnicity or religion. And this would be especially true if the
categorization is dichotomous, dividing the society into two significant sub-
groups. Such segmentation promotes social comparison and perceptions of
conflict of interest that give rise to negative attitudes toward outgroups and
high potential for conflict. By contrast, the potential for intergroup conflict
may be reduced in societies that are more complex and differentiated along
multiple dimensions that are not perfectly correlated. Such a complex social
structure gives rise to cross-cutting category distinctions that mean that, at
the individual level, a person may be attached to one ingroup by virtue of
ethnic heritage, to another by religion, to yet another based on occupation,
34
The Contribution of Identities to Intergroup Conflict


or region of residence, and so forth. With this profusion of social identities
some individuals will be fellow ingroup members on one category distinc-
tion but outgroupers on another. Such cross-cutting ingroup-outgroup dis-
tinctions reduce the intensity of the individual’s dependence on any particu-
lar ingroup for meeting psychological needs for inclusion, thereby reducing
the potential for polarizing loyalties along any single cleavage or group dis-
tinction and perhaps increasing tolerance for outgroups in general.
Summary and Concluding Perspectives
The main point of this chapter has been to argue that any relationship be-
tween ingroup identification and outgroup hostility is progressive and con-
tingent rather than necessary or inevitable. Ingroup formation arises from
processes of social categorization and the need for cooperative interdepen-
dence within groups. At the individual level, identification with ingroups 
is motivated by the need to belong and to participate in bounded coopera-
tive social units. The bounded nature of ingroup attachment means that
trust and cooperation are limited to ingroup relations and withheld from 
outgroups.
Once the self is attached to a distinctive ingroup, additional motives to
achieve positive valuation, maintain secure inclusion, and protect ingroup
boundaries are engaged. It is at this point that outgroups may become rele-
vant to the maintenance or enhancement of ingroup identity. Social compe-
tition for positive valuation shifts initial indifference about outgroup out-
comes to concern for the relative position of the ingroup over outgroups.
Under these conditions, benefiting the ingroup comes at the expense of
outgroup benefits, and positive outcomes for the outgroup may arouse re-
sentment and antagonism if the comparison threatens important bases of
ingroup evaluation. This is the first step in connecting ingroup identity to
negative outgroup attitudes and underlies many forms of social discrimina-
tion and intergroup competition for social status.
At a more extreme point in the progression from ingroup identity to in-
tergroup conflict, perceived threats to ingroup inclusion or differentiation
from outgroups promote perceived conflict between ingroup identity and
the very existence of the outgroup or its goals and values. This is the point
at which preservation of the ingroup justifies direct aggression against 
the outgroup and underlies conflicts over political control, dominance, and
wars of annihilation. Such intense outgroup hostility is not a necessary or
inevitable concomitant of strong ingroup identity, but the basic processes of
social categorization that give rise to differential treatment of ingroup and
outgroup lay the groundwork for distrust and vulnerability that make this
escalation possible.
This analysis of the relationship between ingroup identity and outgroup
attitudes has a number of implications for research on intergroup relations
Ingroup Identification and Intergroup Conflict
35


and the prevention of the more virulent forms of intergroup conflict. First,
it implies that the study of ingroup bias, prejudice, and discrimination
should make a clearer distinction between discriminatory attitudes and be-
havior that are based on ingroup preference and positivity and discrimina-
tion that is based on antagonism and hostility toward outgroups. This is not
to suggest that ingroup-based prejudices are entirely benign or acceptable.
Indeed, these are the prejudices that are most likely to be subtle, uncon-
scious, and pervasive and thus more difficult to detect and address than
“old-fashioned” outgroup hatred. But understanding the motives underly-
ing discrimination based on group membership leads to different prescrip-
tions for preventing or reducing the discriminatory outcomes.
When differential cooperation, positivity, and trust toward others is
based on shared ingroup membership, discrimination can best be reduced
by conditions that foster greater inclusiveness, extending the boundaries of
the ingroup to include former outgroup members. This is the basic ratio-
nale for the common ingroup identity model for reducing intergroup bias
through recategorization of ingroup and outgroup into a superordinate 
category identity (Gaertner et al., 1993). This is most likely to be successful
when symbols of superordinate category membership are highly salient and
individual needs for differentiation are not aroused.
When intergroup attitudes and relations have moved into the realm of
outgroup hate or overt conflict, however, the prospect of superordinate
common group identity may constitute a threat rather than a solution
(Brewer, 2000). When outgroup attitudes are characterized simply by lack
of trust rather than active distrust, there is opportunity for capitalizing on
common interests and shared identities. But when intense distrust has al-
ready developed, common group identities are likely to be seen as threats
(or opportunities) for domination and absorption. In this case, the prescrip-
tion for conflict reduction may first require protection of intergroup
boundaries and distinctive identities.
From this analysis there is an inevitable tension between promoting co-
operative interdependence and shared interests at high levels of inclusive-
ness and the need to maintain and recognize intergroup differentiation and
distinctiveness. The negative effects of divisiveness will be most virulent
when individuals have very limited options for meeting their needs for se-
cure, optimal group identification. When social cleavages are defined along
a single dimension of differentiation, vulnerability to threats to inclusion
and differentiation is very high, and maintenance of ingroup identity is
most likely to be achieved at the cost of high levels of distrust and antago-
nism toward outgroups. When societies are more complex, when individu-
als have multiple avenues for meeting needs for inclusion and distinctive-
ness, intergroup comparisons along any particular division are likely to be
less competitive or critical to member identities. Under these conditions,
the positive benefits of distinctive ingroup identities may be attained with-
out the costs of intergroup hostility and conflict.
36
The Contribution of Identities to Intergroup Conflict


Note
1. Note, we are dealing here with prejudice and hostility at the collective level.
The hatred that a specific individual bears toward a particular type of people is not
at issue unless that hatred is grounded in norms and attitudes shared by the ingroup
as a whole.
References
Abrams, D., & Hogg, M. A. (1988). Comments on the motivational status of self-
esteem in social identity and intergroup discrimination. European Journal of Social
Psychology, 18, 317–334.
Allport, G. W. (1954). The nature of prejudice. Cambridge, MA: Addison-Wesley.
Baron, R. A. (1979). Effects of victim pain cues, victim’s race, and level of prior insti-
gations upon physical aggression. Journal of Applied Social Psychology, 9, 110–114.
Bettencourt, B. A., & Dorr, N. (1998). Cooperative interaction and intergroup bias:
Effects of numerical representation and cross-cut role assignment. Personality and
Social Psychology Bulletin, 24, 1276–1293.
Blanz, M., Mummendey, A., & Otten, S. (1997). Normative evaluations and fre-
quency expectations regarding positive and negative outcome alloctions between
groups. European Journal of Social Psychology, 27, 165–176.
Brewer, M. B. (1972). (Series editor). HRAFLEX Ethnocentrism Interview Series.
New Haven, CN: Human Relations Area Files.
Brewer, M. B. (1979). In-group bias in the minimal intergroup situation: A cognitive
motivational analysis. Psychological Bulletin, 86, 307–324.
Brewer, M. B. (1981). Ethnocentrism and its role in intergroup trust. In M. Brewer
& B. Collins (Eds.), Scientific inquiry in the social sciences (pp. 214–231). San Fran-
cisco: Jossey-Bass.
Brewer, M. B. (1991). The social self: On being the same and different at the same
time. Personality and Social Psychology Bulletin, 17, 475–482.
Brewer, M. B. (1996). In-group favoritism: The subtle side of intergroup discrimi-
nation. In D. Messick & A. T
enbrunsel (Eds.), Codes of conduct: Behavioral research
into business ethics (pp. 160–170). New York: Russell Sage Foundation.
Brewer, M. B. (1997). On the social origins of human nature. In C. McGarty & S. A.
Haslam (Eds.), The message of social psychology (pp. 54–62). Oxford: Blackwell.
Brewer, M. B. (2000). Superordinate goals versus superordinate identity as bases of
intergroup cooperation. In D.Capozza & R. Brown (Eds.), Social identity processes:
Trends in theory and research (pp. 117–132). London. Sage.
Brewer, M. B., & Brown, R. (1998). Intergroup relations. In D. Gilbert, S. Fiske, 
& D. Lindzey (Eds.), The handbook of social psychology. Fourth Edition (Vol. 1, 
pp. 99–142). Boston: McGraw-Hill.
Brewer, M. B., & Campbell, D. T. (1976). Ethnocentrism and intergroup attitudes: East
African evidence. Beverly Hills, CA: Sage.
Brewer, M. B., Manzi, J., & Shaw, J. S. (1993). In-group identification as a function
of depersonalization, distinctiveness, and status. Psychological Science, 4, 88–92.
Brewer, M. B., & Weber, J. G. (1994). Self-evaluation effects of interpersonal versus
intergroup social comparison. Journal of Personality and Social Psychology, 66,
268–275.
Ingroup Identification and Intergroup Conflict
37


Caporael, L. R. (1997). The evolution of truly social cognition: The core configura-
tions model. Personality and Social Psychology Review, 1, 276–298.
Crocker, J., & Major, B. (1989). Social stigma and self-esteem: The self-protective
properties of stigma. Psychological Review, 96, 608–630.
Deschamps, J-C., & Brown, R. (1983). Superordinate goals and intergroup conflict.
British Journal of Social Psychology, 22, 189–195.
Diehl, M. (1990). The minimal group paradigm: Theoretical explanations and em-
pirical findings. In W. Stroebe & M. Hewstone (Eds.), European review of social
psychology (Vol. 1, pp. 263–292). Chichester, England: Wiley.
Doise, W., Deschamps, J-C., & Meyer, G. (1978). The accentuation of intra-
category similarities. In H. T
ajfel (Ed.), Differentiation between social groups
(pp. 159–170). London: Academic Press.
Dovidio, J. F., & Gaertner, S. L. (1993). Stereotypes and evaluative intergroup 
bias. In D. Mackie & D. Hamilton (Eds.), Affect, cognition, and stereotyping
(pp. 167–193). San Diego, CA: Academic Press.
Festinger, L. (1954). A theory of social comparison proccesses. Human Relations, 7,
271–282.
Fiske, S. T., & Ruscher, J. (1993). Negative interdependence and prejudcie: Whence
the affect? In D. Mackie & D. Hamilton (Eds.), Affect, cognition, and stereotyping
(pp. 239–268). San Diego, CA: Academic Press.
Frey, D., & Gaertner, S. L. (1986). Helping and the avoidance of inappropriate in-
terracial behavior: A strategy than can perpetuate a non-prejudiced self-image.
Journal of Personality and Social Psychology, 50, 1083–1090.
Gaertner, S. L., & Dovidio, J. F. (1977). The subtlety of white racism, arousal, and
helping behavior. Journal of Personality and Social Psychology, 35, 691–707.
Gaertner, S. L., Dovidio, J. F., & Johnson, G. (1982). Race of victim, non-respon-
sive bystanders, and helping behavior. Journal of Social Psychology, 117, 69–77.
Gaertner, S. L. , Dovidio, J. F., Anastasio, P., Bachman, B., & Rust, M. (1993). The
common ingroup identity model: Recategorization and the reduction of inter-
group bias. In W. Stroebe & M. Hewstone (Eds.), European review of social psy-
chology (Vol. 4, pp. 1–26). Chichester, England: Wiley.
Greenwald, A. G., & Banaji, M. R. (1995). Implicit social cognition: Attitudes, self-
esteem, and stereotypes. Psychological Review, 102, 4–27.
Heider, F. (1958). The psychology of interpersonal relations. New York: Wiley.
Hewstone, M., Fincham, F., & Jaspars, J. (1981). Social categorization and similarity
in intergroup behaviour: A replication with “penalties.” European Journal of 
Social Psychology, 11, 101–107.
Hinkle, S., & Brown, R. (1990). Intergroup comparisons and social identity: Some
links and lacunae. In D. Abrams & M. Hogg (Eds.), Social identity theory: Con-
struction and critical advances (pp. 48–70). London: Harvester Wheatsheaf.
Hirt, E. R., Zillmann, D., Erickson, G., & Kennedy, C. (1992). Costs and benefits of
allegiance: Changes in fans’ self-ascribed competencies after team victory versus
defeat. Journal of Personality and Social Psychology, 63, 724–738.
Insko, C. A., & Schopler, J. (1987). Categorization, competition and collectivity. In
C. Hendrick (Ed.), Personality and Social Psychology Review. Vol. 8: Group Processes
(pp. 213–251). Beverly Hills, CA: Sage.
Insko, C. A., Pinkley, R., Hoyle, R., Dalton, B., Hong, G., Slim, R., Landry, P.,
Holton, B., Ruffin, P., & Thibaut, J. (1987). Individual versus group discon-
38
The Contribution of Identities to Intergroup Conflict


tinuity: The role of intergroup contact. Journal of Experimental Social Psychology,
23, 250–267.
Insko, C. A., Schopler, J., Hoyle, R., Dardis, G., & Graetz, K. (1990). Individual-
group discontinuity as a function of fear and greed. Journal of Personality and 
Social Psychology, 58, 68–79.
Kelley, H. H., & Stahelski, A. J. (1970). The social interaction basis of cooperators’
and competitors’ beliefs about others. Journal of Personality and Social Psychology,
16, 66–91.
Kramer, R. M., Brewer, M. B., & Hanna, B. A. (1996). Collective trust and collec-
tive action: The decision to trust as a social decision. In R. Kramer & T. Tyler
(Eds.), Trust in organizations (pp. 357–389). Thousand Oaks, CA: Sage.
Levin, S., & Sidanius, J. (1999). Social dominance and social identity in the United
States and Israel: Ingroup favoritism or outgroup derogation? Political Psychology,
20, 99–126.
LeVine, R. A. & Campbell, D. T. (1972). Ethnocentrism: Theories of conflict, ethnic 
attitudes and group behavior. New York: Wiley.
McClintock, C. (1972). Social motivation: A set of propositions. Behavioral Science,
17, 438–454.
Mullen, B., Brown, R., & Smith, C. (1992). Ingroup bias as a function of salience,
relevance, and status: An integration. European Journal of Social Psychology, 22,
103–122.
Mummendey, A., & Otten, S. (1998). Positive-negative asymmetry in social dis-
crimination. In W. Stroebe & M. Hewstone (Eds.), European review of social psy-
chology (Vol. 9, pp. 107–143). Chichester, England: Wiley.
Mummendey, A., & Schreiber, H. (1983). Better or just different? Positive social
identity by discrimination against or by differentiation from outgroups. European
Journal of Social Psychology, 13, 389–397.
Mummendey A., & Schrieber, H. (1984). ‘Different’ just means ‘better’: Some obvi-
ous and some hidden pathways to in-group favoritism. British Journal of Social
Psychology, 23, 363–368.
Mummendey, A., & Simon, B. (1989). Better or different? III: The impact of impor-
tance of comparison dimension and relative ingroup size upon intergroup dis-
crimination. British Journal of Social Psychology, 28, 1–16.
Mummendey, A., Simon, B., Dietze, C., Grunert., M., Haeger, G., Kessler, S.,
Lettgen, S., & Schaferhoff, S. (1992). Categorization is not enough: Intergroup
discrimination in negative outcome allocations. Journal of Experimental Social 
Psychology, 28, 125–144.
Neidenthal, P. M., & Beike, D. (1997). Interrelated and isolated self-concepts. Per-
sonality and Social Psychology Review, 1, 106–128.
Perdue, C. W., Dovidio, J. F., Gurtman, M. B., & Tyler, R. B. (1990). “Us” and
“Them”: Social categorization and the process of intergroup bias. Journal of Per-
sonality and Social Psychology, 59, 475–486.
Pettigrew, T. F., & Meertens, R. W. (1995). Subtle and blatant prejudice in Western
Europe. European Journal of Social Psychology, 25, 57–75.
Platow, M., O’Connell, A., Shave, R., & Hanning, P. (1995). Social evaluations of
fair and unfair allocations in interpersonal and intergroup situations. British Jour-
nal of Social Psychology, 34, 363–381.
Rogers, R. W., & Prentice-Dunn, S. (1981). Deindividuation and anger-mediated
Ingroup Identification and Intergroup Conflict
39


interracial aggression: Unmasking regressive racism. Journal of Personality and 
Social Psychology, 41, 63–73.
Rubin, M., & Hewstone, M. (1998). Social identity theory’s self-esteem hypothesis:
A review and some suggestions for clarification. Personality and Social Psychology
Review, 2, 40–62.
Sachdev, I., & Bourhis, R. (1991). Power and status differentials in minority and ma-
jority group relations. European Journal of Social Psychology, 21, 1–24.
Sears, D. O., & Funk, C. L. (1991). The role of self-interest in social and political
attitudes. In M. Zanna (Ed.), Advances in experimental social psychology (Vol. 24, 
pp. 2–92). New York: Academic Press.
Sherif, M. (1966). In common predicament: Social psychology of intergroup conflict and 
cooperation. New York: Houghton Mifflin.
Sidanius, J. (1993). The psychology of group conflict and the dynamics of op-
pression: A social dominance perspective. In S. Iyengar & W. McGuire (Eds.),
Explorations in political psychology (pp. 183–219). Durham, NC: Duke University
Press.
Smith, E. R. (1993). Social identity and social emotions: T
oward new conceptualiza-
tions of prejudice. In D. Mackie & D. Hamilton (Eds.), Affect, cognition, and
stereotyping (pp. 297–315). San Diego, CA: Academic Press.
Stangor, C., Sullivan, & Ford, T. E. (1991). Affective and cognitive determinants of
prejudice. Social Cognition, 9, 359–380.
Stephan, W., & Stephan, C. (2000). An integrated threat theory of prejudice. In 
S. Oskamp (Ed.), Reducing prejudice and discrimination (pp. 23–45). Mahwah, NJ:
Lawrence Erlbaum.
Struch, N., & Schwartz, S. H. (1989). Intergroup aggression: Its predictors and 
distinctness from in-group bias. Journal of Personality and Social Psychology, 56,
364–373.
Sumner, W. G. (1906). Folkways. New York: Ginn.
T
ajfel, H. (1969). Cognitive aspects of prejudice. Journal of Social Issues, 25, 79–97.
T
ajfel, H. (1970). Experiments in intergroup discrimination. Scientific American,
223(2), 96–102.
T
ajfel, H. (1978). Differentiation between social groups: Studies in the social psychology of
intergroup relations. London: Academic Press.
T
ajfel, H. (1981). Human groups and social categories. Cambridge: Cambridge Univer-
sity Press.
T
ajfel, H., Billig, M., Bundy, R., & Flament, C. (1971). Social categorization and 
intergroup behaviour. European Journal of Social Psychology, 1, 149–178.
T
ajfel, H., & Turner, J. C. (1986). The social identity theory of intergroup behavior.
In S. Worchel & W. Austin (Eds.), Psychology of intergroup relations (pp. 7-24).
Chicago: Nelson-Hall.
T
ajfel, H., & Wilkes (1963). Classification and quantitative judgement. British Jour-
nal of Psychology, 54, 101–114.
T
akagi, E. (1996). The generalized exchange perspective on the evolution of altru-
ism. In W. Liebrand & D. Messick (Eds.), Frontiers in social dilemma research
(pp. 311–336). Berlin: Springer-Verlag.
T
esser, A. (1988). T
oward a self-evaluation maintenance model of social behavior. 
In L. Berkowitz (Ed.), Advances in experimental social psychology (Vol. 21, pp. 181–
227). San Diego, CA: Academic Press.
Triandis, H. C. (1995). Individualism and collectivism. Boulder, CO: Westview Press.
40
The Contribution of Identities to Intergroup Conflict


Turner, J. C. (1975). Social comparison and social identity: Some prospects for 
intergroup behaviour. European Journal of Social Psychology, 5, 5–34.
Turner, J. C. (1978). Social categorization and social discrimination in the minimal
group paradigm. In H. T
ajfel (Ed.), Differentiation between social groups (pp. 101–
140). London: Academic Press.
Turner, J. C. , Hogg, M., Turner, P
., & Smith, P
. (1984). Failure and defeat as deter-
minants of group cohesiveness. British Journal of Social Psychology, 23, 97–111.
Turner, J. C. , Hogg, M., Oakes, P., Reicher, S., & Wetherell, M. (1987). Rediscover-
ing the social group: A self-categorization theory. Oxford: Blackwell.
Weber, J. G. (1994). The nature of ethnocentric attribution bias: Ingroup protection
or enhancement? Journal of Experimental Social Psychology, 30, 482–504.
Wood, J. V. (1989). Theory and research concerning social comparisons of personal
attributes. Psychological Bulletin, 106, 231–248.
Ingroup Identification and Intergroup Conflict
41


Thomas Hylland Eriksen
3
Ethnic Identity, National Identity,
and Intergroup Conflict
The Significance of Personal Experiences
There exists by now an enormous literature on contemporary ethnicity and
nationalism in its various forms. It deals with a wide range of phenomena—
North American multiculturalism and indigenous rights movements, post-
Soviet ethnonationalism in Central and Eastern Europe, urban minority
dilemmas and Islamic revivalism in Western Europe, indigenista movements
in Latin America, and processes of political fission and fusion in contempo-
rary Africa. The analytical focus of this chapter will be on the concept of
identity. The naïve question to be asked at the outset is: what is it about
identity politics that makes it such a formidable force in the contemporary
world? Some important issues will have to be omitted—notably the rela-
tionships among identity politics, globalization, and reflexive modernity
(e.g., Bauman, 1993; Friedman, 1994; Giddens, 1994). Instead, what is of-
fered amounts to an anthropological perspective on the relationship be-
tween personal identity and political identity, using conflicts based on iden-
tity politics as empirical examples. 
This chapter sets out to do three things. First, a brief overview of the
standard social anthropological perspective on the politics of identity is pro-
vided. Identity politics should be taken to mean political ideology, organiza-
tion, and action that openly represents the interests of designated groups
based on “essential” characteristics such as ethnic origin or religion, and
whose legitimacy lies in the support of important segments of such groups.
Membership in such groups is generally ascribed, unlike membership in
other political groups (socialists, liberals, trade unions, etc.). Second, by way
of examples from India, Fiji, and Yugoslavia, parallels and differences be-
tween some such conflicts are highlighted. The examples are chosen mainly
42


for their differences: what they have in common is, apart from their tempo-
ral location in the 1990s, apparently only that they are based on ideologies
of culture and identity. The Yugoslav conflict has been extremely brutal and
tragic; the Fijian one has involved only a few casualties, but has led to im-
portant constitutional changes; while the Indian conflict, although occa-
sionally bursting into violence, is largely contained within the framework of
institutional Indian politics. If it can persuasively be argued that these con-
flicts have important features in common, it is likely that those features will
also be present in other settings where identity politics has a major impact.
The third and final part of the chapter draws on classic political anthropol-
ogy in order to suggest some general features of the relationship between
personal identity and politics.
The Anthropology of Identity
The last decades of the 20th century saw a dramatic reconceptualization 
of core concepts, including culture and society, within the social sciences.
Until the 1960s, the close overlap between culture and ethnicity, or even
culture and nationhood, was generally taken for granted in the scholarly
community. During the past 30 years, however, hardly a single serious con-
tribution to the field has failed to point out that there is no one-to-one rela-
tionship between culture and ethnicity (the seminal text here is Barth,
1969); that cultural differences cut across ethnic boundaries; and that ethnic
identity is based on socially sanctioned notions of cultural differences, not
“real” ones. While ethnic identity should be taken to refer to a notion of
shared ancestry (a kind of fictive kinship), culture refers to shared represen-
tations, norms, and practices. One can have deep ethnic differences without
correspondingly important cultural differences (as in the Bosnian example
below); and one can have cultural variation without ethnic boundaries (as,
for example, between the English middle class and the English working
class).
Several recent debates in anthropology and neighboring disciplines pull
in the same direction: away from notions of integrated societies or cultures
toward a vision of a more fragmented, paradoxical, and ambiguous world.
The currently bustling academic industry around the notion of globaliza-
tion (see Featherstone, 1990, for an early, influential contribution) repre-
sents an empirically oriented take on these issues, focusing on the largely
technology-driven processes that contribute to increasing contact across
boundaries and diminished importance of space. This focus on unbounded
processes rather than isolated communities has contributed to a reconcep-
tualization of the social, which is radically opposed to that of classic Durk-
heimian sociology and anthropology; where flux, movement, and change
become the rule and not the exception in social life (Strathern, 1991; Han-
nerz, 1992; Lash & Urry, 1994). 
Ethnic Identity, National Identity, and Intergroup Conflict
43


Particularly in North America, the classic concept of culture has been
used for what it is worth in domestic identity politics, leading in some cases
to controversial policies of multiculturalism, where individuals have been
endowed with special rights in accordance with their ethnic origins. Critics
might point out that multiculturalism in some of its versions resembles
apartheid; also, that by positing a simple one-to-one relationship between
ethnic origin and culture, it not only encouraged a “disuniting of America”
(Schlesinger, 1992) but also contributed to reifying misleading notions of
culture seen as the commonalities of a bounded set of individuals, like so
many nationalisms writ small. 
A further disruptive tendency has been the so-called postcolonial move-
ment in literary studies, spilling into anthropology and other disciplines,
that has raised the question of who has the right to identify whom; a stan-
dard text in this field of discourse is Edward Said’s Orientalism (1978), al-
though Frantz Fanon developed similar insights two decades earlier. Said
and others argued, briefly, that ethnocentrism was deeply embedded in
Western scholarship dealing with non-Western peoples. Postcolonial critics
also tend to call attention to the multiplicity of voices (an academic cliché
by the late 1990s) present in any society and the general unwillingness of
academic researchers to give all of them the attention they deserve. 
T
wo related debates defined the field for many years. First, there was the
controversy over primordialism and instrumentalism. Was ethnic identity
“primordial”—that is, profoundly rooted in, and generative of, collective
experiences; or did it arise as an ad hoc supplement to political strategies?
An early, powerful defense of the instrumentalist view was represented in
Abner Cohen’s work on urban ethnicity in Africa (1969, 1974), showing the
conscious manipulation of kinship and cultural symbols by political entre-
preneurs seeking political gain. This perspective is still used with consider-
able success in studies of identity politics. Who, then, were (or are) the pri-
mordialists? Clifford Geertz is often associated with this view, arguing as 
he does along hermeneutic lines that cultural systems are more or less self-
sustaining and are thus not subject to the willful manipulation of individuals
(Geertz, 1973), a perspective he retained when writing about nationhood 
in the Third World (Geertz, 1967). Typically, however, ethnicity studies
were—and are—instrumentalist in their basic orientation (Rex, 1997). 
The second debate, usually framed as the opposition between construc-
tivism and essentialism, concerns the question of whether ethnic or national
communities are created more or less consciously, or whether they grow or-
ganically, as it were, out of preexisting cultural communities. In nationalism
studies the most highly profiled antagonists regarding this issue have been
the late Ernest Gellner (1983, 1997) and Anthony D. Smith. Smith (1986,
1991) has developed an intermediate position in arguing the importance of
preexisting ethnies for the development of nationalism while acknowledging
its essential modernity. Gellner, on the other hand, champions the view that
nations are entirely modern creations—the progeny of industrialism and
44
The Contribution of Identities to Intergroup Conflict


the state—that more or less fraudulently invent their past to gain a sem-
blance of antiquity and deep roots; his final statement on the issue is repro-
duced in the posthumously published Nationalism (Gellner, 1997) under the
heading “Do Nations Have Navels?”, a pun on the biblical enigma relating
to Adam’s navel. Regarding definitions, there are also important differences
between theorists. While Gellner holds nations to be ethnic groups who ei-
ther control a state or who have leaders who wish to do so, Anderson (1983)
sees no necessary link between the abstract “imagined community” of the
nation and particular ethnic groups; indeed, several of the main examples in
his famous book on “imagined communities,” including the Philippines and
Indonesia, are multiethnic countries. Yet others have distinguished between
ethnic nations and “civic” ones (Smith, 1991). There is nonetheless general
agreement that nations are by definition linked with states, whether they
are based on a common ethnic identity or not.
In anthropology at least, the recent shift toward the study of identities
rather than cultures has entailed an intense focus on conscious agency and
reflexivity; and for many anthropologists, essentialism and primordialism
appear as dated as pre-Darwinian biology. In addition, there seems to be
good political sense in discarding the old, static view of culture, which is be-
ing used for many political purposes that are difficult to endorse by acade-
mics committed to democratic values, ranging from the Balkan war to dis-
crimination against ethnic minorities in Western Europe. Further, this is an
age when the informants talk back. It could, perhaps, be said that a main
purpose of an earlier anthropology consisted of identifying other cultures.
Representatives of these so-called other cultures are now perfectly able to
identify themselves, which leaves the scholars either out of a job or with a
new mission—that is, to identify their identifications—in other words, to
study reflexive identity politics. 
Not so many years ago, anthropology was still a discipline fueled by a
programmatic love of cultural variation for its own sake, and anthropolo-
gists involved in advocacy tended to defend indigenous peoples’ or other
minorities’ traditional ways of life against the onslaught of modernity. A
main tendency in recent years has, on the contrary, consisted of decon-
structing instrumentalist uses of notions of authenticity and tradition, and
showing not only that the internal variation within a group is much greater
than one would expect but also that traditionalist ideologies are, paradoxi-
cally, direct results of modernization (e.g., Roosens, 1989). This theoretical
shift is a very significant one. It offers a method for investigating extremely
well the strategic action, politics of symbols, and contemporary processes of
identity politics within a uniform comparative framework. 
In other words, cracks in the edifice of mainstream social and cultural
anthropology, some of them directly inspired by events well beyond the
confines of academia, have led to a widespread reconceptualization of soci-
ety and culture. Reification and essentialism have become central terms of
denunciation; multiple voices, situational identification, and cultural flows
Ethnic Identity, National Identity, and Intergroup Conflict
45


are some of the key words delineating the current intellectual agenda. It has
become difficult if not impossible to talk of, say, Nuer culture, Hopi cul-
ture, Dutch culture, and so on, since such terms immediately invite critical
questions of whose Nuer culture, Hopi culture, and so on, intimating that
there are an infinite number of versions of each culture, none of which is
more “true” than the others (Holland, 1997). Ethnicity and nationalism,
then, become the political reifications or constructions of a particular au-
thorized version of a culture, freezing that which naturally flows, erecting
artificial boundaries where they did not exist before, trimming and shaping
the past to fit present needs, and inventing traditions where no organic tra-
ditions exist, or are not adequate, to ensure a sense of continuity with the
past. 
A new kind of political responsibility has entered academia in acute ways
during the last decades. Academic or semiacademic statements about na-
tions, ethnic groups, or cultures may now immediately be picked up, or as-
similated more or less subconsciously, by ideologists and politicians wishing
to build their reputation on national chauvinism, ethnic antagonism, enemy
images, and so on. The liberal academic establishment thus wags a warning
finger at those who dare to talk of culture as the cause of conflicts, shaking
their heads sadly over those lost souls who have not yet heeded the words of
leading theorists such as Barth (1969) and Gellner (1983), criticizing those
who do not realize that culture is chimerical and fleeting, and that reified
culture is a dangerous tool. It is, thus, not only intellectually correct but
also politically correct to reject all forms of essentialism. 
The current scholarly orthodoxy on ethnicity and the politics of identity
can be summed up as follows:
• Although ethnicity is widely believed to express cultural differ-
ences, there is a variable and complex relationship between eth-
nicity and culture; and there is certainly no one-to-one relationship
between ethnic differences and cultural ones. 
• Ethnicity is a property of a relationship between two or several
groups, not a property of a group; it exists between and not within
groups. 
• Ethnicity is the enduring and systematic communication of cultural
differences between groups considering themselves to be distinct. It
appears whenever cultural differences are made relevant in social
interaction, and it should thus be studied at the level of social life,
not at the level of symbolic culture. 
• Ethnicity is thus relational and also situational: the ethnic character
of a social encounter is contingent on the situation. It is not, in
other words, inherent.
This instrumentalist framing of ethnicity, which may appear simply as a
set of methodological guidelines, has firm, although usually untheorized,
philosophical foundations and is, as I have tried to show elsewhere (Eriksen,
46
The Contribution of Identities to Intergroup Conflict


1998b), deeply embedded in empiricist thought. I shall argue that this ap-
proach, notwithstanding its strengths, is limited in overemphasising choice
and strategy (instrumental aspects) when analyzing identity politics. As a re-
sult, the self is taken for granted (A.P. Cohen, 1994), and it is therefore not
shown how it can be possible to mobilize particular aspects of personal
identity for antagonistic identity politics. Yet, anthropology as a discipline is
in a privileged situation to study the dynamics of identity politics, precisely
because of its focus on the ongoing flow of social interaction.
Although an enormous amount of anthropological research has been
carried out on ethnicity and nationalism since around 1970, surprisingly
few studies have dealt with violent conflicts and conflict resolution (see,
however, T
ambiah, 1994; Turton, 1997). The dominant approaches to eth-
nicity have been instrumentalist (with a focus on politics) or constructivist
(with a focus on ideology), and research questions have concentrated on the
establishment and reproduction of ethnically incorporated groups, not on
the circumstances under which ethnicity may become politically less impor-
tant. While my examples (below) and the ensuing discussion will indicate
the fruitfulness of these approaches, it is also necessary to point out the
need for a phenomenological understanding of social identity, which sees it
as emerging from experiences, not as a mere construct of ideology. In this, I
follow scholars such as A. P. Cohen (1994) and Jenkins (1996), who have
called for an anthropology of identity that does not concentrate exclusively
on its political and ideological aspects, but also strives to understand the
self. 
As noted by Holland (1997), anthropologists are generally associated
with a culturalist view of the self, arguing its cultural specificity as against
psychologists, who have been more prone to a universalist view—the self as
something proper not to particular cultures but to humanity, with universal
characteristics lurking below a thin veneer of culture. For the purposes of
the present argument, it is not necessary to take a stance on this contro-
versy, partly since it can be presupposed that modernity creates a particular
kind of selves with important shared characteristics everywhere (Giddens,
1991), but also because this chapter restricts its scope to the relationships
among personal experiences, ideology, and political mobilization. As the 
examples will, it is hoped, show, the similarities are more striking than the
differences here.
Culture and the Breakup of Yugoslavia
No other recent ethnic conflict has been more intensely studied, discussed,
and moralized over than the breakup of Yugoslavia in 1991. This was 
followed by three major wars and a number of smaller skirmishes, and the
situation in many parts of ex-Yugoslavia remains unstable and tense. In 
Europe, the outbreak of war in Yugoslavia has been interpreted by hun-
Ethnic Identity, National Identity, and Intergroup Conflict
47


dreds of commentators. In the press, it was occasionally argued, along 
sociobiological or Hobbesian lines, that humans are driven by aggressive
instincts that emerge when the social fabric falls apart, in this case orga-
nized as factions ultimately based on kinship. Another view, popular in the
European nationalist right, implied that ethnic conflict was inevitable when
different groups are forcibly integrated into one state. Most scholars have,
on the contrary, tended to focus on the cultural logic of feuding in Balkan
society, the deep economic crisis underlying the conflict, or the rise of 
Serbian supremacy during the 1980s, looking for contradictions within Yu-
goslav society rather than into human nature for an explanation of the con-
flict. A widespread view nevertheless sees the cultural differences between
the constituent groups as a basic cause of the conflict (cf., the influential
analysis by Ignatieff, 1994, or Huntington’s controversial model, 1996). 
Yugoslavia was a state that came into being twice: after the First World
War and after the Second World War. The first Yugoslav republic (1918–
41) was for all practical purposes the Serb monarchy writ large, and was rid-
dled with continuous internal tension. Croats and Slovenes reluctantly 
supported King Alexander’s regime, seeing it as a possible defense against
Italian, Austrian, and Hungarian aggressors. It was a precarious state, peri-
odically dictatorial, that had been on the verge of collapse several times 
before it fell apart as Germany invaded the country in 1941. While Serbs
generally resisted the Germans, Croats collaborated and saw the German
intervention as an opportunity to create their own state. 
Before the formation of the second Yugoslav republic in 1945, both of
the largest constituent groups suffered large-scale massacres—Serbs at the
hands of Croatian fascists (Ustasa) in 1941 and Croats killed by Serbian
communists (Partisans) in 1945. The new Yugoslav state (1945–91) was a
nonaligned socialist federation led by a pro-Serbian Croat, Josip Broz Tito,
until his death in 1980. Ethnicity was officially declared a nonissue in so-
cialist Yugoslavia. This official blindness stems from the Marxist view that
class is a more objective and more authentic vessel of social identity than
ethnicity or nationality, which were officially seen as expressions of false
consciousness. This does not, however, imply that Yugoslav policies were
particularly repressive regarding expressions of cultural distinctiveness sym-
bolizing ethnic identity; on the contrary, in this area Yugoslavia was more
liberal than many Western European countries whose leadership feared
separatism and social fission. Ethnic identity was seen as politically irrele-
vant, and partly for this reason, the use of various languages and the prac-
tice of different religions were tolerated in civil society. It is true that the
merging of Serbian and Croatian into one language, Serbo-Croatian, in 
the 1950s signaled an attempt at building a unitary Yugoslav identity, but
the two languages were so closely related that few, except Croat intellectu-
als reacting against the relegation of specific Croat variants as “dialect,”
seem to have taken offense (Schöpflin, 1993). Albanian remained an official
language in Kosovo, as did Slovene in Slovenia. 
48
The Contribution of Identities to Intergroup Conflict


There were nearly twice as many Serbs as Croats in Yugoslavia, and
twice as many Croats as Slovenes or Muslims. Regarding the territorial 
dimension, Slovenes were—and are—largely confined to the nearly mo-
noethnic republic of Slovenia. Both Croatia and Serbia had large minorities
of Serbs and Croats, respectively, as well as Gypsies, while Serbia also in-
cluded nearly a million Albanians (in Kosovo) and smaller numbers of Hun-
garians and others. In Macedonia, most of the population were (and are)
Slav-speaking Macedonians, while the Montenegrins of Montenegro are
culturally close to the Serbs. Bosnia-Herzegovina was the most throroughly
mixed republic, with roughly equal numbers of Serbs, Croats, and Bosnian
Muslims, often living in mixed areas. As is well known, the conflicts of the
1990s have modified this picture somewhat, creating large monoethnic 
territories in formerly mixed areas.
Although the ruling Communist party seems to have believed that a
common Yugoslav identity would eventually supersede the national identi-
ties based on ethnic membership, ethnic identity remained strong in most
parts of the country throughout the postwar era. There were nevertheless
important exceptions, particularly in cities such as Belgrade, Sarajevo, and
Zagreb, where many people increasingly identified themselves primarily as
Yugoslavs and where mixed marriages were common.
Ethnic identities did, in other words, not disappear during the existence
of Yugoslavia. In some urban areas they were arguably weakened, but it
could be—and has been—argued that the nonethnic character of Yugoslav
politics actually led to its strengthening as a vehicle for the political opposi-
tion and made it possible for Serbs to gain control over the armed forces
and state bureaucracy: since political ethnicity officially did not exist (only
cultural ethnicity did), there were no institutionalized ways of preventing
one group from dominating the public sector. 
The wars in the former Yugoslavia have bequeathed to the world the 
neologism ethnic cleansing. It is nevertheless easy to show that the conflicts
involving Serbs, Croats, Bosnian Muslims, Slovenes, and Albanians were
never conflicts over the right to assert one’s ethnic or cultural identity, but
were based on competing claims to rights such as employment, welfare, and
political influence. What needs to be explained is the fact that the conflicts
over these resources were framed in ethnic terms rather than being seen as,
say, regional, class-based, or even ideological. 
The relevant questions are therefore: what is the stuff of ethnic identity
in the former Yugoslavia; in which ways do the groups differ from one an-
other, and why did group allegiances turn out to be so strong? Bosnia-
Herzegovina may be considered as an example. There are three large ethnic
groups inhabiting Bosnia: Serbs, Croats, and Muslims. The main difference
between the groups is religion—Serbs are Orthodox and Croats are
Catholics. (Ironically, religious fervor was not particularly widespread in
prewar Bosnia.) They all have common origins: Slav immigration into 
Illyria from the north took place between c. a.d. 400 and 700, and the cul-
Ethnic Identity, National Identity, and Intergroup Conflict
49


tural differences between Croats and Serbs are perhaps comparable to those
between Norwegians and Swedes. The “objective” difference between
Bosnian Christians and Bosnian Muslims, further, has been compared to
the difference between English Protestants and English Catholics (Cornell
& Hartmann, 1998). Unlike the impression sometimes given in Serbian and
Croatian propaganda, Bosnian Muslims are not the descendants of alien 
invaders, but of locally residing converts. Although each group has its nu-
merical stronghold, many Bosnian regions and villages were mixed before
the war. This implies, among other things, that they went to school to-
gether, worked together, and took part in various leisure activities together.
A Serbian villager in Bosnia had more in common, culturally speaking, with
a Muslim co-villager than with a Serb from Belgrade. This would hold true
of both dialect spoken and way of life in general. However, since religion
turned out to be the central marker of collective identity in the Bosnian
conflict, the effective boundary was drawn not between villagers and city-
dwellers, but between religious categories.
The boundaries between the groups may seem arbitrary. However, 
the large, “national” groups are clearly embedded in smaller, local networks
based on kinship and informal interaction, as well as being culturally
founded in religious schisms, collective myths or memories of treason and
resistance under Ottoman rule, massacres, deception, and humiliations. 
Although it is tempting to argue that any so-called cultural trait can be 
exploited in the formation of national or ethnic groups, it is obvious that
not just anything will do. Nothing comes out of nothing, and strong collec-
tive identities—such as the ones revealed during the war in Bosnia—are 
always embedded in personal experiences. In one of the most detailed 
accounts of ethnicity at the village level in prewar Bosnia, Bringa (1996)
shows that although cultural differences between the groups were perhaps
negligible, and although relations among Serbs, Muslims, and Croats might
be cordial at the local level, there were nevertheless important social prac-
tices of affiliation that created boundaries between them—not in the cos-
mopolitan Sarajevo middle class embraced by Western commentators, per-
haps, but elsewhere. Intermarriage was rare, the close informal networks 
of friends tended to be monoethnic, and the discrete groups maintained 
different, sometimes conflicting myths of origin. The intimate sphere, in
other words, seems to have been largely monoethnic and by this token,
Bosnia was a plural society in the classic sense (Furnivall, 1948); the public
arenas were shared, but the private ones were discrete. 
One may choose not to speak of such features of social reality and every-
day life in terms of “culture,” but they are no more “invented” than any
other social fact. People do not choose their relatives, they cannot choose to
do away with their childhood and everything they learned at a tender age.
These are aspects of identity that are not chosen, that are incorporated and
implicit. People relate to them as reflexive agents, but they do so within
limitations that are not chosen. Such limitations form the objective founda-
50
The Contribution of Identities to Intergroup Conflict


tions of social identification. When analysts such as Cornell and Hartmann
(1998) argue that ethnic identities before the war were weak, sometimes so-
cially irrelevant, and in many cases ambiguous (many had parents from dif-
ferent groups), they refer to particular segments of society—a main example
is the Yugoslav basketball team—whose members were active participants in
pan-Yugoslav contexts. 
The maintenance of ethnic boundaries in socialization and the private
sphere reveals a main cause of the failure of Yugoslav social engineering in
doing away with ethnic identification. It does not explain the outbreak of
war in the early 1990s, but it indicates why the groups that emerged were 
so strong, and why they were based on ethnicity (seen as fictive kinship) 
instead of, say, class or region. Their foundation must be sought not in the
biology of kinship, as some might want to argue, but in the phenomenology
of social experience, the raw material of personal identity. This argument
will be elaborated after an examination of two very different examples of 
intergroup conflict in polyethnic societies.
The Fijian Coup-D’etat
The Pacific island-state of Fiji, located at the ethnographic crossroads of
Melanesia and Polynesia, is perhaps less heterogenous than Yugoslavia, but
it is scarcely less ethnically divided. Its population of about 800,000 is
largely composed of two ethnic categories: Fijians and Indians. The Fijians
are indigenous, largely Christian, speak a Polynesian language, and make
up slightly less than half the population. The Indians are uprooted “over-
seas Indians” whose ancestors were brought to Fiji during colonialism un-
der the British indentureship system described, probably a trifle too grimly,
as “a new system of slavery” by Tinker (1974). They are overwhelmingly
Hindus (with a Muslim minority), speak a locally modified dialect of Hindi,
and were slightly more numerous than the indigenous Fijians until the po-
litical changes in the late 1980s leading to mass migration of Indo-Fijians.
The small minorities of Europeans and Chinese are politically insignificant,
but the economically powerful Europeans, representing the former colonial
regime, have in no small measure shaped the Fijian public sphere, notably
through establishing English as the national lingua franca.
The relatively brief history of democratic Fijian politics up to 1987 has
been described with the metaphors of balance and power sharing (Premdas,
1993; Kelly, 1998). While Indians in practice wielded disproportionate eco-
nomic power, it was tacitly agreed that Fijians should be politically para-
mount. However, there were indications that the “equilibrium” model was
under severe stress in the early- to mid-1980s, and in the 1987 elections, a
coalition supported by most Indians and only a few Fijians won. Although
the new prime minister was a Fijian, many saw his government as a vehicle
for Indian communal politics. In May and September 1987, the military
Ethnic Identity, National Identity, and Intergroup Conflict
51


seized power through two successive coups-d’etat, explicitly doing so to
protect “native” Fijian interests. In the period following the coups, thou-
sands of Indians emigrated. 
The tension between Fijians and Indians had been evident throughout
Fijian history. Unlike the situation in Yugoslavia and especially Bosnia, 
nobody would question the view that there are deep cultural differences 
between Indians and Fijians. Their languages, cultural traditions, religions,
and gender relations differ markedly—indeed, one of their recent ethnogra-
phers has poetically described striking, and culturally potent, differences 
in body language between Fijians and Indians (Williksen-Bakker, 1991).
There seems to be little informal interaction between the groups, most
rural areas are dominated by one or the other, and intermarriage has always
been nearly nonexistent. The effective separation of Fijians from Indians
has always been much deeper than that obtaining between the major groups
(or “nationalities”) in rural Bosnia after the Second World War. Until 1987,
policies of compromise had nonetheless ensured political stability and had
made Fiji a remarkably liberal and relaxed society.
Invoking the slogan “Fiji for Fijians” (which had been launched before
the 1977 elections), the writers of the new constitution, promulgated in
1990, ensured continued indigenous Fijian dominance of the political
sphere, by according that group disproportionate representation in parlia-
ment, ruling that only Fijians can become prime ministers, and giving 
Fijians preferential treatment in other areas as well, such as religion. In ad-
dition, Fijians are guaranteed control over most of the arable land, about 82
percent of which has been communally owned by Fijian kin groups since
the beginning of colonialism in the 19th century. 
The conflict culminating in the military coups can, at one level of analy-
sis, be seen as a clear case of group competition. Indians have done better
economically than Fijians. Ironically, this may partly be explained through
the British colonial policy of indirect rule relating to Fijians, who were al-
lowed to retain important traditional institutions, such as chieftainship and
the rudiments of a caste system, making them in consequence unprepared
to compete with Europeans and Indians in a capitalist economy later. The
Indians, by contrast, had shed important aspects of their traditional social
organization in the process of migration, and were accustomed to economic
individualism through the indentureship system whereby they were made
to work on the European-owned plantations. 
The demographic growth rate among Indians has been higher than
among Fijians (just as the Muslims in Bosnia were more prolific than the
other groups), and there was generally a growing sentiment among Fijian
leaders that they were becoming a minority in the land of their ancestors.
The inegalitarian measures introduced by the military regime, discriminat-
ing between categories of citizens on ethnic grounds, were condemned by
the international community, but less strongly than one might have ex-
pected in a different setting. What is remarkable about the Fijian case is the
52
The Contribution of Identities to Intergroup Conflict


nativist quality of the supremacist rhetoric—how they brought “sons-of-
the-soil” arguments to bear on a national legislation, creating, in effect, a
two-tier society where non-Fijians were relegated to the status of second-
class citizens. They argued that their culture, like that of Maoris in New
Zealand and Aborigines in Australia, was threatened with marginalization
from outside forces. In contrast, nobody in countries like Mauritius,
Trinidad and T
obago, and Guyana—which are in many ways similar to Fiji
but lack substantial indigenous populations—would have been able to in-
voke arguments of cultural authenticity and preservation of traditional cul-
tures in a bid to introduce differential treatment for different ethnic groups
(see Eriksen, 1992, 1998a, for details on Trinidad and Mauritius).
It may well be asked whether contemporary Fijians, being Methodists and
proficient English speakers, are any less culturally uprooted than, say, Trini-
dadians of African descent or, for that matter, Fijians of Indian descent. The
question of cultural authenticity is outside the scope of this chapter; let me
now make a few more pertinent points relating to this example.
Fiji had developed an informal formula for interethnic accommodation
where the largest ethnic groups divided societal sectors between them. In
addition there were—and still are, at least to some extent—developing
fields of shared meaning and cross-cutting alignments, such as the common
use of English as a national language and a shared educational system. Nev-
ertheless the segregation between the groups in both social and cultural do-
mains is more striking to the outside observer than tendencies toward as-
similation. Conflict avoidance would thus have to rely chiefly on group
compromise rather than the development of a hybridized, shared identity—
an option that has occasionally been proposed by politicians and intellectu-
als in other insular, postcolonial plantation societies such as Mauritius and
Trinidad. Finally, owing to historical circumstances and cultural differences,
Indians and Fijians have participated in different ways and have succeeded
to varying degrees in the modern sectors of politics and the economy. It
could indeed be argued that processes of modernization in Fiji, far from 
reducing cultural differences, have deepened them, at least at the socially
operational level. Unlike in Bosnia, it is possible to refer to differences in
local organization, cosmology, and traditional economical practices when
accounting for the ethnic conflict in Fiji, which is nevertheless much less 
violent than the Bosnian one. This is a reminder, against cultural determin-
ists à la Huntington (1996), of the relative unimportance of cultural differ-
ences for ethnic conflict—and it also indicates one of the main strengths of
the constructivist-instrumentalist perspective on identity politics: cultural
differences do not in themselves lead to intergroup conflict, but are invoked
strategically to mobilize support. At the same time, it must be conceded
that the differences in life-worlds and personal identities in Fiji, as in
Bosnia, explain why the political cleavages were given ethnic expression.
The differences were already there before they were exploited for particular
political ends.
Ethnic Identity, National Identity, and Intergroup Conflict
53


Hindutva: An Apparent Anomaly in Contemporary Indian Society
My third and final example differs from the two previous ones in significant
ways. India is a tough case for any scholar trying to develop a general theory
of ethnicity or nationalism, and with few exceptions, it does not figure in
general introductory texts on the field. India is hardly a state based on cul-
tural similarity or even equality in the Western sense; it is a country with
deeply embedded hierarchies and a very considerable degree of internal cul-
tural variation. Its population of nearly a billion is divided by language, reli-
gion, caste, and culture, and it has often been argued that India is culturally
more complex than continents such as sub-Saharan Africa or Europe. Al-
though 80 percent of the population are Hindus in one meaning of the
word or another, India also has the second-largest Muslim population in the
world (after Indonesia) and more Christians than all the Scandinavian
countries put together. Since independence (and partition) in 1947, India
has been defined in Gandhian-Nehruvian terms as a secular, federal country
using English and Hindi as national languages, but with another dozen or
so official regional languages.
Since the early 1980s—but particularly forcefully during the 1990s—a
formerly marginal political movement has steadily increased its influence in
India, culminating in its victories in the successive general elections of 1998
and 1999. This is the movement often referred to as hindutva, meaning
roughly “Hindu-ness,” which rallies behind slogans to the effect that India
should be redefined as a Hindu country. The hindutva movement, led by an
organization called the RSS (Rashtriya Swayamsevak Sangh) began mod-
estly in the interwar years, and its more recent parliamentary wing, the BJP
(Bharatiya Janata Party, “The Indian People’s Party”) is now in power not
only federally, in New Delhi, but in several of the states as well.
The rhetoric of hindutva is strongly reminiscent of European ethnic na-
tionalism. It invokes ancient myths of bitter defeats and noble sacrifices, re-
framing them to fit a contemporary political scene. It quotes liberally from
19th-century poets and sacred texts, and it redefines history to make the
past conform to a redefined present. It advocates a return to the roots, con-
demns Westernization and its adverse moral effects on the young, praises
the family as the key institution of society, and seeks to promote the vision
of India as a hindu rashtra—a Hindu nation. While the late Rajiv Gandhi al-
lowed himself to be photographed wearing a Lacoste shirt and khaki shorts,
BJP leaders always wear traditional Indian clothes. The main enemy image
is nevertheless not the West but Islam, which is depicted as a martial and
cruel religion alien to the subcontinent, and Indian Muslims (the descen-
dants of converts, like Bosnian Muslims) are represented partly as traitors to
Hinduism, partly as foreign invaders. The demolition of a mosque in the
northern town of Ayodhya in December 1992, the ensuing riots in several
Indian cities, and the call for the rebuilding of a Hindu temple allegedly de-
54
The Contribution of Identities to Intergroup Conflict


stroyed by a Mughal ruler four centuries ago marked a climax of sorts in
this respect (see van der Veer, 1994, for details).
The phenomenal rise of this traditionalist movement is a result of several
connected processes of sociocultural change or modernization. First, the
very notion of hindutva, Hindu-ness, is a modern one. Hinduism is not a
“religion of the Book.” It is an noncentralized religion with scores of holy
scriptures, thousands of avatars (incarnations of divinities), and very many
ways of worshiping them. The idea of the Hindu identity as an imagined
community based on cultural similarity is alien to Hinduism as such, which
is a religion based on complementarity, difference, and hierarchy. Regard-
ing political Hinduism, some Indian commentators actually speak of a
Semitization of Hinduism whereby it takes on structural characteristics from
the great religions of West Asia. 
Second, the hindutva movement is explicitly modeled on European na-
tionalism—some early hindutva ideologists were even warm admirers of
Hitler—which has been, for 150 years, an attempt to reconcile change and
continuity by talking of roots and traditions in a situation of industrializa-
tion and urbanization. This is obvious in hindutva practice, whereby issues
regarding national anthems, dress, and foreign foods are given prominence,
while profound social changes continue to affect everyday life as before.
There is a clear connection between the rise of the BJP and the liberaliza-
tion of the Indian economy, the rise of a substantial new middle class with 
a strong consumerist orientation and the rapid spread of new mass media
including the Rupert Murdoch-controlled Star TV network. While liber-
alization of this kind stimulates consumerism (perceived as Westernization),
it also indirectly boosts traditionalism since the new patterns of consump-
tion and the new media scene may indicate that cherished traditions are 
under threat. 
Third, the “contagious” influence from political Islam is obvious; hin-
dutva is the assertion of Hindu identity as opposed to Muslim identity both in
Pakistan and in India itself. Doubly ironic, hindutva has double origins in
European romanticism and West Asian political Islam. When its first ide-
ologist, Dr. Veer Savarkar, wrote in the 1920s that “Hindutva is not the
same thing as Hinduism,” he was therefore right, but not for the reasons he
believed. Savarkar saw hindutva as a wide-ranging social movement emanat-
ing from Hindu faith and practices, while a more historically correct ac-
count sees it as the result of cultural diffusion from Europe and West Asia.
Fourth, and perhaps most important, the hindutva movement can be
seen as a reaction against a growing egalitarianism in Indian society. Al-
ready in the 1950s, policies attempting to improve the conditions of the
“Untouchables,” the lowest castes, were introduced, and during the 1990s,
very radical measures have been proposed to this effect—and in some cases
carried out. About half of India’s population are now defined as being either
Dalits (“Untouchables”), tribals, low-caste people, or “OBCs” (Other Back-
Ethnic Identity, National Identity, and Intergroup Conflict
55


ward Classes), and in theory, all of 49.5 percent of jobs in the public sectors
should be reserved for these groups, following the recommendations of the
government-appointed Mandal commission. Since the early 1990s, this
principle has been enforced in many areas. Naturally, many members of the
“twice-born,” upper castes feel their inherited privileges eroding away, and
hindutva is largely a movement representing the interests of the disen-
chanted upper castes. It is for the most part a reaction against the move-
ment toward greater equality in Indian society. Although hindutva seems to
promote equality among Hindus, an implication of its traditionalist Hin-
duism is the reinvigoration of the caste system, which in effect benefits only
the “twice-born” castes. 
This analysis of hindutva must by necessity be a superficial one (see
Hansen, 1999, for a full treatment). It must be remarked, however, that in-
terreligious marriages (and, indeed, intercaste marriages) are rare outside
certain elite groups. Casual interaction between Hindus and Muslims is far
from unusual, but as in the Bosnian and Fijian cases, the intimate (family)
spheres, as well as personal networks of close friends, rarely cross religious
boundaries. Social classification in India is nonetheless complex, and as will
be indicated later, the Hindu-Muslim divide is only one of several possible
social dichotomies—unlike in Bosnia and Fiji, where religious or ethnic
contrasts tend to be paramount.
Some Comparisons
Previous sections have outlined three contemporary conflicts involving col-
lective identity as a political resource. The differences are obvious; the 
focus here will therefore be on the similarities.
The conflicts have three important sociological features in common.
First, there is in all three cases competition over scarce resources. As Horowitz
(1985) and many others writing about group conflict in contemporary soci-
eties have shown, such conflicts invariably involve perceptions of scarcity
and struggles to retain or attain hegemony or equality. Successful mobiliza-
tion on the basis of collective identities presupposes a widespread belief that
resources are unequally distributed along group lines. “Resources” should
be interpreted in the widest sense possible, and could in principle be taken
to mean economic wealth or political power, recognition, or symbolic
power—although what is usually at stake is either economic or political re-
sources. This feature is easy to identify in all three examples described
above: Fijians and Indians compete over relative political and economic
power; the constituent groups of Bosnia compete over political power
and/or sovereignty; hindutva is an attempt to defend the political and eco-
nomic interests of “Hindus” in secular India.
Second, modernization actualizes differences and triggers conflict. With the
integration of formerly discrete groups into shared economic and political
56
The Contribution of Identities to Intergroup Conflict


systems, inequalities are made visible, as comparison between the groups
becomes possible. In a certain sense, ethnicity can be described as the
process of making cultural differences comparable, and to that extent, it is a
modern phenomenon. The Fijian example, where the increasing integra-
tion of Fijians into the modern sphere made it apparent that Indians were
doing better economically, illustrates this point. In India, the rise of the
Dalit movement struggling for recognition and equal rights on behalf 
of “Untouchable” groups is an expression of the modern value of equality,
and the counterreaction from the Hindu right is an attempt to stop egali-
tarianism from spreading, as well as reflecting—almost with the accuracy of
a mirror image—symbolic competition with Muslims within and (espe-
cially) outside India. The Bosnian example, admittedly, seems less straight-
forward, as socialist Yugoslavia was in many ways no less modern than its
successor countries (some would indeed argue that at least at the level of
ideology, it was infinitely more modern than them). What is clear, and
which also holds true for other Eastern and Central European countries, 
is that the sudden introduction of liberal political rights and a capitalist
economy around 1990—core characteristics of non-socialist moderniza-
tion—created a new dimension of comparison between individuals and new
arenas of competition. 
Third, the groups are largely self-recruiting. Intermarriage is rare in all three
cases (excepting urban Yugoslavia). Although biological self-reproduction is
by no means necessary for a strong collective identity to come about, it
should be kept in mind that kinship remains an important organizing princi-
ple for most societies in the world, and a lot of what passes for ethnicity at the
local level is really kinship. Kinship has an important social dimension in ad-
dition to its symbolic side, which is highlighted in ideologies of fictive or
metaphoric kinship. Symbolic boundaries are never effective unless under-
pinned by social organization. 
Further, there are several important ideological similarities. First, at the
level of ideology, cultural similarity overrules social equality. Ethnic national-
ism in Yugoslavia, political Hinduism in India, and the “sons-of-the-soil”
rhetoric of Fiji all depict the ingroup as homogeneous, as people “of the
same kind.” Internal differences are undercommunicated, and moreover,
in the wider political context, equality values are discarded for ostensible
cultural reasons. (Although it could be argued that hindutva is a Trojan
horse concealing upper-caste interests with all-Hindu rhetoric, the point 
is that it stresses the commonalities of all Hindus irrespective of caste or
language.)
Second, images of past suffering and injustice are invoked. Serbs bemoan the
defeat at the hands of the Turks in Kosovo in 1389; Hindu leaders have
taken great pains to depict Mughal (Muslim) rule in India from the 1500s as
bloody and authoritarian; and indigenous Fijian leaders compare their
plight to that of other indigenous peoples who have suffered foreign inva-
sions. Violence targeting the descendants of the invaders can therefore be
Ethnic Identity, National Identity, and Intergroup Conflict
57


framed as legitimate revenge. Even hindutva leaders, who claim to represent
80 percent of India’s population, complain that Hinduism is under siege and
needs to defend itself with all means available. 
Third, the political symbolism and rhetoric evokes personal experiences. This is
perhaps the most important ideological feature of identity politics in gen-
eral. Using myths, cultural symbols, and kinship terminology in addressing
their supporters, promoters of identity politics try to downplay the differ-
ence between personal experiences and group history. In this way, it be-
comes perfectly sensible for a Serb to talk about the legendary battle of
Kosovo in the first person (“We lost in 1389”), and the logic of revenge is
extended to include metaphorical kin, in many cases millions of people.
The intimate experiences associated with locality and family are thereby
projected onto a national screen. This general feature of social integration
has been noted by Handelman (1990), analyzing national rituals, and much
earlier in Turner’s (1967) studies of ritual among the Ndembu of Zambia. In
showing that rituals have both an instrumental and an emotional (or sen-
sory) dimension—one socially integrating, the other metaphorical and per-
sonally meaningful—Turner actually made a point crucial to the present
analysis—namely, that loyalty to a larger collectivity (such as a tribe or a na-
tion) is contingent on its imagery being personally meaningful. 
Fourth, first-comers are contrasted with invaders. Although this ideological
feature is by no means universal in identity politics, it tends to be invoked
whenever possible, and in the process, historical facts are frequently
stretched. In Fiji, the Fijian population—although genetically a Polynesian-
Melanesian mix—has a strong case here, although it is less obvious that
Indo-Fijians can be immigrants to a country in which they were born, and
therefore legitimately deprived of equal rights. Regarding Bosnia and India,
as mentioned above, there is nothing to suggest that the ancestors of Mus-
lims in the respective countries were more recent arrivals than the ancestors
of Christians or Hindus, although Islam is a relatively recent import. What
is interesting here is how the varying depth of cultural genealogies (“roots”)
is used to justify differential treatment. The historical location of the self
along the dimensions of descent and place is thereby invested with political
significance.
Fifth and finally, the actual social complexity in society is reduced to a set of
simple contrasts. As Adolf Hitler wrote in Mein Kampf, the truly national
leader concentrates the attention of his people on one enemy at the time.
Since cross-cutting ties reduce the chances of violent conflict, the collective
identity must be based on relatively unambiguous criteria (such as place, 
religion, mother-tongue, kinship). Again, internal differences are under-
communicated in the act of delineating boundaries toward the demon-
ized Other. This mechanism is familiar from a wide range of interethnic 
situations, from social classification in Zambian mining towns (Epstein,
1992) to Norwegian–Sami relations in sub-Arctic Scandinavia (Eidheim,
1971), the Sinhalese–T
amil conflict (Kapferer, 1988) and Quebecois nation-
58
The Contribution of Identities to Intergroup Conflict


alism (Handler, 1988): the Other is reduced to a minimal set of “traits,” and
so is the collective Self.
These similarities do not necessarily indicate that there are universal
mechanisms linking personal selves and larger collectivities, but they do
suggest that there is a universal “grammar” common to contemporary iden-
tity politics everywhere. In the final sections of this chapter this argument
will be pursued slightly further, and it will be suggested that universal con-
nections between the self and the collective exist, which must be under-
stood not only to account for traditional societies but to make sense of the
present. Far from being an “atavistic” or “primitive” counterreaction to
globalization or modernization, identity politics is a special case of some-
thing more general—namely, collective identity anchored in personal 
experiences.
Where Is the Identity of Identity Politics?
Social scientists have proposed many typologies of ethnic conflict, dividing
the groups involved into categories such as majority, minority, irridentist,
and separatist, using variables such as division of labor, relative political
power, and historical intergroup relations as criteria of classification. In my
view, this kind of exercise can at best generate a limited understanding of
the dynamics of group conflict. T
o begin with, the very adjective “ethnic” is
hardly appropriate to describe all conflicts based on identity politics. Indian
Hindus are not an ethnic group in any meaningful sense, and it is a matter
of definition whether Serbs, Croats, and Muslims in Bosnia should be con-
sidered ethnic groups (they have shared origins only a few centuries back).
Many contemporary conflicts displaying some or all of the features listed
above cannot be seen as ethnic. I note a few African examples: The Su-
danese civil war is partly fought over religion (northeners are Muslims try-
ing to Islamicize the south), partly over culture and language—neither
northern nor southern Sudanese are ethnic groups. Hutus and Tutsis in
Rwanda and Burundi, like the constituent groups of Bosnia, are culturally
very close; they speak the same language and have the same religion. The
Somali civil war presents an even more puzzling case, as Somalia is one of
the few sub-Saharan states that are truly ethnically homogeneous and so far
the only one that seems to have relinquished the trappings of statehood
completely, having dissolved into warring clans (an intermediate level of so-
cial organization, between the family and the ethnic group) since the early
1990s. T
o the northwest of Somalia, one of the great forgotten wars of
Africa is being fought over a contested border area between Eritrea and
Ethiopia. Eritrea, which seceded from Ethiopia in 1991, has never been
based on religion or ethnic identity, but has a vague legitimacy as a histori-
cal nation in the brief period of Italian colonialism before the Second
World War. The current war is being fought between Tigrinya speakers on
Ethnic Identity, National Identity, and Intergroup Conflict
59


both sides of the borders, who are united through religion, language, cus-
toms, history, and even kinship ties, but they are no less bitterly divided po-
litically. This conflict in turn creates its strange bedfellows in the alliance
between Ethiopian Tigrinyas and Amhara speakers from the highlands,
who are traditional enemies.
In other words, the concept “ethnic conflict” is misleading, whether it is
used to classify phenomena or to explain hostilities. Several of the alterna-
tive terms one might consider are, however, no less misleading: “Cultural
conflict” will clearly not do, as it is obviously not what is usually thought of
as cultural differences that lie at the heart of the conflicts. At the village
level, even Hindus and Muslims in India hold many of the same beliefs and
worship in similar ways. The low-intensity conflict in Fiji involves groups
that are by any criterion more culturally different than, say, the Bosnian
groups. All the conflicts considered here are over resources perceived as
scarce: territory, political power, economic gain, employment, recogni-
tion—rights in a wide sense. What they have in common is their successful
appeal to collective identities perceived locally as imperative and primor-
dial, identities associated with a deep moral commitment, whether ethnic
(based on notions of kinship and descent), regional (based on place), or reli-
gious (based on beliefs and forms of worship). For these reasons, the term
“identity politics” is preferable as a generic term for all such political move-
ments, whether nonviolent or violent.
This final section will therefore amount to an attempt to unravel the
identity of identity politics. What is it that makes it so powerful? What is
the “identity” that such political movements can draw upon? 
Benedict Anderson proposes an answer in the introduction to his seminal
Imagined Communities (1983), where he points out that nationalism has
more in common with phenomena such as religion and kinship than with
ideologies like liberalism and socialism. He argues that nationalism (and,
one might add, any form of identity politics) expropriates personal identity,
transforming intimate experiences into the raw material of politics. I owe
my existence to my parents, and by metonymical extension they represent
the larger, abstract collective. I harbor tender feelings for my childhood,
which by extension becomes my group’s glorious and tragic history. I feel
attached to the place where I grew up, which was not just any arbitrary
place but the nation (or, as the case might be, the sacred land of Hinduism,
the traditional territory of the Fijians, the tormented country of the brave,
but sadly misunderstood Serbs). Indeed, this argument can profitably be
seen as echoing Turner’s aforementioned argument on the instrumental and
emotional dimensions (or “poles”) of ritual. In both cases, the integrative
strength of the imagined community (be it a tribe or a nation) depends on
its ability to mobilize emotions proper to the intimate sphere of kinship and
personal experience. 
The conditions for this transformation to take place—the move from an
interpersonally anchored identity to an abstract national, ethnic, or reli-
60
The Contribution of Identities to Intergroup Conflict


gious identity—are usually tantamount to certain general conditions of
modernity (cf. Gellner, 1997). It is through school and mass media that
people are taught to identify with an abstract, mythically rooted community
of people “of the same kind.” Through the replacement of traditional
economies with an abstract labor market, they become participants in a
large-scale system of subsistence. Through the implementation of a bureau-
cratic system of political management, their allegiances are at least partly
moved from the concrete to the abstract community. 
It is important to remember, as theorists of nationalism and ethnicity
have pointed out time and again, that identification is relational, situational,
and flexible, and that each person carries a number of potential identities,
only a few of which become socially significant, making a difference in
everyday life. Even fewer gain political importance, forming the basis of
power struggles and group competition. This is not, however, to say that
collective identities can be created out of thin air. They have to be con-
nected, in credible ways, to people’s personal experiences. These experi-
ences in turn are flexible—not only historians but everybody else as well 
selects and interprets events to make a particular kind of sense of the past—
but not indefinitely so. Regarding our main examples, in Fiji virtually no-
body doubts whether he or she is Indian or Fijian, and politics—whether
based on compromise or conflict—will have to take this into account for the
foreseeable future. In Yugoslavia after the breakup, cross-cutting ties and
cultural hybridity were undercommunicated. Cosmopolitanism was in-
creasingly seen as a suspect, unpatriotic attitude, and people of mixed an-
cestry were forced to choose a bounded, unambiguous identity: they had 
to select past experiences that made them either Serbian, Croatian, or 
Muslim—more or less like the proverbial North African mule, who speaks
incessantly about his uncle, the horse, but never mentions his father, the
donkey. In India, finally, some of the strongest scholarly arguments against
the lasting influence of hindutva have actually pointed toward people’s per-
sonal experiences (Frøystad, 1999). Since Indian everyday life is still perme-
ated by caste distinctions, and caste continues to define the very fabric of
social integration, these scholars argue that hindutva—the idea that all Hin-
dus have something profound in common—is so counterintuitive to most
Indians that it can unite Hindus only as long as the enemy image of Mus-
lims can be kept ablaze. 
All the basic components of political identity familiar from classic politi-
cal anthropology can be identified in contemporary identity politics: it is
based on a sometimes ambiguous mix of kinship and locality; it has well de-
veloped myths of origin and myths of past suffering; and it distinguishes
clearly between “us” and “them.” The main difference between, say, the no-
madic Nuer society studied by E. E. Evans-Pritchard (1940) in the 1930s
and Serbian (or Croatian) nationalism today is probably that of scale: while
the Nuer rarely imagined themselves as members of larger groups than the
clan, a Serb in Vojvodina can readily identify him or herself with a Serb in
Ethnic Identity, National Identity, and Intergroup Conflict
61


Kosovo. The act of transformation from personal, concrete social experi-
ences to the abstract community is naturally much more demanding in a
large-scale society than in a village-based one, hence the importance of
modern institutions of communication, economic transactions, and political
rule for the growth of abstract communities.
Some Final Lessons from Political Anthropology
Having long ago abandoned the early ambition of becoming “a natural sci-
ence of society,” social and cultural anthropology has for decades been re-
luctant to formulate lawlike propositions about the functioning of society.
The constructivist turn of recent years seems to confirm that contemporary
anthropology is less concerned with absolute truths than with the analysis
of local cultural constructions. This need not be so, and the study of current
identity politics may illustrate the power of comparative anthropology in
generating general hypotheses.
Early instrumentalist research on ethnic groups, particularly in Africa (as
in A. Cohen, 1974), searched for the logic of group cohesion, which they
assumed to be roughly the same everywhere. The related, actor-based per-
spective developed by Barth (1969) and his colleagues assumed the logic of
action to be quite universal—people act to maximize benefits. Later analy-
ses of the constructedness of ethnic and national ideologies (the seminal
text is Hobsbawm and Ranger, 1983; see Chapman, Malcolm, McDonald,
& T
onkin, 1989, for an overview) also emphasize universal characteristics of
a particular kind of societal formation (the modern state) and its relation-
ship to group identities based on notions of culture. The canonical texts on
nationalism (such as Gellner, 1983; Anderson, 1983; Smith, 1991) also have
clear universalist ambitions. As this chapter has made clear, these ap-
proaches have obvious strengths, but they need to be supplemented by de-
tailed research on the experiential world of the everyday—the Lebenswelt
(life-world) of the actors. A renewed focus on the informal, intimate, and
often noninstrumental dimensions of everyday life reveals that terms such
as “ethnicity” by themselves explain little. The parallels between a sup-
porter of the BJP and a supporter of Serbian supremacists should not be lo-
cated to their respective “ethnic” identities or “civilizational” membership,
but to the fact that their everyday life, social networks, and personal obliga-
tions connect them to particular groups that may be exploited politically,
given the right circumstances. It should also be kept in mind that class poli-
tics can sometimes be a form of identity politics (Shore, 1993), which can
profitably be understood along the same lines as ethnic or religious identity
politics. The cause of group allegiance lies in the everyday, not in the over-
arching ideology.
In order to complete this analysis, it is necessary to go a few decades
back, to classic political anthropology, in order to see how the perspectives
62
The Contribution of Identities to Intergroup Conflict


developed earlier in the chapter can be enriched by the work of previous
generations. 
The integration of persons into groups can be described as the work of
an inverted refrigerator: the function of a refrigerator is to generate inward
coldness, but in order to do so it more or less inadvertently, as a side effect,
creates outward warmth. Conversely, groups form to create warmth for
their members, but they necessarily create some outward coldness in order
to be able to do so. Under particular circumstances, the outward coldness is
more readily perceptible than the inward warmth. A sociological principle
originally formulated by Georg Simmel, known as “Simmel’s Rule,” simply
states that the internal cohesion of a group is contingent on the strength of exter-
nal pressure. This principle may explain why group integration generally is
so much stronger in small groups, especially if they are oppressed, than in
large ones—why, for example, Scots seem to have fewer difficulties defining
who they are than do the English. 
An interesting corollary of Simmel’s Rule is the fact that what kind of
group emerges depends on where the perceived pressure comes from. Both
gender-based and class-based social movements have periodically been suc-
cessful, given that the perceived threat was seen, not as alien religion or for-
eign ethnic groups, but as male supremacy and ruling classes, respectively.
In accordance with this, some inhabitants of Sarajevo during the war felt
that the conflict was really an urban–rural one, since city-dwellers had a lot
in common, irrespective of religion, that they did not share with rural peo-
ple. Strong opposition groups in India, similarly, argue against a view of 
Indian politics as divided between Hindu communalists and liberals because
they see the main problem of Indian society as one of poverty and distribu-
tion of resources, to which neither of the parties seems to give priority. In
Fiji, finally, the immediate reason for the 1987 coup was the establishment
of a government of national unity promising to address issues of social wel-
fare and economic development rather than intergroup issues. In other
words, redefinitions of societal cleavages are entirely possible insofar as they
do not contradict people’s everyday experiences too obviously. 
In the course of this chapter, cross-cutting ties and conflicting loyalties
have already been mentioned as mitigating forces in situations of inter-
group conflict. Phrased within the terminology used here, one might say
that shared experiences across boundaries reduce the risk of conflict. In
Max Gluckman’s reinterpretation of Evans-Pritchard’s Nuer material from
the 1930s (Gluckman, 1956–1982), this point was made forcefully. The
Nuer were organized along kinship lines across villages, but they were also
locally integrated in villages. The women married out of the village and the
lineage, so that everybody had affines (in-laws) in other villages. Further-
more, men were tied to nonrelatives through trade, initiation rituals, and
friendship. All of these factors led to a reduction in the incidence of vio-
lence among the Nuer. In contemporary identity politics, it can easily be
seen—and has been remarked above—how political leaders emphasize in-
Ethnic Identity, National Identity, and Intergroup Conflict
63


ward similarity and outward boundaries in order to reduce the potentially
mitigating impact of cross-cutting ties. They are true to people’s everyday
life, but try to emphasize certain experiences at the expense of others (in-
ward solidarity and similarity, outward conflict and difference).
Another, even more time-honored principle from political anthropology
is the twin notions of fusion and fission in tribal societies. When the sole
organizing principle for a group lacking hierarchies and formal political of-
fice is kinship, there are limits to the group’s growth; at a certain point, it
splits into two. Without such a fission, internal conflicts would soon be-
come overwhelming given the simple social organization of such societies,
and the effects would be destructive. Fusion of discrete groups has also been
studied extensively, but in many acephalous societies it is seasonal (nomadic
groups fuse in the dry season or in winter) and fragile.
A more dynamic view of contraction and expansion of tribal groups 
was developed, especially by Africanists, from the 1940s (Fortes & Evans-
Pritchard, 1940). In studies of feuding and political competition, they
showed how two or several local groups that might be periodically involved
in mutual feuding united temporarily when faced with an external enemy.
This form of organization, described as segmentary by Evans-Pritchard, fol-
lows the proverb often cited in recent years to explain the logic of the So-
mali civil war: “It’s me against my brother, my brother and I against our
cousins, and our cousins, my brother and I against everybody else.” A form
of segmentary logic is apparent in politics nearly everywhere; a distinguish-
ing mark of modern nation-building has nevertheless been its attempt to
channel loyalties away from various subnational levels of identity in order to
monopolize the political loyalty of individual citizens. 
The segmentary logic creates a fluid, relational political organization
that, in its pure form, is impracticable in modern state societies given their
requirements for stability, centralized power, and reified systems of political
representation. This does not, however, mean that segmentary identi-
fication does not continue to exist, and one of the causes of oppositional
identity politics in modern nation-states is their not providing subnational
identity groups appropriate political arenas, thereby encouraging counter-
reactions in the form of identity politics directed against the state. 
The formation of identity-based political groups generally entails both
an expansion and a contraction of the focus for identification. At the time of
the breakout of conflict in Bosnia, the federal or even state level was in-
creasingly seen as irrelevant—the process was one of fission. At the same
time, internal conflicts and schisms within each constituent group were
minimized, and as a result each group became more coherent and united
than before. In the cases of Fiji and India, this is even more obvious: among
Fijians, rivalry between chiefs and clans has diminished in importance as 
Fijian politics has grown increasingly ethnic; similarly, rifts within the In-
dian population on the basis of regional origin, which could formerly lead
to Indian subgroups supporting Fijian-dominated governments, have be-
64
The Contribution of Identities to Intergroup Conflict


come much less important since the military coup. In the Indian case, the
very idea of hindutva implies an enormous expansion of the ingroup for
Hindus. Trying to bridge differences based on language, caste, region, and
culture, hindutva tries to create a morally committing all-Indian Hindu
identity based on symbolic equality. This is, in the Indian context, a very
radical move. Simultaneously, Indian citizenship and (overarching, supra-
religious) national identity become less important since the federal Indian
state includes millions of non-Hindus, who are depicted as internal enemies
by hindutva spokespersons.
In all the examples considered here, group segmentation at a higher
level, and the ensuing formation of imagined communities larger than the
locality but smaller than the state, is immediately related to the need for a
firm boundary in a situation of conflict within the state. The alternative
identity of national citizenship, which encompasses the other as well, no
longer functions. Interestingly, contemporary identity politics is very simi-
lar to nationalism—for example, in its appeal to mythical foundations, its
abstract postulation of similarity and equality, its rejection of segmentary
identity formation, and its attempt to reduce a world of many small differ-
ences to a world of only a few, major ones. In many cases, it is more success-
ful than nationalism, particularly in postcolonial, multiethnic states. This 
is not only because identity politicians promise its adherents that they will
win zero-sum games against political competitors, but also because they 
are able to represent themselves as natural extensions of people’s personal,
experience-based identities. 
A challenge for modern states, thus, consists in coping with the fact that
personal identity can be exploited politically not only by the state itself but
by others as well, not because the self is infinitely multifaceted (it is not),
but because the experiences and relationships that make up the self can be
expanded symbolically in several, often conflicting ways. Processes of seg-
mentary fusion and fission, the formation of different kinds of groups in-
volving overlapping personnel owing to the functioning of Simmel’s Rule,
group-based antagonism, and competition: these ways of expressing politi-
cal interests, underpinned by shared meaning within the ingroup, are no
more eradicated by modernity than is personal identity. The challenge thus
consists in laying the foundations for “a sense of belonging to a community
larger than each of the particular groups in question” (Laclau, 1995, p. 105),
and this can be done only by first acknowledging both the richness and the
variability of personal identities.
In the face of violent identity politics, “ethnic cleansing,” and the strong
attention to “roots” and historically based identities (ungenerously de-
scribed as “the narcissism of small differences” by Michael Ignatieff) char-
acterizing many societies in recent years, it is not surprising that intellectu-
als have recently tried to think essentialism away, emphasising the endlessly
flexible and fluid character of human identification. A typical expression of
this position can be found in a recent text by the influential sociologist 
Ethnic Identity, National Identity, and Intergroup Conflict
65


Zygmunt Bauman, where he states: “If the modern ‘problem of identity’ is
how to construct an identity and keep it solid and stable, the postmodern
‘problem of identity’ is primarily how to avoid fixation and keep the options
open” (Bauman, 1996, p. 18).
In today’s world it may be ethically imperative to endorse Bauman’s posi-
tion, but it is equally important to keep in mind that humans are not free-
floating signifiers, and no amount of benevolent intentions will be able to
change people’s life-worlds overnight. Rather than trying to think them
away, it is necessary to understand them and come to terms with their 
enduring power. Notwithstanding globalization and the universalization 
of modernity, cultural differences continue to exist, within and between
places, within and between nations and ethnic groups. It is also, however,
doubtlessly true that carbon can be turned into graphite as well as dia-
monds, and the ways in which cultural differences become socially relevant
vary importantly. But to pretend they do not exist outside ethnic and na-
tionalist ideologies would be intellectually indefensible; people’s personal
experiences are the very raw material of such ideologies. Here lies an im-
portant limitation in constructivist models of identity. Collective identities
are constructed, consciously or not, but nothing comes out of nothing. In
locating the universal not in the workings of identity politics (it changes
historically and varies geographically), nor in the eternal sovereignty of 
the state (the same objection applies), but rather in the social life-worlds in
that individuals make sense of the world, we may have found a basis for
comparison that will outlive academic fads and contemporary politics. 
Acknowledgments
The author would like to express his thanks to the editors and three anonymous
readers for their critical readings and useful comments on an earlier version of the
chapter.
References
Anderson, B. (1983). Imagined communities: Reflections on the origin and spread of na-
tionalism. London: Verso.
Barth, F. (Ed). (1969). Ethnic groups and boundaries. Oslo: Scandinavian University
Press.
Bauman, Z. (1993). Postmodern ethics. Cambridge: Polity.
Bauman, Z. (1996) From pilgrim to tourist; or a short history of identity. In S. Hall
& P. Du Gay (Eds)., Questions of cultural identity (pp. 18–36). London: Sage.
Bringa, T. (1996). Being Muslim the Bosnian way: Identity and community in a central
Bosnian village. Princeton, NJ: Princeton University Press.
Chapman, M., McDonald, M., & T
onkin, E. (1989). Introduction—history and so-
cial anthropology. In E. T
onkin, M. McDonald, & M. Chapman (Eds.), History
and ethnicity (pp. 1–21). London: Routledge.
66
The Contribution of Identities to Intergroup Conflict


Cohen, A. (1969). Custom and politics in urban Africa: A study of Hausa migrants in a
Yoruba town. London: Routledge & Kegan Paul.
Cohen, A. (1974). Two-dimensional man: An essay on power and symbolism in complex 
society. London: Routledge & Kegan Paul.
Cohen, A. P. (1994). Self consciousness: An alternative anthropology of identity. London:
Routledge.
Cornell, S., & Hartmann, D. (1998). Ethnicity and race: Making identities in a chang-
ing world. London: Pine Forge Press.
Eidheim, H. (1971). Aspects of the Lappish minority situation. Oslo: Scandinavian Uni-
versity Press. 
Epstein, A. L. (1992). Scenes from African urban life: Collected Copperbelt essays. Edin-
burgh: Edinburgh University Press.
Eriksen, T. H. (1992). Us and them in modern societies: Ethnicity and nationalism in
Trinidad, Mauritius and beyond. Oslo: Scandinavian University Press.
Eriksen, T. H. (1998a). Common denominators: Ethnicity, nationalism and compromise in
Mauritius. Oxford: Berg.
Eriksen, T. H. (1998b). Culture and ethnicity: A second look. Lecture at 6th SIEF
conference, Amsterdam, 21 April 1998.
Evans-Pritchard, E. E. (1940). The Nuer. Oxford: Clarendon.
Featherstone, M. (Ed). (1990). Global culture. London: Sage.
Fortes, M., & Evans-Pritchard, E. E. (Eds.). (1940). African political systems. London:
Oxford University Press.
Friedman, J. (1994). Global culture and local process. London: Sage.
Frøystad, K. (1999). Personal communication.
Furnivall, J. S. (1948). Colonial policy and practice: A comparative study of Burma and
Netherlands India. New York: New York University Press.
Geertz, C. (Ed.). (1967). Old societies and new states: The quest for modernity in Africa
and Asia. New York: The Free Press.
Geertz, C. (1973). The interpretation of cultures. New York: Basic Books.
Gellner, E. (1983). Nations and nationalism. Oxford: Blackwell.
Gellner, E. (1997). Nationalism. London: Weidenfeld & Nicolson.
Giddens, A. (1991). Modernity and self-identity. Cambridge: Polity.
Giddens, A. (1994). Beyond left and right. Cambridge: Polity.
Gluckman, M. (1956/1982). Custom and conflict in Africa. Oxford: Blackwell.
Handelman, D. (1990). Models and mirrors: Towards an anthropology of public events.
Cambridge: Cambridge University Press.
Handler, R. (1988). Nationalism and the politics of culture in Quebec. Madison: Wiscon-
sin University Press.
Hannerz, U. (1992). Cultural complexity: Studies in the social organization of meaning.
New York: Columbia University Press.
Hansen, T. B. (1999). The saffron wave: Democracy and Hindu nationalism in modern
India. Princeton, NJ: Princeton University Press.
Hobsbawm, E., & Ranger, T. (Eds.) (1983). The invention of tradition. Cambridge:
Cambridge University Press.
Holland, D. (1997). Selves as cultured: As told by an anthropologist who lacks 
a soul. In R. Ashmore & L. Jussim (Eds.), Self and identity: Fundamental issues
(pp. 160–190). Oxford: Oxford University Press.
Horowitz, D. L. (1985). Ethnic groups in conflict. Berkeley: University of California
Press.
Ethnic Identity, National Identity, and Intergroup Conflict
67


Huntington, S. (1996). The clash of civilizations and the remaking of a world order. New
York: Simon and Schuster.
Ignatieff, M. (1994). Blood and belonging: Journeys into the new nationalism. London:
Vintage.
Jenkins, R. (1996). Social identity. London: Routledge.
Kapferer, B. (1988). Legends of people, myths of state. Baltimore, MD: Smithsonian 
Institution Press.
Kelly, J. D. (1998). Aspiring to minority and other tactics against violence. In D. C.
Gladney (Ed.), Making majorities: Constituting the nation in Japan, Korea, China,
Malaysia, Fiji, Turkey and the United States (pp. 173–197). Stanford, CA: Stanford
University Press.
Laclau, E. (1995). Universalism, particularism and the question of identity. In 
J. Ratchmann (Ed.), The identity in question. London: Routledge.
Lash, S., & Urry, J. (1994). Economies of signs and space. London: Sage.
Premdas, R. (1993). Balance and ethnic conflict in Fiji. In J. McGarry & B. O’Leary
(Eds.), The politics of ethnic conflict regulation (pp. 251–274). London: Routledge.
Rex, J. (1997). The nature of ethnicity in the project of migration. In M. Guibernau
& J. Rex (Eds.), The ethnicity reader: Nationalism, multiculturalism and migration.
Cambridge: Polity.
Roosens, E. E. (1989). Creating ethnicity. London: Sage.
Said, E. W. (1978). Orientalism. New York: Pantheon.
Schlesinger, A. M. Jr. (1992). The disuniting of America. New York: Norton.
Schöpflin, G. (1993). The rise and fall of Yugoslavia. In J. McGarry & B. O’Leary
(Eds.), The politics of ethnic conflict regulation (pp. 172–203). London: Routledge.
Shore, C. (1993). Ethnicity as revolutionary strategy: Communist identity construc-
tion in Italy. In S. Macdonald (Ed.), Inside European identities (pp. 27–53). Oxford:
Berg.
Smith, A. D. (1986). The ethnic origins of nations. Oxford: Blackwell.
Smith, A. (1991). National identity. Harmondsworth: Penguin.
Strathern, M. (1991). Partial connections. Savage, MD: Rowman and Littlefield.
T
ambiah, S. J. (1989). The politics of ethnicity. American Ethnologist, 16, 335–349.
Tinker, H. (1974). A new system of slavery: The export of Indian labour overseas 1830–
1920. Oxford: Oxford University Press.
Turner, V. (1967). The forest of symbols: Aspects of Ndembu ritual. Ithaca, NY: Cornell
University Press.
Turton, D. (Ed.) (1997). War and ethnicity: Global connections and local violence. Wood-
bridge, Suffolk: University of Rochester Press.
van der Veer, P. (1994). Religious nationalism: Hindus and Muslims in India. Berkeley:
University of California Press.
Williksen-Bakker, S. (1991). Fijians in business: A study in the transformation of
symbols. Doctoral dissertation, University of Oslo.
68
The Contribution of Identities to Intergroup Conflict


PART II
THE CONTRIBUTION 
OF ETHNIC AND 
NATIONAL IDENTITIES 
TO POLITICAL CONFLICT 
IN THE UNITED STATES


This page intentionally left blank 


Jack Citrin
4
Cara Wong
Brian Duff
The Meaning of American 
National Identity
Patterns of Ethnic Conflict and Consensus
National identity is one among many often co-existing and overlapping so-
cial identities, including territorial, racial, religious, linguistic, and gender
identities. The psychological task for individuals is to order and integrate
their different collective “selves” (Ashmore & Jussim, 1997). The political
task for government is to balance the need for national unity with the com-
peting claims of other group identities. In modern society, emphasizing loy-
alty to the “nation” and making this the essence of the individual’s political
self-definition emerged as the dominant way to boost social solidarity
(Greenfield & Chirot, 1994). As an immigrant nation, the United States has
always faced the problem of coping with ethnic diversity; the motto e
pluribus unum expresses the desire for a strong sense of common American
identity without indicating the proper balance between the national “one”
and the ethnic “many.”
The contemporary meaning of American national identity is the focus of
this chapter. Immigration and differences in fertility rates have changed the
ethnic and religious composition of the United States (Farley, 1996;
Warner, 1993). Compared to 50 years ago, the country is more diverse, with
many more people today from Latin America, Asia, and Africa. Moreover,
this demographic change has occurred in the context of technological trans-
formations eroding national sovereignty in economic life and of the emer-
gence of intellectual trends legitimating the primacy of ethnicity in political
life.
Scholars disagree about whether globalization and multiculturalism have
weakened Americans’ sense of attachment to the “nation,” particularly
among minority groups (Reich, 1992; Barber, 1995; Hollinger, 1997;
71


Wolfe, 1998; de la Garza, Falcon, & Garcia 1996). There also are conflict-
ing views about the implications of a stronger sense of ethnic, as opposed to
national, identity. Some scholars (Raz, 1994; Young, 1990) regard ethnic
pride as a source of individual self-esteem among cultural minorities that
contributes to the achievement of their group’s goals. Others (Schlesinger,
1992; Miller, 1995) claim that the emphasis on ethnic distinctiveness (in-
group favoritism) inevitably weakens common bonds and intensifies group
conflict (outgroup hostility), raising the specter of cultural and political
Balkanization. 
This chapter employs survey research to analyze the meaning and politi-
cal consequences of American national identity in the context of the demo-
graphic, economic, and intellectual changes noted above. The conceptu-
alization of national identity formulated here draws on social identity
theory (T
ajfel & Turner, 1986; Thoits & Virshup, 1997) to encompass both
self-categorization (or identification as) and affect (or identification with).
However, it goes beyond merely assessing the individual’s positive (or nega-
tive) feelings about the nation to determine the normative foundations of
their sense of American identity. 
The content of American identity is the set of ideas and sentiments that
form the conceptual framework of nationhood. At the cultural level, this
refers to a particular collective representation of the nation indicating the
subjective criteria for membership in the national community (Citrin, Haas,
Muste, & Reingold, 1994). A given conception of national identity thus in-
cludes normative assumptions about how members of different ethnic and
cultural groups should relate to one another. For example, must minority
groups assimilate to a dominant language, religion, or political outlook in
order to achieve full acceptance as fellow-nationals, or can a nation accom-
modate loyalty to a variety of cultural traditions? Historical experience sug-
gests that there is no single answer to this question, so it is important to
conceive of national identity in a way that recognizes the possibility of sev-
eral conceptions of “Americanness” vying for popular support. 
The next section of the chapter provides a more detailed account of the
tripartite definition of national identity that distinguishes among the di-
mensions of self-categorization, affect, and normative content. The second
section summarizes the content of competing ideological conceptions of
American identity, with particular attention to their normative assumptions
regarding the articulation of strong ethnic identities. The analysis of survey
data, presented in the third section, addresses three main questions: 
1. What is the degree of national versus ethnic identification in
American public opinion, as assessed by several different measures
of self-categorization and affective support? 
2. What is the pattern of consensus and cleavage among white, black,
Hispanic, and Asian respondents? Specifically, do minority ethnic
groups have a weaker sense of national identity than whites, and
72
The Contribution of Ethnic and National Identities to Political Conflict


are there ethnic group differences in how people order their na-
tional and ethnic identities? 
3. What is the impact of national and ethnic identities on policy pref-
erences? Drawing on symbolic politics theory (Sears, 1993), the 
final section of the chapter tests hypotheses about the relative in-
fluence of these identifications on preferences regarding foreign
policy, racial policies, and “multicultural” issues such as immigra-
tion and group representation. 
These analyses shed light on the impact of a superordinate “American”
identity on feelings about “outgroups,” both domestic and foreign, and pro-
vide a basis for speculating about the implications of different patterns of
national and ethnic identity for conflict and cooperation among America’s
racial and ethnic groups.
The Dimensions of Social Identity
Identity is a slippery concept. It is an assertion of both sameness and differ-
ence. One answers the question “What is your identity?” by naming who
one is like. This entails drawing boundaries: one is the same as some others
and different from everyone else. For this reason, a social identity both in-
tegrates and divides. 
The first step in measuring a social identity is to determine the basis of
self-categorization; one identifies the characteristic(s) shared by those with
whom one psychologically belongs. We each possess multiple potential 
social identities whose degree of overlap and whose relative significance 
for our self-concept and behavior may vary (Thoits & Virshup, 1997).
Moreover, while self-categorization may be the initial step in the formation
of a psychologically meaningful social identity, the range of one’s options is
finite. Biology, the life cycle, social structure, the economy, and the govern-
ment create the categories that demarcate social identities with potential
political consequences. Thus, some social identities are optional and others
imposed. Furthermore, the borders dividing identity groups are permeable,
but not entirely open. It is easy to be a Bulls fan and then to “exit” psycho-
logically for the Celtics when Michael Jordan retires. It is harder to shed
one’s nationality or ethnicity and almost impossible to change one’s race or
sex. Finally, even when one can adopt a particular social identity by self-
categorization (Turner, Hogg, Oakes, Reicker, & Blackwell, 1987), this
choice frequently must be affirmed by the agreement of others in the group
that you do indeed possess the criteria for membership. The daughter of
Chinese immigrants may call herself an American, yet might be told by 
fellow-citizens of European origin that she does not share their national
identity.
Self-categorization refers to “identification as,” connoting a perceived
The Meaning of American National Identity
73


self-location in a group, and should be distinguished from “identification
with,” which indicates positive affect toward others in the group and must be
measured separately. Social identity theory (T
ajfel, 1978) posits that mere
awareness of belonging to a group engenders positive feelings about the
group and a tendency to act on behalf of other group members, even if they
are unknown to one personally. In the same vein, Miller (1995) has argued
that the moral value of a strong sense of national identity is that it fosters
diffuse feelings of sympathy and obligation toward fellow citizens. 
In addition, there is a difference between membership in a group and psy-
chological attachment to it. One can identify with members of a minority
group, agreeing that they deserve more access to jobs or political power,
without categorizing oneself as a member of that group. This example un-
derscores the potential significance of identities based on shared values
rather than sociological similarities and the need to determine the condi-
tions under which “identification as” and “identification with” are closely
connected. 
“Identification with” entails learning about a group’s defining customs,
expectations, and values, and making them one’s own. These common val-
ues or ideas define the third component of a social (e.g., national) identity—
its content. The specific content of a group identity is socially constructed, in
the sense that consensual decisions shape and reinforce the normative crite-
ria for membership. Since these criteria are vulnerable to challenge and
subject to change, it is important to assess the extent of their legitimacy. 
Still, social identities are not automatically political. They are politicized
when feelings of identification with a group are combined with a belief in
advancing its goals through collective action (Miller, Gurin, Gurin, &
Malanchuk, 1981). During the British Raj, Indians undoubtedly differenti-
ated themselves cognitively from the English. The emergence of an ide-
ology and organized movement demanding change in the cultural, eco-
nomic, and legal status of Indians infused this social identity with political
content. 
The Affective Dimension: Patriotism versus Chauvinism
A nation denotes a group of people seeking or possessing a common home-
land. Nationalism as a doctrine asserts that a group of people sharing char-
acteristics that differentiate them, in their own minds, from others should be
politically autonomous. The affective dimension of national identity (iden-
tification with) refers to feelings of closeness to and pride in one’s country
and its symbols. Many scholars go further, however, and distinguish love of
one’s own country from a sense of superiority to other countries (Schaar,
1981; Staub, 1997; Kosterman & Feshbach, 1989; Sullivan, Fried, & Dietz,
1992; Connor, 1993; T
aylor 1995). They invoke the concept of patriotism
to refer to the positive emotion of love for one’s own people and homeland
74
The Contribution of Ethnic and National Identities to Political Conflict


and use “nationalism” pejoratively, not technically, to designate arrogance
and contempt toward other countries. Writing in this vein, for example,
George Orwell described patriotism as defensive and nationalism as aggres-
sive. For Schaar (1981, p. 285), nationalism is a perversion of patriotism,
“its bloody brother.” Psychologists developing separate measures for these
concepts tend to give the label “nationalism” or “blind patriotism” to name
disdain for other countries and a drive to dominate them (Bar-T
al, 1993;
Kosterman and Feshbach, 1989; Staub, 1997). 
Drawing on social identity theory’s distinction between ingroup fa-
voritism and outgroup hostility clearly is useful when studying national and
ethnic identity. The emotive use of the term “nationalism,” however, is un-
necessarily confusing. “Chauvinism,” not nationalism per se, is the term
used here to refer to an extreme and bounded loyalty, the belief in one’s
country’s superiority, whether it is right or wrong. Whether patriotism, an
ideology of mutual affection among those with a common sense of national
consciousness (Kelman, 1997, p. 170), and chauvinism empirically slide into
each other, such that caring for one’s own typically is associated with hos-
tility toward others, is a controversial issue to be explored below (see also
Brewer, this volume). 
Normative Conceptions of American Identity
History shows that the attributes used to define the idea of a “people” or
“nation” are numerous and malleable (Greenfield, 1992). Historians distin-
guish “civic” nationalism, where criteria for belonging are identical to citi-
zenship and so, in principle at least, open to all, from “ethnic” nationalism,
where membership is based on a sense of shared blood (Connor, 1993).
There is a modern tendency to endorse the principle of ethnicity as the le-
gitimate basis for political autonomy, but there are prominent counterex-
amples of multiethnic states such as the United States (Greenfield, 1992).
As Max Weber put it, nationality ultimately is based on the principle of ter-
ritory, ethnicity on lineage. 
Whatever the specific foundations of a sense of national identity, nation-
alism, in its neutral, technical meaning, implies that membership in the na-
tion is the citizen’s overriding group loyalty, taking precedence over other
available foci of affiliation, including ethnicity, in circumstances where they
conflict. If a strong national identity means consciousness of special bonds
to one’s fellow-nationals, so that they become, at least symbolically, a large
extended family, then constructing a nation may involve psychological
losses as well as gains. The development of a national identity may require
diminishing or abandoning one’s attachment to other groups, including
one’s ethnic group. 
The third dimension of our conception of national identity refers to its
normative content. This means the particular set of ideas about what makes
The Meaning of American National Identity
75


the nation distinctive—ideas about its members, its core values and goals,
the territory it ought to occupy, and its relations to other nations (Citrin et
al., 1994). In a multiethnic society, these ideas lay down the nature of le-
gitimate commitments to subnational communities of descent. In contem-
porary American politics, liberalism, nativism, and multiculturalism are 
alternative political theories about how society should be organized and
how national identity should be defined. Each proposes a different solution
for how properly to balance national and ethnic identities, with conflicting
implications for policymakers. The analysis below therefore assesses the 
degree of popular support for these competing ideas. 
The liberal conception of American identity is “civic,” not “ethnic.”
Commitment to the national “creed” of democracy and individualism
(Huntington, 1981; Gleason, 1980) is what makes one an American. Since
belonging to the nation is equated not with shared blood but with common
beliefs and customs (Gleason, 1980; Lind, 1995), anyone, regardless of an-
cestry, can become American through adherence to the dominant set of
ideals, which include equality of opportunity and respect. The liberal image
of nationality is ethnically inclusive, in principle, if not always in practice
(Smith, 1997). 
When it comes to ethnic diversity, the liberal conception of national
identity is optimistic about the ability of contemporary American society to
assimilate newcomers. This means that ethnic Balkanization is not a serious
threat, so there is no reason to extirpate the traditions of new immigrants as
attempted by the Americanization program in the 1920s. Over time, the
largely Asian and Hispanic newcomers will blend into the cultural main-
stream, just like their European predecessors. Thus, people are free to
honor their ethnic heritage if they wish, without undermining America’s
ability to create one people out of many. Still, according to the liberal con-
ception of national identity, when the claims of nationality and ethnicity
conflict, the former should take precedence. In psychological terms, liberal-
ism rejects the notion of primordial identities, the idea that differences
among communities of descent are fundamental and enduring, exercising an
inherent dominion over the individual’s political conduct. 
Rogers Smith (1997) documents the persistent gap between the liberal
theory of national identity and more nativist political practices. From the
beginning of the United States, the official definition of “Americanness” ex-
cluded blacks and Native Americans and consigned women to a lesser role.
Then, in the 19th and 20th centuries, restrictive immigration and natural-
ization laws discriminated on a racialist basis. Asians were denied citizen-
ship, Mexicans were repatriated, and in 1924, immigration was limited to a
small number of people of Northern European origin. 
The nativist response to ethnic diversity (Higham, 1988) is to insist on
cultural conformity. If there is to be immigration, the nation’s policy should
favor the admission of people who already are familiar with American po-
litical culture (Brimelow, 1995). Once here, newcomers should undergo a
76
The Contribution of Ethnic and National Identities to Political Conflict


program of indoctrination that cleanses them of their traditional loyalties
and imparts knowledge of cultural practices. Nativism thus gives an ethno-
centric cast to American national identity. Here, nationality does not simply
trump one’s ethnicity; rather, nationality fuses with a particular Anglo-
Saxon ethnicity. 
Scholars disagree about the extent and the voluntary nature of cultural
assimilation of immigrants to America (Gerstle, 1997; Hollinger, 1997;
Gordon, 1964; Glazer & Moynihan, 1959). While the passage of the 1964
Civil Rights Act and the 1965 Immigration and Nationality Act marked the
ascendancy of the liberal conception of citizenship, in previous decades na-
tivism and racism strongly influenced public policy. Gerstle (1997, p. 556)
argues that government support for nativism and the end of mass immigra-
tion suppressed the hyphenated American identities that had thrived earlier
and narrowed the range of acceptable cultural behavior. Even so, nativism
never entirely succeeded (Glazer & Moynihan, 1959); every immigrant did
not truly “melt” into a single, unchanging political culture. 
Multiculturalism is an alternative ideological response to the presence of
ethnic diversity in America. The “soft” or “liberal” version of this concep-
tion of national identity emphasizes the virtues of cultural pluralism and
calls on the government to help ensure tolerance toward minorities (Miller,
1995; Appiah, 1994; Raz, 1994). “Hard” or “radical” multiculturalism goes
further, conceiving of the nation as a confederation of ethnic groups with
equal rights and construing ethnicity as the preferred basis of one’s political
identity (Citrin, Sears, Muste, & Wong, forthcoming; Sears, Citrin, Chele-
den, & van Laar, 1999). 
Unlike liberalism, then, multiculturalism assumes that one’s ethnic iden-
tity is essential to one’s personal dignity and self-realization (Raz, 1994).
This conception of national identity does hold that differences among com-
munities of descent are basic and persisting. Since no culture is superior to
any other, none should be privileged in a multiethnic polity (Young, 1990;
Goldberg, 1994). T
o assure that minority cultures survive and flourish—
something that is viewed as indispensable for the well-being of their mem-
bers—“hard” multiculturalism therefore demands group rights and govern-
ment efforts to preserve minority cultures (Okin, 1997). Without such
support, it is argued, the standing of minority ethnic groups inevitably will
decline in the face of the cultural and economic power of the numerical 
majority, and members of minority ethnic groups will lose the core of their
social identity. 
Multiculturalism emphasizes ethnic consciousness and pride, but pro-
vides no apparent basis for social solidarity among America’s ethnic groups
save the principle of mutual tolerance (Raz, 1994). Because of this, critics
(Schlesinger, 1992; Kateb, 1994) warn that elevating the psychological
salience of ethnic identity, as Yugoslavia’s tragic recent history illustrates,
will increase group conflict or make it difficult to mobilize citizens on be-
half of national goals. In the vocabulary of social identity theory, they claim
The Meaning of American National Identity
77


that legitimating ingroup favoritism will lead inexorably to hostility and ag-
gression toward outgroups, including other ethnic groups within the
American national community.
Research Questions and Method
Discussion about the consequences of changing the balance between na-
tional and ethnic identities often proceeds in an empirical vacuum, without
much reference to the prevailing attitudes and beliefs of the American pub-
lic. This chapter employs survey data to present evidence about the follow-
ing questions: What is the level of patriotism in the United States, and does
it differ substantially across ethnic groups? Do white, black, Hispanic, and
Asian Americans differ in how they order their national and ethnic identi-
ties or in how they conceive of membership in the national community?
And does one’s particular conception of national identity influence opinions
about government policies?
Psychological theories regarding intergroup relations differ in their pre-
dictions about the preeminence of ethnicity in how people define themselves
politically (Sidanius, Levin, Rabinowitz, & Frederico, 1999; Sears et al.,
1999). Our own expectations are guided by symbolic politics theory (Sears,
1993; Sears et al., 1999). This approach assumes that individuals possess sta-
ble predispositions established in early socialization (e.g., through parents or
mass media) and reinforced by later experience (e.g., through education or
work). Ethnicity has no inherent priority in identify formation. Ethnic iden-
tification is one potentially significant predisposition, national attachment 
is another. The relative strength of these identities varies according to one’s
socialization experiences, and their influence on current behavior depends
upon the nature of environmental cues. For example, feelings of national
identity are likely to be evoked by international terrorist activities because
these raise the salience of one’s status as an American, but not by proposals to
increase the tax on tobacco.
By emphasizing the role of social learning, the symbolic politics perspec-
tive is compatible with either ethnic antagonism or harmony in multiethnic
societies. In the American context, it appears that national pride and a sense
of the country’s exceptionalism are strongly socialized attitudes, and this 
socialization may mitigate ethnic conflict by integrating diverse groups in an
overarching identity (Horowitz, 1985). All the country’s ethnic groups are
exposed to patriotic norms and the country’s liberal political tradition (Hartz,
1955; Lipset, 1996) that transmits belief in individual rights, not group
rights. The potency of these cultural messages would predict only minor 
ethnic differences in national identity and support for multiculturalism. 
Symbolic politics theory conceives of national and ethnic identities as
structures of valenced beliefs stored in memory (Lau & Sears, 1986). Which
of a person’s political predispositions influences her political choices partly
78
The Contribution of Ethnic and National Identities to Political Conflict


depends on the nature of the stimuli she encounters, such as events, poli-
cies, and candidates’ messages (Zaller, 1992). When a particular attitude
such as national pride is evoked, the theoretically predicted response is to
act in a way that reinforces one’s preexisting orientation. Accordingly, sub-
jective conceptions of national identity are hypothesized to govern opinions
about specific issues, such as foreign policy, immigration, or teaching
American history, which evoke symbols of nationhood, but not about “neu-
tral” issues, such as protecting endangered species.
The survey method relies on self-reports as the basis for classifying social
identities. The main body of evidence here comes from the 1996 General
Social Survey (GSS) of the American public aged 18 and older conducted
by the National Opinion Research Center (Davis & Smith, 1997). This sur-
vey employed a split-sample design and the results are based on the 1,367
respondents of the total sample of 2,094 who were asked questions about
national identity. In part because the 1996 GSS national survey included
very few Hispanic and Asian respondents, comparable evidence from the
1994 and 1995 Los Angeles County Social Surveys (LACSS), conducted by
UCLA’s Institute for Social Science Research (Sears et al., 1999) also is re-
ported. The Los Angeles data have special interest since they delineate the
degree of ethnic consensus and conflict in an extraordinarily diverse local
community where the salience of multiculturalism is a political constant.
The 1994 LACSS sample included 279 whites, 231 blacks, 264 Latinos, and
47 Asian respondents, and the corresponding numbers in the 1995 sample
were 259, 71, 202, and 47.
The 1996 GSS included a variety of items regarding beliefs and feelings
about America. From this set the analyses in this chapter use items covering
the different aspects of patriotism and chauvinism, as well as other ques-
tions assessing subjective beliefs about the normative criteria for American
nationality. Other questions about the respondent’s level of pride in distinct
features of American society and history were used to provide convergent
evidence regarding national attachment. In addition, questions concerning
one’s political self-categorization and one’s level of identification with one’s
ethnic group provided evidence about the balance between national and
ethnic identities.
Public Attachment to America
The present survey evidence shows clearly that most Americans retain a
strong, positive sense of national identity. T
able 4.1 presents the responses
of the 1996 GSS national sample, broken down by ethnicity, to an array of
relevant items, with the relatively few “don’t know” answers excluded. (For
the full wording of the questions, see Davis & Smith, 1997). T
able 4.1
groups these questions to reflect the tripartite definition of national identity
outlined above and to indicate the composition of multi-item attitude
The Meaning of American National Identity
79


table 4.1 Beliefs about American National Identity (1996 GSS)
T
otal
Whites
Blacks
Hispanics
(1367)
(1018)
(156)
(61)
Political Self-Categorization
Think of self mainly as “just American” in
social and political issues*
90%
96%
66%
79%
Affective Attachment to America
Patriotism scale items:**
How important is being an American to you?
(% saying ‘most important thing in their life’)
46
47
49
26
How close do you feel to America? (% saying 
‘close’ or ‘very close’)
81
83
69
83
Agree ‘I would rather be a citizen of America 
than any other country’
91
92
89
79
Pride in Aspects of America
% saying ‘very proud’ or ‘somewhat proud’
The way democracy works
83
85
77
81
Its political influence in the world
80
81
77
75
Its history
88
92
69
81
Its fair and equal treatment of all groups in 
society
57
60
42
51
Chauvinism scale items:
Agree ‘World would be a better place if people 
from other countries were more like Americans’
40
40
47
47
Agree ‘America is a better country than most 
other countries’
81
83
74
73
Agree ‘America should follow its own interests, 
even if this leads to conflicts with other nations’
44
44
50
39
Agree ‘People should support their country even 
when it’s in the wrong’
32
30
41
33
Normative Conceptions of American Identity
% Saying very or fairly important for a True 
American:
Nativism scale items:
T
o have been born in America
69
67
82
42
T
o be a Christian
54
50
77
50
T
o have lived in America for most of one’s life
73
71
84
77
Assimilationism scale items:
T
o be able to speak English
93
94
91
85
T
o feel American
87
89
81
72
T
o have American citizenship
93
92
93
93
Multiculturalism
Agree ‘It is impossible for people who do not 
share American customs and traditions to 
become fully American’
35
34
47
34
Agree ‘Ethnic minorities should be given 
govt assistance to preserve their customs 
and traditions’
17
11
44
45
Think it is better if groups adapt and blend into 
larger society rather than maintain their 
distinct customs and traditions
58
61
56
50
*This question was not asked in the 1996 GSS.  The data we report is from the 1994 GSS.
**The items listed under Patriotism, as well as those under the headings Chauvinism, Assimilationism, and
Nativism, are the items that were used in constructing these scales for the analyses appearing in other tables.


scales. The Political Self-Categorization item asks respondents to consider
which identity is primary to them when they think about political and social
issues. The Patriotism and Chauvinism items assess the affective dimension
of national identity. These items are similar in content to other measures of
psychological attachment to the nation (Kosterman & Feshbach, 1989;
Staub, 1997). Another set of “pride” questions taps the respondent’s sense
of pride in regards to specific aspects of American society. The “true Ameri-
can” questions capture normative conceptions of national identity and are
modeled on an earlier Traditional Americanism scale (Citrin, Reingold, &
Green, 1990). In addition, several items assess support for the maintenance
of diverse cultures within the political community, a core principle of the
multiculturalist conception of American identity. 
Clearly, table 4.1 shows that patriotic sentiment remains pervasive in the
American public. One poll is just a snapshot in time, of course, and the level
of positive affect one records will vary with the response categories used to
measure the underlying attitude. Nevertheless, the consistency in expres-
sions of pride and belonging is impressive. When asked to rate the impor-
tance of being an American on a 10-point scale, fully 46 percent of the 1996
GSS sample said it was “the most important thing in their life.” In response
to similar questions, 81 percent said they felt “close to America” and 91 per-
cent agreed that they would “rather be a citizen of America than any other
country.” Responses to the items in the “Pride” subheading in table 4.1
show that this generalized feeling of attachment carries over to judgments
about most specific features of national experience. For example, 83 percent
of the 1996 GSS sample expressed pride in “the way democracy works in
America.” However, many fewer, 57 percent, said they were proud of the
country’s record in treating all groups in society equally. Clearly, one can
have a strong sense of patriotism without believing that the nation is per-
fect. 
Nor does everyone extend pride in America to a conviction in the coun-
try’s superiority. While 81 percent of the 1996 national sample did agree
that America is a better country than most others, just 40 percent thought
the world would be a better place if people from other nations were more
like Americans. Only 32 percent agreed that one should support their coun-
try ”even when it is wrong,” compared to 50 percent who disagreed. Chau-
vinism, in the sense of a latent hostility toward other nations, is present, but
the dominant outlook seems to be just a preference for one’s own people
and place. 
The pattern of answers to the questions about what is important for 
being “truly American” reinforces this conclusion. Each of the attributes
listed was deemed either “very” or “fairly” important, suggesting that most
Americans believe that there is a distinct national identity. The more inclu-
sive political criteria—American citizenship (87 percent) and simply feeling
American (93 percent)—were chosen more often than the ascriptive quali-
ties of being native born (69 percent) or Christian (54 percent). The ability
The Meaning of American National Identity
81


to speak English (93 percent) might appear as an ethnocentric criterion, but
this is an achieved rather than an ascribed trait and research indicates that
within two generations most immigrants do learn the country’s common
language (Portes, 1996). In fact, language minorities and foreign-born resi-
dents are as likely as white or black Americans to emphasize the linguistic
criterion for national identity (Citrin et al., 1990). 
Multiculturalism asserts the need to maintain minority cultures within
America and regards the symbols of assimilation with suspicion, if not out-
right hostility. The 1996 GSS asked respondents whether they believed it is
“better for a country if different racial or ethnic groups maintain their dis-
tinct customs and traditions” or if they “adapt and blend into the larger so-
ciety.” Of those who expressed a definite opinion, 58 percent said that
“blending in” would be better. (In the case of this question, however, fully
27 percent of the respondents did not express a definite opinion, suggesting
substantial ambivalence about the choice as posed. When the respondents
without a definite opinion are included, 43 percent of the total sample are
in favor of “blending in,” while 31 percent favor the maintenance of cultural
diversity). Even fewer, only 17 percent of the GSS sample, agreed with the
position of many advocates of multiculturalism that ethnic minorities
should receive government support to preserve their traditions.
At the same time, there was considerable confidence about the possibility
of assimilation. More respondents (42 percent) disagreed than concurred
(35 percent) with the proposition that “it is impossible for people who do
not share American customs and traditions to become fully American.” Per-
haps because of this optimism, many Americans may seem to feel that as-
similation and maintaining connections to one’s ethnic heritage are mutu-
ally compatible. Public opinion endorses the importance of a common
national identity without insisting on a forced march to cultural conformity.
For example, a 1994 GSS poll found that only 26 percent of the public said
that ”ethnic history is getting too much attention in the public schools,” an-
other indication of the tendency to view the existence of a unifying Ameri-
can culture as compatible with respect for pluralism (Merelman, Streich, &
Martin, 1998). 
Using exploratory and confirmatory factor analysis, we combined the
items listed in table 4.1 to construct Patriotism (alpha = .64) and Chauvinism
(alpha = .66) indices and summary measures of Nativist (alpha = .79) and 
Assimilationist (alpha = .64) conceptions of Americanism. Although there is
statistical evidence for treating these dimensions as distinct, clearly they are
interrelated. Nativism and Assimilationism correlate .58 (Pearson’s r) in the
1996 GSS sample, and Patriotism and Chauvinism also are associated 
(r = .37). An interesting note is that Assimilationism was more strongly re-
lated to Patriotism (r = .47) than to Chauvinism (r = .38), whereas the obverse
was true for Nativism (r = .29 and .45, respectively), a pattern pointing to a
“psycho-logical” bond between an exclusionary definition of American iden-
tity and a sense of superiority to other peoples. However, beliefs about the
82
The Contribution of Ethnic and National Identities to Political Conflict


value of cultural diversity are not strongly related to these variants of positive
attachment to the nation. Responses to the question about whether or not
groups should blend into the mainstream culture correlated positively, but
weakly, with the Patriotism (.13), Chauvinism (.13), Nativism (.07), and As-
similationism (.13) scales. Admittedly, these correlations may be attenuated
owing to the use of just a single-item to measure support for assimilation into
the mainstream culture and due to the low variance on the scales. 
Normative Consensus or Ethnic Conflict? 
Although strong feelings of national attachment are the norm, the possi-
bility of ethnic conflict in outlook remains. The nativist prediction that 
diversity erodes national unity assumes that minority groups are less com-
mitted to the idea of a common American identity. The multiculturalist
perspective yields a similar expectation. If ethnicity is the dominant crite-
rion of one’s social identity, then minority group members should be less
likely than those in the dominant ethnic group to view the entire nation as
“theirs” and, therefore, to be less likely to express the usual forms of pride
and affection. 
Despite frequent claims that there is a vast racial divide in American
opinion (Kinder & Sanders, 1996), our surveys indicate that ethnic differ-
ences are slight when it comes to patriotic sentiment (table 4.1). There is a
strong similarity in the outlook of blacks and whites regarding the personal
importance of American citizenship. Blacks (69 percent) are somewhat less
likely to say they feel “very close” or “close” to America than whites (83
percent) or Hispanics (83 percent), but, again, the overwhelming majority
of all three groups express a strong sense of attachment to the nation. In the
1994 LACSS, 79 percent of the whites and 73 percent of the black respon-
dents said they were “extremely” or “very” proud to be an American.
The opinions of Hispanics point to a continuing connection to their
countries of origin among recent immigrants. As table 4.1 shows, they are
less likely than either blacks or whites to say being an American is impor-
tant to them, more willing to say they would move to another country, and
less likely to say that being born in America or feeling American is an im-
portant characteristic of being a true American. The same general pattern
exists among the Hispanic and Asian respondents in the Los Angeles data
(reported in full in Sears et al., 1999). Only 44 percent of the Hispanic re-
spondents and 58 percent of the Asians in the Los Angeles study felt “ex-
tremely” or “very” proud to be an American, but these differences from
white opinion are largely because many of these respondents are recent ar-
rivals in the country and are not yet citizens. The findings of the Latino
National Political Survey (de la Garza, Falcon, & Garcia, 1996, p. 346) also
indicate that after controlling for demographic characteristics, including
length of time in the United States, Mexican Americans “express patriotism
The Meaning of American National Identity
83


at levels equal to or higher than do Anglos.” Somewhat surprisingly, the
Los Angeles survey found virtually no ethnic differences on the question of
whether “it is better if groups change so they blend into the larger society
rather than maintain their distinct cultures.” 
The pervasive patriotic and assimilative tendencies among all ethnic
groups frays, however, when one moves from the realm of ideals to judging
historical experience. In the 1996 GSS survey, as reported in table 4.1,
blacks and, to a lesser extent, Hispanics were less likely than whites to ex-
press pride in America’s history, in the way its democracy works, and in the
treatment of all social groups. Members of minority groups also are less
likely than whites to agree that America has provided “people of my ethnic
group a fair opportunity to get ahead in life” (LACSS) and more likely to
agree that the government should give special assistance to ethnic minori-
ties (GSS). Minority groups, then, clearly display a heightened sensitivity to
their collective status, largely agreeing with whites about the goal of e
pluribus unum, but, as previous research also found (Hochschild, 1995), 
asserting more often that discrimination and unequal opportunities for 
minority groups are ongoing problems in the United States.
The normative conceptions of what it means to be a “true American”
also are very similar in all ethnic groups. In this regard, blacks are somewhat
more likely to endorse what we have termed ascriptive or nativist criteria of
identity. They are significantly more likely (by the chi-square test) than ei-
ther white or Hispanic respondents in the 1996 GSS to agree that being
born in America, living in America for most of one’s life, and being a Chris-
tian are “very important” qualities of national identity. In an important
sense, these opinions express the visceral quality of their American roots for
the black respondents. African Americans constitute the oldest group of
nonwhite immigrants in the country, despite the forcible nature of their ar-
rival. Thus, the symbolic meaning of their invocation of nativist criteria for
national identity might well be a positive statement about their own “true
American-ness” rather than prejudice toward outsiders. 
Symbolic politics theory emphasizes the role of socialization in the for-
mation of social identities and would predict that common exposure to the
patriotic themes so dominant in American life would engender high levels
of national attachment across all ethnic groups over time. Of course, so-
cialization is ongoing and multifaceted, and there generally are group dif-
ferences in prior learning and current experience that have an impact on
political predispositions (Kinder & Sears, 1985). T
able 4.2 summarizes the
ethnic group differences in feelings of national identity by comparing their
mean scores on the multi-item measures of Patriotism, Chauvinism, Na-
tivism, and Assimilationism Indices, with each variable scored from 0 (low)
to 1 (high). The results show that in all three ethnic groups, positive attach-
ment to American nationality prevails. Nevertheless, ANOVA indicates that
several of the ethnic differences that do appear around the highly favorable
norm are statistically significant. Blacks do have lower mean scores than
84
The Contribution of Ethnic and National Identities to Political Conflict


whites on Patriotism, though higher scores on Nativism. Hispanics also
have lower Patriotism scores than whites, as well as less support for the As-
similationist conception of national identity. Whether one should empha-
size the element of consensus or these minor and spotty differences is a
matter of debate (compare Sidanius, this volume). But whatever one makes
of the divergences among whites, blacks, and Hispanics, clearly the glass of
patriotism is much more than half full in every group. The real ethnic di-
vide in American politics emerges on specific policies bearing directly on
the status of minority groups, not on conceptions of national identity. 
The results of a multiple regression analysis in which race, ethnicity,
gender, age, income, education, region, and immigrant generation were
employed as predictors of the Patriotism, Chauvinism, Nativism, and As-
similationism scales, respectively, confirm this conclusion. (For space rea-
sons, the full results are not reported here, but will be provided on request.)
With the imposition of statistical controls, there is no “race effect” on Pa-
triotism or Chauvinism scores. Blacks remain significantly more likely than
whites to endorse the nativist conception of American identity, largely be-
cause of their greater support for Christianity as a legitimate defining crite-
rion of nationality. This suggests that another social identity, religion, is im-
portant to understanding ethnicity, national identity, and intergroup
conflict in America. 
More generally, there is a striking similarity in the demographic under-
pinnings of the four measures of national attachment: the elderly, those
with lower levels of education, and those residing in the South were signifi-
cantly more likely to have high, “pro-America” scale scores. Finally, first-
generation immigrants were significantly less likely to have high scores on
the Patriotism, Nativism, and Assimilationism indices. However, neither
second- nor third-generation Americans differ from those with an even
The Meaning of American National Identity
85
table 4.2 Ethnic Differences in National Identity1 (1996 GSS)
Mean Score on a 0 to 1 Scale
T
otal
Whites
Blacks
Hispanics
Affective attachment to America
Patriotism
.80
.81
.76
.75
Chauvinism
.59
.59
.62
.56
Normative conceptions of American identity
Assimilationism
.86
.86
.85
.79
Nativism
.64
.62
.79
.60
1. The four attitude measures are additive indices created by summing responses to the several groups of
items designated in table 4.1.  The response options for each item were first recoded to range from zero to
one.  After summation, index scores were also recoded to range from zero to one with higher scores indicat-
ing higher levels of patriotism, chauvinism, etc.  Reliabilities (Chronbach’s alpha) for the respective scales
are patriotism (.64), chauvinism (.66), assimilationism (.64), and nativism (.79).


longer family history in the country, once again suggesting that socializa-
tion into a sense of national pride and belief in a distinctive American iden-
tity does not take long. 
National, Hyphenated, and Ethnic Identities
Since citizens can have multiple loyalties (or identities), it is important to
assess the intensity of ethnic and national identifications separately.
Whether these attachments are complementary or competitive is an em-
pirical matter. The 1996 GSS asked respondents separate questions about
how close they felt to the United States and to their own racial or ethnic
group. A cross-tabulation of answers to these items show that 63 percent of
the white, 65 percent of the black, and 66 percent of the Hispanic respon-
dents expressed feelings of closeness to both the nation and their own ethnic
group. Among whites, only 11 percent felt close attachment to their own
ethnic group, but not to the country as a whole, whereas 24 percent of the
black and 14 percent of the Hispanic respondents fell into this category.
This suggests a stronger salience of an ethnic identity among blacks, with
the possible implication that they would be more likely than whites to 
privilege ethnicity over nationality when trade-offs must be made. Still, the
dominant result is that a large majority among both races expresses af-
finity to both identity groups. Subjectively, national and ethnic identities in
the United States tend to be experienced as complementary rather than
competing.
T
able 4.1 reports the responses to the political self-categorization ques-
tion: “When you think of social and political issues, do you think of yourself
mainly as a particular ethnic, racial, or nationality group, or do you think of
yourself mainly as just an American?” (The data come from the 1994, not
the 1996 GSS poll, which did not include the item). This question wording
calls for a ranking of identities and explicitly places the choice in a political
context. Of the white respondents, 96 percent answered “just an American,”
compared to 66 percent of the black and 79 percent of the Hispanic respon-
dents (see table 4.1). The uniformity of opinion among whites is consistent
with the view of the various groups of European immigrants having
“melted,” so that their most salient political identity is no longer cultural or
religious, but simply a superordinate identity as Americans (Alba, 1990;
Waters, 1990; Roediger, 1994). 
The 1994 and 1995 LACSS studies asked a similar self-categorization
question with equivalent results. In the pooled samples, 93 percent of the
white respondents gave the overarching “just an American” self-designa-
tion. Blacks (69 percent), Hispanics (70 percent), and Asians (71 percent)
were significantly less likely (by the chi-square test) than whites to say they
thought of themselves as just Americans. When forced to choose, however,
a large majority in all three minority ethnic groups opted for a common na-
86
The Contribution of Ethnic and National Identities to Political Conflict


tional identity, not ethnic particularity, as their first choice. Moreover,
among Hispanics and Asians, those who categorized themselves in terms of
ethnicity tended to be first-generation immigrants, suggesting, once again,
that the absorption of a sense of national identity is the normal outcome of
socialization in this country. 
The 1995 LACSS allowed respondents to say whether they identified as
just Americans or as both Americans and as members of an ethnic, racial, or
nationality group. As reported in table 4.3, only 17 percent of the whites
opted for a dual or hyphenated identity. But, when given this opportunity to
choose, a majority in all three minority groups preferred to categorize
themselves as both Americans and members of an ethnic group and not as
exclusively “American.” Unfortunately, there are no national data to repli-
cate this result, but it is consistent with other small-scale studies testing so-
cial identity theory that suggest the relatively greater salience of an ethnic
identity for members of smaller, easily demarcated, homogeneous, and 
disadvantaged groups (Prentice & Miller, 1999; Brewer, this volume). The
historically dominant group in a multiethnic state often fuses rather than
hyphenates their ethnic and national identities (Connor, 1993). It is beyond
the scope of this chapter to explore the antecedents of these alternative self-
categorizations by minority group members. Among immigrant groups, it
appears that length of residence in the United States and speaking English
diminish the likelihood of identifying oneself primarily in ethnic terms 
(de la Garza et al., 1996), but these factors could not account for the varied
responses of blacks. 
Identities and Policy Preferences
The political relevance of social identities rests partly on how they influence
attitudes and behavior toward one’s own and other groups. A strong sense
of group identity is expected to promote conduct that favors one’s fellow
group members, even in a “minimal group” (T
ajfel, 1978). But since there
are multiple dimensions of self-identification, how one defines the “I” as
“we” when a political decision arises is likely to be significant. Symbolic
politics theory holds that broad attitudes, including feelings of national and
ethnic identity, are more likely to determine preferences on specific issues
when those predispositions are central to an individual’s self-definition and
when they are cued by stimuli associated with the objects of these attitudes
(Sears, 1993; Zaller, 1992). For example, a strong sense of Jewish identity is
more likely to govern opinions about American policy toward Israel or a
mandatory school prayer than the future of the spotted owl. In addition, as
shown above, ethnic identity may indeed be more significant for the con-
struction of the social self among some groups than others. 
Nationalism is an integrating ideology that overrides the claims of less
comprehensive group loyalties. In fact, liberal theorists (T
amir, 1997;
The Meaning of American National Identity
87


table 4.3 Political Self-Categorization: American, Dual, or Ethnic Identity (1995 LACSS)
Whites
Blacks
Hispanics
Asians
(n = 254)
(70)
(174)
(46)
Just an American Identity
78%
16%
11%
4%
Both American and ethnic identity
17
54
59
59
Ethnic identity
5
30
31
37
Whites
Blacks
Hispanics
Asians
US born 
Not US born 
US born 
Not US born
US born 
Not US born 
US born 
Not US born
(n = 223) 
(31)
(68)
na
(59)
(115) 
(8)
(38)
Just an American identity 
80%
65%
16%
na
24%
4% 
13%
3%
Both American & ethnic identity 
17
16
54
na
53
62 
63
58
Ethnic identity 
3
19
29
na
24
34 
25
40


Hollinger, 1997; Miller, 1995) justify holding on to a national identity on
the grounds that this sentiment generates feelings of special obligation to
and responsibility for the other members of one’s national community. If
this surmise is correct, then people should be less likely to feel a sense of
duty to members of groups who define themselves as somehow outside or
different from the national “family.” Accordingly, since multiculturalism
validates the maintenance of group differences, the stronger one’s attach-
ment to an overarching American national identity, the more likely one
should be to oppose recognition and support for groups emphasizing their
cultural distinctiveness. Alternatively, if the majority group views ethnic mi-
norities as disadvantaged members of a community with a shared identity,
feelings of national attachment could accommodate, even facilitate, support
for policies aimed at helping them. 
In considering the political implications of ethnic consciousness, Tyler,
Lind, Ohbuchi, and Sugawara (1998) found that assimilated and “bicultur-
alist” members of minority groups (that is, those who emphasize their
American identity and dual identity, respectively) were equally likely to ex-
press trust in government, to regard existing institutions as fair, and to com-
ply with government policy. Both groups were significantly more likely
than “separatist” respondents from the same ethnic group (those emphasiz-
ing their ethnic identity) to express these attitudes. The present data show
that the relative salience of an ethnic as opposed to national self-identifica-
tion affects how people evaluate multiculturalist principles (see table 4.4).
Minority group members for whom ethnic identification takes priority over
national identity generally are less likely than those who say they view poli-
tics mainly as an American to express pride in America and to agree that
their group has had a fair chance to get ahead, and are more likely to believe
that groups should maintain their own distinct cultures, to deny that politi-
cal organizations based on ethnicity promote separatism, and to think that
people are best represented by leaders with the same ethnic background.
(See table 4.4 for results of chi-square tests for statistical significance.) 
T
able 4.4 also suggests that the association between an ethnic self-identi-
fication and support for multiculturalism is stronger and more systematic
among blacks than among either Hispanics or Asians. For example, among
blacks, 61 percent of those who identified themselves as “just American” felt
it is better for minority groups to blend into the larger society rather than
maintain their distinct cultures, compared to only 25 percent of those who
defined themselves primarily as members of an ethnic group. Among His-
panics, the equivalent figures were 60 percent and 54 percent, a far smaller
difference, and among Asians, they were 57 percent and 48 percent, respec-
tively. The future trend in how the different minority groups balance their
ethnic and national identities thus has clear implications for the pattern of
intergroup conflict over issues of cultural recognition and representation.
Ordinary least-squared regression analysis was employed to delineate the
connections between the affective and normative dimensions of national
The Meaning of American National Identity
89


table 4.4 The Impact of Political Self-Categorization on Beliefs about the American National Community (1994 LACSS)
Blacks
Hispanics
Asians
Just
Ethnic
Just
Ethnic
Just
Ethnic
American
Identity
American
Identity
American
Identity
(n = 156)
(68)
(182)
(66)
(36)
(8)
Feel extremely or very proud to be an American
81%
54%
53%
22%
61%
33%
Agree that American society has provided people of my ethnic
50
22*
70
63
69
63
group a fair opportunity to get ahead in life
Think it is better if groups change so that they blend into the larger
61
25
60
54
57
48#
society (as in the idea of a melting pot) rather than maintain their
distinct cultures1
Agree that political organizations based on race or ethnicity
70
57#
71
52*
70
38#
promote separatism and make it hard for all of us to live together
Leader background doesn’t matter for representation
68
40*
67
50*
60
38
Chi-square test of differences between “just American” and “ethnic identity” respondents within each racial/ethnic group:
*p<.05
#p<.10
1The percentages in this row represent pooled data from the 1994 and 1995 LACSS.


table 4.5 The Relationship between Attitudes T
oward National Identity  and Policy Preferences (1996 GSS)
Affective Attachment
Normative Conceptions
(3)
(6)
Patriotism
Assimilationism
(1)
(2)
& Chauvinism
(4)
(5)
& Nativism
Patriotism
Chauvinism
(patrsm) (chauv)
Assimilationism
Nativism
(assim) (nativ)
Policy Preferences
Multiculturalism
(A) Need to share American customs
.05
.38**
–.08
.42**
.17**
.27**
–.07
.29**
(B) Support melting pot as ideal
.20*
.24**
.18
.17
.25**
.01
.33**
–.10
(C) Reduce immigration levels
.09
.18**
.06
.17**
.26**
.17**
.21**
.09*
(D) No cultural benefit from immigrants
.04
.20**
–.01
.21**
.11**
.15**
–.01
.16**
Racial Policies
(E) Oppose affirmative action
.02
–.01
.01
.00
–.05
–.03
–.05
–.01
(F) Oppose govt. help for blacks
.06
.14*
.03
.13
.12*
.06
.08
.04
(G) Oppose more spending on blacks
.17
.06
.14
.00
–.01
.06
.03
.05
Domestic spending
(H) Oppose social spending
.03
.03
–.03
.05
.03
.00
.05
–.02
Foreign Policy
(I) Support more spending on defense
.23*
.31**
.12
.25*
.19*
.13*
.10
.10
(J) Support protectionism
.18**
.42**
.06
.40**
.29**
.23**
.16**
.16**
Entries are unstandardized regression coefficients for national identity variables in equations with controls for race, ethnicity, age, education, income, gender, region (South), generation, 
ideology, and party identification.
*p<.05
**p<.01


identity on the one hand and policy preferences on the other. The statistical
model estimates the effects of the Patriotism, Chauvinism, Assimilation,
and Nativism Scales, (columns 1, 2, 4, and 5 in table 4.5), respectively, on 
issues related to cultural diversity (rows A–D), racial issues (rows E–G), 
domestic spending (row H), and foreign policy (rows I–J). The equations
included race, ethnicity, age, education, income, gender, region, immigrant
generation, partisan affiliation, and liberal-conservative self-identification
as control variables. Owing to the small number of Hispanic respondents in
the 1996 GSS sample, this analysis is confined to whites and blacks. For
simplicity of presentation, the table omits the effects for the control vari-
ables and reports only the unstandardized regression coefficients for the
measures of national identity (Patriotism, Chauvinism, Assimilationism,
and Nativism). All variables were recoded to vary between 0 and 1, and the
dependent variables all are scored in the “nationalist” or “conservative” 
direction. Accordingly, the positive coefficient of .24 for the association 
between Chauvinism and the Melting Pot item (column 2, row B) indicates
that believing that America is superior to other countries is strongly related
to the idea that ethnic minorities in the United States should blend into the
mainstream culture. 
T
able 4.5 also reports the results of models that simultaneously included
both Patriotism and Nationalism Scales (column 3) and the Assimilationism
and Nativism Scales (column 6) as predictors. These equations were de-
signed to provide a tentative estimate of the relative degree of association
between each of these affective orientations and normative conceptions of
national identity. The results (which are reviewed by comparing the coeffi-
cients across columns on a row-by-row basis) do provide some evidence of
tension between a strong sense of national identity and multiculturalism.
Although the respondent’s race has no relationship to these multiculturalist
policy positions in the multivariate model, three of the four indicators 
of national identity predict support for the belief that it is impossible for
people who do not share American customs to become fully American (row
A). With regard to the question asking whether ethnic and racial groups
should blend into mainstream society, Patriotism, Chauvinism, and Assimi-
lationism had statistically significant relationships with giving the melting-
pot response (row B). Immigration is the major source of ethnic diversity
and new claimants for cultural recognition in America, and Chauvinism,
Assimilationism, and Nativism, but not Patriotism, had statistically signifi-
cant relationships with support for reducing the level of immigration (row
C) and disagreement that immigrants have a positive cultural impact on
America (row D). 
When Patriotism and Chauvinism are included as predictors simultane-
ously, the more aggressive outlook tends to have the stronger association
with opposition to increased cultural diversity in the United States. And
when the effects of Assimilationism and Nativism are compared, it is the
latter more exclusionary attitude that generally appears to have the stronger
92
The Contribution of Ethnic and National Identities to Political Conflict


negative relationship to support for multiculturalism. An ethnocentric con-
ception of American national identity, therefore, may intensify internal po-
litical conflict by hardening resistance to the demands of cultural minorities
for recognition and support. 
T
able 4.5 also reports on opinions about government efforts targeted at
helping blacks (rows E–G) and about spending on domestic social programs
(measured by a Social Spending Index combining beliefs about whether the
government should spend more on health, education, welfare, and the envi-
ronment—row H). The results are a virtual mirror image of the earlier
findings about the factors influencing opinions about cultural diversity and
immigration. In the case of the domestic policy issues, race is strongly asso-
ciated with preferences, with blacks consistently more favorable toward
governmental activism. But none of the four measures of national identity
has a statistically significant relationship with scores on the Social Spending
Index, opinions about affirmative action, or attitudes about whether the
government should spend more to help blacks. The only coefficients in the
array that attain conventional levels of statistical significance are the weak
associations of Chauvinism and Assimilationism with opposition to special
efforts to improve the living conditions of blacks. The results of these
analyses indicate that generalized feelings of national attachment neither
boost support for assisting one’s fellow citizens through government spend-
ing nor enhance opposition to policies, such as affirmative action, aimed at
assisting blacks or other ethnic minorities. 
This may not be surprising, since the controversies over domestic spend-
ing and racial policy tend not to be framed in terms that engage attitudes
toward the nation’s integrity or power. As hypothesized above, the impact
of social identities (and other predispositions) on specific policy preferences
should vary across political domains. Feelings of national attachment
should be strongly engaged only when the specific issues clearly touch on
the values of American power, sovereignty, or purity. The data presented in
table 4.5 consistently support this proposition. In addition to their effects
on beliefs about cultural diversity and immigration, all four measures of na-
tional attachment had statistically significant effects in predicting opinions
about whether military spending should be increased and on scores on an
American Protectionism Index (alpha = .65), constructed by summing an-
swers to items about limiting foreign imports, prohibiting foreigners from
buying land in America, and requiring television stations to give preference
to American programs. 
A final regression analysis, reported in table 4.6, explored potential dif-
ferences in the connections of ethnic as opposed to national identities on
racial policy preferences. Here, responses to both the question “How close
do you feel to the United States?” and the question “How close do you feel
toward your own racial or ethnic group?” were included as predictors of
opinions regarding affirmative action, support for government efforts to
help blacks, and government spending for blacks. In this analysis, the vari-
The Meaning of American National Identity
93


able coding and controls are identical to those reported for table 4.5. How-
ever, to distinguish between the impact of ethnic identifications among
whites and blacks, these equations were estimated separately for whites and
blacks. (Because of the paucity of Hispanic respondents, the analysis could
not be replicated for that group.) 
T
o highlight the results, table 4.6 reports only the coefficients for na-
tional attachment and ethnic closeness. Confirming the results reported
above, positive affect for the nation as a whole has no impact one way or the
other on the racial policies queried. There are obviously significant racial
differences in policy preferences, but among blacks, feeling close to one’s
own ethnic group had no great additional influence on support for policies
aimed at improving their collective status. The fear of some political com-
mentators that black racial identification leads to a highly distinctive politi-
cal outlook seems overstated. 
Among whites, however, a stronger sense of ethnic identity does have a
significant effect across the three policies; white identification seems consis-
tently to increase opposition to government policies designed to help
blacks. Given our concern about the impact of ethnic identification on in-
tergroup conflict, these results are sobering. “Identity politics,” by which
we mean the tendency to judge issues and events in terms of how they affect
the standing of one’s own group, can juxtapose one’s ethnic and national
94
The Contribution of Ethnic and National Identities to Political Conflict
table 4.6 National Attachment and Ethnic Identity as Predictors of Opinion 
on Race-T
argeted Policies (1996 GSS)
Oppose
Oppose Special
Oppose More
Affirmative
Govt. Help
Spending on
Action
for Blacks
Blacks
Among White Respondents
Close to own ethnic group (close =1)
.07**
.05*
.08*
(white identification)
Close to the US
–.02
–.02
–.03
(national attachment)
Sample size (n)
543
546
347
Among Black Respondents
Close to own ethnic group (close =1)
.12
–.05
.09
(black identification)
Close to the US
–.11
.16
.01
(national attachment)
Sample size (n)
72
74
50
Entries are unstandardized regression coefficients for national and ethnic identity variables in equations with
controls for age, education, income, gender, region (South), generation, ideology, and party identification.
*p<.05
**p<.01


“selves.” While minorities may be more likely to express a strong ethnic
identity, counterintuitively, it appears that the effects of ethnic identification
are more prominent among whites. In a disturbing way, the results in table
4.6 are a reminder that the evocation of white consciousness, whether
through nativist mobilization or through reaction to the assertion of ethnic
pride by other groups, can lead to divisiveness and prejudice. 
Conclusion
A large body of survey evidence confirms that a positive sense of national
identity is pervasive in the United States, as is the belief that there is a
unique American identity. In terms of affective attachment, patriotism—the
sense that America is “best for me”—was more widespread than chauvin-
ism—the sense that this country is inherently superior to all others. Signifi-
cantly, too, we found no evidence of a deep ethnic divide in feelings about
American nationality. The slightly lower level of patriotism among blacks,
as a group, than whites is less compelling than the fact that the large major-
ity of all ethnic groups express pride in and closeness to their country and
its symbols. In addition, there is a broad consensus among all ethnic groups
regarding the main criteria for American identity. 
The strong sense of attachment to an American identity among ethnic
minorities should calm fears about the consequences of the nation’s in-
creased ethnic diversity. Nearly as much as whites, minority respondents in
our surveys rejected the idea of organizing political life along ethnic lines.
Furthermore, most of them do not express their primary political identity
in purely ethnic terms. 
These results do not fit easily into an image of public attitudes as driven
primarily by interest-based conflict between dominant and subordinate eth-
nic groups. The symbolic politics perspective outlined here presumes that
attitudes and values are acquired through a process of reinforcement, and
that the role of purely instrumental motivation often is secondary (Sears &
Funk, 1990). The survey evidence suggests that members of all ethnic
groups may be attracted to an American identity because of its traditional
liberal premises. In both California and national surveys (Citrin et al., 1990,
1994), the criteria of a “true American” chosen as important most often by
all four major ethnic groups were egalitarianism and tolerance (“treating
people of all races equally” and “respecting people’s freedom to say what
they want”).
The staying power of this liberal conception of American identity may
help explain why many minority respondents viewed attachment to the
country and a sense of closeness to their own ethnic group as complemen-
tary rather than competitive. The results presented above show that mi-
nority group members frequently conceive of themselves in dual terms,
defining themselves as both Americans and part of a distinctive racial or
The Meaning of American National Identity
95


ethnic group. Whites, by contrast, generally see themselves as “just Ameri-
cans”; for them, national and ethnic identities are merged. What is critical,
both psychologically and politically, is how the dual identity—national and
ethnic—is accommodated intrapersonally. That is, when does the con-
sciousness of being simultaneously American and black lead to a sense of
ambivalence that pits the self against either the nation or the group, and
how is this conflict among loyalties resolved? The present data reinforce
the familiar point that blacks identify with American ideals, not American
realities. As James Jones suggests (personal communication), their dual
identity reflects recognition of the persistence of what Myrdal (1944) called
the “American dilemma.”
In the political domain, a strong sense of national identity consistently
was related to opposition to liberal immigration policies and official support
for maintaining minority cultures. There is, then, an ideological tension be-
tween multiculturalism and nationalism. Strong feelings of national attach-
ment also were associated with approval for increasing America’s military
strength, as well as a willingness to take quite drastic measures to reduce the
potential impact of foreign interests and influences. These issues, of course,
focus on the boundaries between citizens and strangers, a distinction that is
less relevant for opinion formation when it comes to domestic spending or
racial policies. 
The strength of national attachment was unrelated to how whites felt
about policies targeted at helping blacks, their fellow-citizens. However, a
strong ethnic identification among whites was associated with greater an-
tagonism to affirmative action and more government spending to assist
blacks. In this instance, ingroup favoritism does seem to lead to greater 
inter-group conflict. 
The portrait of normative consensus among the general public seems
strangely discordant with the seeming prevalence of ethnic competition in
American political life. The division of political offices, government con-
tracts, and other benefits among ethnic groups is closely watched by ac-
tivists. Racial issues deeply divide the country’s two political parties. Given
all this and more, what accounts for the failure of the surveys to detect in-
tense disagreements between the “majority” and “minority” groups when it
comes to attitudes about national identity?
One answer is that the existence of a gulf between the views of political
activists and ordinary citizens is not uncommon. Another possibility is that
the current approach to measuring national and ethnic identifications is in-
adequate. Surveys may pay too much attention to abstract norms and too
little to responses to distributive conflicts or actual social encounters among
members of different groups. In this regard, researchers should frame issues
in a way that asks respondents to make trade-offs among their multiple
identities when it comes to allocating costs and benefits. This would enable
a more comprehensive assessment of intragroup, as well as intergroup, con-
flict in how these choices are made. 
96
The Contribution of Ethnic and National Identities to Political Conflict


Another important issue concerns the enduring quality of the tendency
of recent immigrants to prefer a dual or hyphenated political identity. The
present data indicate that a developing sense of patriotism and a sense of
identity as an American are part of the process of acculturation that His-
panic and Asian groups are experiencing. More generally, today’s immi-
grants, like their predecessors, become “desocialized from their native cus-
toms over time, even if the emotional significance of attachment to the
group persists” (Glazer & Moynihan, 1975, p. 8). Because of this continu-
ing affective tie, appeals to ethnicity still can be an effective strategy for mo-
bilizing political support for the pursuit of group or individual interests
(Horowitz, 1985). The utility of ethnicity in political combat, then, may 
reinforce the tendency of political leaders to emphasize the salience of this
aspect of identity and strive to sustain its psychological significance within
their ethnic group. 
In addressing the unresolved issue concerning the conditions under
which hostility toward outgroups accompanies ingroup pride, Gaertner,
Dovidio, Nier, Ward, & Banker (1999) argue for the efficacy of a superordi-
nate common identity in mitigating intergroup conflict. Applied to the
larger realm of American national politics, the question posed by social
identity theory becomes whether an enhanced sense of ethnic identification
among minorities leads toward satisfactory conflict resolutions or toward
entrenched Balkanization. Clearly, the mobilization of group consciousness
among the disadvantaged can force movement toward less inequality and
more social justice. On the other hand, the enhanced expression of ethnic
identification in one group tends to be emulated, often provoking resistance
to change and the hardening of group boundaries. Whatever the direction
of future change in public attitudes, the evidence summarized here points to
the advantages of a common sense of American identity founded on the 
realization of equal citizenship. 
References
Alba, R. D. 1990. Ethnic identity: The transformation of white America. New Haven:
Yale University Press.
Appiah, K. A. (1994). Identity against culture: Understandings of multiculturalism.
Berkeley, CA: Doreen B. T
ownsend Center for the Humanities.
Ashmore, R. D., & Jussim, L. (1997). Self and identity: Fundamental issues. New York:
Oxford University Press. 
Barber, B. (1995). Jihad vs. McWorld. New York: Times Books.
Bar-tal, D. (1993). Patriotism as fundamental beliefs of group members. Politics and
the Individual, 3, 45–62.
Brimelow, P. (1995). Alien nation. New York: Random House.
Citrin, J., Sears D., Muste C., & Wong, C. (Forthcoming). Multiculturalism in
American public opinion. British Journal of Political Science.
Citrin J., Haas E. B., Muste C., & Reingold B. (1994). Is American nationalism
The Meaning of American National Identity
97


changing?: Implications for foreign policy. International Studies Quarterly, 38,
1–31.
Citrin J., Reingold B. & Green. D. P. (1990). American identity and the politics of
ethnic change. Journal of Politics, 52, 1124–1154.
Connor, W. (1993). Beyond reason: The nature of the ethno-national bond. Ethnic
and Racial Studies, 163, 373–389.
Davis J., & Smith. T. (1997). The general social survey codebook. Chicago: National
Opinion Research Center.
de la Garza, R. O., Falcon A., Garcia C. (1996). Will the real Americans please stand
up: Anglo and Mexican-American support of core American political values.
American Journal of Political Science, 40, (2), 335–350.
Farley, R. (1996). The new American reality : Who we are, how we got here, where we are
going. New York: Russell Sage Foundation.
Gaertner, S. L., Dovidio J.F., Nier J. A., Ward C. M., & Banker, B. S. (1999). Across
cultural divides: The value of superordinate identity. In D. Prentice & D. Miller.
(Eds.), Cultural divides: The social psychology of cultural contact (pp. 173–212). New
York: Russell Sage Foundation.
Gerstle, G. (1997). Liberty, coercion, and the making of Americans.” Journal of
American History, 84, 524–558.
Glazer, N. & Moynihan D. P. (1963). Beyond the melting pot. Cambridge: MIT Press.
Glazer, N., & Moynihan. D. P. (1975). Ethnicity: Theory and experience. Cambridge,
MA: Harvard University Press.
Gleason, P. (1980). American identity and Americanization. In S. Thernstrom, A.
Orlov, & O. Handlin (Eds.), Harvard Encyclopedia of American Ethnic Groups
(pp 31–58). Cambridge, MA: Harvard University Press. 
Goldberg, D. T. (1994). Introduction: Multicultural conditions. In D. T. Goldberg,
(Ed.), Multiculturalism: A critical reader (pp. 1–41). Cambridge, MA: Blackwell.
Gordon, M. (1964). Assimilation in American life: The role of race, religion, and national
origins. New York: Oxford University Press.
Greenfeld, L. (1992). Nationalism : Five roads to modernity. Cambridge, MA: Harvard
University Press.
Greenfield, L., & Chiriot, D. (1994). Nationalism and aggression. Theory and Soci-
ety, 23, 79–130.
Hartz, L. (1955). The liberal tradition in America: An interpretation of American political
thought since the revolution. New York: Harcourt, Brace.
Higham, J. (1988). Strangers in the land: Patterns of American nativism, 1860–1925.
2nd ed. New Brunswick, NJ: Rutgers University Press.
Hochschild, J. L. (1995). Facing up to the American dream: Race, class, and the soul of the
nation. Princeton, NJ: Princeton University Press. 
Hollinger, D. (1995). Postethnic America. New York: Basic Books.
Hollinger, D. A. (1997). National solidarity at the end of the twentieth century: 
Reflections on the United States and liberal nationalism. Journal of American 
History, 84, 559–569.
Horowitz, D. (1985). Ethnic groups in conflict. Berkeley: University of California
Press.
Huntington, S. P. (1981). American politics: The promise of disharmony. Cambridge,
MA: Belknap Press.
Kateb, G. (1994). Notes on pluralism. Social Research, 61, 511–537. 
98
The Contribution of Ethnic and National Identities to Political Conflict


Kelman, H. (1997). Nationalism, patriotism, and national identity: Social-psycho-
logical dimensions. In D. Bar-tal & E. Staub (Eds.), Patriotism (pp. 165–189).
Chicago: Nelson-Hall.
Kinder, D. R., & Sanders, L. M. (1996). Divided by color: Racial politics and democratic
ideals. Chicago: University of Chicago Press.
Kinder, D., & Sears, D. O. (1985). Public opinion and political action. In 
G. Lindzey & E. Aronson (Eds.), Handbook of Social Psychology. 3rd ed., vol 2., 
(pp. 659-741) Reading, MA: Random House.
Kosterman, R., & Feshbach, S. (1989). T
owards a measure of patriotic and national-
istic attitudes. Political Psychology, 10, 257–274.
Lau, R., & Sears, D. (Eds.). (1986). Political cognition : The 19th Annual Carnegie
Symposium on Cognition.. Hillsdale, NJ: Lawrence Erlbaum Associates.
Lind, M. (1995). The new nationalism and the fourth American revolution. New York:
The Free Press.
Lipset, S. M. (1996). American exceptionalism: A double-edged sword. New York: W. W.
Norton. 
Merelman, R. M., Streich, G. & Martin, P. (1998). Unity and diversity in American
political culture: An exploratory study of the national conversation on American
pluralism and identity. Political Psychology, 19, 781–808.
Miller, A. H., Gurin, P., Gurin G., & Malanchuk O. (1981). Group consciousness
and political participation. American Journal of Political Science, 25, 494–511.
Miller, D. (1995). On nationality. Oxford: Clarendon Press.
Myrdal, G. (1944). An American dilemma: The Negro problem and modern democracy.
New York, London: Harper & Brothers.
Okin, S. M. (1997). Is multiculturalism bad for women? Boston Review, 22, 5.
Portes, A. (Ed.) (1996). The new second generation. New York: Russell Sage Foundation.
Prentice, D. A. & Miller, D. T. (Eds.) (1999). Cultural divides: Understanding and
overcoming group conflict. New York: Russell Sage Foundation.
Raz, J. (1994). Multiculturalism: A liberal perspective. Dissent, 41, 67–79.
Reich, R. B. (1992). The work of nations: Preparing ourselves for 21st century capitalism.
New York: Vintage Books.
Roediger, D. R. (1994). Towards the abolition of whiteness: Essays on race, politics, and
working class history. London: Verso.
Schlesinger, A. M. Jr. (1992). The disuniting of America: Reflections on a multicultural
society. New York: Norton.
Schaar, J. H. (1981). Legitimacy in the modern state. New Brunswick, NJ: Transaction
Books.
Sears, D. O. (1993). Symbolic politics. In S. Iyengar & W. J. McGuire, (Eds.), Explo-
rations in political Psychology (pp. 113–149). Durham: Duke University Press.
Sears, D. O., Citrin J., Cheleden S., & Van Laar C. (1999). Cultural diversity and
multicultural politics: Is ethnic Balkanization psychologically inevitable? In D.
Prentice & D. Miller, (Eds.), Cultural divides: The social psychology of cultural con-
tact. (pp. 35–79) New York: Russell Sage Foundation. 
Sears, D. O., & Funk, C. L. (1990). Self-interest in Americans’ political opinions. 
In J. J. Mansbridge, (Ed.), Beyond self-interest (pp. 147–170) (1999). Chicago:
University of Chicago Press.
Sidanius J., Levin S., Rabinowitz J. L., & Frederico C. M. (1990). Peering into the
jaws of the beast: The integrative dynamics of social identity, symbolic racism
The Meaning of American National Identity
99


and social dominiance. In D. A. Prentice & D. T. Miller, (eds.), Cultural divides:
Understanding and overcoming group conflict (pp. 80-132) New York: Russell Sage
Foundation. 1999.
Smith, R. M. (1997). Civic ideals: Conflicting visions of citizenship in U.S. history. New
Haven: Yale University Press.
Staub, E. (1997). Blind versus constructive patriotism: Moving from embeddedness
in the group to critical loyalty and action. In D. Bar-tal & E. Staub, (Eds.), Patri-
otism (pp. 213–228).Chicago: Nelson Hall.
Sullivan, J., Fried A., & Dietz M. (1992). Patriotism, politics, and the presidential
election of 1988. American Journal of Political Science, 36, 200–234.
T
ajfel, H. (1978). Social categorization, social identity and social comparison. In H.
T
ajfel (Ed.), Differentiation between social groups (pp. 61–76). London: Academic
Press. 
T
ajfel, H., & Turner J. (1986). The social identity theory of intergroup behavior. 
In S. Wrochel & W. Austin, (Eds.), Psychology of intergroup relations. Chicago:
Nelson-Hall.
T
aylor, C. (1995). Cross purposes: The liberal-communitarian debate. In Philosophi-
cal arguments. (pp. 181–203). Cambridge, MA: Harvard University Press.
Thoits, P. A., & Virshup, L.K. (1997). Me’s and we’s: Forms and functions of social
identities.” In R. D. Ashmore & L. Jussim, (Eds.), Self and identity: Fundamental
issues. New York: Oxford University Press. 
Turner, J. C., Hogg, M.A. Oakes, P. J. Reicher, S. D. & Blackwell, M. S. (1987). 
Rediscovering the social group: A self-categorization theory. Oxford: Basil Blackwell.
T
yler, T., Lind E. A., Ohbuchi K., & Sugawara, I. (1998). Conflict with outsiders:
Disputing within and across cultural boundaries. Personality & Social Psychology
Bulletin, 24, 137–147.
Warner, R. S. (1993). Work in progress: T
oward a new paradigm for the sociological
study of religion in the U.S. American Journal of Sociology, 98, 1044–1093.
Waters, M. (1990). Ethnic options: Choosing identities in America. Berkeley: University
of California Press. 
Wolfe, A. (1998). One nation after all. New York: Viking.
Young, I. M. (1990). Justice and the politics of difference. Princeton, NJ: Princeton
University Press.
Zaller, J. (1992). The nature and origins of mass opinion. New York: Cambridge Uni-
versity Press. 
100
The Contribution of Ethnic and National Identities to Political Conflict


Jim Sidanius
5
John R. Petrocik
Communal and National Identity
in a Multiethnic State
A Comparison of Three Perspectives
As recent episodes of “ethnic cleansing” in Rwanda, Bosnia, and Kosovo
make painfully clear, some multiethnic states seem as inclined to interethnic
violence at the dawn of the 21st century as they were at the start of the 20th.
Indeed, the national mobilization of a 50-year-long cold war may have tem-
porarily restrained hostilities that lost none of their underlying virulence.
Grievances that were set aside by a need to face a common enemy have now
become priority issues for organized interests, parties, and public policy.
Furthermore, ethnic conflict is not peculiar to nation-states mobilized into
the stand-off between the cold war alliances, but appears quite generalized
across the globe (see Gurr & Harff, 1994).
In many cases, intergroup conflict within multiethnic states turns on the
compatibility of subgroup identities and commitments to the nation-state.
Thus, one wonders if Israel can rely only upon Jewish citizens? Can a
Gypsy also be Hungarian? Are Finns of Swedish extraction loyal to Fin-
land? Can Francophones be as interested in the welfare of Canada as En-
glish speakers? Can a Turkish immigrant (and his children) be as German as
an ethnic German? Can Chinese, Filipino, and Mexican immigrants also be
good Americans? The national loyalty question arises because “social ho-
mogeneity and political consensus are regarded as prerequisites for, or fac-
tors strongly conducive to, stable democracies” (Lijphart, 1977, p. 1). These
questions are crucial because the modern nation-state is predominantly 
(if not universally) organized around ethnic definitions of citizenship
(Horowitz, 1985). Yet few nations are homogeneous and, as migration in-
creases, ethnic diversity creates more opportunities for conflict within them
along ethnic, racial, religious, and linguistic divides (hereafter referred to
101


collectively as “communal” differences). A better understanding of the
manner in which people are able to negotiate loyalty to their ethnic particu-
larisms while adopting or retaining the civic culture and loyalty demanded
by the nation-state seems as worthwhile today as it ever was in the past.
This is especially true given that it is the institutions of the state that pro-
vide the principal framework within which communal differences are re-
solved. This chapter examines the relationship between ethnic and national
identities within the United States.
Diversity as a Source of Conflict
While we recognize a history of conflict among communal groups, the
ethos of American society treats such behavior as aberrant, largely idiosyn-
cratic individual choices. Communal conflict is seen as a “breakdown” of a
normal condition. We argue the contrary: communal conflict is a manifes-
tation of a systematic tendency for: (1) groups to form social identities
around communal differences, (2) there to be an inverse relationship be-
tween the strength of communal and common identities, and (3) both mate-
rial and symbolic differences between groups to become occasions for con-
flict. The thesis underlying the following analysis is that social diversity
shapes social identities that provide material and symbolic reasons for com-
munal conflict in the United States. Further, the thesis asserts that conflict
is an unavoidable, systematic by-product of social diversity itself. Material
interests that coincide with social identities (especially when they are redis-
tributive) will make communal conflict more common and severe. Institu-
tions can moderate its severity. A civic and social culture can be constructed
that pronounces communal conflict illegitimate, thereby reducing it to a
minimum. But it will always be present because social identities will always
be salient to some fraction of the society, and any material interest that co-
incides with a communal distinction will insert that communal difference
into the conflict because of the sensitivity of people to communal distinc-
tions (Lijphart, 1977; Powell, 1982).
Ethnic, Racial, and National Loyalty
This chapter analyzes ethnic conflict by focusing on certain relationships
between communal identity and national loyalties. We have two “nation-
ality” groups—Hispanic Americans and Asian Americans. The first is asso-
ciated with the countries of Latin America (principally Mexico), even
though many are the progeny of families that have been in the United
States since the early 18th century. Although some Asian Americans also
come from families that are long-term residents of the United States, most
are recent immigrants. Consequently both fit into the conditions that are
102
The Contribution of Ethnic and National Identities to Political Conflict


presumed to create the “divided loyalty” often attributed to identifiable
subnational communities. 
African Americans have a singular history. Even though many can trace
their time in the United States to the 17th century, slavery and segregation
have made them strangers in the land. Africa may not exert a sense of iden-
tity for many, but the separation from the larger society that was forced on
them for almost three centuries created a group experience that is, we be-
lieve, functionally equivalent to being placed on the margins of the society. 
Specifically, this chapter examines the relationship between national loy-
alty and a preference for one’s ethnic and “racial” group. Surprisingly little
empirical research has been directed at this issue. Within the American
context, most thinking and research directed at this question falls into two
categories: the classical “melting pot” model of incorporation and an alter-
native “pluralist” perspective, with the latter giving rise to multicultural 
visions of the social order. 
The rarely examined assumption of these images of American society is
that there is no inherent conflict between subgroup and national identity. A
few attempts to examine the relationship have usually reported a compati-
bility, sometimes an easy one, between these identities (de la Garza, Falcon,
& Garcia, 1996). However, this assumption and the empirical evidence in
favor of it flies in the face of emerging work on social dominance, which
finds incompatibility between subgroup and national social identities. In
general, this work suggests that national identity is generally associated with
communal identity for one group and the exclusion of others (Sidanius,
1993; Sidanius, Feshbach, Levin, & Pratto, 1997; Sidanius & Pratto, 1999).
The purpose of this chapter is to further examine the linkage between com-
munal identity and national loyalties in the United States.
The United States is a hard case because disaffection is limited and disaf-
fected groups are few in number. The extreme group cleavages that assure a
negative association between commitment to common and subgroup iden-
tities are absent in the United States. Moreover, the prevailing civic culture
emphasizes the compatibility of subgroup and common identities and loy-
alty. Yet, as the following data show, such negative relationships actually ex-
ist. The nuances of this relationship and their implications for multiethnic
states are elaborated below.
A Thematic History of American Identity and Ethnic Diversity
The earliest comparative studies of mass politics quickly established that
conflict in even the most well-ordered democratic societies closely followed
communal divisions. Religious, ethnic, racial, linguistic, and regional differ-
ences shaped social and political debate more strongly than most other dis-
tinctions (class, for example; see Lipset & Rokkan, 1967; Horowitz, 1985;
Lijphart, 1979; Petrocik, 1981, 1998). The history of the United States fol-
Communal and National Identity in a Multiethnic State
103


lows this pattern (see, e.g., Kleppner, 1970; Kelley, 1979). Communal con-
flict was fueled by cycles of immigration that transformed the nation from a
population of largely English Protestants (with some dissenter religions) to
one that included continental nationalities, Catholics, and Jews. Making
even the most modest assumption about fertility patterns, it seems likely
that immigrants and first-generation Americans (who could be expected to
be more like their immigrant parents than the children of “old stock”
Americans) represented upwards of a third of the population of the country
by 1900.
American politics was affected by this diversity from the earliest days of
the republic. The Alien Act of 1798, for example, was passed by the declin-
ing Federalists to limit the ability of immigrants to vote in the hope that the
growth of the Democratic-Republicans would be slowed. Their lack of suc-
cess did not forestall many further attempts to limit the impact of immigra-
tion and population diversity on an originally English, Protestant society.
The Irish, and Catholics in general, were the first to redefine what America
could represent; central and north Europeans came next; southern and east-
ern Europeans were the major group for about 60 years in the late 1800s
and early 1900s; immigration from Asia during this same period was impor-
tant in the west. In different ways and at different times, this population di-
versification led to new political parties and legislation to limit the number
of immigrants and their influence on American life. 
The surge in immigration in the late 1980s—the immigration rate was
2.3 per thousand in 1980 but 6.1 per thousand in 1990—reawakened latent
sensitivity to the new “strangers in our midst.” This reawakening has been
especially pronounced in California, where the number of new immigrants
with a non-European “racial” and ethnic heritage is very large. The new
immigrants, like their predecessors, are perceived by the native population
to be substantially different, raising familiar questions about how they fit
into the society, where their loyalties and commitments lie, what changes
should be made to accommodate them, and what costs they will impose on
the society. School boards, zoning boards, city councils, utilities regulators,
local and state elections, and presidential politics have all been forums for
conflicts between Americans and the new immigrants. Some of the disputes
are material, turning on taxes and social welfare costs; others are symbolic
(e.g., Chinese residents of the San Gabriel Valley objected to a switch in the
telephone area code from 818 to 626 because of the numerological signifi-
cance of the numbers for the Chinese). Unresolved problems between
white and black Americans have acquired renewed significance with the 
rising tide of immigration.
Americanization and the Melting Pot
Specific disputes between natives and immigrants have always been con-
flicts over the identity of the national community: about what values, view-
104
The Contribution of Ethnic and National Identities to Political Conflict


points, and interests were compatible, and which could not be abided be-
cause they served and promoted loyalties that were hostile to a common
American identity. It rested on the expectation that different groups em-
braced significantly different values, beliefs, and perceptions. A de facto
policy of “Americanization” was the first way in which the nation dealt with
diversity. Everybody was expected to adopt the language, dress, customs,
beliefs, and loyalties of the dominant English culture that defined the
United States. By the end of the 19th century, the polity replaced “Ameri-
canization” with the “melting pot.” 
The melting pot replaced the expectation of submersion in the preexisting
English Protestant culture with a preference (not necessarily universal) for a
continuous reformulation of what constituted American culture. Immi-
grants, in the melting pot formulation, had valuable things to offer America,
and this value was to be recognized as American society incorporated the lan-
guage, cuisine, social sensibilities, and economic and political values of the
immigrant cultures. This reformulation of American culture, coupled with
marriage across the lines of national origin and religion, became the alterna-
tive resolution of ethnic diversity and the suspicions it produced. It resolved
(at least intellectually) the struggle over whether immigrants were real Amer-
icans by loosening (if not severing) the connection between WASP and
American. The result was a continually changing American culture and a
continually changing American ethnicity still expressing the basic theme
“from many, one” (see e.g., Salins, 1997). Restrictions on European (and es-
pecially German) nationality groups during World War I, immigration lim-
its after 1920, and “Americanization” campaigns (largely tied to learning En-
glish and shedding the most obvious habits and custom of the immigrant
culture) through World War II produced substantial assimilation and largely
eliminated the salience of nationality and ethnic differences by the middle
1960s (Walch, 1994; Freedman, 1996).
The Unmelted Pot of Pluralism
A more widespread recognition of the pluralism of American society is of-
ten attributed to the reemergence of a white ethnic identity in response to
the civil rights movement (Greeley, 1971, 1976; Novak, 1971), but, as Mil-
ton Gordon (1964) demonstrated almost four decades ago, and others con-
firmed (e.g., Wolfinger, 1965; Parenti, 1967), resistance to melting-pot
pressures was strong and seems to be to the present day (Alba, 1990). While
the notion of the melting pot has been the popular metaphor for most of
this century, it co-existed with a “pluralist” image of “salad bowls,” “quilts,”
or “glorious mosaics” to explain how one could be an American while 
retaining a distinct “racial” or ethnic identity.
While there is variation in the precise manner in which the term “plural-
ism” is understood, the central idea is that subgroup identities are retained.
In a pluralist society, communal segments defined by religion, language, na-
Communal and National Identity in a Multiethnic State
105


tional origin, or “race” remain salient to individuals. Catholics recognize
Protestants; the progeny of Italian immigrants see “Yankees” and Polish
Americans; blacks and whites are aware of a “racial” divide; Louisiana Ca-
juns recognize differences between themselves and their Protestant neigh-
bors. While the pluralist image expects social views and policy preferences
to vary by ethnic and “racial” segment, there is no notion that the differ-
ences are unbridgeable because the pluralist political culture expects in
dividuals in the subgroups to feel an identity and loyalty to the national
community, and accept a set of obligations to it. Some formulations of plu-
ralism expect positive supra-communal effects from the communal identity.
Rather than contributing to “disuniting” and fractionation (Schlesinger,
1992), these distinct ethnic loyalties are thought to co-exist and even pro-
mote a common national identity through the communal identity (Powell,
1995; de la Garza et al., 1996).
The Melting Pot, Pluralism, and Conflict 
A melting-pot, immigrant society expects virtually no communal conflict
because the melting pot society has virtually no communal divisions around
which material or symbolic grievances can be politicized. Pluralism, al-
though it has a strong expectation of communal conflict and expects groups
to be sensitive to how social resources (e.g., income, education, employ-
ment opportunities, and social standing) are distributed, assumes a political
framework within which affected groups have different advantages and are
able to bring them to bear in ways that ensure that their core interests are
promoted. (Dahl, 1961, offers the classic exposition; but also see Polsby,
1963, for a significant elaboration of the basic theory that guided virtually
all research on the topic.) Every turn to government may energize inter-
group conflicts because such collective action enforces contributions that
some groups will not prefer to make, especially when the outcome is redis-
tributive. In the textbook model of political pluralism, all groups participate
equally and have a similar positive regard for the basic institutions that
structure the political competition because no group is permanently ex-
cluded or permanent losers to any other groups.
Diversity, Social Dominance, and Communal Conflict
Central to the pluralist argument is the notion that all ethnic groups be re-
garded as co-equal partners in the pursuit of the American dream. However,
in contrast to this vision, there are a cluster of theories of intergroup rela-
tions that suggest that “co-equality” never describes relationships among
salient groups within multiethnic societies. These theoretical models have
been referred to as group dominance models, where social dominance
theory is the most explicit and recent statement of this general position (see
106
The Contribution of Ethnic and National Identities to Political Conflict


Sidanius & Pratt. 1999; for related models see Blumer, 1960; Jackman,
1994). These group dominance models suggest that societies tend to be 
organized as group-based hierarchies, with dominant groups enjoying dis-
proportionate power, prestige, and privilege. This group-based hierarchy
applies to both democratic and nondemocratic states, despite differing dis-
courses concerning inclusivity or pluralism. 
Social dominance theory, as it has been developed by Sidanius and Pratto
and their colleagues (see, e.g., Pratto, Sidanius, Stallworth, & Malle, 1994;
Sidanius, 1993; Sidanius & Pratto, 1999) shares with pluralism an expecta-
tion that communal identities will not be submerged into a larger national
consciousness, but, contrary to pluralism, it expects some groups to be relatively
consistent losers in the struggle. The foundation for this expectation is social
dominance theory’s assertion that social systems are structured as group-
based social hierarchies with one or few groups at the top of the hierarchy,
with all others—sometimes in a carefully ordered fashion—arrayed below
them. The mechanisms and processes that establish and maintain the ex-
pected social hierarchy can be informal—group-based associational pat-
terns that create and limit opportunities—or formal—political and eco-
nomic institutions that promote the interest of dominant segments (see
especially Sidanius & Pratto, 1999). 
In time, institutions can become identified with specific groups rather
than as common institutions of the society, and considerable political con-
flict can turn on perceptions of who is served by them. Examples of this in
the United States are easy to observe. Social welfare agencies are popularly
identified with minority ethnic groups. White Americans often resist finan-
cial pleas from urban school districts because the school systems in big cities
are identified with minority children. And while police forces are staffed
with “officer friendlies” for whites, they are commonly regarded as hostile
and dangerous by African Americans and other minorities (see Sidanius &
Pratto, 1999, chap. 8). 
Exclusionary Patriotism
Not only do such group dominance models have much to say about how
positive and negative social value will be distributed across the group-based
social hierarchy, but these models also have implications for the general in-
terface between national and group identity. Since public institutions are
closely linked to dominant groups, evaluations of them should vary with the
salience of an individual’s identification with the dominant as compared to
the subordinate groups. Specific intergroup differences in evaluations of
political figures, the police, or particular government agencies are well
known. The interest here is in demonstrating a more general relationship
between communal identity and a common national identity. This can have
several manifestations, but our data allow us to examine only one aspect: re-
gard for the nation or patriotism.
Communal and National Identity in a Multiethnic State
107


We expect the dynamics of communal conflict in an environment of 
social dominance to create exclusionary patriotism: a situation in which 
communal and national identities are supportive and positively correlated
among socially dominant groups, but in conflict among those in subordi-
nate communal groups. Specifically, we expect the following: 
1. The prominence of the society’s public institutions in maintaining
communal hierarchies will cause dominant communal groups to be
more identified with the nation’s symbols.
2. We expect a differential relationship between the intensity of com-
munal identities and a common national identity across groups.
Among dominate groups we expect communal and national identi-
ties to be fused and positively correlated; among subordinate
groups we expect them to be at odds and negatively correlated.
3. We expect the negative relationship to be moderated by the experi-
ences of the subordinate group. That is, those further down the
communal hierarchy are likely to see more of a conflict between
their communal identity and national symbols that are associated
with the communal identity of the dominant group. The subjective
social distance among the groups predicts some of the effect. Of
course, a real history of significant conflict between the dominant
and subordinate group would lead to even stronger effects. 
This relationship should be observable even though “exclusionary pa-
triotism” is specifically contrary to the American civic culture. The key
variable is not the avowed norms but the actual operation of the society. 
Because the society functions in a way that allows and even encourages a
perception that the political order is more closely identified with European-
Americans than it is with other major racial and ethnic groups, the dynam-
ics of exclusionary patriotism will shape the relationship between subgroup
identities and attitudes toward American national identity.
Empirical Implications of the Interface Between 
National and Ethnic Attachment
The melting-pot, pluralist, and group dominance models of American soci-
ety lead to a different set of expectations concerning the interface between
national and communal identity. This chapter will explore three expecta-
tions for each model. 
• The Salience of Ethnic Identity. The melting-pot model predicts that
ethnic or “racial” identification will be a trivial social category for
most Americans. Both the pluralist and group dominance perspec-
tives predict that these social identities will remain important and
salient social identities. The salience of ethnicity should express it-
108
The Contribution of Ethnic and National Identities to Political Conflict


self not only in terms of social identity but also in terms of a gener-
alized ethnocentric bias against Americans from other ethnic
groups.
• Patriotism and Ethnicity. Both the melting-pot and pluralist models
predict that patriotic attachment to the United States will be essen-
tially equal across ethnic groups. In contrast, and for reasons ex-
plained above, the group dominance model predicts that dominant
European-Americans will have significantly higher levels of patrio-
tism than will members of subordinate groups (e.g., Asian, Latino,
or African Americans).
• The Interface Between National and Communal Identity. The melting-
pot and pluralist models predict no correlation between patriotism
and communal identity. The melting-pot model predicts no corre-
lation because it expects essentially no ethnocentrism in the society.
The variance in ethnocentrism is trivial and idiosyncratic. The plu-
ralist model of the society predicts no correlation between commu-
nal identity and regard for national institutions because the wins
and losses of one’s ethnic group are not predetermined by an inher-
ently subordinate position to dominant groups who use public in-
stitutions to enforce their preferences. Indeed, the results of de la
Garza et al. (1996) even suggest that ethnic identity can actually
promote a stronger sense of regard for the nation. However, social
dominance theory (Sidanius et al., 1997) expects the correlation 
between national and communal identity to vary across the social
status continuum. It should be positive among dominants (e.g., 
European-Americans) and negative among subordinates (e.g.,
African-Americans). T
able 5.1 summarizes the expected relation-
ships between ethnic and national attachment for each model.
Thus far, we know of only two studies that have attempted to explore the
relative plausibilities of these conflicting models (but see also Citrin et al.,
this volume). Using a nationally representative sample of Mexican Ameri-
cans, de la Garza and his colleagues (1996) found clear empirical support
for the pluralist perspective. They found that Mexican Americans who were
strongly attached to their ethnic heritage were no less patriotic than 
Mexican-Americans with weak ethnic attachments. Furthermore, their data
showed virtually no relationship between the strength of this ethnic identity
and core American values such as individualism. Even more striking was
their finding that Mexican Americans appeared slightly more patriotic than
native-born white Americans. 
A later study by Sidanius and colleagues (1997) found only partial 
support for the de la Garza findings. When comparing the relationships 
between ethnic and national identity between European- and African-
Americans, the results showed clear support for the group dominance per-
spective in two respects. First, European-Americans showed significantly
Communal and National Identity in a Multiethnic State
109


higher levels of patriotism than African-Americans. Second, there was clear
evidence of asymmetry in the correlation between patriotism and commu-
nal identity across dominant and subordinate ethnic groups. Among domi-
nants (i.e., whites), ethnocentrism was positively correlated with patriotism,
while among subordinates (i.e., blacks) ethnocentrism was negatively corre-
lated with patriotism, as the social dominance model expects. While the
dominant-subordinate asymmetry was quite consistent for the contrast be-
tween European- and African Americans; it was less consistent with respect
to the contrast between European and Latino Americans. Moreover, while
the asymmetry held among a sample of European- and Latino-American
college students, it did not hold among a national probability sample of
white and Latino adults (see Sidanius et. al., 1997). In Sidanius et al.’s
(1997) analysis of an American national probability sample, the results were
essentially consistent with the de la Garza et al. (1996) findings.
How might we explain the somewhat conflicting findings between the de
la Garza et al. (1996) and the Sidanius et al. (1997) studies? First and most
obviously, the de la Garza et al. study only compared whites and Latinos,
whereas the Sidanius et al. study also included blacks. Second, even the
slightly different findings with respect to Latinos might have something to
do with major differences in political context. In recent years, immigration
from Mexico and Central America has become a salient and contentious po-
litical issue in California, where the Sidanius et al (1997) data were col-
lected. It is quite possible that Southern California’s highly racialized public
discourse concerning illegal Latino immigration (e.g., Proposition 187) in-
duced Latinos to feel under attack and as political outsiders as African
Americans have felt for the bulk of American history. Finally, we think that
their analysis does not measure their key conceptual variable very well. The
acculturation variable, the centerpiece of their analysis, is only a weak ap-
110
The Contribution of Ethnic and National Identities to Political Conflict
table 5.1 Hypotheses Derivable from the Melting-Pot, Pluralism, and Group
Dominance Models
Theoretical Model
Hypotheses
Melting-pot
Pluralism
Group Dominance
1. Ethnicity remains strong 
No
Yes
Yes
social identity
2. Dominants will be more 
No
No
Yes
patriotic than subordinates
3. There will be a correlation 
No
No, or positive 
Yes. Positive 
between ethnocentrism and 
correlation
correlation among 
patriotic attachment to
dominant, negative 
the nation
correlation among 
subordinates


proximation of subgroup attachment. Their measure categorized individu-
als by place of birth and language facility: U.S.-born Mexican Americans
who were English speakers were regarded as the most acculturated while
foreign-born Spanish speakers were classed as the least acculturated. Accul-
turation, however, is not the relevant variable. Although there is probably a
relationship between the individual’s level of acculturation and the strength
of his or her preference for Mexicans over non-Mexican-Americans, there
is no reason to assume that it is strong. It may, in fact, be extremely weak.
Among European Americans (the Irish, Italians, Poles, etc.), identity with
national origin is poorly related to rates of endogamy, the observance of
ethnic customs, and facility with the immigrant language (Alba, 1990). A
priori, there is no compelling reason to expect Mexican Americans to be 
different. 
T
o be sure, the null results also appear with the one approximate mea-
sure of identity used in the analysis. Respondents classed as Latinos were
those who, when asked their “racial” identity, chose the option “something
else” rather than white. This group, slightly less than half of their sample,
clearly thought of themselves in ethnic terms. In our experience, those who
self-describe as Latinos are more ethnically identified than those who do
not use this appellation—and this subgroup was not less patriotic or less in-
dividualistic in their study. However, an unwillingness of the respondent to
choose the category “white” as their ethnic or racial group is a weak mea-
sure of identity, and the null results probably should not be surprising.
Consider that it is fairly common for white Americans to respond with a na-
tional origin distinction when asked their nationality, but there is tremen-
dous variability in the strength of attachment to the national origin of
Americans who call themselves Italian, Polish, or Lithuanian when report-
ing their nationality. This variance almost certainly exists among Mexican
Americans. It is at least reasonable to suspect that it does, and the weak rela-
tionship de la Garza and his colleagues report for the measure may only re-
flect the indeterminate property of the categorization.
The finding that Mexican Americans are more patriotic and individualis-
tic than native-born whites is even less persuasive. As de la Garza and his
colleagues note, the whites with whom they draw the contrast were avail-
able for the analysis only because they lived in areas with large Mexican-
American populations. These whites are almost certainly not representative
of the white population as a whole, given the patterns of residential segre-
gation that exist in the United States. We would expect such whites to be
poorer and less educated. Their social class is likely to orient them toward a
preference for government action (the exact opposite of the “individualism”
expected of the median American in the de la Garza et. al. formulation).1
The social values that permit these whites to live comfortably in Mexican-
American neighborhoods may also have negative implications for expres-
sions of patriotism, depressing, or even reversing, therefore, the expected
relationship between it and nativity and ethnicity.
Communal and National Identity in a Multiethnic State
111


Method and Design
We used five different data sets in these analyses and a total of 7,400 
respondents. 
Data set 1: UCLA Student Sample
This data set consisted of 725 randomly sampled UCLA undergraduates
collected in the fall of 1993. The sampling frame stratified a list of all regis-
tered students into four “ethnic” strata (whites, blacks, Latinos, and Asians),
and randomly sampled from each stratum. The subjects were enticed to
participate by the offer of four $50 prizes. The ethnic breakdown of all of
the samples is found in table 5.2. The students were given a questionnaire
primarily assessing their attitudes toward patriotism and their affect toward
various American ethnic groups. However, unlike the Sidanius et. al. (1997)
study, in this chapter we will explore the responses of resident-aliens as well.
Patriotism Measures.
Kosterman and Feshbach (1989) and Sidanius et al.
(1997) have shown that patriotism could be well measured by an eight-item
scale consisting of items such as: “I am proud to be an American,” and
“Every time I hear the national anthem, I feel strongly moved.” Not only
did this scale have a high degree of face validity, but it also had a fairly high
level of reliability (with a Cronbach a = .91). Because the patriotism mea-
sures were assessed using different scales across the different studies, in or-
der to aid in cross-study comparisons, we transformed all of the patriotism
measures to the same 0.00 to 1.00 scale, where 0 represents the lowest pos-
sible score a respondent can get and 1.00 represents the highest possible
score a respondent can get. 
Ethnocentrism. While communal identity can be regarded as expressing the
general salience of one’s “racial,” ethnic, or religious identity, we will use
the term “ethnocentrism” to refer to preference for communal ingroups over
communal outgroups. The ethnocentrism measures were operationalized
by use of thermometer rating scales. The respondents were asked to indi-
cate their degree of positive or negative feeling toward each of four ethnic
groups using a seven-point rating scale (7–very positive, 4–neutral, 1–very
negative). The groups were whites, African Americans, Latinos, and Asians.
T
wo different types of ethnocentrism scores were computed: (1) general-
ized ethnocentrism and (2) hierarchical ethnocentrism. Generalized ethno-
centrism (GEI) was used to measure the respondents’ degree of ethnocen-
trism against the “generalized other,” regardless of whether this generalized
other had relatively high or low ethnic status. GEI was computed simply by
taking the thermometer rating given to one’s ethnic ingroup minus the
thermometer rating given to the average ethnic outgroup. Hierarchical eth-
nocentrism (HEI), on the other hand, was designed to be sensitive to the 
112
The Contribution of Ethnic and National Identities to Political Conflict


hierarchical nature of the status differences among the ethnic groups. Thus,
HEI was measured slightly differently, depending upon whether one was a
member of a dominant or subordinate ethnic group. For whites, hierarchi-
cal ethnocentrism (HEI) was defined in precisely the same way as for gener-
alized ethnocentrism (GEI). However, consistent with the notion of sub-
ordination to a dominant social group, HEI for the each of the three
subordinates groups was defined as the positive rating given to one’s ethnic
ingroup minus the rating given to higher-status whites. 
Data Set 2: Los Angeles County Social Survey (LACSS)
These data are drawn from two years (1997 and 1998) of the Los Angeles
County Social Survey (LACSS). The LACSS is a countywide random-digit
dial telephone survey of adults living in households conducted by the Com-
puter Assisted T
elephone Interviewing unit of UCLA’s Survey Research
Center. The LACSS employs a 12 call-back procedure, systematically vary-
ing the day of the week and the time of day, before dropping any numbers
from the sample. A total of 1,232 respondents were interviewed (see table
5.2).
Patriotism Measures. T
wo items were used to operationalize patriotism: (1)
“I find the sight of the American flag very moving,” and (2) “I have great
love for the United States.” The questions were answered on a four-point
scale from “4–strongly agree” to “1–strongly disagree.” The reliability of
the scale was considered adequate (Cronbach a = .68).
Ethnocentrism.
Generalized and hierarchical ethnocentrism were defined
by use of affect differentials in the same manner as used in Data set 1, the
UCLA student sample. 
Communal and National Identity in a Multiethnic State
113
table 5.2 Data Sets and Ethnic Groups
Samples
Data Set 1
Data Set 2
Data Set 3
Data Set 4
Data Set 5
Ethnic
UCLA Students
LACSS
NES
NES
GSS
Group
1993
1997–98
1992
1996
1996
T
otal
Whites
153
539
1,905
1,325
1,015
4,937
Blacks
113
139
312
191
170
925
Latinos
98
221
207
149
54
729
Asians
72
68
28
22
10
200
Jews
32
—
—
—
—
32
Immigrants
257
265
—
—
55
577
T
otal
725
1,232
2,452
1,687
1,304
7,400


Ethnic Grievances. Another series of questions assessed respondents’ “sense
of ethnic grievance”—that is, the extent to which their ethnic or racial group
has been disadvantaged by those in power. These questions assessed whether
respondents: (1) believed that their group should have more influence in gov-
ernment; (2) believed that their group was often held back by other people;
(3) believed that the government helps particular groups and does not pay
enough attention to people in general; (4) believed that public officials from
their own group would be more likely to do things respondents approved of;
(5) would not vote for someone who was not likely to help their group; (6)
oppose tax increases that help groups other than their own; and (7) would
vote for a candidate from their own group, even if that candidate did not
agree with them on every issue. All questions except the first two used an
agree or disagree format (which also included options for don’t know and re-
fused, and neither agree or disagree if this was volunteered by the respon-
dent). The response options for the first question were should have more in-
fluence, less influence, or just about right (as well as don’t know and refused).
The response options for the second question were often held back, some-
times held back, and treated as well as anybody else (as well as don’t know and
refused). For this question, both held back and sometimes held back were
coded as “agreement” (i.e., agreeing that their group has been held back).
Data Set 3: National Election Study—1992 (NES)
The third data set was a national probability sample of Americans collected
as part of the 1992 National Election Study (NES) conducted by the Cen-
ter for Political Studies at the Institute for Social Research at the University
of Michigan. Altogether, the 1992 study interviewed 2,452 respondents and
the analysis is limited to the four major ethnic groups that are the focus of
this study (i.e., whites, African Americans, Latinos, and Asians; see table 5.2
for details.
Patriotism. Patriotism was defined by two questions that had a content that
was virtually identical to the patriotism questions used in data set 2: (1)
“When you see the American flag flying does it make you feel extremely
good, very good, somewhat good, or not very good?” (2) “How strong is
your love for your country? Extremely strong, very strong, somewhat
strong or not very strong?” (Cronbach a = .78). 
Ethnocentrism. Ethnocentrism was defined in precisely the same way in data
sets 2 and 3.
Data Set 4: National Election Study—1996 (NES)
The fourth data set was the 1996 National Election Study (NES), also con-
ducted by the Center for Political Studies at the Institute for Social Re-
114
The Contribution of Ethnic and National Identities to Political Conflict


search at the University of Michigan (N = 1,687; see table 5.2). This data
set did not contain measures of patriotism, but it did contain feeling ther-
mometers that were used to compute ethnocentrism in the same way as in
data sets 2 and 3. The 1996 NES permitted a replication of the communal
identity patterns observed in the 1992 study, and a check against a one-time
pattern.
Dataset 5: General Social Survey—1996 (GSS)
This data set (N = 1,304 ) was used to replicate earlier findings and examine
dimensions of national identity beyond what was possible with the patrio-
tism index used in the 1992 NES study (data set 3) and the LACSS samples
(data set 2). While the dimension of patriotism, or love of the nation, is a
central dimension of national attachment, it is clearly only one of several
different possible ways in which citizens can be psychologically attached to
the nation. Previous research (Kosterman & Feshbach, 1989; Sidanius et
al., 1997; Hofstetter, Feierabend, & Klicperova-Baker, 1999) has shown
that there are at least two important dimensions of national attachment: pa-
triotism (love of country and its symbols) and nationalism (the desire that
one’s country dominate other countries). Previous research also suggests
that while dominants are significantly more patriotic than subordinates,
they are not more nationalistic than subordinates (see Sidanius et al., 1997). 
The 1996 GSS permitted a further exploration (as well as replication) of
the communal-national trade-off because it contained 34 questions con-
cerning Americans’ feelings of national attachment. While the 34 national
attachment items formed a reasonable coherent and homogeneous scale
(Cronbach a = .84), examination of the items revealed distinctly different
types of national attachment within this scale. T
o further explore this issue
and before any further analyses were attempted, we first factor-analyzed
these items using an alpha factor extraction approach. While there were as
many as 10 factors with eigen values greater than 1.00, the scree test sug-
gested that only four of these factors were meaningful. Altogether these
four factors accounted for approximately 37 percent of the total variance.
The factors were interpreted as: 
1. Patriotism or pride in one’s nation (51 percent of the common vari-
ance) was primarily defined by nine items such as: “Proud of the
way democracy works in America,” “Proud of America’s economic
achievements,” “Proud of America’s political influence in the
world,” Proud of America’s scientific and technical achievements.”
2. Nationalism (20 percent of common variance) essentially embraced
a chauvinistic and belligerent attitude toward other nations and
peoples (see also Bar-tal & Staub, 1997). The items most strongly
defining this dimension were: “TV should give preference to
American films,” “Foreigners should not be allowed to buy land,”
Communal and National Identity in a Multiethnic State
115


“America should follow its own interests,” “People should support
their country even when the country is wrong,” and “People would
be better off if they were more like Americans.”
3. Unwillingness to leave America (16 percent of common variance) pri-
marily expressed people’s unwillingness to move away from
America (e.g., “Willing to move outside of America,” “Willing to
move outside of North America”).
4. Importance of American identity (13 percent of common variance), as
distinct from patriotism or pride in one’s nation, expressed people’s
sense of social identity as being American. The dimension was pri-
marily defined by items such as: “It is important to have lived in
America for life,” “It is important to have been born in America,”
“It is important to have American citizenship.”
Factor scores were computed for each respondent on each factor. Not
surprisingly, these factor scores were positively correlated, with the
strongest correlation found between “Patriotism” and “American identity”
(r = .42), and the weakest between “nationalism” and “unwilling to leave
America.” ( r = .10).
The 1996 GSS included one question that we used to assess the salience
of ethnic identification. This question read “Compared to people who are
not (your ethnicity), would you say that you feel much closer to people who
are (your ethnicity), somewhat closer to them, or not closer at all?” The
term “your ethnicity” did not actually appear—it was always replaced by the
respondent’s own ethnicity (e.g., if they said they were Asian, it read
“Asian”; if they said they were “Latino” etc.).
Results
The alternative models of American society posit ethnic differences (or in
the case of the melting-pot model, no ethnic differences) in: the salience 
of ethnic identity, group levels of patriotism, and the relationship between
patriotism and ethnocentrism.
Issue 1: The Continued Salience of Ethnic Identity
Within the melting-pot model, ethnic identification should not be a salient
social identity. We explored the validity of this assumption by examining: 
(1) the degree to which American citizens also think of themselves in
“racial”/ethnic terms; (2) ethnocentric bias against other American ethnic
groups, and (3) people’s political grievances as a function of their own 
ethnic/racial ingroups. First T
able 5.3 shows ethnicity to be a salient social
identity for most Americans. For example, some 76 percent of European-
Americans in the nation as a whole feel close (34 percent are very close) to
116
The Contribution of Ethnic and National Identities to Political Conflict


other whites. Among, African, Latino and Asian Americans, this percentage
is even greater at 89 percent, 81 percent, and 84 percent, respectively. The
same general pattern was found for Americans in Los Angeles County. How-
ever, not surprisingly and in both cases, the salience of ethnicity varied sig-
nificantly as a function of ethnic group membership. In both cases ethnicity
was more salient for the ethnic minorities than for European Americans (na-
tional data set: X2(6) = 81.71, p < .001; LA dataset X2(6) = 61.75, p < .001).
Second, using the Los Angeles and national probability samples, we ex-
amined the generalized ethnocentrism scores of each ethnic group. If one’s
ethnic/racial particularism is no longer a salient feature of people’s social
identity, as a strict reading of the melting-pot model would expect, then
Americans from one particular ethnic or “racial” group should not exhibit
ethnocentric bias against other American “racial”/ethnic groups. However,
as the data in table 5.4 show, all four major American ethnic/racial groups
display an ethnocentric bias against other groups. While European Ameri-
cans in Los Angeles appeared to be slightly less ethnocentric than members
of the other major ethnic groups (the thermometer difference was approxi-
mately 3 degrees compared to differences of 8, 10 , and 8 degrees among
blacks, Hispanics, and Asians, respectively), the level of general ethnocen-
tric bias among European Americans in the nation as a whole appeared to
be no more or less extreme than among other major ethnic groups in the
nation as a whole (i.e., an 11 degree average difference compared to an av-
erage difference of 17 degrees for African Americans, 10 degrees for Lati-
nos, and 8 degrees for Asians). This very broad and generalized level of eth-
nocentric bias is consistent with the thrust of social identity theory (T
ajfel &
Turner, 1986).
Finally, ethnic group membership also has political relevance and politi-
cal overtones. Using the 1997 LACSS, we asked respondents from each of
the major ethnic groups a series of questions concerning whether they felt
Communal and National Identity in a Multiethnic State
117
table 5.3 Feelings of Closeness to One’s Own Ethnic/Racial Group (Citizens
Only)
Whites
Blacks
Hispanics
Asians
USA
LA
USA
LA
USA
LA
USA
LA
Respondent feels:
Very close
34%
27%
70%
58%
47%
43%
46%
37%
Some closeness
42
52
19
39
34
40
38
41
Not close
24
21
11
10
19
14
17
22
T
otal
100% 100%
100% 100%
100% 100%
100% 100%
Number of cases
975
559
166
140
343
432
14
95
Source: The U.S. national data are drawn from the 1996 General Social Survey. The Los Angeles data are
from the 1997 Los Angeles County Social Survey.


that members of their own ethnic group were being fairly treated by the po-
litical system, whether public officials from their own ethnic group would
treat them more fairly, and whether they would vote for a member of their
own group, even if they disagreed with the respondent (see table 5.5). For
example, while members of the ethnic minorities are significantly more
likely than whites to feel that their own ethnic group needs more influence
in the political system (X2(3) = 15.82, p < .01), a majority of all groups feel
that their ethnic ingroups need more political influence. Similarly, a ma-
jority of all ethnic groups believe that some groups get too much help and
not enough attention is paid to what is good for all people. The only issue
where one’s ethnic identity does not seem to play a major role concerns re-
spondents’ willingness to vote for a member of one’s own ethnic group,
even when the respondent disagrees with the candidate. T
able 5.5 shows
that a distinct minority of respondents is willing to vote strictly on the basis
of ethnicity, varying from 23 percent for blacks to 6 percent for Asians.
Altogether, ethnicity remains an important and salient social identity
with clear political overtones. In general, the data give little support to the
melting-pot image, at least to the extent that the melting pot assumes that
Americans are blended and substantially unaware of their “racial” and eth-
nic differences. T
able 5.5 also makes clear that the pluralist notion of con-
tented collaboration among competing groups is at least contested. Ethnic
minorities are particularly likely to believe that they suffer at the hands of
others because of their communal identity and that members of their group
are likely to be more sympathetic and protective.
118
The Contribution of Ethnic and National Identities to Political Conflict
table 5.4 Thermometer Differences Among Ethnic/Racial Groups (Citizens
Only)
Whites
Blacks
Hispanics
Asians
USA
LA
USA
LA
USA
LA
USA
LA
Rating of:
Whites
71
81
73
79
71
75
67
74
Blacks
62
77
85
87
69
65
63
66
Hispanics
60
78
70
79
78
79
59
66
Asians
58
79
61
78
63
66
71
77
Own group
71
81
85
87
78
79
71
77
Average of
other groups
60
78
68
79
68
69
63
69
Difference
+11
+3
+17
+8
+10
+10
+8
+8
Number of cases
3151
559
495
140
343
432
61
95
Source: The U.S. national data are drawn from the 1992 and 1996 National Election Study surveys of those
years. The Los Angeles data are from the 1997 and 1998 Los Angeles County Social Survey.


Issue 2: Ethnicity and Patriotism
Our second question concerns whether the degree of patriotic attachment
to the nation is the same across all major ethnic groups. While both the
melting-pot and pluralist models would expect no differences among the
groups, the group dominance model of American society expects consider-
able differences. Dominant European Americans should have significantly
higher levels of patriotism than Hispanics, Asians, and African Americans.
In addition, since African Americans have long been, and remain, the group
at the very bottom of America’s ethnic hierarchy, the group dominance ap-
proach would expect the mean level of patriotism among African Americans
to be relatively low and no greater than that found among resident-aliens,
or people barred from participation in the political system. T
o explore these
questions, we compared patriotism as a function of ethnicity across four of
the five data sets (UCLA students, Los Angeles County residents, and two
representative samples from the nation as a whole).
The results in table 5.6 are inconsistent with both the melting-pot and
the pluralism models. Contradicting the melting-pot and the pluralist per-
spectives on America, patriotic attachment is not uniform across the ethnic
status hierarchy. However, the differences are not substantially large. More-
over, the absolute scores indicate that all groups have relatively high patrio-
tism scores. The bulk of the intergroup difference, reflected in the correla-
tion ratios (varying between .14 and .35), is due to the relatively low
Communal and National Identity in a Multiethnic State
119
table 5.5 Sense of Grievance Among Those Identifying with an Ethnic/Racial
Group (Citizens Only)
White
Black
Hispanic
Asian
Percent agreeing:
Believe that ones group needs more 
influence
58
86
71
67
Believe that he/she has been “held back” 
because of their ethnic/racial identity
27
77
88
67
Believe that some groups are helped too 
much and not enough attention is paid 
to what is good for all people
54
55
59
67
Believe that public officials from own 
group are more likely to do preferred 
things
30
41
59
56
Oppose programs that help some group 
but not mine
47
46
44
33
Would not vote for a candidate if he/she 
would not help people like respondent
55
73
61
28
Would vote for a candidate from my group 
even if I disagreed about some things
21
23
17
6
Number of cases
151
53
162
46
Source: The 1997 Los Angeles County Social Survey.


patriotism scores of the African Americans. But even their scores are not
particularly low, except in the UCLA student data set. In an absolute sense,
then, no group is substantially alienated from the American polity, but there
is a clear tendency for patriotism to be highest among European Americans
and unequivocally lower among America’s most subordinate group, African
Americans.2 
These differences survive our tests for the confounding effects of other
social attitudes. The test used demographic information such as gender, age
and socioeconomic status as proxies for confounding effects that might be
correlated to a general sense of disaffection resulting from material disad-
vantages tied to social status. The analysis was done with multiple regres-
sion that controlled for gender, age, and SES while testing for the partial 
effect of a binary coded ethnic group membership variable. In all of the re-
gression analyses, European-American ethnicity was treated as the contrast
category. Use of this coding scheme had the distinct advantage of providing
a direct test of the patriotism differences between the dominant ethnic
120
The Contribution of Ethnic and National Identities to Political Conflict
table 5.6 Patriotic Attachment to America as a Function of Ethnicity Across Four
Samples.
Ethnic Group
Resident Alien in US
More
than 5
Less than
Whites
Asians
Latinos
Jews
Blacks
yrs
5 yrs
Sample 1 (UCLA Students, 1993): F(6,703) = 16.64, p<.001, h = .35
Mean patriotism
.693
.628
.617
.555
.432
.585
.477
N
153
72
98
32
106
219
30
SD
.23
.20
.23
.23
.24
.20
.22
Sample 2 (L.A. County, 1997–98): F(4,1203) = 9.53, p<.001, h = .18
Mean patriotism
.849
.806
.818
—
.757
.763*
—
N
530
66
217
—
138
257
—
SD
.20
.19
.22
—
.25
.23
—
Sample 3 (NES-1992): F(3,2206) = 66.47, p<10–10, h = .29
Mean patriotism
.794
.712
.789
—
.589
—
—
N
1,736
26
167
—
281
—
—
SD
.22
.22
.22
—
.29
—
—
Sample 5 (GSS-1996): F(3,1204) = 8.72, p<.01, h = .14
Mean patriotism
.531
—
.505
—
.488
.480*
—
N
973
—
48
—
38
49
—
SD
.11
—
.13
—
.13
.16
—
*Indicates that length of residency for this alien group is unknown.


group (i.e., whites) and each of the subordinate ethnic groups, net of the 
effects of all other factors.
The results of these analyses for the four different data sets are found in
table 5.7. The table reports only the coefficients for the ethnic group con-
trasts (i.e., white vs. ethnic minority). The coefficients for gender, age, and
various SES measures are not individually reported because they are only
proxies for unexamined control variables that have no theoretical impor-
tance.3 The tables show a moderate degree of consistency across data sets:
after considering the effects of the major demographic factors, the patrio-
tism differences among ethnic groups in table 5.6 persist in table 5.7. This
was especially true among the college students. Without exception, the
contrasts between socially dominant whites and all of the subordinates
groups showed whites to have significantly higher patriotism scores, with,
as before, the most powerful contrast being blacks versus whites (t = –8.60, p
< .001, b = –.39). The contrasts between whites on the one hand, and Asian
and Latino Americans on the other hand, tended not to be significant
within the other general population samples. 
African Americans, though exhibiting relatively high absolute levels of pa-
triotic attachment and attaching relatively high importance to their Ameri-
can identity (see figure 5.1), express less patriotic sentiment than any of the
other ethnic groups, and they are more likely to exhibit the exclusionary con-
flict between their communal identity and national sentiment. Indeed,
African Americans tended to be no more patriotic than resident-aliens (see
table 5.6). In the student data set, the patriotism scores among African Amer-
icans were significantly lower than those found among resident-aliens in the
United States for more than five years (Scheffé post-hoc, p < .001). The effect
is apparent even after controlling for demographic differences between
groups: European Americans have a significantly greater level of patriotic 
attachment than resident-aliens (data set 1: t = –4.79, p < .001; t = –4.54, 
p < .001; dataset 2: t = –2.02 , p < .05; dataset 5: t = –1.83, p < .05). African-
Communal and National Identity in a Multiethnic State
121
table 5.7 Patriotism and Ethnocentrism Among Dominant and Subordinate-
Groups
European- vs. 
European- vs. 
European- vs. 
African-Americans
Hispanic-Americans
Asian-Americans
B
b
B
b
B
b
UCLA
–.25**
–.39**
–.06*
–.10*
–.08*
–.10*
LACSS 1997/1998
–.07**
–.11**
.01
.01
–.01
–.01
NES 1992
–.22**
–.30**
–.00
–.00
–.00
–.02
GSS 1996
–.04**
–.10**
.00
.00
NA
NA
Note: Numbers under the columns labeled “B” are standardized coefficients from a regression that controls
for gender, age, and socio-economic status. Numbers in the “b” columns are the unstandardized coefficients.
**p<.05; *p = .05


Americans, controlling for demographic differences, are never significantly
more patriotic than resident-aliens. Among university students, African
Americans had significantly lower patriotism scores than aliens residing in
the United States five years or longer (data set 1: –t = 5.20, p < .001).
Issue 3: Patriotism and Communal Identity
Both the melting-pot and pluralist models of American society expect eth-
nocentrism and national attachment to be uncorrelated. This expectation is
not met in these data. Feelings about the nation are related to communal
identity. As one would infer from social dominance theory, this relationship
is asymmetrical across the social status hierarchy. Ethnocentrism and na-
tional assessments are positively correlated for whites but negatively related
for others, especially African Americans.
The following analysis regressed patriotism on the hierarchical ethnocen-
trism (HEI) index within each major ethnic group within the data sets
where we had appropriate measures (data sets 1, 2, and 3). The results were
reasonably consistent across all three data sets, but were, again, most con-
sistent among the university students (see table 5.8). For example, the more
white students were ethnocentric against African Americans, the more pa-
triotic they were (b = .37, p < .05; see table 5.8). The same general tendency
122
The Contribution of Ethnic and National Identities to Political Conflict
figure 5.1 Importance of American Identity as a Function of Ethnicity
Whites
Latinos
Resident-aliens
Blacks
Ethnicity
Importance of American Identity


held with respect to whites’ HEI against both Latinos and Asians (i.e., b =
.44, p < .01, and b = .41, p < .05, respectively). Among subordinates, on the
other hand, the exact opposite trend was found. Without exception, ethno-
centrism among subordinates was negatively related to patriotism. Thus,
while the regression of patriotism upon ethnocentrism (HEI) among whites
against Latinos was significantly positive (i.e., b = .44, p <. 01), the corre-
sponding regression of patriotism on HEI against whites among Latinos
was negative (i.e., b = -.49, p <.01). T
o test whether or not these asymmetries
in the relationships between ethnocentrism and patriotism were statistically
significant across the status continuum, we used simple slopes analyses to
test for interaction between ethnic group and ethnocentrism (see Aiken &
West, 1993). As can be seen for the student dataset in table 5.8, these slope
differences were significant in all cases (i.e., F = 19.63, p < .001, F = 20.80, p
< .001, F = 17.78, p < .001 for the white vs. black, white vs. Latino, and
white vs. Asian contrasts, respectively).
The intergroup asymmetries were substantially weaker in the general
population but generally consistent with expectations. Only two results are
patently contrary to the exclusionary patriotism hypothesis: in the Los An-
Communal and National Identity in a Multiethnic State
123
table 5.8 Patriotism Regressed upon Hierarchical Ethnocentrism Within Domi-
nant and Subordinate Groups (Entries are unstandardized regression coefficients.)
Ethnocentrism Index among 
Ethnocentrism Index among
Interaction
Dominant Group (i.e., Whites)
Subordinate Groups
Statistics
Among
Among
Among
Among Whites
Blacks
Latinos
Asians
F-statistics
White-
White-
White-
Black-
Latino-
Asian-
for Slope
Black
Latino
Asian
Whites
White
White
Differences
UCLA Sample
.37*
–.47***
19.63**
.44**
–.49***
20.80**
.41*
–.51*
17.78**
Los Angeles Sample 1997–98
–.00n.s.
–.09n.s.
0.20n.s.
–.00n.s.
–.19*
2.05n.s.
–.20**
–.10n.s.
4.94*
NES 1992 Sample
.15**
–.21**
30.09***
.16**
–.12n.s.
8.64**
.16**
–.19n.s.
0.73n.s.
*p<.05, **p<.01 ***p<.001


geles sample white ethnocentrism toward blacks and Hispanics is unrelated
to patriotism. Every other coefficient is properly signed. The insignificant
difference between whites and Asians in the 1992 NES is a product of a
small number of cases; however, the numerical value of the differences is
substantial. Altogether, six of the nine interactions showed that the slope of
the relation between ethnocentrism and patriotism differed between domi-
nants and subordinates. For whites, seven of nine correlations between eth-
nocentrism and patriotism were positive and significant. For the other eth-
nic groups, five of nine correlations between ethnocentrism and patriotism
were negative and significant (see table 5.8).
Summary and Discussion
This paper offered three basic ways of thinking about the linkage between
communal identity and national attachment in American society: (1) the
traditional American melting-pot approach, (2) the pluralist model, and (3)
the group dominance approach. In an effort to assess the relative validity 
of these different ways of thinking about the intersection of national and
ethnic identity, we explored three issues (see table 5.1). These issues con-
cerned: (1) the salience of ethnic identity and the presence of ethnocentrism
against other groups of Americans, (2) the relationship between patriotic at-
tachment to the nation and ethnic group membership, and (3) the nature of
the interface between communal identity and patriotic attachment.
The results are mixed, but on balance more supportive of expectations
proffered by social dominance theory. Ethnic subgroup identities remain a
salient social identity for each of the major groups we examined and, consis-
tent with the expectations of social identity theory (T
ajfel & Turner, 1986),
average Americans display clear ethnocentric bias against other ethnic
groups of American citizens. Further, and inconsistent with both the melt-
ing-pot and pluralism perspectives, patriotic attachment to the nation was
generally not uniform across ethnic groups. European Americans have
stronger patriotic attachment than those from subordinate communal
groups. This tendency was particularly strong and consistent with respect
to the white/black contrast. There was not a single instance in which this
white/black contrast failed to be statistically reliable.
Finally, the pluralism image of the society was contradicted by the corre-
lation between ethnic and national attachment across ethnic groups. Within
the pluralism model of American society, communal attachments should
have no implication for national attachment. That is not the case, especially
for black Americans. Here, the data were clearly more consistent with the
group dominance thesis than with the pluralism perspective. The clear
asymmetry in the relationship between communal identity and patriotism
across the social status continuum can be regarded as a special case of ideo-
124
The Contribution of Ethnic and National Identities to Political Conflict


logical asymmetry discussed by social dominance theory and demonstrated in
a number of studies (Fang, Sidanius, & Pratto, 1998; Levin & Sidanius,
1999; Levin, Sidanius, Rabinowitz, & Federico, 1998; Sidanius, Pratto, &
Rabinowitz, 1994; Sidanius & Pratto, 1999).4 The robustness of these basic
asymmetrical findings is supported not only by their consistency with work
generated by social dominance research but also by recent work outside of
this tradition (Hofstetter et. al, 1999). 
The distinctive relationships observed for black Americans raises the
likelihood that the degree to which communal and national identity are
connected depends upon contextual and situational factors that cannot be
systematically considered in this chapter. For example, the degree to which
the group dominance perspective accurately describes intergroup status re-
lationships may well be the critical variable that links communal and extra-
communal identities. African Americans, almost uniquely, have a history of
subordination that makes it easy and reasonable to associate the institutions
of the nation-state (and sentiments such as traditional patriotism) with
abuse and oppression. In such an environment it is simple to understand
why they, relative to Hispanics, Asians, or any other group of immigrants,
are inclined to see a conflict between their communal identity and their
sentiments about the nation. Many other groups, including European im-
migrants in their first decades in this country, also suffered (Freedman,
1996), but the severity of the oppression of American blacks is relatively
unique. 
The issue that policymakers and interpreters of the American social or-
der need to address is whether the patterns characteristic of white/black
communal and national sentiments can be approximated symbolically—in
the absence of the distinctive history of blacks. We suspect that it can. Peter
Skerry’s (1993) analysis of Mexican Americans in San Antonio and Los An-
geles offers affirmative evidence on this point. He persuasively argues that
the ethnic politics practiced by Mexican Americans in San Antonio replicate
the pluralist experiences of European immigrants, with the result that San
Antonio Mexican Americans participate in and feel committed to the insti-
tutions of the society. In Los Angeles, by contrast, Skerry found an ethos
that defined Mexican Americans as oppressed, a mind-set that created an
insurgent political style, and a dramatically less supportive view of the soci-
ety’s institutions. In brief, in objectively similar conditions, an ideologically
shaped political ethos in Los Angeles created among L.A. Mexican Ameri-
cans a subordinate group perspective that established the conditions for ex-
clusionary patriotism. We suspect, although the data to demonstrate it are
lacking, that the strong asymmetry in the UCLA student data set is a reflec-
tion of the “oppressed minority” ethos of Hispanic and Asian politics in the
region. This speculation suggests that future research in this area should ex-
plore the material and perceptual situational and contextual factors associ-
ated with the linkage of communal and national identity.
Communal and National Identity in a Multiethnic State
125


Notes
1. We have doubts about the appropriateness of measuring individualism with
the questions used in their study. They are, more appropriately, viewed as mea-
sures of some attitude about the government’s responsibility for the social welfare
of the population. This is almost certainly related to an individualism-communi-
tarian attitude, but it is a related orientation, not individualism itself. The failure
of the three questions to form a reliable scale (de la Garza, 1996, p. 342) is a fur-
ther indicator of the weakness of the questions as indicators of some underlying
social orientation.
2. However, as mentioned before, unlike the other data sets examined, the 1996
GSS data set contained a much richer array of national attachment questions. These
additional questions afforded us the opportunity to examine the connection between
ethnic and national attachment across additional dimensions of national attachment.
Thus, besides the dimension of patriotism, we were also able to identify three addi-
tional dimensions of national attachment: (1) nationalism—i.e., an aggressive and
dominance-oriented stance toward other nations; (2) unwillingness to leave America,
and (3) importance of American identity. Most interestingly, the ethnic differences with
regard to these three additional dimensions of national attachment were very differ-
ent from those found for patriotism. First, there were no significant ethnic groups
differences with respect to nationalism. Second, while there was an overall difference
with respect to Unwillingness to leave America (i.e., F(3,1249) = 10.68, p < .001, h =
.16), inspection of Scheffé post-hoc comparisons showed that this difference was es-
sentially due to the fact that resident-aliens were significantly less opposed to leav-
ing the United States than all other American groups—not surprisingly.
The one national attachment dimension where the group differences were quite
different from previous results was importance of American identity. Not only did the
different ethnic groups show significant overall differences here (F(3,1290) = 13.68, p
< .001; h = .18), but the nature of these differences were qualitatively different from
those found with respect to patriotism (see figure 5.1). As seen in figure 5.1, while
one should expect resident-aliens to have the lowest level of American identity,
given what we have already seen, it is not immediately obvious that African Ameri-
cans should have the highest level of American identity. Furthermore, controlling
for demographic differences did not substantially change this picture. The analyses
showed that while the resident-aliens had significantly lower American identity
scores than the dominant European Americans (t = –3.74, p < .001), African Ameri-
cans had unambiguously greater American identity scores than the dominant Euro-
pean Americans (t = 3.50, p < .001). Rerunning these regression analyses and recod-
ing ethnicity using blacks as the contrast group revealed that, everything else being
equal, blacks also had significantly greater American identity than all groups except
Latinos.
3. Gender had a significant relationship to patriotism in all the data sets; age was
significant in every data set except the UCLA data (where age differences were at-
tenuated). Patriotism was higher among males and older people. Income, education,
and similar class measures were usually of trivial importance.
4. The ideological asymmetry hypothesis posits that while the correlations be-
tween ingroup bias and certain system-justifying ideologies will be positive among
dominants, these correlations will be negative among subordinates.
126
The Contribution of Ethnic and National Identities to Political Conflict


References
Aiken, L. S., & West, S. G. (1993). Multiple regression: Testing and interpreting interac-
tions. Newbury Park, CA: Sage Publications. 
Alba, R. D. (1990). Ethnic identity: The transformation of white America. New Haven:
Yale University Press.
Allen, J. P., & Turner, E. (1997). The ethnic quilt. Population diversity in Southern 
California. The Center for Geographical Studies, Department of Geography,
California State University, Northridge.
Bar-tal, D., & Staub, E. (1997). Patriotism in the lives of individuals and nations.
Chicago: Nelson-Hall.
Blumer, H. (1960). Race prejudice as a sense of group position. Pacific Sociological 
Review,1, 3-5.
Dahl, R. A. (1961). Who governs? New Haven: Yale University Press.
de la Garza, R. O., Falcon, A., & Garcia, F. C. (1996). Will the real Americans
please stand up: Anglo and Mexican-American support of core American politi-
cal values. American Journal of Political Science, 40, 335–351.
Fang, C. Y., Sidanius, J., & Pratto, F. (1998). Romance across the social status con-
tinuum: Interracial marriage and the ideological asymmetry effect. Journal of
Cross Cultural Psychology, 29, 290–305.
Fredrickson, G. M. (1981). White supremacy: A comparative study in American and
South African history. Oxford: Oxford University Press.
Freedman, S. G. (1996). The inheritance: How three families and America moved from
Roosevelt to Reagan and beyond. New York: Simon and Schuster.
Gordon, M. (1964). Assimilation in American life: The role of race, religion, and national
origins. New York: Oxford University Press.
Greeley, A. M. (1971). Why can’t they be like us. New York: Dutton. 
Greeley, A. M. (1976). Ethnicity: A preliminary reconnaissance. New York: Wiley.
Gurr, T. R., & Harff, B. (1994). Ethnic conflict in world politics. Boulder, CO: 
Westview.
Hofstetter, C. R., Feierabend, I. K., & Klicperova-Baker, M. (1999). Nationalism
and ethnicity: A community study. Paper delivered at the Annual Meeting of the
International Society for Political Psychology, Amsterdam, The Netherlands,
July 17–21, 1999.
Horowitz, D. L. (1985). Ethnic groups in conflict. Berkeley: University of California
Press.
Huckfeldt, R. & Kohfeld, C. W. (1989). Race and the decline of class in American poli-
tics. Urbana: University of Illinois Press.
Jackman, M. (1994). The velvet glove: Paternalism and conflict in gender, class and race
relations. Berkeley: University of California Press.
Kelley, R. (1979). The cultural pattern in American politics: The first century. New York:
Knopf.
Kleppner, P. J. (1970). The cross of culture. New York: Free Press.
Kosterman, R., & Feshbach, S. (1989). T
oward a measure of patriotic and national-
istic attitudes. Political Psychology, 10, 257–274. 
Levin, S., & Sidanius, J. (1999). Social dominance and social identity in the United
States and Israel: Ingroup favoritism or outgroup derogation? Political Psychology,
20, 99–126.
Communal and National Identity in a Multiethnic State
127


Levin, S., Sidanius, J., Rabinowitz, J. L., & Federico, C. (1998). Ethnic identity, 
legitimizing ideologies and social status: A matter of ideological asymmetry. 
Political Psychology, 19, 373–404.
Lijphart, A. (1977). Democracy in plural societies: A comparative exploration. New
Haven: Yale University Press.
Lijphart, A. (1979). Religious vs. linguistic vs. class voting: the “crucial experiment”
In comparing Belgium, Canada, South Africa, and Switzerland. American Political
Science Review, 73 (June), 442–458.
Lipset, S. M., & Rokkan, S. (1967). Cleavage structures, party systems, and voter
alignments: An introduction. In S. M. Lipset & S. Rokkan, (Eds.), Party systems
and voter alignments (pp. 1–64) New York: Free Press.
Novak, M. (1971). The rise of the unmeltable ethnics. New York: Macmillan.
Parenti, M. (1967). Ethnic politics and the persistence of ethnic voting. American
Political Science Review, 61, 717–725.
Petrocik, J. R. (1981). Party coalitions: Realignments and the decline of the New Deal
party system. Chicago: University of Chicago Press.
Petrocik, J. R. (1998). Reformulating the party coalitions: The “Christian Democra-
tic Republicans.” Paper delivered at the 1998 annual meeting of the American
Political Science Association. Sep 3-6. Boston Marriot Copley Place Hotel. 
Polsby, N. W. (1963). Community power and political theory. New Haven, Yale Uni-
versity Press.
Powell, C. L. (1995). My American journey. New York: Random House.
Powell, G. B. Jr. (1982). Contemporary democracies: Participation, stability, and violence.
Cambridge: Harvard University Press.
Pratto, F., Sidanius, J., Stallworth, L. M., and Malle, B. F. (1994). Social dominance
orientation: A personality variable predicting social and political attitudes. Jour-
nal of Personality and Social Psychology, 67, 741–763.
Salins, P. D. (1997). Assimilation, American style. New York: Basic Books.
Schlesinger, A. M., Jr. (1992). The disuniting of America. New York: Norton.
Sidanius, J. (1993). The psychology of group conflict and the dynamics of oppres-
sion: A social dominance perspective. In S. Iyengar & W. McGuire (Eds.), Explo-
rations in political psychology (pp. 183–219). Durham, NC: Duke University Press. 
Sidanius, J., Feshbach, S., Levin, S., & Pratto, F. (1997). The interface between eth-
nic and national attachment: Ethnic pluralism or ethnic dominance? Public Opin-
ion Quarterly, 61, 103–133.
Sidanius, J., & Pratto, F. (1993). The inevitability of oppression and the dynamics of
social dominance. In P. Sniderman & P. T
etlock (Eds.), Prejudice, Politics, and the
American dilemma. (pp. 173–211). Stanford University Press. 
Sidanius, J., & Pratto, F. (1999). Social dominance: An intergroup theory of social hierar-
chy and oppression. New York: Cambridge University Press.
Sidanius, J., Pratto, F., & Rabinowitz, J. (1994). Gender, ethnic status, ingroup at-
tachment and social dominance orientation. Journal of Cross-Cultural Psychology,
25, 194–216.
Skerry, P. (1993). Mexican-Americans: The ambivalent minority. New York: Free Press.
Smith, T. W. (1991). What do Americans think about Jews? Working papers on contempo-
rary anti-Semitism. New York: The American Jewish Committee, Institute of
Human Relations.
T
ajfel, H., & Turner, J. C. (1986). The social identity theory of intergroup behavior.
128
The Contribution of Ethnic and National Identities to Political Conflict


In S. Worchel & W. G. Austin (Eds.), Psychology of intergroup relations (pp. 7–24).
Chicago: Nelson-Hall.
Walch, T. (1994). Immigrant America: European ethnicity in the United States. New
York: Garland Publishing.
Wolfinger, R. E. (1965). The development and persistence of ethnic voting. Ameri-
can Political Science Review, 59: 896–908.
Communal and National Identity in a Multiethnic State
129


This page intentionally left blank 


PART III
THE CONTRIBUTION OF 
SOCIAL IDENTITY TO 
VIOLENT INTERGROUP 
CONFLICT


This page intentionally left blank 


Robert W. White
6
Social and Role Identities 
and Political Violence
Identity as a Window on Violence in Northern Ireland
Much of the violence in the contemporary world appears, on the surface, to
involve two groups of people engaged in “them versus us” behavior. From
this view, group members develop a social (group) identity in comparison
with members of some other group. This comparative self-definition leads
to heightened differences across some social category, for example, across
tribes or ethnic groups, and, ultimately, violence. Examples of intergroup
violence based on social identity might include Tutsi versus Hutu, Arab ver-
sus Israeli, and Serb versus Croat. I suspect, however, that much of this vio-
lence does not reduce to a straightforward “them versus us” situation.
It has been argued that violence in Northern Ireland, as an example, in-
volves individuals with competing social identities based on religion—of
Irish Protestants (the majority community) in conflict with Irish Catholics
(the minority community); for example, Whyte, 1990, p. 97; Cairns, 1982.
This view is subject to question. The Irish Republican Army (IRA) has
been the most violent actor in Northern Ireland since 1969 (as measured
by deaths caused). Typically, the IRA is viewed as a Catholic defense orga-
nization. Yet, the origins of the IRA lie in the Irish Protestant community,
and there have been some Protestant members of the organization (see
McKittrick, Kelters, Feeney, & Thornton, 1999, pp. 462–463; White,
1997a). 
The complexity of the political violence in Northern Ireland may be
demonstrated by a relatively straightforward examination of who has killed
whom since 1969.1 The examination shows that there are three broadly de-
fined violent actors: Irish Republican paramilitaries, Protestant (or “Loyal-
ist”—loyal to the British crown) paramilitaries, and the security forces.2
133


T
able 6.1 presents the number of persons killed by each actor for each 
“status” of victim in the Northern Ireland conflict (Sutton, 1994). This is
one of the world’s most studied conflicts, and these data are consistent with
other attempts to enumerate the dead (see White, 1997a; McKittrick, et al.,
1999, pp. 15–19, 1473–1493; Whyte 1990, p. viii).
The primary target of Protestant paramilitaries is Catholic civilians. Six
hundred and twelve (67 percent) of their 911 victims were Catholic civil-
ians. Only 26 (of 911—3 percent) of their victims were Irish Republicans.
Irish Republicans killed the most people. And, Irish Republicans killed a
large number of Protestant civilians (20 percent, or 377/1,926). However,
the single largest group of Irish Republican victims was members of the
British army. Adding in members of the Royal Ulster Constabulary (RUC),
134
The Contribution of Social Identity to Violent Intergroup Conflict
table 6.1 Agency Causing Deaths and the Status of the Victims, Northern 
Ireland, 1969–1993
Status of Agent
Security Forces
Status of
Irish Republican
Protestant
British
RUC-
Victim
Paramilitaries
Paramilitaries
Army 
UDR/RIR1
Other
T
otal
British Army
494
1
3
2
3
503
RUC
280
6
2
1
3
292
UDR/RIR
196
3
2
0
1
202
Republican
154
26
123
18
5
326
Paramilitaries
Protestant
27
61
11
2
1
102
Paramilitaries
Catholic
181
612
134
30
30
987
Civilians
Protestant
377
111
17
6
27
538
Civilians
Political
25
31
1
2
1
60
Activists
Prison
23
2
0
0
0
25
Officers
Former
62
0
0
0
0
62
Security
Officers
Other
107
58
3
1
7
176
T
otals
1926
911
296
62
78
3273
Source: Sutton, 1994.
1. The Royal Ulster Constabulary and the Ulster Defence Regiment/Royal Irish Regiment. The RUC is
the local police force in Northern Ireland. The UDR, a locally recruited regiment of the British Army, was
replaced by the Royal Irish Regiment.


the police force in Northern Ireland, and the Ulster Defence Regiment
(UDR; a locally recruited regiment of the British Army, later replaced by
the Royal Irish Regiment; RIR) shows that about half (970 of 1926) of the
victims of Republican paramilitaries were members of the security forces.
Irish Republicans do not exclusively, or even primarily, target the Protestant
community in Northern Ireland (see also White, 1997a, 1997b; Bruce,
1997; Drake, 1998).
State agents have killed a relatively large number of people in Northern
Ireland. The British Army killed 296 people. The Royal Ulster Constabu-
lary and the Ulster Defence Regiment/Royal Irish Regiment killed 62 peo-
ple. The two communities have not suffered equally at the hands of the
state. It is not surprising that state agents have killed a large number of Irish
Republican paramilitaries, given that Republicans shoot and bomb them.
However, state agents killed 164 Catholic civilians compared to 23 Protes-
tant civilians (and 131 Republican paramilitaries). In Northern Ireland the
state is not a neutral actor (see also White & White, 1995; White, 1999;
Thatcher, 1993, pp. 384–385). T
able 6.1 suggests that between 1969 and
1993, Irish Republicans and the security forces confronted each other with
violence, while Protestant paramilitaries directed their violence against
Catholic civilians (see also note 3).
My objective is to draw on the concepts of social and role identity to pro-
vide insight into the complexity of violence in places like Northern Ireland.
This violence does not reduce to a “them versus us” model, but rather in-
volves a complex mix of identities. I suspect that several conflicts in the
world today, upon in-depth examination, will be shown to be as complex as
the Northern Ireland conflict. With respect to Northern Ireland, I argue
that the violence of Protestant paramilitaries may be understood in “them
versus us,” or in Irish Protestant versus Irish Catholic, terms. Protestant
paramilitaries target their opponents because they are Catholic or Irish
Catholic. Irish Republican paramilitary activity is best understood as a com-
bination of behavior based on a social identity and behavior based on a role
identity. However, for Republicans the social identity dimension of the con-
flict is not Irish Catholic versus Irish Protestant in Northern Ireland, but
instead Irish Nationalist versus the British government and its agents, in
Britain and Ireland. Also, an Irish Republican role identity (described be-
low) complements the Irish social identity. Finally, the British view the con-
flict as an Irish problem—of Irish Protestants versus Irish Catholics (them
vs. them). This has important implications for the behavior of British elites,
British soldiers, and Protestant and Republican paramilitaries, as demon-
strated below.
Like many other conflicts in the world, the violence in Northern Ireland
is centuries in the making. Historical background is therefore presented
first. An in-depth examination of the various identities at work in Northern
Ireland follows.
Social and Role Identities and Political Violence
135


The Historical Setting: Background to the Complexity
Norman/English soldiers invaded Ireland in the 12th century. From then
until the mid-sixteenth century, political violence in Ireland was primarily
motivated by what today is referred to as national identity. The English
found a people different from themselves in language, customs, and outlook
on life. For the next four hundred and so years there were numerous violent
conflicts between native Irish and invading English (Cronin, 1980).
In the 16th century, with the Reformation, religion became important.
The Irish and Old English in Ireland remained Catholic. Most people in
Great Britain embraced Protestantism. In 1601, Irish Chiefs rebelled and
lost. A few years later they fled to the European continent. Their lands—in
Ulster in the North of Ireland—were confiscated and given to English and
Scottish settlers. In the 1640s, the native Irish in Ulster rebelled. They were
supported by the Old English in Ireland. Both groups were Catholic. From
this point religion and national identity became intertwined in Anglo-Irish
politics. When Oliver Cromwell invaded in 1649 he did so with particular,
Protestant religious zeal. His zeal was also a function of the fact that it was
Irish Catholics who were being pacified; Cromwell never treated Catholics
in England as harshly as he did Catholics in Ireland (Fraser, 1973/1997;
Cronin, 1980, pp. 488–489). There was a further confiscation of Irish land
and property, which were placed in trustworthy, Protestant hands. In the
late 1680s, William of Orange seized the throne in England. James II, a
Catholic, fled to Ireland. He was supported by his fellow Catholics in Ire-
land, who sought relief from Protestant domination. William’s army de-
feated James’s army at the Battle of the Boyne, and for the third time in a
century the Irish suffered the results of a lost war (e.g., O’Leary & Mc-
Garry, 1993). 
These 17th century wars established a Protestant ascendancy in Ireland.
By the end of the century roughly 10 percent of Irish land was in Catholic
hands, though perhaps 90 percent of the population was Catholic (i.e., “na-
tive Irish”; O’Leary & McGarry, 1993, pp. 66–69). The wars were about 
religion and political and social control of Ireland. The native Irish were
virtually excluded from Irish social and political affairs. In the process, a
siege mentality was generated among the Protestant elite. Placed in control
of a conquered people were a Protestant Ascendancy that was vastly out-
numbered and constantly fearful.
In the late 18th century the political philosophy of Republicanism was
introduced into the mix. The United Irishmen, the first Irish Republicans,
were founded by Presbyterian merchants and manufacturers in Belfast and
Dublin. They were influenced by the French and American Revolutions,
and by a desire for economic and political freedom. They found allies in
Irish Catholics, who sought political and social relief. Republican ideology
is a political philosophy, not a religious or ethnic one. In the 1790s, it was a
radical philosophy. The United Irishmen sought such reforms as universal
136
The Contribution of Social Identity to Violent Intergroup Conflict


(male) suffrage and no property qualification (see Cronin, 1980). The
United Irishmen rebelled in 1798. Unlike the republicans in the United
States and France, they failed.
In response to the violence in Ireland, the Act of Union (effective Janu-
ary 1, 1801) disbanded an independent Irish parliament and created the
United Kingdom of Great Britain and Ireland. This did not end political 
violence in Ireland. Irish Nationalists rejected the Union and aspired for an
independent Irish nation. Irish Republicans sought to achieve this nation
through force of arms. There were Republican rebellions in 1803, 1848,
and 1867. The rebellions failed, but there remained in Ireland persons
committed to an independent Irish Republic. Throughout the 1870s and
1880s there was Republican-inspired violence. Throughout the 19th cen-
tury, the general Protestant community in Ireland—“Unionists”—sup-
ported the Union. This was especially so in northeast Ireland, where the
descendants of the 17th-century settlers feared that “Home Rule” would
lead to “Rome Rule.” In Belfast in the 1880s, for example, there was wide-
spread rioting in opposition to a Home Rule Bill for Ireland (see Cronin,
1980; Lyons, 1973).
Northern Ireland dates from events in 1916–1922. In 1916, there was a
rebellion the day after Easter. The rebels lost, but they reorganized as the
Irish Republican Army (IRA) and the political party Sinn Féin (founded
1905). The two organizations worked to break the connection with the
United Kingdom. They established a revolutionary Irish government in
Dublin and carried out military activities against the British presence in Ire-
land. The IRA and Sinn Féin were successful throughout most of Ireland.
In the northeast, the IRA and Sinn Féin were opposed by both the state’s
security agents—for example, the Royal Irish Constabulary—and by a
Protestant paramilitary organization, the Ulster Volunteer Force. In re-
sponse to the political unrest, the British passed the Government of Ireland
Act (1920), which partitioned Ireland. The Act created Northern Ireland,
which comprised six of the nine counties of the province of Ulster, and the
Irish Free State. The parliamentary headquarters for Northern Ireland was
established in Belfast. Dublin became the seat of power for the Free State.
At its creation, Northern Ireland was approximately two-thirds Protestant
and one-third Catholic. The Irish Free State (declared a republic in 1949),
was (and is) approximately 95 percent Catholic.
In 1921, negotiations between Republicans and British representatives
led to a treaty. The IRA and Sinn Féin split over the Anglo-Irish Treaty
(1921), and a civil war ensued. Those Republicans opposed to the treaty
maintained the IRA and Sinn Féin, and fought against former comrades
who accepted the treaty and the new Irish Free State. Pro-treaty forces won
the war, and, among other things, this cemented partition. Those who re-
mained in Sinn Féin and the IRA refused to recognize partition and the
governments in Dublin and Belfast. They also maintained their organiza-
tions. From 1939 to 1945 and from 1956 to 1962, the IRA engaged in un-
Social and Role Identities and Political Violence
137


successful paramilitary campaigns in England and Northern Ireland, in an
attempt to create a united Ireland (see Bell, 1979).
In Northern Ireland, the Protestant community embraced the new
province. The Catholic community, both in Northern Ireland and in the
Irish Free State, largely rejected it. Thus, partition created a double minority
situation. In Northern Ireland Protestants dominated a minority Catholic
community who believed they had been gerrymandered into a Protestant
state. Within Ireland as a whole the Protestant community was in the minor-
ity. The government of the Free State/Republic rejected Northern Ireland
and claimed the entire island for itself.
In the 1960s, there was a civil rights campaign in Northern Ireland. The
campaign sought to redress grievances that Nationalists/Catholics had
faced since the founding of the province. Civil rights protests led to coun-
terdemonstrations by Protestant extremists, which led to widespread riot-
ing in Northern Ireland in August 1969. The British government sent in
troops to restore order. It is out of events in the mid-to late 1960s that the
three actors of table 6.1 are derived: Irish Republican paramilitaries, Protes-
tant paramilitaries, and the security forces (Bell, 1979, 1993).
Organizationally, Irish Republicans can legitimately trace themselves to
the rebellion in 1848. Veterans of 1848 were involved in rebellious activity
in the 1860s, and some of those involved in the 1860s (and 1870s and 1880s)
took part in the 1916 Easter Rising (see Bell, 1979, 1987, p. 136). In 1969
and early 1970, the IRA and its political wing, Sinn Féin, split again and the
Provisional IRA (hereafter, IRA) and Provisional Sinn Féin (hereafter, Sinn
Féin) were created. The founders of the Provisionals were Republican vet-
erans, some with activist histories that dated to 1916. Contemporary
Protestant paramilitary organizations date from the mid-1960s and the cre-
ation of the Ulster Volunteer Force (named for the UVF of the 1910s). In
1971 the Ulster Defence Association (UDA) was started as a coordinating
body for vigilante groups. The Ulster Freedom Fighters (UFF) became the
paramilitary wing of the UDA (see Flackes & Elliott, 1989, pp. 272–277,
278).
Political violence in Northern Ireland is hundreds of years in the mak-
ing. It is not simply about religion, but instead involves a complex mix of
national identity, religion, and political aspirations. Social identities and
role identities, as described below, provide an insight into this complexity.
Identities: Social (Group) and Role
According to T
ajfel (1981, p. 255), social identity is that “part of an indi-
vidual’s self-concept that derives from his knowledge of his membership of
a social group.” T
ajfel typically refers to groups based on race, ethnicity, or
religious identification. For T
ajfel, the need to view oneself positively may
heighten differences between groups and lead to conflict. While T
ajfel fo-
138
The Contribution of Social Identity to Violent Intergroup Conflict


cuses on intergroup relations, Turner, Hogg, Oakes, Reicher, & Blackwell
(1987) extend the social identity perspective to emphasize intragroup be-
havior. They define social identity as self-categorizations that “define the
individual in terms of his or her shared similarities with members of certain
social categories in contrast to other social categories” (Turner, Oakes,
Haslam, & McGarty, 1994, p. 454). Thoits and Virshup (1997) summarize
these two perspectives by noting that they focus on “we’s,” in the sense that
social identities are “collective identities.” The focus is on how the indi-
vidual and the group mesh together. While the social groups around which
a collective “weness” develops are often described in terms of race, gender,
and ethnicity, a social identity may also develop around a political group-
ing—for example, the Ulster Volunteer Force or the Irish Republican
Army. My use of the term “social identity” is consistent with Thoits and
Virshup’s (1997) usage, in which they refer to a collective “weness” that ex-
ists in comparison to a collective “other.”
Role identities focus on “me’s” (Thoits & Virshup, 1997). Role identities
act as “internalized positional designations” (Stryker, 1980, pp. 60–61; see
also McCall and Simmons, 1978) that locate and define individuals within a
group. Role identities refer to individual-level self-conceptions of the re-
quirements attached to social positions. Individuals internalize these re-
quirements, which become identified as a part of the self. Role identities
have “salience” for individuals (Stryker, 1980, 1968). In certain situations
one role identity may be more salient for an individual than another role
identity. In addition, there is the potential for significant role conflict and
role strain—when the expectations associated with roles conflict or compete
(Stryker, 1980). My use of the term “role identity” follows the definition of
Stryker and refers to self-conceptions of the requirements attached to a so-
cial position that develop from social interaction—for example, by learning
to be a Volunteer in the Irish Republican Army from other IRA Volunteers.
The social or group (collective) identity is conceived of as a group-level
phenomenon, where the individual conceives of the self as one of “us”—
often in contrast to “them.” The focus is on cognitive processes, comparing
one’s group to some other group. As Thoits and Virshup (1997, p. 115)
state, “Group- or category-based identities are collective-level self-concep-
tions; they are identifications of the self with a collectivity.” In contrast, role
identities focus on who the individual is. Instead of comparing “us” vs.
“them,” the individual considers “Who am I?” by drawing on social interac-
tion and incorporating the expectations of other people into their self-
conceptualization. Both kinds of identity are social, but in different ways
(see also Stryker, 2000; Klandermans and de Weerd, 2000). For those who
employ the concept of social identity, the self is a product of comparison
and categorization processes in which “the perceiver appraises the self in re-
lation to others.” For those who employ the concept of role identity, the self
is a product of incorporating others’ expectations and appraisals, of apprais-
ing the self from the perspective of others—for example, in role-taking 
Social and Role Identities and Political Violence
139


exercises (see Thoits & Virshup, 1997, pp. 120–122). In terms of violent 
political activists in Northern Ireland, persons may kill on behalf of their
group, and/or they may kill because they have internalized such behavior as
part of the role of being a particular political activist. Indeed, activists may
kill because they are motivated simultaneously by both a social identity and
a role identity.
In the following, I argue that the violence of Protestant paramilitaries is
best understood from a social identity perspective. Protestant paramilitaries
act in defense of their social group by killing members of the other (Catholic)
social group. In contrast, I argue that Irish Republican paramilitaries are mo-
tivated by both social and role identities. Irish Republicans target the security
forces in Ireland because these forces represent an oppressor of their “Irish-
ness” (the British government) and because they have internalized the role
requirements of being an Irish Republican. Finally, I argue that the British
perspective may be understood in social identity terms. However, instead of
taking a “them versus us” perspective—for example, British forces versus
Irish Republican forces—the British take a “them versus them” perspective
and view the violence as an Irish problem, involving Irish Protestants versus
Irish Catholics.3
Data Sources: Accounts from Activists
The accounts below are from very different sources. None of the data were
collected with this particular research topic in mind. The accounts from
Irish Republicans are from my ongoing research on the causes and conse-
quences of small-group political violence. These accounts are taken from
interviews that centered on why the respondents became involved in the
Irish Republican movement. The Republican accounts represent my inter-
pretation of how Irish Republicans may be understood in terms of social
and role identities. This interpretation is based on interaction with Irish
Republicans over a 17-year period. I have formally interviewed perhaps 100
Irish Republicans, and have informally met with an unknown number of
others. Many of the respondents were low-level Sinn Féin workers, such as
advice center staffers. Included among the respondents are several promi-
nent Republicans and several veterans of the IRA. Many of the respondents
were born and raised in families where a parent or other significant adult
was active in Republican politics or had strong Republican sympathies (see
White, 1993; White & Fraser, 2000).
The accounts from Protestant paramilitaries, British elites, and British
soldiers are taken from secondary sources. These sources include interviews
with journalists, memoirs, and published accounts of speeches. I did not
collect these data. My interpretation of the accounts of Protestant paramili-
taries, British elites, and British soldiers, therefore, is qualified. With re-
spect to Protestant paramilitaries, I present the accounts to illustrate the
140
The Contribution of Social Identity to Violent Intergroup Conflict


conventional wisdom with respect to these people—that is, that the motive
of these paramilitaries is sectarian and is based on a strong identification
with being Protestant. This conventional wisdom is based on solid social
science research (e.g., Bruce, 1992). The statements from British elites and
British soldiers are consistent with survey data, noted below, which shows
that Northern Ireland is not a high-priority item for the British public.
This is one indicator that British people may interpret the conflict as one of
“them versus them.”
Northern Ireland: The Actors and Their Identities
In Northern Ireland, political actors represent two communities that reside
there and one external interest group: the Protestant community, the Irish
Nationalist community, and British representatives. Within the Protestant
community, there are Unionists and Loyalists. Unionists want Northern
Ireland to remain part of the United Kingdom. Loyalists are Unionists who
take an extreme position on the Union and are willing to defend it with
paramilitary activity. Irish Nationalists, who are predominantly Catholic,
seek an end to the Union and a united Ireland. Among Irish Nationalists are
Irish Republicans, who are willing to use paramilitary activity to bring
about a united Ireland. British interests are represented by political ap-
pointees, for example, the secretary of state for Northern Ireland, and per-
sons involved with the security apparatus, such as the British army.
The Protestant Community: Unionists.
In many ways, the Unionist mentality is a siege mentality. In spite of
their political power in Ireland, Unionists have always been in the minority,
and they have, therefore, always perceived their position to be threatened.
When it was created, Northern Ireland offered Unionists a gerrymandered
majority situation, but this was not guaranteed to last forever. Historically,
the Protestant ascendancy in Ireland maintained their position only with
the help of the English and British state. The British parliament created
Northern Ireland as a haven for pro-British Protestants; the Protestant
community in Northern Ireland did not do this on its own.
T
oday, most Protestants in Northern Ireland are Unionists (there are
some Catholic Unionists). In addition, many Protestants in Northern Ire-
land identify not with Ireland but with Britain. When asked in 1968, on the
eve of the current violence, which term best describes themselves (Irish,
British, Ulster, Sometimes British-Sometimes Irish, Anglo-Irish), only 39
percent of Protestants chose “British.” An additional 32 percent chose “Ul-
ster.” Perhaps surprising, the first choice of 20 percent of Protestants to this
question was “Irish” (see Rose, 1971, p. 485). In 1978, in the midst of a vio-
lent conflict, the question was repeated on a survey by Edward Moxon-
Social and Role Identities and Political Violence
141


Browne. Moxon-Browne found that the number of Protestants whose first
identity was “British” increased significantly, to 67 percent. The number
who chose “Ulster” fell to 20 percent, and only 8 percent chose “Irish”
(Moxon-Browne 1983, pp. 5–7).
The Protestant Community: Loyalists
Loyalists (loyal to the British crown) tend to be working-class Unionists
who view violent opposition to Irish nationalism as both justified and ne-
cessary; they support, or are, Protestant paramilitaries (see Bruce, 1992, 
pp. 14–15; Nelson, 1984). There is general agreement that Protestant para-
militaries are motivated by sectarianism, by deep-seated mistrust, and by
hatred of Catholics. Steve Bruce, an authority on Protestant paramilitaries,
stated in 1994:
Although leaders of the UDA [Ulster Defence Association] and UVF [Ulster
Volunteer Force] have periodically tried to confine the violence of their mem-
bers to some notion of “legitimate targets” (that is, republicans) or have
claimed to be trying so to restrict the deaths, very many of their victims were
chosen simply because they were Catholics. Even in these more publicity-
conscious days there are loyalist paramilitary leaders who say that they like to
kill republicans but an ordinary Catholic will do. (1994, p. 125)
The data in table 6.1 support this interpretation. The majority of victims
of Protestant paramilitaries are Catholic civilians. Because Catholics tend to
be Irish Nationalists, they are by (Protestant paramilitary) definition trai-
tors and subject to violence. Protestant paramilitaries punish the general
Catholic community when they are threatened or want to hit back. Because
Irish Republicans are hard to target, the objective is to create enough fear in
the Catholic community that they, in turn, will force the Republican para-
militaries to quit. At its most basic level, Loyalists view themselves as fight-
ing for tribal survival (Bruce, 1992, p. 55). 
In identity terms, Protestant paramilitaries view the world from a
Protestant social identity, in opposition to Catholics. Consider the follow-
ing from a former Protestant paramilitary, John White. White was a lead-
ing member of the Ulster Freedom Fighters (UFF). He describes his up-
bringing:
I remember my mother, an uneducated working-class woman, saying to me,
Son, we don’t go into a united Ireland, and me saying, why not, Mammy, and
I remember her saying, because you’ll not get a job, and you’ll probably not
get a house, and they’ll murder us, you know. And my Da, he just had a con-
firmed hatred for Roman Catholicism. I remember one time my brother
brought a Roman Catholic to the house, a girlfriend, and my Da told her to
get out of the house and not come back again. So, you know, I thought there
must be some genuine fears there, when even my Da was saying, no bringing
T
aigs home to this house. (Stevenson, 1996, p. 62)4
142
The Contribution of Social Identity to Violent Intergroup Conflict


In 1976, White confessed to the 1973 murders of Paddy Wilson and
Irene Andrews. Wilson was a member of the Social, Democratic, and
Labour Party (SDLP), a moderate, middle-class political party associated
with the Catholic community in Northern Ireland. The SDLP then, and
now, advocated a united Ireland, but through peaceful means. Now out of
prison, John White commented on why Wilson, who was publicly known
to be unarmed, was a target:
My analysis then was that he [Wilson] supported the concept of a united Ire-
land; the IRA were trying to pursue a united Ireland; therefore, by his
rhetoric, he was supporting the IRA. It may have been warped at that time,
but that was the way I viewed it when I was twenty two years of age. (Steven-
son, 1996, pp. 59–61)
John White’s account is not unique among former Protestant paramili-
taries. The following is from Alex Calderwood, a veteran of the Ulster De-
fence Association (UDA). He states, “I grew up hating Catholics.” 
I admit that I was a sectarian bigot. . . . It was my background and the soci-
ety that programmed me in a sense, but I was responsible for what I done. I
was a bigot and I hated Catholics. (Stevenson, 1996, p. 65)
A central theme in the accounts of Protestant paramilitaries is a sense of
“us versus them,” Protestant versus Catholic. The account of another for-
mer Protestant paramilitary, William Smith, is consistent with this:
In the seventies these communities were close-knit communities. . . . You
still had the extended family situation, and word-of-mouth and rumor was far
more valid than printed matter or television. But there’s no one thing, it was
the whole situation at the that time. You had segregated schools, you had seg-
regated areas, you had a whole history dealt down to you by word-of-mouth
about troubles, the IRA, and all that—that was all part of your dogma the
whole way through your life. So whenever the bogeyman appeared in 1970,
this was the bogeyman we were told about. The bogeyman appeared and the
rest became part of life—there was thousands and thousands, I mean, the vast
majority of people were involved in it. So when your community felt threat-
ened, you felt threatened. (Stevenson, 1996, p. 67)
Based on these accounts, it appears that persons who became Protestant
paramilitaries were motivated by an “us versus them” perspective. These
accounts suggest that the paramilitaries learned this perspective because
they were encouraged, by family, friends, and the Protestant community
around them, to view things comparatively—Protestant versus Catholic.
Thus, Catholics are people to be hated because “they’ll murder us.” 
Beginning in the early 1970s, large numbers of people were recruited
into Protestant paramilitary organizations like the Ulster Volunteer Force,
the Ulster Defence Association, and the Ulster Freedom Fighters. It can be
expected that members of these groups developed social identities that were
particular to these groups—that is in addition to their Protestant social
Social and Role Identities and Political Violence
143


identity they probably also developed social/collective identities based on
their membership in a particular paramilitary group. As members of these
groups they may also have developed role identities (see below). However,
the key to understanding these groups and their members remains their
Protestant social identity. This social identity is common to most if not all
Protestant paramilitaries, and is probably the most important motivator of
their behavior.
The Irish Community: Nationalists
An Irish Nationalist is a person who wants all of Ireland to be ruled by Irish
people (Irish people being defined as people living in Ireland, even if they
identify themselves as British). Irish Nationalism predates the Reformation
and the Act of Union (effective January 1, 1801; Cronin, 1980). T
oday, most
Irish Nationalists are Catholic. With respect to Northern Irish politics, the
terms “Catholic” and “Nationalist” are often used interchangeably, but they
are not synonymous. Catholics in Northern Ireland do not exclusively iden-
tify themselves as Irish. When Richard Rose (1971) surveyed the Northern
Irish public in 1968, he found that when asked which term “best describes
the way you think of yourself” (British, Irish, Ulster, sometimes British-
sometimes Irish, Anglo-Irish), 76 percent of Northern Irish Catholics chose
the “Irish” option. Only 15 percent of Catholics identified themselves as
“British,” and even fewer (5 percent) identified themselves as “Ulster.” T
en
years later, in the midst of violence, when Edward Moxon-Browne included
the question on his survey, the percentage had fallen; 69 percent of
Catholics chose “Irish” as their first option. Again, only 15 percent of
Catholics identified themselves as “British,” and few (6 percent in 1978)
identified themselves as “Ulster.” Also in 1978, when asked “which of these
would you least like to call yourself? British, Irish, Ulster, sometimes
British-sometimes Irish, Anglo-Irish, Other?”, less than 2 percent (1.7 per-
cent) of Catholics chose “Irish”—that is, very few Catholics rejected the
Irish national identity (see Moxon-Browne, 1983, pp. 6–9). In contrast to
the Protestant community, the national identity of Catholics/Nationalists in
Northern Ireland appears to be more stable (see also Moxon-Browne,
1983). 
The Irish Community: Irish Republicans
At its most basic level, an Irish Republican is someone who “seeks to sepa-
rate Ireland from England by force” (Cronin, 1980, p. 23). At their core, all
Republicans agree on two things: (1) the British government is the cause of
political strife in Ireland, and (2) a British withdrawal from Ireland can be
gained only through force of arms (see White & Fraser, 2000). These be-
liefs are evident in accounts from Republicans and may be interpreted in
both social and role identity terms. 
144
The Contribution of Social Identity to Violent Intergroup Conflict


Consider the following from Respondent 1. He describes events that led
him into the Irish Republican Army. Recognizing that he is Irish and that
the British were killing “our people” was essential for his recruitment into
the IRA. In the early 1970s he went to school in England:
RESPONDENT 1 [Hereafter, R]: I went to England and as I say everything
then was at a distance, it wasn’t as immediate. . . . But then Bloody Sunday
[in which British troops shot dead 14 Nationalist civilian protesters, in 1972]
happened and again this incredible sense of outrage. And then, just through-
out the year, watching the British, some of the British people—their attitudes
to us, the Irish. And it was just—fuck sake. The attitudes were just that—
you’re “Paddy.” They automatically call you Paddy when you go over there.
I’d say, “My name’s Ruairí.” “Sure, you’re Paddy. You’re all called Paddy and
Mick.” I said, “No, my name’s Ruairí.” I took it from my friends there because
it was usually a joke, there was no malevolence in it. But, there were people
who used it as a derogatory term. And just throughout the year, watching
things going at home. And following along Bloody Sunday I decided that
what I was doing, which I wasn’t happy with anyway, that I was wanting to go
home more and more. And then when I left England to travel for a couple of
months, hitched around, and I came home. And as soon as I came home I got
into Sinn Féin and it was a natural progression from there. Because my beliefs
were fairly well formulated at the time. The British were killing our people, they
were locking them up and they were nothing more than Stormont [the
Northern Ireland parliament]. (emphasis added) 
Similar comments were offered by a Protestant Republican (see also Hynd-
man, 1996). He was born and raised in East Belfast. By his background he
was an unlikely candidate for involvement with Irish Republicans. When he
moved to England, he discovered that he was Irish.
INTERVIEWER [Hereafter, I]: You mentioned our people, you said they
were killing our people or shooting our people. Do you consider yourself
Irish? 
R: Oh yes, I am Irish you see. 
I: But I mean, the typical—
R: (Laughs) I know it’s uh—I understand what you mean. Given my back-
ground, right? For a long time I thought I was British. I didn’t really know
what I was, actually. When I was younger I went to England to work due to
the high unemployment at home. And, then I was made very aware of my iden-
tity, and it was Irish. I was left in no doubt that I was a Paddy. And uh, you
know the old stories about people going to try to find flats, you know, some-
where to live in London, where the signs are in the window, “No Irish,” “No
Coloureds.” It was a good thing to experience because it did give me a per-
ception, it did give me an education which I might not necessarily have gotten
if I had stayed in the North [of Ireland]. But certainly that made me very
aware of my identity. Then moving from my Loyalist area, through other lo-
cations, finally to end up in a Republican area was a great experience. Living
there which had always—had been a place which I had visited quite a lot, but
Social and Role Identities and Political Violence
145


living in it is a lot different from visiting. T
o experience the continual daily
harassment, the constant harassment. It’s a great education as well. T
o see the
poverty as well. [RESPONDENT 2, emphasis added]
These two respondents acquired an Irish social identity by personally expe-
riencing discrimination and devaluation of their Irishness. The develop-
ment of this social identity helped to make them potential recruits to Irish
Republicanism. Their recruitment was also influenced by their acquiring a
Republican role identity.5
Respondents 1 and 2 became Republicans when they were adults. Many
Republicans, however, are born and raised in Republican families and 
they are influenced by Republicanism at an early age. One result of this in-
fluence is an Irish social identity. Respondent 3 was raised in a Republican
family: 
R: the only solution to our problems, and it has been my opinion all down
through the years, is a united Ireland. It’s the only way forward for us. There
is no other way. The other way is a creeping, cringing, crawling spirit of the
Irish nation. We desire freedom, we seek freedom, and it’s not too much to
demand. All we’re seeking is our own country. . . . I think it only makes
good common sense that if you are the owner of a house, I always make this
analogy, that if you’ve got a four-roomed house, and there’s one person occu-
pying the house. And there’s quite a large family of you. And you make a de-
mocratic decision time after time for the person that’s occupying your house,
undemocratically, for to remove . . . to get out of the house. That democra-
tic decision is taken. But the man won’t move. What’s your alternative? Get
him by the scruff of the neck and throw him out. It’s the only way. Our people
has taken the democratic decision in this country here several times to ask Britain to
leave this country. So as we can apply our own national government. But she
hasn’t adhered to that democratic decision, and that’s why this struggle is still
going on. (Respondent 3, emphasis added)
Respondent 4 was also raised in a Republican family. He stated the following:
R: [When I joined, my goal for the Movement] was to see the complete free-
dom of this country. In other words, what the Republican goal always was, to
break the connection with England and to establish an Irish Republic. An Irish So-
cialist Republic. And that was my reason for joining the Movement and that is
why I got involved. (Respondent 4, emphasis added)
Like Protestant paramilitaries, Republicans—both from Republican fami-
lies and not—develop a social identity that places them in an “us versus them”
situation. The “them,” however, is the British government, not the Protes-
tant community in Northern Ireland. 
There is evidence that Irish Republicans develop a role identity that
complements their Irish social identity, and that this role identity develops
through social interaction with family members and/or peers who are Re-
publicans. Consider the following from the three respondents raised in 
Republican families:
146
The Contribution of Social Identity to Violent Intergroup Conflict


I: Why in general did you get involved in the Movement?
R: Well, Bob, I was involved with the Republican Movement at a very early
age because of my father being a member of the Republican Movement as
well. From 1916 ‘til the Civil War my home was a place for anybody who
came on the run, or to seek help. It was always given to them. So I was sort of
brought up, you know what I mean, within Republican circles, my father be-
ing a Republican. And uh, I suppose it rubbed off on me. I always had Repub-
lican ideals, you know? (Respondent 3)
R: Well, the reason why I decided to get involved in the Republican Move-
ment was that I was born into the Republican tradition. Both on my father’s
and my mother’s side. And that is one of the reasons. And reading Irish his-
tory I realized and came to the conclusion that the Republican Movement
were the true inheritance of the Republican faith. That is the reason why I 
became [involved]. 
. . .
R: [Reading the question] Did any particular person have a strong influence—
eh, to get involved? Well, I wouldn’t just say to get involved, but, there was
one woman, she was my grandmother. Incidentally my four grandparents
were teachers. But this particular—my mother’s mother, she was an outstand-
ing republican sort of thing, you know? When she died in 1969—She never
wavered. . . . She was strong in her Republican [beliefs] to the day she died.
And when she—her house, as a matter of fact, was the headquarters of the
Irish Republican Army [in her area]. . . . she was a very highly respected
woman sort of thing there. . . . And she had a big influence on my life there
because as you may know you’re—well, in this country anyway your grand-
children go on holidays with their grandparents. . . . She never wavered, as
I said. And she had a big influence. And even up to this day, if any questions
come up [I would refer to her Republicanism]. (Respondent 4)
R: Republicanism in our house was a—sort of what would be termed now as
pure Republicanism, in that my father wouldn’t of been very sympathetic to-
ward a socialist line of thinking. It would of been more of a Republican prin-
cipled directly. And, uh, seen the great danger in socialist tendencies within it.
And, to some extent that would of influenced my direction for a number of
years. And even yet to some extent that you would fear a socialist—too big a
socialist involvement within it. And uh, he was very much just a straight line
Republican. 
I: Is that why you would have gone with Provisional rather than Official
[when the Republican Movement split in 1969/70]?
R: That’s right. That would of been a big influence actually. And uh, not only
from hearing it in the past but actually hearing him talking about it at that
particular time, eh, that the split come. My direction like would of been influ-
enced a big part by his sort of talking to me. (Respondent 5)
It appears that these children of Republicans were raised such that they ac-
quired Republican principles and learned the role of being a Republican—
for example, to help other Republicans “on the run” and to never “waver”
in your commitment. This acquisition and learning occurred through inter-
Social and Role Identities and Political Violence
147


action with other Republicans. Accounts from Republicans not born and
raised in Republican families suggest that they also acquire a role identity
through interaction with both their family members (even though they are
not Republican) and their peers who are Republicans. The following is
from Respondent 1. He is not from a Republican family. He joined the IRA
in the early 1970s. He was strongly influenced by his personal experiences
with peers who were involved in the Irish Republican Movement. His in-
volvement stems from a family background that stressed being a responsible
person, and his social interaction with people who he believed were being
responsible—that is, persons who were willing to risk their lives in a just
struggle against a much stronger opponent (and his development of an Irish
social identity, as noted earlier).6
R: The main reason that I got involved with the Republican Movement is that
I was brought up to believe that you should treat people as equals and that no-
body had the right to walk over you, nor did you, more importantly, did you
have the right to walk over anybody else. And the Republican Movement, it
may have historical background, yes, but essentially it got off the ground be-
cause it came out of the community and it was a community that accepted it
because it was its own people. And therefore I suppose you could say it was
my contemporaries that, if anybody influenced, would have influenced me to
join. But it was just a natural thing because it was the time of ferment, of po-
litical ferment that was in it. I felt that I couldn’t close my eyes to it and that I
had a responsibility to do something. I’m not sure, mind you, that would be
everybody’s reply—it sounds terribly altruistic [laughs]. But the point is that
that’s just the way I felt about it. I just felt that I had had a pretty good life
compared to most of the lads that would have been in the movement in Derry.
I come from a middle class background and I never really had to worry about
anything and I just felt there was a responsibility for me to do something. But
it wasn’t as clear cut as that. 
. . .
R: I didn’t join because my friends were in the movement. That’s not the rea-
son I joined. I was motivated to join the Republican Movement because I
knew that the people who were in it were sincere. I mean, a 19-year-old kid or
an 18-year-old kid who makes a conscious decision to go out and fight a vastly
superior army on the streets, to take on a government and all its forms, and to
change society—He may not think in terms of vast political thought about
how society’s going to be changed by him. But if he makes that conscious de-
cision instead of happily waiting until he get his welfare check, and then goes
off and plays pool or drinks beer and goes to dances—I mean they’re de-
nounced as being mindless thugs but that is just not true. It could never for
me be a description of them. (Respondent 1)
As it was for Protestant paramilitary organizations, in the early 1970s
Irish Republican organizations received large numbers of recruits. As with
Protestant paramilitaries, the new Republican paramilitary recruits were in-
fluenced by a social identity. Unlike Protestant paramilitaries, the Republi-
148
The Contribution of Social Identity to Violent Intergroup Conflict


can recruits did not interpret their activity as Catholic versus Protestant,
but rather as Irish versus British. In addition, and also unlike Protestant
paramilitaries, the Republican paramilitary recruits joined a movement with
a lengthy history. Along with this history came established roles for mem-
bers of the Republican Movement—“she never wavered.” Out of interac-
tion with persons who were already Republican (be they family members 
or peers) the Republicans acquired a role identity that complemented their
social identity.
The British Perspective:“Them vs.Them”
The British interpretation of the violence is that it is an Irish problem in
which members of two Irish communities, Protestant and Catholic, cannot
get along. This is implied in the accounts of Respondent 1 and Respondent
2, one Catholic and the other Protestant, but both “Paddy” in Britain. In ef-
fect, the British deny the social identity of Protestant paramilitaries, who
view themselves as British, and they deny the political aspirations of Irish
Republican paramilitaries, who do not want to be citizens of the United
Kingdom. 
This is not a new phenomenon. In 1921, on a visit to Belfast, King
George V said:
I speak from a full heart when I pray that my coming to Ireland today may
prove to be the first step towards an end of strife amongst her peoples, what-
ever their race or creed. In that hope I appeal to all Irishmen to pause, to
stretch out the hand of forgiveness and conciliation, to forgive and forget and
to join in making for the land which they love a new era of peace, content-
ment and goodwill. (see Lyons 1973, p. 426)
Fifty years later, Reginald Maudling was the British Home Secretary re-
sponsible for Northern Ireland affairs at Westminster (from 1970-1972). 
In his memoirs, he described how the “Irish” problem perplexed him, an
Englishman.
It is very hard for an Englishman to understand the feelings of those who live
in Northern Ireland. The history of their struggles is a long one, and they
tend to cherish every moment of hatred in it. The deep divide between
Catholic and Protestant, which, incidentally, is more racial than religious, had
been handed down faithfully from generation to generation. (Maudling 1978,
p. 180)
Traditionally, Ireland is not that important for British elites. This makes
the violence that much more an “Irish problem.” After helping execute
Charles I, Oliver Cromwell spent nine and one-half months in Ireland,
pacified the country, and then returned to the real business at hand—
running England (Fraser, 1973/1997). In the early part of this century,
there were a series of debates and proposals associated with the Home Rule
[for Ireland] Bill in 1914. William Manchester, in his biography of Winston
Social and Role Identities and Political Violence
149


Churchill, describes the Irish situation as being, “foremost in everybody’s
mind, so certain to burst into flames at any moment” (1983, p. 463). But, as
a meeting concluded Sir Edward Grey read out a document from the For-
eign Office. It was an ultimatum from Austria to Serbia, part of the begin-
ning of what became World War I. The war developed, and the Irish issue
was “deferred.” Churchill later recalled, 
The parishes of Fermanagh and Tyrone faded back into the mists and squalls
of Ireland, and a strange light began immediately, but by perceptible grada-
tions, to fall and grow upon the map of Europe. (Manchester, 1983, p. 463)
The “Irish question” also has not had an impact on the voting British
public. Northern Ireland constitutes less than 3 percent of the population
of the United Kingdom. Since 1983, the province has sent only 17 (of 635)
MPs to Westminster. In response to Gallup polls on important issues in the
United Kingdom, Northern Ireland traditionally has “fluctuated between
nil mention and coming at the bottom of a list of ten or a dozen issues” (see
O’Malley, 1990, p. 226). 
Because Northern Ireland has not been a high priority for the British
public, British soldiers, recruited from that public, appear to be relatively
unaware of the politics of the situation (at least in the early years of the re-
cent conflict). Consider the following from two soldiers sent to Northern
Ireland in 1969. 
BRITISH SOLDIER 1: The boys knew where they were supposed to go in
terms of sealing off streets, so it was then a question of getting men down to
the corners and road junctions and giving them cover in case there was shoot-
ing. We hadn’t a clue what was going on really, but the feeling I got within half
an hour was one of relief, on both sides. The Irish are terribly curious and they
came out of all sorts of corners, and I think both Protestants and Catholics
were relieved that there might be some stability. . . . (emphasis added;
Arthur, 1987, p. 7)
BRITISH SOLDIER 2: For the first week I slept underneath an awning by a
baker’s shop. I’d chosen the baker’s shop because I knew we’d get fresh bread
and buns each day. This was the period of complete daze, immediately after
the initial rioting. No one knew what was going on, so we had to start talking to the
people to find out what might happen. We felt we were terribly welcome, particu-
larly by the Protestants, but they kept on saying: ‘Why have you got your ri-
fles pointing at us, and your backs to the Catholics, when they’re the ones
who started it?’ So immediately we came up against the sectarian problem:
which way do you face? Who do you defend? Which side do you protect? (emphasis
added; Arthur, 1987, pp. 9–10)
The perspective on Ireland of British elites, citizens, and soldiers may
best be described as viewing the conflict as “them versus them.” The degree
to which Britishers identify with Britain and adopt potential role identi-
ties—“As a British soldier my job here is to keep the peace”—may be less
important than the fact that they view what happens in Northern Ireland as
150
The Contribution of Social Identity to Violent Intergroup Conflict


external to themselves, or at least not of their own making, that it is a con-
flict caused by “others” and it is not a reflection of British policies or 
actions. 
By denying the identity of those Protestants in Ulster who identify
themselves as British, British elites contribute to the fears of these people.
By refusing to recognize the political claims of the Republicans, and by ac-
tively denying these claims, British elites leave Republicans with violence as
a means of expressing their political desires. Perhaps most important, view-
ing the conflict as external, or involving “them versus them,” contributes to
policy mistakes on the part of the British government, as described below. 
Implications
In Ireland there are three violent political actors: Protestant paramilitaries,
Irish Republican paramilitaries, and state agents. Each perceives the vio-
lence in Northern Ireland from a different perspective. These perceptions
have real consequences. This is demonstrated in the British policies of
criminalization and Ulsterization, which were implemented in the mid-
1970s. Their impact is still felt today.
In 1972, IRA prisoners led by Billy McKee, an IRA veteran of the 1930s
and 1940s, went on a hunger strike in order to gain political status. In the
midst of negotiations that led to a truce between the British government
and the IRA in late June, political status was granted to both Republican
and Protestant paramilitary prisoners. Criminalization revoked this status.
After March 1, 1976, persons convicted of paramilitary crimes were treated
as ordinary criminals. The British government denied the political nature
of the IRA’s activity and equated IRA violence with the sectarian violence of
Protestant paramilitaries. Ulsterization dates from 1977 and is designed to
make the Royal Ulster Constabulary the primary policing agent in North-
ern Ireland (e.g., Flackes & Elliott, 1989, p. 392 ; see also Bell, 1993, 
p. 517). As a result, the violence became that much more an Irish problem
(of “them vs. them”).
The implications of criminalization and Ulsterization were profound.
Irish Republican prisoners rejected criminalization. They refused to wear
prison clothes and wrapped themselves only in blankets. The protest esca-
lated, ultimately resulting in the hunger-strike of 1981. In an attempt to re-
gain political status, 10 Irish Republican prisoners fasted to death (see Bell,
1993). The behavior of the Republican prisoners on the strike may be un-
derstood in terms of a role identity and role behavior. Hunger-striking has a
lengthy history among Republicans; the hunger-strike weapon is part of the
Republican repertoire. The first Republican to die on a hunger strike in this
century was Thomas Ashe, in 1917. There were deaths by Republicans on
hunger strike in the 1910s, 1920s, 1940s, and the 1970s. People like Ashe, it
can be argued, were role models. One dimension of the role identity of be-
Social and Role Identities and Political Violence
151


ing an Irish Republican is to confront the state while in prison with the
weapon of the hunger strike. Billy McKee, the senior Republican hunger
striker in Belfast in 1972, was active in the IRA in the 1940s when then 
senior Republicans were on hunger strike. Some of those on the hunger
strike in 1981 were under the command of people like McKee (if not Mc-
Kee himself) when they were new recruits to the Republican Movement in
the early 1970s. 
The hunger strike in 1981 led to a mass mobilization of the general 
Nationalist community in Ireland. One of the strikers, Bobby Sands, was
elected as an abstentionist MP to Westminster. The chief beneficiary of the
hunger strike was the Republican movement. Both Sinn Féin and the IRA
attracted large numbers of new recruits (see Bell, 1993; White, 1993). For
the British government, who held firm against the hunger strike, it was a
disaster. Ulsterization probably saved the lives of British soldiers, but it, like
criminalization, did not stop Irish Republicans, who claim that they ‘target
the uniform’ rather than the religion of members of the security forces (see
Flackes & Elliott, 1989, p. 414, table 4). While killing a British soldier may
have more of an impact, Republicans were quite willing to kill members of
the RUC (see White, 1997a, 1997b; Bruce, 1997)). 
Criminalization and Ulsterization had very different results for Protes-
tant paramilitaries. Criminalization did not lead to deaths from hunger
strikes by Protestant paramilitary prisoners. Although there have been
Protestant paramilitary prisoners on hunger strike, none have maintained a
fast to the death. There are a number of reasons for this and two are impor-
tant here. First, it must be acknowledged that the conditions for Protestant
paramilitary prisoners have never been as severe as they have been for Irish
Republican paramilitaries. Pro-British Protestant paramilitary prisoners are
less of a threat to the state than are anti-state Republican prisoners (see
White, 1999). Second, and very important for present purposes, there is no
history of hunger striking to the death by Protestant paramilitaries. The
historic and contemporary “role models” (martyrs) that Irish Republicans
have—Thomas Ashe in 1917, Sean McCaughey in 1946, Michael Gaughan
in 1974, and Bobby Sands in 1981, for example—are not available to
Protestant paramilitaries. The role identity of a Protestant paramilitary
prisoner confronting the British government by dying on hunger strike
does not exist.
Indeed, at the beginning of the violence a Protestant paramilitary role
identity could not exist because such people, and such organizations, had
not yet formed. When the Ulster Volunteer Force and the Ulster Freedom
Fighters were created in the mid-1960s and early 1970s, a previous genera-
tion of Protestant paramilitaries was not there to serve as paramilitary lead-
ers (role models). The original UVF was largely recruited into the security
forces of Northern Ireland in the 1920s. That organization ceased to exist,
and its former members became police agents of Northern Ireland (see Far-
rell, 1980, pp. 35, 44–45). While some of the parents of the Protestant para-
152
The Contribution of Social Identity to Violent Intergroup Conflict


military recruits of the 1960s and 1970s were members of the RUC and the
Ulster Defence Regiment (UDR; e.g., T
aylor, 1999, p. 82), this is qualita-
tively different from being raised in a family with a tradition of involvement
in illegal paramilitary activity. A leader of a vigilante group described their
activity at the beginning of the current violence:
We never planned to go on the kill. There was no time that we sat down and said
‘That’s it. Stiff a T
aig.’ Mind, we planned doing something to the Ardoyne af-
ter the three Scots boys [three British soldiers] were killed [in 1971] but for
some reason it never came off. No, it was ground up. One or two volunteers
just started doing it. (taken from Bruce, 1992, p. 54; emphasis added)
Contrast this with the description of Billy McKee’s activities in the early
1970s. The interviewer, the journalist Peter T
aylor, asked a Republican who
was recruited back into the IRA by McKee to comment on his activities in
the early 1970s:
What did he [McKee] tell you about the policy towards British soldiers?
He didn’t indicate that there was going to be an immediate offensive
against the British army. He said, “These things take time. People have to be
trained. People have to be motivated. People have to be equipped. All this won’t 
happen overnight.” But the intention was there and it sounded good to me.
(T
aylor, 1998, p. 74; emphasis added)
For Protestant paramilitaries, criminalization reinforced their interpre-
tation that Republicans are out to destroy their community. While Republi-
cans state that they only target uniforms, not the ethnic/religious belief of
the occupant, Protestants (paramilitaries and in general) view attacks on the
RUC as sectarian because the force is 90 to 95 percent Protestant. In re-
sponse to the killing of members of the RUC, Protestant paramilitaries
killed Catholic civilians (see White, 1997a, 1997b; Bruce, 1997).
The accounts from Protestant paramilitaries suggested that they do not
interpret their behavior in terms of a role identity. However, these accounts
are from secondary sources and social and role identities were not the focus
of those who collected the accounts. Also, the accounts are from activists
who were not born and raised in Protestant paramilitary families. The vio-
lence in Northern Ireland has spanned a generation. Many of the Protes-
tant (and Republican) paramilitaries of the 1970s and early 1980s now have
adult children. It is very possible that with the continuity of the Protestant
struggle there has developed among Protestant paramilitaries a role identity
that has been passed to another generation in the same manner as Irish Re-
publicans pass on their role identity. T
oday, the leader of the Ulster Demo-
cratic Party, the political wing of the Ulster Defence Association, is Gary
McMichael. Gary McMichael is the son of John McMichael, a leading
Protestant paramilitary who was killed by the IRA in 1987 (Cusack & Mc-
Donald 1997, p. 267; Bruce, 1992, pp. 232–233). If a Protestant paramili-
tary role identity has developed, this identity should be consistent with the
Social and Role Identities and Political Violence
153


Protestant social identity, and the basic social identity differences among
Protestant paramilitaries, Irish Republican paramilitaries, and the British
should remain. 
Conclusion
Much of the violence in the world today appears to be caused by heightened
social identities, by groups that define themselves as “us versus them.” On
the surface, violence in Northern Ireland fits this pattern. The above case
study reveals, however, that there are three significant violent actors in
Northern Ireland, not two. These actors are Protestant paramilitaries, Irish
Republican paramilitaries, and the state’s security forces. Each of the actors
interprets the violence in different social identity terms. Protestant para-
militaries see themselves as British and they define themselves as Protes-
tants in contrast to Irish Catholics. Further, they perceive that Irish Repub-
licans act in a manner consistent with an Irish Catholic identity. Irish
Republicans, in contrast, act in a manner consistent with an Irish social
identity. Their conflict is not with Irish Protestants but with British parlia-
ments that lay claim to Irish territory. They view themselves as being in a
war of national liberation against the British government. Protestant para-
militaries interpret Republican activity differently. For Protestant paramili-
taries, targeting the British in Northern Ireland means targeting Protes-
tants in Northern Ireland—they view themselves as the British in Northern
Ireland. In the midst of these different interpretations, British people have
traditionally viewed the violence in Ireland from an outsider’s perspective.
The violence is between Protestant Irish and Catholic Irish, and the British,
as the lawful government, are caught in the middle. This adds to the com-
plexity of the situation, especially because British government actions are
subject to different interpretations by Protestant paramilitaries and Irish
Republican paramilitaries.
In addition to their Irish social identity, the data suggest that there is an
Irish Republican role identity and that this influences Republicans’ behav-
ior. This role identity develops through social interaction with other Re-
publicans. Although the data did not necessarily indicate the development
of a Protestant paramilitary role identity, there is no reason to believe that
such an identity has not developed over time. Similarly, British elites and
British soldiers probably have role identities. As with Republicans, these
role identities probably develop through social interaction. 
At the moment, there is a peace process underway in Northern Ireland.
It has been a long time coming, with origins that date from events in the
mid- to late-1980s. As of yet, peace has not been fully confirmed. Involved
in this peace process are the three general actors that are the subject of this
paper—Protestant paramilitaries, Irish Republican paramilitaries, and the
British government (also involved are several other actors, including non-
154
The Contribution of Social Identity to Violent Intergroup Conflict


violent political activists and the Irish Government in Dublin). The differ-
ent perceptions that the actors have of themselves and their counterparts
contribute to the difficulty of achieving a lasting peace in Ireland.
The Northern Ireland conflict is hundreds of years in the making. The
conflict is not unique. Many of the conflicts that attract our attention today
are decades, if not centuries, old. Serb versus Croat and Hutu versus Tutsi
are centuries old conflicts. There are more than two violent actors in
Northern Ireland. Other violent settings also involve more than two actors,
if not currently then historically. The British were key actors in the carving
up of Palestine. T
oday, the United States plays a key role in that conflict, as
Arab states and Palestinian paramilitaries are aware. Similarly, for the vic-
tims of the bombing of Kosovo (whether or not it was justified), the United
States and its allies were significant actors. An in-depth investigation of
many conflicts that appear to involve two actors engaged in “them versus
us” behavior may show them to be as complex as the Northern Ireland 
conflict.
Acknowledgment
This research was supported by grants from the Harry Frank Guggenheim Founda-
tion, the National Science Foundation (grant SES-8318161), and the IUPUI Office
of Faculty Development. I thank Stephen Heathorn, Shel Stryker, Peggy Thoits,
Eric Wright, the participants in the Symposium, and the reviewers for their com-
ments.
Notes
1. My use of the term “political violence” follows Tilly’s conception of the term
“any observable interaction in the course of which persons or objects are seized or
physically damaged in spite of resistance” (Tilly, 1978, p. 176). Political violence in
Ireland is not confined to Northern Ireland. The Irish Republican Army recruits
and carries out operations on an all-Ireland basis. Protestant paramilitaries have
killed people in the Irish Republic, and the IRA has been active throughout Ireland,
England, and on the European continent. Included in the Sutton (1994) data in
table 6.1 are persons who were killed outside of Northern Ireland, in events related
to the conflict there. By far the vast majority of persons killed because of the North-
ern Ireland conflict were killed in Northern Ireland (see White, 1997a). 
2. These actors are “broadly” defined because there are in fact several different
Republican, Protestant, and security force organizations active in Northern Ireland.
These include the Irish Republican Army, the Irish National Liberation Army, the
Ulster Volunteer Force, the Ulster Freedom Fighters, the British army, and the
Royal Ulster Constabulary. 
3. This is too simple a description of a complex situation. While Irish people are
in general “thems” to Britishers (as described below), there are probably degrees of
“themness.” In particular, one group of “them”—Ulster Protestants—identifies with
Britain, is devoted to the British crown, proudly flies the Union Jack, etc., while an-
Social and Role Identities and Political Violence
155


other group of “them”—Irish Nationalists—identifies with Ireland, rejects the
British crown, flies the Irish flag, etc. While Irish people may be “them,” some of
“them” are likely to be more tolerable than others. This helps to explain differences
in state repression, whereby Republican paramilitaries experience more repression
at the hands of the British than do Protestant paramilitaries (e.g., White, 1999). It
also helps explain the much larger number of Catholic civilians versus the relatively
small number of Protestant civilians killed by the security forces. Catholic civilians
come from the dis-loyal “them” community. They are subject to much more state
repression (see White & White, 1995).
4. “T
aig” is a slang term for a Catholic person. It is typically used by Protestants
in a derogatory fashion. It is a corruption of the Irish first-name, “T
adhg.”
5. In order to demonstrate the development of an Irish identity I present infor-
mation from two respondents who lived in England prior to their joining the IRA.
The life stories of these two respondents are such that their understanding of their
Irishness was practically forced on them, which makes for a good presentation. The
other Republican respondents (Respondents 3, 4, and 5), developed an Irish social
identity without first living in Britain.
6. Determining when a person acquires a role identity is difficult. Some people
may acquire role identities only after joining an organization and formally moving
into the role. Others may acquire role identities from interaction with persons in
roles prior to their actually joining a given organization. In a sense, they may find
themselves in the role, and then join the organization. Imagine the potential recruit
who goes out on an operation or two with his friends and quickly finds herself a 
regular member of an IRA unit. Similarly, consider the child of a Republican who, at
the age of 18, is interned under the assumption that he is a member of the IRA, but
is not. Interred without trial with other Republican prisoners, having been brought
up in Republican surroundings, this individual probably knows the role—or learns it
very quickly—of being an IRA prisoner, adapts quickly to prison, and joins the IRA.
References
Arthur, M. (1987). Northern Ireland: Soldiers talking. London: Sidgwick & Jackson.
Bell, J. B. (1979). The secret army: The IRA 1916–. Dublin: Irish Academy Press.
Bell, J. B. (1987). The gun in politics: An analysis of Irish political conflict, 1916–86. New
Brunswick, NJ: Transaction.
Bell, J. B. (1993). The Irish troubles: A generation of violence. New York: St. Martin’s.
Bruce, S. (1992). The red hand: Protestant paramilitaries in Northern Ireland. Oxford:
Oxford University Press.
Bruce, S. (1994). The edge of the Union: The Ulster Loyalist political vision. Oxford: 
Oxford University Press.
Bruce, S. (1997). Victim selection in ethnic conflict: Motives and attitudes in Irish
Republicanism. Terrorism and Political Violence, 9, 56–71.
Cairnes, E. (1982). Intergroup conflict in Northern Ireland. In H. T
ajfel (Ed.), Social
identity and intergroup relations (pp. 277–297) Cambridge: Cambridge University
Press.
Cronin, S. (1980). Irish nationalism: Its roots and ideology. Dublin: Irish Academy
Press.
Cusack, J. & McDonald, H. (1997). UVF. Dublin: Poolbeg Press.
156
The Contribution of Social Identity to Violent Intergroup Conflict


Drake, C. J. M. (1998). The role of ideology in terrorists’ target selection. Terrorism
and Political Violence, 10 (Summer), 53–85.
Farrell, M. (1980). Northern Ireland: The orange state. London: Pluto Press.
Flackes, W. D., & Elliott, S. (1989). Northern Ireland: A political directory. Belfast:
Blackstaff Press.
Fraser, A. (1973/1997). Cromwell: Our chief of men. London: Weidenfeld and Nicolson.
Hyndman, M. (1996). Further afield: Journeys from a Protestant past. Belfast: Beyond
the Pale Publications.
Klandermans, B., & de Weerd, M. (2000). Group identification and political
protest. In S. Stryker, T. J. Owens, & R. W. White, Self, identity, and social move-
ments. (pp. 68–90) Minneapolis: University of Minnesota Press.
Lyons, F. S. L. (1973). Ireland since the famine. London: Fontana.
Manchester, W. (1983). The last lion: Winston Churchill, 1874-1932, Visions of glory.
New York: Little, Brown and Co.
Maudling, R. (1978). Memoirs. London: Sidgwick & Jackson.
McCall, G., & J. L. Simmons. (1978). Identities and interactions. New York: Free
Press.
McGarry, J., & O’Leary, B. (1995). Explaining Northern Ireland. Oxford: Blackwell
Publishers.
McKittrick, D., Kelters, S. Feeney, B., & C. Thorton. (1999). Lost lives: The stories 
of the men, women, and children who died as a result of the Northern Ireland troubles.
Edinburgh: Mainstream Publishing.
Moxon-Browne, E. (1983). Nation, class and creed in Northern Ireland. Aldershot:
Gower Publishing.
Nelson, S. (1984). Ulster’s uncertain defenders: Protestant political, paramilitary, and
community groups and the Northern Ireland conflict. Belfast: Appletree Press; 
Syracuse, NY: Syracuse University Press.
O’Leary, B., & McGarry, J. (1993). The politics of antagonism: Understanding Northern
Ireland. London: Athlone Press.
O’Malley, P. (1990). The uncivil wars: Ireland today. Boston: Beacon.
Rose, R. (1971). Governing without consensus: An Irish perspective. Boston: Beacon
Press.
Stevenson, J. (1996). “We wrecked the place”: Contemplating an end to the Northern Irish
troubles. London: The Free Press.
Stryker, S. (1968). Identity salience and role performance: The relevance of 
symbolic interaction theory for family research. Journal of Marriage and the
Family, 30, 558–564.
Stryker, S. (1980). Symbolic interactionism: A social structural version. Menlo Park, CA:
Benjamin/Cummings.
Stryker, S. (2000). Identity competition: Key to differential social movement partic-
ipation? In S. Stryker, T. J. Owens, & R. W. White, Self, identity, and social move-
ments (pp. 21–40). Minneapolis: University of Minnesota Press.
Sutton, M. (1994). An index of deaths from the conflict in Northern Ireland, 1969–1993.
Belfast: Beyond the Pale Publications.
T
ajfel, H. (1981). Human groups and social categories: Studies in social psychology. Cam-
bridge: Cambridge University Press.
T
aylor, P. (1999). Loyalists. London: Bloomsbury.
Thatcher, M. (1993). The Downing Street years, 1979–1990. New York: Harper-
Collins
Social and Role Identities and Political Violence
157


Thoits, P. A., and Virshup, L. K. (1997). Me’s and we’s: Forms and functions of 
social identities. In R. Ashmore & L. Jussim (Eds.), Self and identity: Fundamental
issues. New York: Oxford University Press.
Tilly, C. (1978). From mobilization to revolution. Reading, MA: Addison-Wesley.
Turner, J. C., Oaks, P. J. Haslam, S. A. & McGarty. C. (1994). Self and collective:
Cognition and social context. Personality and Social Psychology, Bulletin 20,
454–463.
Turner, J. C., with Hogg, M. A., Oakes, P. J., Reicher, S. D., & Blackwell, M. S.
(1987). Rediscovering the social group: A self-categorization theory. Oxford: Basil
Blackwell.
White, R. W. (1993). Provisional Irish Republicans: An oral and interpretive history.
Westport, CT: Greenwood Press
White, R. W. (1997a). The Irish Republican Army: An assessment of sectarianism.
Terrorism and Political Violence, 9 (Spring), 20–55.
White. R. W. (1997b). The Irish Republican Army and sectarianism: Moving 
beyond the anecdote. Terrorism and Political Violence, 9, 120–131.
White, R. W. (1999). Comparing state repression of pro-state vigilantes and anti-
state insurgents: Northern Ireland, 1972-75. Mobilization, 4, 189–202.
White, R. W., & Fraser, M. R. (2000). Identities and long term movement 
activism: Republican Sinn Féin. In S. Stryker, T. J. Owens, & R. W. White 
(Ed.), Self, identity, and social movements. (pp. 324–346). Minneapolis: University
of Minnesota Press.
White, R. W. & White, T. F. (1995). Repression and the liberal state: Northern 
Ireland, 1969-72. Journal of Conflict Resolution, 39, 330–352.
Whyte, J. (1990). Interpreting Northern Ireland. Oxford: Clarendon Press.
158
The Contribution of Social Identity to Violent Intergroup Conflict


Ervin Staub
7
Individual and Group Identities in
Genocide and Mass Killing
Mass killing and genocide, in which a group turns against another with sys-
tematic and widespread violence, has been a tragic aspect of life in the 20th
century. How does such violence, especially its most extreme form, geno-
cide, come about? What is the motivation for such violence and how does it
evolve? What are the social conditions, characteristics of cultures, and psy-
chological processes of individuals that make it more likely? Most impor-
tant for the present volume, what role do individual and group identities
play in genocide and mass killing?
The U.N. Genocide Convention has defined genocide as “acts commit-
ted with intent to destroy, in whole or in part, a national, ethnic, racial or
religious group.” I see the words “in part” in the definition above as prob-
lematic. When some people in a group are killed, for example, to intimidate
the whole group, or when members of the population are killed in the belief
that this will eliminate support for guerrillas, as in Guatemala, this may best
be considered mass killing, not genocide. However, genocide and mass
killing have similar origins and prevention requires similar actions. Because
some nations objected to the inclusion of political groups, they have not
been included in the Genocide Convention. Some authors have called vio-
lence that aims to eliminate a political group, “politicide” (Harff & Gurr,
1990) while others, including myself, prefer the term “genocide”. Influ-
ences leading to genocide against political and other kinds of groups and
the preventive efforts required seem to be the same (Staub, 1989b).
I define “mass killing” as killing many members of a group without the
intention to eliminate the group (Staub, 1989b). Such killings can be part of
an evolution toward genocide. It is important for both understanding and
159


prediction, and therefore prevention, that similar influences lead to mass
killing and genocide. While the likelihood of group or collective violence is
predictable, its exact nature, whether it will be mass killing or genocide,
may not be.
In this chapter I first outline a conception of the origins of genocide and
other collective violence that I have developed over the last 20 years and
have applied to the analysis of a number of specific instances of genocides
and mass killing (see Staub, 1989b, 1996a, 1999b). I next examine elements
of this conception in detail as I extend it by exploring the role of identity in
the generation of such violence. I will look at different types of and aspects
of identities: individual identity and varied forms of it; social identity, the
extent to which a person’s identity is based on membership in the group;
group self-concept, or how the members see and define their group. I will
also discuss the psychological and social processes through which people
strengthen their identities in difficult times, often in destructive ways.
One of the important reasons for understanding the origins of genocide
and mass killing is to find ways to prevent them. Thus, in the last section 
of the chapter I will consider prevention and the role of identities in 
prevention.
The Origins of Genocide and Mass Killing
Over the years I have developed a model of the origins of genocide and
mass killing (Staub 1989b, 1996b, 2000). This is depicted schematically in
figure 7.1. Genocide or mass killing usually begins with some form of “in-
stigation.” Instigation may have little to do with the group that ends up as
the victim. Difficult life conditions in a society are such instigators: severe
economic problems, such as inflation, depression, and unemployment; po-
litical disorganization or political chaos; and rapid social changes.
Difficult life conditions usually frustrate fundamental human needs—
what I have called “basic needs.” All of the following psychological needs
may be frustrated by difficult life conditions: the need for security, for a
positive identity, for effectiveness and control, for connection to other peo-
ple, and for a meaningful comprehension of reality and of one’s own place
in the scheme of things. These needs are viewed here as universal, even
though culture may shape their strength, manifestation, and available av-
enues for fulfillment. Especially when these needs are reasonably fulfilled,
another important need also enters—the higher need for spirituality or
transcendence, the need to go beyond a focus on one’s own self, needs, and
concerns (Staub, 1989b, 1996a, 1998a, 1999a, 1999b, 1999c).
Conflict between groups is another instigator. One type of conflict has to
do with vital needs, such as territory needed for living space. Conflicts can
be intractable, often for psychological reasons, such as a group’s identity be-
160
The Contribution of Social Identity to Violent Intergroup Conflict


ing tied to living in a certain territory. Another type of conflict occurs when
a powerful group limits the opportunities and resources of a subordinate,
powerless group.
Group conflict also frustrates basic needs. What needs it affects will de-
pend on specific conditions, but it is likely to affect security and identity.
Conflict between superordinate and subordinate groups, as well as the un-
even effects of difficult life conditions on different groups, may give rise to a
sense of injustice. While justice may not be a fundamental need in the same
way as the needs described to this point, it is nonetheless a powerful human
motive.
The instigation and the resulting frustration of basic needs leads to psy-
chological and social processes that turn a group against others. These in-
clude individuals turning to some group for identity; people elevating the
group by devaluing other groups and taking action against their members;
the ingroup scapegoating outgroups; and ideologies or visions of social
arrangements that promise a better life but also identify enemies.
One example is Rwanda. All the instigators were present there. There
were increasing economic problems, starting in the late 1980s. The inva-
sion of a Tutsi-led group of rebels in 1990, consisting of refugees from vio-
lence against Tutsis in previous decades or the children of refugees, began a
civil war. Political demands by varied groups added to political chaos and
disorganization. The Tutsis were scapegoated for all the problems. An in-
tense ideology of “Hutu power” was developed and propagated, with the
Individual and Group Identities in Genocide and Mass Killing
161
figure 7.1 Influences and Processes Contributing to Genocide and Mass Killing


Tutsis defined as enemies to be destroyed (des Forges, 1999; Prunier, 1995;
Staub, 1999b).
These psychological and social processes often lead to action against the
devalued, scapegoated group, or a group that is seen as an ideological en-
emy, and thus is the starting point for an evolution toward possible mass
killing or genocide. Individuals and groups change as they harm others.
They justify their actions by devaluing their victims more and more. They
become desensitized to the suffering of their victims. The standards of 
acceptable social behavior change, allowing and even encouraging violence.
Institutions change or new ones are established that serve discrimina-
tion and violence. In the end, there may be a reversal of morality: killing 
the scapegoat or ideological enemy may become a “higher good” (Staub,
1989b).
Unless halted by bystanders, the violence evolves in intensity. Unfortu-
nately, both internal bystanders (members of the involved groups) and ex-
ternal bystanders (other nations and the international community) tend to
remain passive. This affirms and encourages the perpetrators. At times 
outsiders—groups and nations—support the perpetrators because of past
ties or a desire to promote what they define as their own national interests.
When the kind of psychological and social processes here described do not
lead to genocide, usually bystanders are involved in a positive way—they in-
tervene in some way to alter the processes leading to collective violence
(Staub, 1989b, 1999b, 2000).
Certain characteristics of cultures make the genocide or mass killing
process more likely. A history of devaluation of a group of people is a pri-
mary characteristic. Very strong respect for authority and a monolithic soci-
ety with a limited range of values, in contrast to a pluralistic society, are also
important. Unhealed wounds in a group owing to past victimization, which
lead members of the group to feel easily threatened and respond with what
they see as defensive aggression, make it more likely that such a genocidal
process unfolds in response to instigation. The nature of individual identi-
ties that the society has fostered and the nature of group self-concept are
also important. 
The influences here described do not act separately and independently.
They interrelate and join in meaningful ways. I will examine here some as-
pects of this interrelation; others can be seen and deduced from the subse-
quent discussion (see figure 7.1 and Staub, 1989b). It is the frustration of
basic needs by instigating conditions that leads group members, whose indi-
vidual identity is shaken, to turn to the group for identity, to focus more on
their social identity, or to “give themselves over” to an identity group. This
frustration also leads to scapegoating and the creation of destructive ideolo-
gies (which identify enemies), that turns the group against another group.
The kind of ideology that is adopted, and the scapegoat and ideological en-
emy that is selected, very much depend on characteristics of the culture. 
Whether the group turns against another depends in part on whether 
162
The Contribution of Social Identity to Violent Intergroup Conflict


the devaluation of some group is part of the culture, the nature of that de-
valuation, and how intense it is (Staub, 1989b, 1999b). It depends on past
victimization of the group and unhealed wounds, which make it more vul-
nerable to threat. It depends on the strength of “authority orientation,”
how deeply and unconditionally people are accustomed to be led, to respect
and obey authorities and the extent they are accustomed to gain support
from their group. The stronger this orientation, the more they will be af-
fected by difficult life conditions and other instigators, and the more diffi-
cult it will be to stand on their own. Their need to turn to new authorities
who might be able to guide and support them will be greater.
A history of devaluation of a group, which all members of a society tend
to absorb to some extent, a strong authority orientation, and a monolithic
rather than pluralistic-democratic culture, as well as other cultural charac-
teristics (Staub, 1989b), all make it less likely that people will oppose initial
actions against a victim group, or initial “steps along the continuum of de-
struction.” The passivity of bystanders has the effect of allowing and even
encouraging the evolution of violence. 
The psychological processes that initiate and maintain the increasing 
violence are fostered both by intense instigators and the presence of cul-
tural characteristics. The social conditions that constitute the instigators af-
fect both perpetrators and internal bystanders. As the violence evolves, it
constitutes culture change; further evolution happens in a changing and
changed cultural context. 
Identity and the Origins of Mass Killing and Genocide
I now describe how individual and group identities are involved in the paths
to genocide or mass killing depicted in figure 7.1. T
o begin, severe eco-
nomic problems deeply impact people. Unemployment, or the inability to
provide for oneself and one’s family owing to severe inflation, makes people
feel insecure and ineffective. Such conditions call into question people’s
personal identity and self-worth: Who am I if I have no work and cannot
take care of the basic material needs of myself and my family? At the same
time that identity is questioned, ties to one’s group can also weaken because
the problems originate in society and social conditions.
Intense political conflict and disorganization also affect people’s experi-
ence of their group and perception of the world. Rapid social change, which
often occurs at times of economic and political crises, adds to social chaos
and disorganization. Since substantial social change requires adjustment,
even positive change places demands on people. 
The different elements of difficult life conditions, which may occur sepa-
rately but frequently are present in some combination, frustrate many of
the basic needs. In difficult times, focusing on their own needs, people be-
come disconnected from each other. At times when they need it most, peo-
Individual and Group Identities in Genocide and Mass Killing
163


ple lack connection and support. The social upheaval makes their old
worldview inadequate in providing a comprehension of reality: an under-
standing of how people, society, and the world operate, and of their own
role or place in the scheme of things. 
As a result, when a society experiences persistent, intense life problems,
its members desperately need a renewed basis for identity, security, a sense
of effectiveness, connection to other people, and comprehension of reality.
They turn for identity and the fulfillment of other basic needs to a new
group, or to leaders who strengthen the group through scapegoating out-
siders or offer it new identity, meaning, and hope through ideologies that
promise a better life and offer the fulfillment of basic needs.
Identity, Scapegoating, and Ideology
Three types of “identities” are discussed in this chapter: first, personal identi-
ties, or the ways in which individuals answer the question, who am I?; sec-
ond, social identities, or the extent to which individual identity is rooted in,
or connected to, the group; third, group self-concepts, or the socially shared
way members perceive and experience their group. Relevant to all of them
is a group’s culture and the social and political institutions that both express
and shape that culture. The nature of individual identity is shaped by cul-
ture; in turn, the identities that characterize individuals maintain the cul-
ture. Social identity theory has focused on how people categorize them-
selves (and others) as members of groups, how they identify themselves
with certain groups, and the tendency they have to enhance themselves by
comparing their group favorably with other groups (T
ajfel, 1982; Turner,
1987; Myers, 1999). It emphasizes the psychological experience of rooted-
ness in and connection to the group. 
But social identities also have content, as defined by the culture. “We”
are intelligent and hardworking, or easygoing and ready to enjoy life, or
brave and ready to fight for our rights or for maintaining important values.
The content of social identity is what I refer to as group self-concept.
While social identity theory stresses the tendency to favorably evaluate
one’s group, group self-concept can be positive or negative (superior or
weak and vulnerable). As in the case of individual identity, consciously held
views and evaluations of the group can exist side by side with feelings about
and valuations of the group that may not be consciously held, such as a
sense of vulnerability or weakness.
Individual identity is, in part, a social identity rooted in groups to which
one belongs ranging from family to nation (T
ajfel, 1978, 1982; Turner,
1987; Bar-T
al & Staub, 1997). It is not surprising, therefore, that in difficult
times people try to strengthen their identity through personal identification
with some group. A good social identity—an identity based in a group that
is seen in a positive light, for example, as effective and powerful—can also
164
The Contribution of Social Identity to Violent Intergroup Conflict


be a good source of security, feelings of effectiveness and control, and con-
nection. The values and worldview of the group also provide a comprehen-
sion of reality. 
Scapegoating some other group can strengthen both individual and
group identity and also help fulfill other needs. If another group is viewed
as responsible for life problems, then neither I nor my group are to blame.
Scapegoating at such times is usually a social process in that it is done to-
gether with like-minded others. This not only builds identity but strength-
ens security and connection. 
Ideologies are almost always a part of the genocidal process. I see ideolo-
gies as visions of social arrangements that tell people how to live life. They
are blueprints for the organization of societies and relations among indi-
viduals. The ideologies created or adopted in difficult times offer the satis-
faction of basic needs. They promise people a better life. Sometimes their
stated purpose is to improve life for all human beings—what I have called
“better world ideologies.” Communism has been such an ideology. Nation-
alistic ideologies promise better life to people in a particular group by en-
hancing the group’s wealth, power, or purity (Staub, 1989b).
Often, ideologies have both better life and nationalistic components.
This was true of the Nazi ideology. The Khmer Rouge ideology was osten-
sibly a better world ideology, focusing on total social equality. But it was
group based, looking to the group’s history as a simple agricultural society
for ways to create this social equality. Without explicit elaboration it also
included a strongly nationalistic agenda. This expressed itself in the espe-
cially intense violence against the Vietnamese in Cambodia, in hostility to
Vietnam, an ancient enemy, and in genocidal violence against other minori-
ties in Cambodia. 
Legitimizing ideologies (Levin, Sidanius, Rabinotitz, & Frederico, 1998;
Staub, 1989b, 1999b) offer justifications for the power and privilege of
dominant groups. Violent reactions by powerful groups to demands by sub-
ordinate groups for rights and privilege have been one of the most frequent
sources of mass killing since World War II (Fein, 1993). In my view, this 
response of dominant groups is not simply to a threat to their power and
privilege. They also respond to a threat to their identity and worldview
(Staub, 1989b). They have come to see themselves as rightfully occupying
positions of power and influence, owing to their background, history, dili-
gence, intelligence, bravery, or other aspects of their presumed nature. The
world is arranged the “right way” with them in power (see Sidanius &
Petrocik, this volume).
People usually hold or turn to and adopt ideologies as part of a group or
a movement. At times some people create or put together ideas that form
the ideology, as did Hitler (1925) or the leaders of the Khmer Rouge in
Cambodia (Staub, 1989b). However, it is when shared among a group of
people that an ideology comes to life. The ideology defines the group and
gives the group and its individual members a positive social identity. It also
Individual and Group Identities in Genocide and Mass Killing
165


provides a basis for connection. Participating in the movement can make
people feel effective and in control. The goals of the movement become the
motivating force for members. The ideology and the movement that arise
from it often become central sources of people’s identity and comprehen-
sion of reality.
While human beings need positive visions in difficult times, the ideolo-
gies that emerge usually identify another group as an enemy. This group is
seen either as the cause of problems, or as standing in the way of the ide-
ology’s fulfillment, or both. They, therefore, become the objects of hostility.
Harmful actions begin to be directed at them. 
The Role of Individual Identities in Genocide
Certain individual identities that a culture promotes make the genocidal
process more likely. Research and theory have identified differences in the
type of self-concept or individual identity that is shaped by individualist and
collectivist cultures (Triandis, 1994). Differences in women’s and men’s
identities in Western societies, including the United States, have also been
noted (Sampson, 1988). It has been proposed, for example, that male iden-
tity is more autonomous, more like the identity shaped by individualistic
cultures, while female identity is more relational, like identities shaped by
collectivist cultures.
Considering conditions that give rise to genocide, both relational/collec-
tivist and autonomous/individualist identities may have vulnerabilities. Dif-
ficult life conditions may especially frustrate autonomous persons. They
have learned to stand on their own, but now they are unable to effectively
fend for themselves or their families. American soldiers who became pris-
oners of war in the Korean War had more difficulty in resisting brainwash-
ing than, for example, Turkish soldiers, seemingly because the former were
trying to face it alone while the latter supported each other (Kincaid, 1959).
Relational identities may also create difficulties. People with such identi-
ties may feel interconnected with other people in a way that makes it espe-
cially difficult to speak out against and oppose the direction the group is be-
ginning to take. As their group or society begins to scapegoat, identify an
enemy, discriminate, or take other harmful actions against another group,
they are more likely to remain silent and go along. This makes more proba-
ble the evolution of increasing discrimination and violence. 
I have suggested that relational/collectivist identities may be differenti-
ated into embedded and connected (Staub, 1993a). The embedded identity,
which I just described, is characterized by a strong connection to other peo-
ple that also embodies dependence on the group and an inability to separate
from it. In contrast, people who have developed connection to others as in-
herent to their identities, but who also have learned to stand on their own,
to be separate, have connected selves or identities. 
The nature of personal identity has implications for peoples’ relationship
166
The Contribution of Social Identity to Violent Intergroup Conflict


to their group. People with connected selves may be less likely to give
themselves up to the group in difficult times. They may be less likely to re-
linquish their own identity and give themselves over to a social or group
identity than people with embedded selves. Seemingly paradoxically, be-
cause they can use others for support more effectively, those with connected
self-concepts may be even less likely to do this than people with autono-
mous/individualist identities. As a result, people with connected selves may
be less likely than those with either embedded or autonomous identities to
remain passive as the evolution of harm-doing begins. They may be more
likely to speak out and oppose actions contrary to their own values, or to
the values they believe the group ascribes to. This is extremely difficult 
to do but is essential in halting the evolution of increasing violence. 
I am suggesting that identity-related differences in individuals’ relation-
ship to the group have implications for group violence. Related to this, I
have proposed that some individuals are “constructively” patriotic while
others are “blindly” patriotic (Staub, 1997). In research exploring this we
find that constructive patriots express love for their group, but also a will-
ingness to question and criticize the policies and practices of the group that
are contrary to caring values and concern for human welfare, or contrary to
what they see as important group values (Schatz & Staub, 1997; Schatz,
Staub, & Lavine, 1999). Although this has not been assessed in the research,
I expect that constructive patriots would have connected identities. In con-
trast, blind patriots, who I expect to have either individualist or embedded
self-concepts, love their group but are unwilling to criticize it or to question
its policies or practices. Different identity types and associated relationships
to the group have clear implications for supporting or opposing destructive
views and group practices. 
Other classifications of individual identities that are relevant to genocide
are possible. Some people may have secure identities (Staub, 1993a). Pre-
sumably their basic need for security has been fulfilled in the course of their
lives, especially early lives. Perhaps they also have had experiences with tak-
ing effective action to protect themselves. As a result they feel reasonably
secure in the world. They would be less affected by instigating conditions,
such as difficult life conditions, and less moved to the psychological and so-
cial processes that were described as a consequence of such conditions.
Others, perhaps due to the frustration of the need for security early in life,
including victimization or other traumas, may develop insecure identities,
making it more likely that these psychological processes leading to genocide
unfold in response to instigating conditions.
Authority Orientation and Identity
Observation and research indicate that in many cases of genocide or mass
killing the society has been characterized by strong respect for authority
(Gourevich, 1998; Kressel, 1996; Smith, 1998; Staub, 1989b,1999b). “Au-
Individual and Group Identities in Genocide and Mass Killing
167


thority orientation” in the culture is a term I prefer to “authoritarian per-
sonality” because it focuses purely on respect for authority and hierarchy,
without implications of repression of sexual feelings and other characteris-
tics that have been described as aspects of the authoritarian personality.
Authority orientation affects the nature of identity in a number of ways.
Individuals who have embedded selves are connected to others through net-
works of dependent relationships. People in an authority-oriented culture
also seem deeply embedded in the group, but this is based on their orienta-
tion to authority. The experiences they have as this orientation is developed
in them, normally from early childhood on, diminishes the evolution of a
strong independent self. 
In the face of difficult life conditions or group conflict, people who have
relied on leaders for guidance and protection will find it more difficult to
bear the threat, anxiety, and frustration of basic needs. When policies and
practices are instituted in a group that harm others, they will be less willing
to speak out and to oppose the authorities and the rest of the group. Strong
respect for authority also makes obedience to immoral orders by authorities
more likely.
The Nature of Group Self-concept
Weakness, Superiority, and a Positive Self-Concept. The nature of group self-
concept (by which I mean the way members consensually see and experi-
ence their group) is also important. How do members of a group, people
who identify with a group, see their group? What is their conception and
experience of the group? And, how do differences in group self-concept
predispose toward mass killing or genocide? I have hypothesized (Staub,
1989b) that both a sense of weakness and vulnerability, and feelings and be-
liefs of superiority, may make it more likely that groups will turn to and en-
gage in systematic violence against others. When a group feels weak and
vulnerable, life problems in the society or conflict with other groups will in-
tensify vulnerability and the need to defend oneself. Elevating the group,
scapegoating another group, creating “positive visions” that identify ene-
mies, and self-defense through aggression become more likely. 
A sense of superiority, when frustrated, may also become a powerful
source of psychological processes and actions that contribute to genocide.
Feeling superior makes individuals and groups vulnerable to disconfirma-
tion of their superiority and, as a result, likely to experience narcissistic
wounds (Baumeister, 1997). However, a self-concept of superiority is differ-
ent from a high self-esteem. Feelings of superiority may be the result of a
history of success by a society in many realms combined with the usual eth-
nocentrism of groups that leads to interpretation of this success as evidence
of the group’s superiority. But more often, a belief in one’s superiority is ex-
aggerated and may represent a compensatory identity. In the lives of most
individuals and groups there are enough small and large experiences of inef-
168
The Contribution of Social Identity to Violent Intergroup Conflict


fectiveness, failure, and loss that, under normal conditions, a more moder-
ate sense of self would develop.
A positive identity in individuals is likely to arise from experiences of
nurturance and affirmation and from success in overcoming obstacles.
Comparable experiences on the group level can give rise to positive group
self-concepts. The absence of these may lead to low self-concept. Their ab-
sence, or negative experiences, overlaid with experiences of success but
without incorporating the difficult, painful experiences into one’s identity,
can lead to a feeling of superiority, which is what I regard as a compensatory
identity. In children a defensive self-esteem or defensive egotism, which was
found to be different from high self-esteem, was associated with more bul-
lying of others. Defensive self-esteem was ascribed to children who de-
scribed themselves and were described by others as having positive self-
esteem, but were also described by others as “always wants to be the center
of attention; thinks too much of himself or herself; can’t take criticism”
(Salmivalli, Kaukiasnen, Kaistaniemi, & Lagerspetz, 1999, p. 1271).
Germany may be a good example of having a superior group self-
concept. The group self-concept of Germans early in this century was ex-
tremely positive, of a kind that may be described as a superior compen-
satory identity. This was due to recent military successes, economic
strength, a history of cultural achievement, effective civic institutions, and
an internal climate that propagated self-satisfaction with German culture
and society beyond what is “normal” ethnocentrism (Staub, 1989b). 
These positive experiences might account for what seems was an early
20th-century German sense of superiority. But Germans also had a history
of problems and suffering. Many Germans died in wars in previous cen-
turies. During the Thirty Years’ War (1618–1648), a third of Germany’s
population died (Craig, 1982). After struggles among many smaller units,
Germany became a single country only late in the 19th century. History
leaves its mark, and perhaps underlying the outward sense of group superi-
ority there was also a deep-seated national vulnerability. Germany was then
frustrated by a series of events beginning with the loss of World War I.
These included a revolution that overthrew the monarchy, hyperinflation,
economic depression, political chaos, and actions by other countries that
Germans felt were humiliating to them (e. g., the French occupying the
Ruhr district to ensure the receipt of compensation Germany was forced to
pay after World War I). Whether it was defensive egoism, or normal ethno-
centrism, or even a “genuine” sense of superiority, the nature of their group
identity was likely important in Germans’ reactions to the intense life prob-
lems following World War I. 
A decade ago I wrote: “in groups and in individuals very high self-evalu-
ation often masks self-doubt. Persistent life difficulties may contradict high
self-evaluation and bring self-doubt to the surface. Even when there is no
underlying self-doubt, a very high self-evaluation may be associated with
limited concern for others. Among individuals, a moderately positive self
Individual and Group Identities in Genocide and Mass Killing
169


concept is most strongly associated with sensitivity and responsiveness to
other people” (from Staub, 1989, p. 55; see also Jarymowitz, 1977; Reese,
1961). This still seems a reasonable conclusion. Both individuals and groups
have to value themselves in order to value other people, but not value them-
selves so strongly that others do not matter, or so highly and unrealistically
that setbacks are experienced as devastating blows to collective and personal
self-worth. 
Cultural Devaluation and Ideologies of Antagonism 
A form of group identity that I have called “ideology of antagonism” is es-
pecially important in genocide and collective violence. Inherent in group
identity is a differentiation between “us” and “them.” In my view a devalua-
tion of “them” is not inherent in, but easily arises out of, this differentiation
(see Brewer, this volume). There are many factors that develop a persistent
devaluation of one group by another, which then becomes part of the
group’s culture. They include justification of past mistreatment of the other
group, fear of the other’s difference, and the need to create identity when
one group has split off the other. (The latter may be an important origin of
Christian anti-semitism [Staub, 1989b]). Devaluation embedded in a cul-
ture is a crucial cultural source of a group turning against another.
At times, two groups have had a long history of conflict with mutual vio-
lence. The evolution of increasing violence has already occurred, and as a
result new violence easily flares up. In the course of this past history, not
only the view of the other but the identity of each group as well has come to
be defined around this enmity: the other is my enemy, and I am the enemy
of the other. In short, group self-concept comes to center on violent con-
flict with the other group (see Kelman, this volume).
The Hutus and Tutsis in Rwanda seem to be an example of ideologies of
antagonism. Up to the 20th century, the Tutsis were dominant, but group
membership was defined by a complex combination of ethnic origin and
class, with rich Hutus sometimes acquiring a classification as Tutsis. Under
colonial rule by the Belgians, Tutsi dominance was intentionally enhanced,
with Hutus receiving increasingly bad treatment. They became an under-
class with limited rights and privilege. In 1959, in a violent revolt, they
killed about 50,000 Tutsis. The Hutus ruled the country after it gained in-
dependence in 1962. There were significant massacres of Tutsis in the
1960s and 1970s, with continued discrimination and more limited, sporadic
violence against them after that.
In the 1990s, a Tutsi-led group, consisting mainly of descendants of
refugees over the past few decades, invaded Rwanda. In the course of the
following few years, a complex political and military situation developed in
the country. Dominant Hutus increasingly advocated “Hutu power,” with
media propaganda explicitly or implicitly promoting the destruction of 
Tutsis. The genocide of 1994 followed. It was stopped by the victory of the
170
The Contribution of Social Identity to Violent Intergroup Conflict


Tutsi-led army over the government (see des Forges, 1999; Prunier, 1995;
Staub, 1999b).
Cultural devaluation is extremely persistent, and ideology of antagonism
is even more so. Neither tend to go away on their own, unattended, simply
with the passage of time. They remain part of the culture, expressed in lit-
erature, in everyday relations, and in the life of institutions. They remain
part of the identity of individuals and groups. Active social processes are re-
quired to change them.
Victimization, Healing, and Identity
When a group has been victimized in the past, without significant healing,
it has an increased likelihood of becoming a perpetrator of mass violence
(see Eriksen, this volume). People who experience trauma are deeply af-
fected (Herman, 1992; Janoff-Bulman, 1992; McCann & Pearlman, 1990).
Group victimization diminishes the different selves—the individual self, 
social identity, and group self-concept. Who am I and who are we, if we are
so treated? Since individual identity is partly based on group identity, even
those members of the group who are not present and have not been indi-
vidually in danger are deeply affected.
Victimized groups are more likely to respond to threat with violence. In
case of conflict, their understandable self-focus makes it difficult for them
to consider the needs of others. Feeling insecure, their self vulnerable, and
perceiving the world as a dangerous place, they experience threat as more
intense than it is. They are more likely to engage, therefore, in aggression
which they see as defensive (Staub, 1998b). 
Healing following victimization makes it less likely that a group turns
against another and perpetrates violence. Acknowledgment of the group’s
suffering by others and the expression of caring and empathy contribute to
healing. The group itself can also promote healing by directly engaging
with its past experience, including certain kinds of memorials and rituals of
mourning and remembrance (Staub, 1998b). Healing from both individual
trauma and group victimization involve repair of identity, as well as connec-
tions to other people (Pennebaker & Keough, 1999). The two are inher-
ently related. The reestablishment of trust in people and the capacity to ex-
perience connection to them help repair identity.
Perpetrators also need to heal. At times, their woundedness has con-
tributed to their actions. Their own violence further wounds them. It seems
impossible to inflict tremendous harm on others, to kill many people and
not become wounded or damaged. When perpetrators and victims continue
to live together, as in Bosnia and Rwanda, for reconciliation to take place,
some healing by both groups has to precede it. Reconciliation is required,
in turn, to stop a continuing cycle of violence.
In the case of the Armenians who survived the genocide by the Turks in
1915–1916, healing was interfered with by the denial of Turkey that a geno-
Individual and Group Identities in Genocide and Mass Killing
171


cide had taken place and by the complicity in this denial of many other
countries. The relationships of the recently created Armenia to other na-
tions, especially Turkey and Azerbaijan populated by ethnic Turks, have
been affected by this history and the identity it has helped create. That
identity has components of vulnerability, including a view of the world as
dangerous and, perhaps, an ideology of antagonism in relation to Turks. 
It is difficult to identify victim and perpetrator groups that have healed
and thus see the consequences of such healing. While there has been aware-
ness of the need for individuals to heal following the mass killing in Bosnia
and the genocide in Rwanda, there has been little prior consciousness of the
need for whole groups to heal. There have certainly been natural processes
that have moved groups toward healing, but there have probably been no
intentional efforts specifically aimed at this by victimized groups themselves
or by the community of nations or by other bystanders. 
A consciousness about the need for healing seemed to have emerged, at
least in contemporary times, with respect to individual survivors of the
Holocaust, and later Vietnam veterans, as well as victims of individual vio-
lence such as sexual or physical abuse (Herman, 1992; McCann & Pearl-
man, 1990). The aim of such healing has been to improve individual lives,
not reducing the likelihood of later group violence.
Germany might serve as an example of healing, even if far from com-
plete. Germany was wounded by the sequence of events that contributed to
the Holocaust and by its own horrific actions during World War II and the
Holocaust. Afterwards, the Germans were led to hesitantly engage with
their past through the Nuremberg trials, other trials, and through the rest
of the world reminding Germans of their conduct. At the same time, in-
stead of acting in a generally punitive and harsh way, the world offered sup-
port, help, and connection to Germany. This ranged from the Marshall
plan, to the willingness by French leaders to make connections with Ger-
many, to Germany’s becoming a member of the European Community, and
to other demonstrations of the world accepting Germany into the interna-
tional community. In different ways and for different reasons, this was true
of both West and East Germany. Reconnecting with others is an avenue to
healing for both victims and perpetrators. 
Changes in Identities as Harm Doing Evolves
As noted above and depicted in figure 7.1, genocide and mass killing are
usually the outcome of an evolution of increasing harm doing by perpetra-
tors. There may be one victim group, although in the course of this evolu-
tion the boundaries expand and violence toward other groups also becomes
probable. How do individual and group identities change along the way?
The perpetrators come to devalue members of the victim group more
and more. They justify their actions, in part, by the nature of the victim,
who is not only increasingly devalued but also increasingly excluded from
172
The Contribution of Social Identity to Violent Intergroup Conflict


the moral realm, the realm of people to whom moral values and moral con-
siderations apply (Opotow, 1990; Staub, 1989b). Perpetrators also justify
their actions by the “higher” values of their ideology. Along the way the
perpetrators may come to see themselves as people who are willing to make
the sacrifices required, including the killing of the victims, to fulfill impor-
tant goals. Thus, they may see themselves, as Himmler encouraged the SS
to see themselves, as bearing noble sacrifice for their higher cause (Hilberg,
1961). 
In addition to their identity changing in relation to the victims, perpetra-
tors are likely to become desensitized to human suffering in general. Deval-
uation tends to extend from the original victims to new groups, which 
prepares the way for the extension of violence (Staub, 1989a, 1989b). In
Germany, for example, after the euthanasia program, the killing of physi-
cally handicapped, mentally ill, and other “inferior” Germans, after the
genocide of the Jews and gypsies, and after violence against people in occu-
pied countries, the authorities began to consider the killing of physically
ugly German prison inmates (Lifton, 1986). 
As boundaries weaken and new groups are devalued and become poten-
tial or actual victims, the perpetrators’ identity as people who value human
beings and human welfare is likely to change. As this process unfolds perpe-
trators are likely to distance themselves from all human beings not a part of
their group, however that is defined. At the extreme, might it be the case
that even the welfare of one’s ingroup and of friends and family become of
less concern?
The identity of bystanders also changes. Internal bystanders—members
of the perpetrator group who are not themselves perpetrators—usually re-
main passive. They are also affected by instigating conditions. They have
grown up in a culture that imbues members with respect for authority and
devalues the victim group. Thus, they share a “cultural tilt” with perpetra-
tors. This cultural background, often in combination with fear of a repres-
sive system, reduces the motivation to oppose the group as it moves along a
continuum of increasing violence.
Passivity changes bystanders. It is difficult to see others suffer and do
nothing. T
o justify their passivity, reduce their guilt, and reduce empathy
with victims that would make them suffer, passive bystanders tend to dis-
tance themselves from victims, in part by increasingly devaluing them. This
makes it even more likely that they will remain passive. It also leads some
bystanders to join perpetrators.
For example, a group of psychoanalysts in Berlin remained passive as
their fellow Jewish analysts were the object of harassment and persecution.
They began to use racially appropriate Nazi terms to refer to Jewish col-
leagues, like “not pure Germans.” These served both as euphemisms and as
means to distance and objectify their fellow analysts. They accepted a rela-
tive of Herman Goring, untrained in psychoanalysis, as the head of their In-
stitute. They began to rewrite psychoanalytic theory to fit Nazi ideology
Individual and Group Identities in Genocide and Mass Killing
173


and doctrine. Some of them became participants in the euthanasia program,
and in the end some got involved in the extermination of Jews (Staub,
1989a).
The process in this specific instance may be an example of how by-
standers distance themselves, even from familiar others and from close asso-
ciates. Adopting at first an “objective,” disinterested stance, and introducing
ideological thinking change both the view of and emotional relationship to
the other. Through this process bystanders, as well as perpetrators, change
the nature of their own self. Those changes in identity are intertwined with
changes in society and culture. As violence evolves, the norms and standards
of society change, allowing and often promoting actions that were previ-
ously inconceivable. For instance, in Rwanda, neighbors killed neighbors
and people gave up members of their families to the killers (des Forges,
1999; Gourevich, 1998; Staub & Pearlman, in press).
I have suggested that these changes in perpetrators and bystanders repre-
sent changes in identity. Although identity is a broad concept, we may see it
as including a person’s self-concept and orientation to and evaluation of the
self, as well as values relating to and orientation toward and evaluation of
other people. Thus, I am proposing here that a person’s view of the world is
also part of personal identity (Epstein, 1980). Some or all of these elements
of identity will change in the course of a societal evolution toward genocide.
The changes in a person’s self-concept and in the group’s collective sense of
self mutually reinforce each other.
However, the negative changes in the view of the other, and the changes
in the self, as individuals and a society move along a “continuum of destruc-
tion” (Staub, 1989a, 1989b), are not the whole story. Even in very simple
societies there is variation in personality and, presumably, self-concept.
Only a small percentage of people have the “modal identity type” that may
be seen as characteristic of the society (Konner, 1978). In genocidal soci-
eties, in addition to perpetrators and passive bystanders, there are usually
some people who attempt to help victims. The rescuers of Jews in Nazi Eu-
rope have been the primary group of such helpers who have been exten-
sively studied (Fogelman, 1994; Oliner & Oliner, 1988; T
ec, 1986). But in
other societies as well, there have been heroes who have endangered them-
selves to save lives. 
I have heard several such stories in Rwanda, both in conversation and in
formal interviews I conducted with people whose lives were saved. In one
case, for example, a Hutu man who had worked for a woman sent another
man to help her after her husband was killed. The man who came, a
stranger carrying a Bible, repeatedly endangered himself as he faced off
people who came to the house wanting to take her away. The woman telling
this story was unclear about his motivation to help: she suggested his reli-
gion, his being a good man. Having some difficulty accepting such self-
sacrificial behavior, she wondered about possible selfish motives but could
not identify any.
174
The Contribution of Social Identity to Violent Intergroup Conflict


Once people begin to help in some limited way, they are further shaped
by their own actions. Like violence, so helping evolves. People who have
agreed to help for a short time, or to a limited extent, change as a result of
their actions. They become more committed to the welfare of those they
have helped and often to the welfare of human beings in general. They
come to see themselves as people who are willing to endanger themselves to
help others. Their values and identity change (Staub, 1989b, 1993b). 
The Role of Identity in the Prevention of Genocide and Mass Killing
The Role of Leaders 
How can genocide and mass killing be prevented and what is the role of dif-
ferent kinds of identities in prevention? How can people gain positive iden-
tities in difficult times without turning to destructive groups or making
their groups destructive? One way is for leaders to offer positive visions that
are constructive, visions that offer hope but are realistically aimed at solving
life problems without identifying enemies, visions that connect rather than
separate people. The New Deal under Roosevelt was one example of this.
This can be difficult and at times even dangerous for leaders when there is
already strong enmity against a potential victim group.
Healing by Victimized Groups 
An important contributor to prevention is healing by victimized groups.
Healing can help repair damaged identities and worldviews (as well as fulfill
other frustrated needs). It can recreate some sense of security and trust in
the world. As a result, it makes it less likely that a victim group becomes a
perpetrator. Acknowledgement of the group’s suffering, expressions of em-
pathy and sorrow, and other forms of affirmation by the rest of the world
can contribute to healing. Punishment of perpetrators can also help victims.
It can communicate to victims that what was done to them is not regarded
as normal and acceptable by the world. That message, in turn, makes it
seem less likely that it will happen again (Staub, 1998b, 1999c). 
Engaging with their painful experiences can also help survivors of group
violence heal. Trauma victims frequently avoid such engagement even
though they are often affected by what has happened to them in many ways:
through intrusive imagery, hyper arousal, negative feelings about the self,
mistrust and difficulty in relating to people (Herman, 1992; McCann &
Pearlman, 1990). Engagement can lead to integration of the self and recon-
nection with people (Pennebaker & Keough, 1999). It can lead to finding
meaning, perhaps in helping other people, working for the prevention of
similar violence against people, or living life in a different way. 
Healing makes reconciliation possible. In places like Rwanda and Bosnia,
Individual and Group Identities in Genocide and Mass Killing
175


where victims and perpetrators live together, or where mutual violence has
occurred, reconciliation is essential to prevent continued cycles of violence.
The self is to a significant extent defined by relationships to others. Ideo-
logies of antagonism, intense devaluation and hostility toward another, 
or strong fear of a neighbor are central aspects of identity. Reconciliation,
which when genuine requires forgiveness and overcoming devaluation and
antagonism, changes identity in central ways. As a result of healing and rec-
onciliation, feelings of diminished self-worth and vulnerability can lessen.
Mistrust in people and in the world can also be reduced. Both the desire for
revenge as a form of justice and the readiness to use force to defend oneself
and one’s group can decrease. Trust in people, trust in the world as a livable
place, and hope for the future can reemerge.
How can healing, reconciliation, and positive connection to others in
general be promoted in previously victimized groups? We have so far only a
limited repertoire of skills and techniques for this. T
o begin with, victimized
groups have to start with at least some desire for change, which, in case of
intense antagonism, may be lacking. When there is some openness, con-
texts for healing can be created by bringing members of different groups to-
gether. People can write about and talk to each other about what has hap-
pened to them (Pennebaker & Keough, 1999). They can talk about their
pain, sorrow, loss, and anger. As they listen to each other and respond with
empathy, pain, which has been a source of disconnection, can become a
source of connection (Staub & Pearlman, 1996).
This is one aspect of a project that we have been conducting in Rwanda
on “healing, forgiveness, and reconciliation.” In addition to such experien-
tial aspects of promoting healing and reconciliation, the training we have
used also had psycho-educational aspects. Participants learned about and
discussed the impact of trauma on people in the hope that this would help
them understand their own experience better. They learned about avenues
to healing. They heard about the fulfillment and frustration of basic needs
and they received information and engaged in discussion about the origins
of genocide (Staub, 2000; Staub & Pearlman, in press).
This last element of the training has had seemingly dramatic effects.
Learning about a number of other genocides and about how genocide
comes about seemed to reaffirm for participants their own humanity. If
these terrible things have been done to others as well, and if there are un-
derstandable human processes that give rise to them, then the victims in
Rwanda remain part of the human realm in spite of what happened to them.
In addition, understanding the forces that led to genocide seemed to hu-
manize the perpetrators, both in the survivors’ eyes and, perhaps, in their
own, thereby contributing to reconciliation. The type of training described,
which we conducted in Rwanda, can help not only victims but also mem-
bers of the perpetrator group. Genocide and mass killing, inflicting violence
on others, deeply wounds the perpetrators. It also wounds members of the
perpetrator group who have remained passive bystanders. They also are in
176
The Contribution of Social Identity to Violent Intergroup Conflict


need of healing. Healing, in turn, can open the way to reconciliation. As
reconciliation enhances a feeling of security, it furthers the possibility of
healing.
Some of the participants in our project used the approach provided in
the training, integrated with their own traditional approach, with groups 
in the community. Formal, experimental assessment showed that trauma
symptoms decreased among community members who participated in
meetings led by people we trained, both over time, and in comparison to
groups led by different trainers who used their organizations’ traditional
approach, as well as in comparison to no treatment control groups. Orien-
tation to members of the other ethnic group became more positive among
community members who participated in the meetings that used this 
integrated approach, both over time and in comparison to people in the
other groups (Staub, Pearlman, Gubin and Hagengimana, unpublished 
research).
Healing by leaders is also important. Like the other members of a vic-
timized group, so the people who emerge as leaders have been affected by
past victimization. In Bosnia, a number of the Serb leaders have had mem-
bers of their families killed by the Croat regime that killed many Serbs and
Jews during World War II. Outsiders who engage with such leaders need to
be aware of and attempt to address their wounds to individual and group
identity, if they are to effectively influence policies and practices (Staub,
1999c).
Hostile groups can also benefit from joint efforts in behalf of shared
goals. Depending on the nature of hostility and past injuries, this may re-
quire prior dialogue, problem solving in small groups (see Kelman, in this
book), or healing. But significant engagement by members of groups with
each other in the course of working together for shared goals can bring
benefits in overcoming devaluation and antagonism (Allport, 1954; Petti-
grew, 1997; Staub, 1989b, 1999c).
Culture Change
Another group-level issue is culture change. Cultural devaluation and au-
thority orientation shape identities. Monolithic cultures, with a limited set
of values and limited access of at least some groups of people to the public
domain, also make genocide more likely, while pluralistic cultures make
genocide less likely. When respect for authorities is moderate and there are
varied values and points of view that can be expressed in the public domain,
people are more able to tolerate uncertainty and therefore social upheaval
(Soeters, 1996).
Democracies, which are pluralistic, are unlikely to engage in genocide
(Rummel, 1994). This, however, seems especially true of “mature” democ-
racies. (Staub, 1996b) These are democracies in which laws are enforced, in
which all groups can participate in public life, and which have a “civil soci-
Individual and Group Identities in Genocide and Mass Killing
177


ety “ (i. e., well-established civic institutions). Germany was not this kind of
democracy at the time of the Weimar Republic, nor was Argentina at the
time of “the disappearances” (Staub, 1989b), when elected governments
were regularly replaced by military rule. Working to create democracy and
a civil society is one way to bring about culture change that makes genocide
less likely. Such a culture is likely to form individuals whose relationship to
the group is more akin to constructive, rather than blind, patriotism and
who are less likely to be embedded in their group.
Multiple group identities seem important in preventing genocide. When
people can gain identity and fulfill other basic needs through connection to
a variety of groups, they are less likely to give themselves over to and lose
themselves in a group that turns against others. Their ability to oppose de-
structive ideologies and practices will increase. 
However, these groups need to be reasonably independent of each other.
In many cases of genocide, subgroups of society, which could have been the
source of at least somewhat independent identities, have subordinated
themselves to or have become embedded in the state or in an ideological
movement. This was recently the case with the churches in Rwanda. Many
church leaders, both in the dominant Catholic Church and in the minority
Protestant churches, were long connected to the ruling elements of the
country. They became accomplices during the genocide, facilitating the
killing of Tutsis (des Forges, 1999; Gourevich, 1998).
Inclusive Caring, Moral Courage, and the Behavior of Bystanders
What type of socialization will lead children to speak out against harmful
practices within their groups? What is required for people to exert influ-
ence within their group so that it does not remain a passive bystander in the
face of violence against people outside the group? Warmth, affection, effec-
tive but not punitive guidance, allowing children reasonable autonomy,
helping them develop a sense of significance and a “voice,” and guiding
children to help and not harm others make it likely that they will develop
inclusive caring, connected identities and constructive patriotism, as well as
moral courage (Staub, 1996a). Such characteristics would make genocide
less likely.
Constructive patriotism has a value basis in caring about human welfare.
Holding such a value motivates criticism and opposition to harmful policies
and practices of the group. Caring about human welfare is a value that has
to develop in children through experiential learning. Without love and af-
fection and positive guidance in their own lives, without feeling valued,
children are unlikely to learn to value others.
T
o publicly oppose harmful practices and advocate benevolent ones of-
ten requires moral courage. This is not necessarily physical courage, but the
ability and willingness to tolerate conflict. The kind of individual identities
people develop and the nature of social identities, what the group requires
178
The Contribution of Social Identity to Violent Intergroup Conflict


of its members, are central to this. Even when there is no physical danger,
people often won’t speak out against harmful and potentially violent prac-
tices out of concern for ostracism or exclusion from the group. Is it possible
to create cultures in which opposition for the sake of human welfare is 
valued? 
It is relevant in this context that, according to one researcher, about one-
third of rescuers of Jews in Nazi Europe were marginal to their group (T
ec,
1986). They had a different religion, or one foreign parent, or moved to
their community from some other place, or may have been simply person-
ally different and therefore less taken in by or less invested in their commu-
nity. This marginality may have made them less identified with their group,
perhaps fearing less the possibility of ostracism and enabling them to have a
separate, independent perspective.
Inclusive caring means the development of caring for people beyond the
group. It has to do with the relationship of “us” and “them,” with the devel-
opment of an identity that makes individuals and groups not draw a line at
the boundary of their group that diminishes concern for the welfare of
those beyond the group. (Staub, 1996a). Many of the rescuers grew up in
families where this line between their group and minorities was not as
sharply drawn as in the rest of the group. The parents and the family en-
gaged with people outside the “ingroup,” often including Jews (Oliner &
Oliner, 1988).
Individual identities based on inclusive caring would lead people to want
their countries to be active bystanders, who would respond when there is 
violence against people outside their borders. Self-concepts that embody
inclusive caring would be likely to move individuals to exert influence on
their governments not to remain passive. This is essential for the preven-
tion of mass killing and genocide. While this has rarely happened, the 
limited examples that exist strongly suggest that if the community of na-
tions responds early, as harm doing just begins, the evolution of increasing
violence can usually be stopped. 
Even groups of individuals can exert significant influence. After the revo-
lution in Iran the Bahai, for a long time a devalued and persecuted group,
were immediately intensely persecuted. Over 200 Bahai were executed. The
international Bahai community immediately made representations to the
various governments where Bahai lived and to the international community.
The United Nations, as well as individual nations, passed resolutions con-
demning the killing and persecution. While the persecution did not stop,
the killings did (Bigelow, 1993), with some limited resumption, unfortu-
nately, in the 1990s. 
In places where violence could have been much greater, as in Northern
Ireland, or between Israelis and Palestinians, usually external bystanders
had some role in limiting violence. Internal bystanders, members of differ-
ent groups reaching out to each other, perhaps enabled by the behavior of
external bystanders, also seems to have had a role (Staub, 1999b). In North-
Individual and Group Identities in Genocide and Mass Killing
179


ern Ireland, the British military and political presence provided support to
political and social leaders who have worked for opportunities in jobs and
more social equality for Catholics, as well as to community organizations
that brought members of the two groups together for dialogue and other
forms of engagement with each other (Cairns & Darby, 1998; Staub, 1999b;
see, however, White, this volume).
With early, committed, and resolute actions by the community of na-
tions, the evolution of violence can probably be inhibited much of the time
without the use of military force (Staub, 1999b). But for this to happen re-
quires important changes in identities: in individuals’ views of themselves
and their nations and in nations’ views of themselves and of the interna-
tional community. It requires committed bystanders and the creation of an
international system that leads to action when action is required. 
Summary and Conclusions
Identities of individuals, the nature of the relationship to the group in
which they invest their identity, and group-self concept are all important in
the generation of genocide, mass killing, and other harmful behavior by
groups toward other groups. They are also important in prevention. How-
ever, identities must be considered in the context of other aspects of indi-
vidual and group psychology. Identity is a concept that refers to certain
structures in individual personality with related processes. In reality, of
course, human beings (and cultures and societies), are unitary wholes with
interrelated structures and processes.
Important concepts related to identity that have relevance to genocide
and mass killing include values, the way the self and others are divided (“us”
and “them”), orientation to authority, the individual’s relationship to the
group, and basic human needs (figure 7.1). One may conceptualize all these
as part of identity, or separate them and see them as concepts with their
own structures and processes. I find it useful to separate these and then con-
sider the relationships among them. As I suggested in this chapter, the na-
ture of a person’s identity and group self-concept will certainly affect orien-
tation to authority, divisions between “us” and “them,” and so on. 
I have proposed a series of steps that can take members of a group from
instigating factors (left side of figure 7.1) to full-fledged genocide or mass
murder (right side of figure 7.1). Instigating factors include intergroup con-
flict and difficult life circumstances (e.g., poverty) that threaten basic needs
(e.g., need for security, sense of self-worth, connections to others). Behav-
iors that elevate the ingroup and derogate outgroups (e.g., scapegoating)
provide means of coping with these threats to the self. Derogation of the
outgroup can encourage initial acts of violence against the outgroup. Initial
violence, in turn, sets in motion a cycle of further violence and polarization
180
The Contribution of Social Identity to Violent Intergroup Conflict


that becomes tied to self and group identity. This cycle may eventually esca-
late to mass killing and genocide. The development of this chain of events is
moderated by both cultural variables (top of figure 7.1) and the behavior of
bystanders (bottom of figure 7.1). For example, ideologies that legitimize
violence and a group’s history of being a victim of violence are two cultural
factors that can increase the likelihood that initial instigators will lead to
mass killing or genocide. On the other hand, involvement of bystanders
(whether ingroup members or outsiders, such as the international commu-
nity) can mitigate the march from instigators to eventual genocide.
One of the important relationships suggested by this chapter is that be-
tween identity and basic needs (figure 7.1). Security, connection, and effec-
tiveness are important from birth on, and their fulfillment and frustration
begin at birth. When we consider motivation for action that may have to do
with identity, under many conditions a complex of motives and needs will
be involved. When people engage in destructive ways of fulfilling their need
for identity (for example, by joining an ideological movement that aims to
destroy “enemies”), they are also usually seeking to satisfy other basic needs
that I have proposed: security, connection, a sense of effectiveness and con-
trol, and the need for comprehension of reality. The same is true when peo-
ple act in positive ways. Reconciliation with others may arise from and has
effects on identity, but also fulfills varied basic needs and expresses and af-
firms certain values.
Individual identity is a storehouse of and is shaped by all the experiences
a person has, at least all significant experiences. For example, it is shaped by
a history of fulfillment and frustration of basic needs. Some may regard it,
therefore, as superordinate to other aspects of personality and their related
processes. In my view, however, the most useful theory will maintain and
develop conceptual distinctions. The important task for basic theory build-
ing is to identify the interrelations of concepts and related processes.
How are these concepts, underlying structures, and processes interre-
lated in the tragic domain of human life this chapter addresses? Is identity a
more basic need than others? When a basic need such as security or con-
nection is fulfilled, does that automatically strengthen identity? How can
basic needs be fulfilled constructively for whole groups of people? Will 
answers to these and other questions help in halting and preventing group
violence? How can we apply gains we make in theoretical understanding to
help us in the practical realm, in preventing violence?
These questions point to an obligation of psychologists. The obligation
is even deeper than applying knowledge to the great practical issues of hu-
man life. It is to be purposeful in our efforts to gain knowledge, whether in
the laboratory or through experience in real-life settings. The obligation is
to be guided by a vision of reducing human suffering and enhancing human
welfare, especially when the knowledge we seek relates to such important
matters as the origins and prevention of genocide and mass killing.
Individual and Group Identities in Genocide and Mass Killing
181


References
Allport, G. W. (1954). The nature of predjudice. Reading, MA: Addison-Wesley.
Bar-T
al, D., & Staub, E. (1997). Introduction: The nature and forms of patriotism.
In D. Bar-T
al & E. Staub (Eds.), Patriotism in the lives of individuals and groups.
Chicago: Nelson-Hall.
Baumeister, R. F. (1997). Evil: Inside human violence and cruelty. New York: W.H.
Freeman.
Bigelow, K. R. (1993). A campaign to deter genocide: The Baha’i experience. In H.
Fein, (Ed.), Genocide watch. New Haven: Yale University Press.
Cairns, E., & Darby, J. (1998). The conflict in Northern Ireland. American Psycholo-
gist, 53, 754–760.
Craig, G. A. (1982). The Germans. New York: New American Library.
des Forges, A. (1999). Leave none to tell the story: Genocide in Rwanda. New York:
Human Rights Watch.
Epstein, S. (1980). The self-concept: A review and the proposal of an integrated
theory of personality. In E. Staub (Ed.), Personality: Basic aspects and current re-
search. Englewood Cliffs, NJ: Prentice-Hall.
Fein, H. (1993). Accounting for Genocide after 1945: Theories and some findings.
International Journal of Group Rights, 1, 79–106.
Fogelman, E. (1994). Conscience and courage: Rescuers of Jews during the Holocaust.
New York: Anchor Books.
Harff, B., & Gurr, T. R. (1990). Victims of the state genocides, politicides and group
repression since 1945. International Review of Victimology, 1, 1–19.
Gourevich, P. (1998). We wish to inform you that tomorrow we will be killed with our
families. New York: Farrar, Straus and Giroux.
Herman, J. (1992). Trauma and recovery. New York: Basic Books.
Hilberg, R. (1961). The Destruction of the European Jews. New York: Harper & Row.
Hitler, A. (1925). Mein Kampf. Translated by Ralph Manheim. Boston, MA:
Houghton Mifflin Company.
Janoff-Bulman, R. (1992). Shattered assumptions. New York: The Tree Press.
Jarymowitz, M. (1977). Modification of self-worth and increment of prosocial sensi-
tivity. Polish Psychological Bulletin, 8, 45–53.
Kincaid, E. (1959). In every war but one. New York: Norton.
Konner, M. (1978). Social and personality development: An anthropological 
perspective. In M. Lamb (Ed.), Social and personality development. New York:
Holt.
Kressel, N. J. (1996). Mass hate: The global rise of genocide and terror. New York:
Plenum Press.
Levin, S., Sidanius, J., Rebinowitz, J. L., & Frederico, C. (1998). Ethnic identity, 
legitimizing ideologies and social status: A matter of ideological asymmetry.
Political Psychology, 19, 373–404.
Lifton, R. J. (1986). The Nazi doctors: Medical killing and the psychology of genocide.
New York: Basic Books.
McCann, L. I., & Pearlman, L. A. (1990). Psychological trauma and the adult survivor:
Theory, therapy, and transformation. New York: Bruner/Mazel.
Miller, A. (1982). For your own good: Hidden cruelty in child-rearing and the roots of 
violence. New York: Farrar, Straus and Giroux.
Myers, D. (1999). Social psychology. New York: McGraw-Hill.
182
The Contribution of Social Identity to Violent Intergroup Conflict


Oliner, S. B., & Oliner, D. (1988). The altruistic personality: Rescuers of Jews in Nazi 
Europe. New York: Free Press.
Opotow, S. (Ed.). (1990). Moral exclusion and injustice. Journal of Social Issues, 46 (1).
Pennebaker, J. W., & Keough, K. A. (1999). Revealing, organizing and reorganizing
the self in response to stress and emotion. In R. J. Contrada, & R. D. Ashmore
(Eds.), Self, social identity, and physical health. (pp. 101-121) London: Oxford. 
Pettigrew, T. F. (1997). Generalized intergroup contact effects on prejudice. Person-
ality and Social Psychology Bulletin, 23, (2), 173–185.
Prunier, G. (1995). The Rwanda crisis: History of a genocide. New York: Columbia
University Press.
Reese, H. (1961). Relationships between self-acceptance and sociometric choices.
Journal of Abnormal and Social Psychology, 62.
Rummel, R. J (1994). Death by government. New Brunswick, NJ: Transaction 
Publishers.
Salmivalli, C., Kaukiainen, A., Kaistaniemi, L., & Lagerspetz, K. M. J. (1999). Self-
evaluated self-esteem, peer-evaluated self-esteem, and defensive egotism as pre-
dictors of adolescents’ participation in bullying situations. Personality and Social
Psychology Bulletin, 25, 1268–1278.
Sampson, E. E. (1988). The debate on individualism. American Psychologist, 47,
15–22.
Schatz, R., & Staub, E. (1997). Manifestations of blind and constructive patriotism.
In D. Bar-T
al & E. Staub (Eds.), Patriotism in the lives of individuals and groups.
Chicago: Nelson-Hall.
Schatz, R., Staub, E. & Lavine, H. (1999). On the varieties of national attachment:
Blind versus constructive patriotism. Political Psychology, 20. 151–175.
Simpson, C. (1993). The splendid blond beast. New York: Grove Press.
Smith, N. S. (1998). The psychocultural roots of genocide. American Psychologist, 53,
743–753.
Soetters, J. L. (1996). Culture and conflict: An application of Hofstede’s theory to
the conflict in the former Yugoslavia. Peace and Conflict: Journal of Peace Psychol-
ogy, 2 (3), 233–244.
Staub, E. (1989a). Steps along the continuum of destruction: The evolution of by-
standers, German psychoanalysts and lessons for today. Political Psychology, 10,
39–53.
Staub, E. (1989b). The roots of evil: The origins of genocide and other group violence. New
York: Cambridge University Press.
Staub, E. (1993a). Individual and group selves, motivation and morality. In T. Wren,
and G. Noam, (Eds.), Morality and the self. (pp. 337–359) Cambridge: MIT Press.
Staub, E. (1993b). The psychology of bystanders, perpetrators and heroic helpers.
International Journal of Intercultural Relations, 17, 315–341.
Staub, E. (1996a). Altruism and aggression in children and youth: Origins and cures.
In R. Feldman, (Ed.), The psychology of adversity. Amherst: University Massachu-
setts Press.
Staub, E. (1996b). Cultural-societal roots of violence: The examples of genocidal 
violence and of contemporary youth violence in the United States. American 
Psychologist, 51, 117–132.
Staub, E. (1997). Blind versus constructive patriotism: Moving from embeddedness
in the group to critical loyalty and action. In E. Staub & D. Bar-T
al (Eds.), Patri-
otism. Chicago: Nelson-Hall.
Individual and Group Identities in Genocide and Mass Killing
183


Staub, E. (1998a). Basic human needs and their role in altruism and aggression. Un-
published manuscript. Department of Psychology. University of Massachusetts
at Amherst.
Staub, E. (1998b). Breaking the cycle of genocidal violence: Healing and reconcilia-
tion. In J. Harvey (Ed.), Perspectives on loss: A sourcebook. Philadelphia: T
aylor and
Francis.
Staub, E. (1999a). A brighter future: The evolution of caring and nonviolent chil-
dren. Book in preparation.
Staub, E. (1999b). The origins and prevention of genocide, mass killing, and other
collective violence. Peace and Conflict: Journal of Peace Psychology, 5 (4), 303–336.
Staub, E. (1999c). The roots of evil: Social conditions, culture, personality and basic
human needs. Personality and Social Psychology Review, 3, 179–192.
Staub, E. (2000a). Genocide and mass killing: Origins, prevention, healing and 
reconcilation. Political Psychology, 21, 367–382.
Staub, E. (2000b). Mass murder: Origins, prevention and U.S. involvement. Encyclo-
pedia of violence in the United States. New York: Charles Scribner’s Sons, Reference
Books.
Staub, E., & Pearlman, L. (1996). Trauma and the fulfillment of the human poten-
tial. Workshop presented of the meetings of the International Society for Trau-
matic Stress Studies. November, San Francisco.
Staub, E., & Pearlman, L. A. (in press). Healing, reconciliation and forgiving after
genocide and other collective violence. In Forgiveness and Reconciliation, Radnor,
PA. T
empleton Foundation Press. 
Staub, E., Pearlman, L. A., Gubin, A., and Hagengimana, A. Healing, forgiveness
and reconciliation in Rwanda: Intervention and experimental evaluation. Manu-
script in preperation. Department of Psychology, University of Massachusetts at
Amherst.
T
ajfel, H. (1978). Social categorization, social identity and social comparison. In H.
T
ajfel (Ed.), Differentiation between social groups (pp. 61–76). London: Academic
Press.
T
ajfel, H. (1982). Social psychology of intergroup relations. Annual Review of Psy-
chology, 33, 1–39.
T
ec, N. (1986). When light pierced the darkness: Christian rescue of Jews in Nazi-occupied
Poland. New York: Oxford University Press.
Triandis, H. C. (1994). Culture and social behavior. New York: McGraw-Hill.
Turner, J. C. (1987). Rediscovering the social group: A self-categorization theory. New
York: Basic Blackwell.
184
The Contribution of Social Identity to Violent Intergroup Conflict


PART IV
THE ROLE OF SOCIAL 
IDENTITY IN REDUCING 
INTERGROUP CONFLICT


This page intentionally left blank 


Herbert C. Kelman
8
The Role of National Identity in
Conflict Resolution
Experiences from Israeli-Palestinian 
Problem-Solving Workshops
My colleagues’ and my work as scholar-practitioners has focused on analysis
and resolution of protracted, seemingly intractable conflicts between na-
tional, ethnic, or other kinds of identity groups, best exemplified by inter-
communal conflicts, such as those in Cyprus, Northern Ireland, Sri Lanka,
Bosnia, and apartheid South Africa. My own most intensive and extensive
experience, over the past quarter-century, has been with the Israeli-Pales-
tinian conflict, and my analysis draws primarily on that experience.1
Using the Israeli-Palestinian conflict as a case in point, this chapter ex-
amines the way in which issues of national identity can exacerbate an inter-
national or intercommunal conflict and the way in which such issues can be
addressed in conflict-resolution efforts. The chapter starts out with a brief
history of the Israeli-Palestinian conflict, setting the stage for the identity
issues at the heart of the conflict. It then proceeds to describe the struggle
over national identity between the two peoples, which has led them to per-
ceive their conflict in zero-sum terms, with respect to not only territory and
resources but also national identity and national existence. Next, it argues
that long-term resolution of this and similar deep-rooted conflicts requires
changes in the groups’ national identities, such that affirmation of one
group’s identity is no longer predicated on negation of the other’s identity.
Such identity changes are possible as long as they leave the core of each
group’s national identity intact. Furthermore, the chapter proceeds to ar-
gue, such changes need to be and can be “negotiated” between the two
groups. One venue for negotiating identity, described in the next section, is
provided by the problem-solving workshops between Israeli and Palestinian
elites that my colleagues and I have convened for many years. Finally, the
187


paper concludes with an illustration of the possibilities and limits of the ne-
gotiation of identity, based on a joint Israeli-Palestinian exploration of the
problem of Palestinian refugees.
The Israeli-Palestinian Conflict
The Israeli-Palestinian conflict is now more than a century old (see T
essler,
1994, for a comprehensive account of the history of the conflict, or
Mendelsohn, 1989, and Gerner, 1991, for shorter accounts). Its origins go
back to the birth of political Zionism at the end of the 19th century (see
Halpern, 1969, and Hertzberg, 1973). The early decades of the 20th cen-
tury brought to Palestine waves of Jewish immigrants who purchased land,
built settlements and social institutions, and clearly signaled their intention
to establish a Jewish homeland and ultimately a Jewish state in Palestine.
The growing Jewish presence was soon perceived as a threat by the Arab
population of the land, which was itself influenced by the development of
Arab nationalism and the construction of a specifically Palestinian identity
(see Muslih, 1988, and Khalidi, 1997). Violence first erupted in the 1920s
and has continued to mark the relationship between the two peoples ever
since.
During the period of the British mandate, which was established after
World War I, various formulas for the political future of Palestine were ex-
plored, including partition and establishment of a federal state, but none
was found to be acceptable to both the Arab and the Jewish populations (or
indeed to either one of them). In November 1947—in the wake of World
War II and the decimation of European Jewry—the United Nations Gen-
eral Assembly voted to end the British mandate over Palestine (on May 15,
1948) and to partition the land into a Jewish and an Arab state. The Zionist
leadership accepted the partition plan, with reservations. The Arab leader-
ship, both within Palestine and in the neighboring states, rejected it. Fight-
ing between the two sides broke out immediately after adoption of the UN
resolution and turned into all-out war after May 15, 1948, when the British
forces withdrew, the Jewish leadership in Palestine declared the indepen-
dent state of Israel, and regular armies from the neighboring Arab states
joined the fray. Fighting continued until early 1949.
In July of 1949, Israel and the Arab states signed armistice agreements
(though the state of war continued). The armistice lines became the official
borders of the State of Israel. These borders included a larger portion of
Palestine than the UN partition plan had allotted to the Jewish state. The
Arab state envisioned by the partition plan did not come into being. T
wo
parts of mandatory Palestine remained under Arab control: the West Bank,
which was eventually annexed by Jordan, and the Gaza Strip, which came
under Egyptian administration. The establishment of Israel and the war of
1948–49 also created a massive refugee problem, with the flight or expul-
188
The Role of Social Identity in Reducing Intergroup Conflict


sion of hundreds of thousands of Palestinian Arabs from their homes in the
part of Palestine that became the State of Israel. I shall return to the refugee
problem later in this chapter.
The map changed radically as a result of the Arab-Israeli war of June
1967—and, along with it, the political atmosphere in the Middle East. By
the end of the Six-Day War, as Israelis called it, Israel occupied the West
Bank and the Gaza Strip, thus extending its control over the entire territory
of mandatory Palestine. It also occupied the Sinai Peninsula and the Golan
Heights—Egyptian and Syrian territories, respectively. The new geopoliti-
cal and strategic situation created by the 1967 War led to the Palestinianiza-
tion of the Arab-Israeli conflict, bringing it back to its origin as a conflict
between two peoples over—and increasingly within—the land they both
claimed (Kelman, 1988).
The Palestinianization (or re-Palestinianization) of the conflict has mani-
fested itself in the actions of the Arab states, of the Palestinian community it-
self, and of Israel. Israel’s neighboring Arab states gradually withdrew from
the military struggle against Israel—though not before another major war in
1973—leaving it, essentially, to the Palestinians themselves. The disengage-
ment of the Arab states became dramatically clear with the 1977 visit to
Jerusalem of President Anwar Sadat of Egypt, the largest and most powerful
Arab state, an initiative that led to the Camp David accord of 1978 and the
Egyptian-Israeli peace treaty of 1979. The Palestinians took repossession of
their struggle, which in the years between 1949 and 1967 had been mostly in
the hands of the Arab states. Fatah, under the leadership of Yasser Arafat, and
other Palestinian guerrilla organizations grew in strength and eventually
took over the Palestine Liberation Organization (PLO), which was originally
a creature of the Arab League. Between the 1960s and the 1980s, the Pales-
tinian movement gradually shifted its emphasis from the liberation of all 
of Palestine through armed struggle against Israel to the establishment of an
independent Palestinian state in the West Bank and Gaza through largely 
political means. The end of the occupation became the immediate goal of the
movement and, with the onset of the intifada—the uprising in the West Bank
and Gaza—in December 1987, the occupied territories became the focal
point of its struggle.
On the Israeli side, the intifada further underlined the Palestinianization
of the conflict in the wake of the 1967 War. What had been largely an inter-
state conflict between 1948 and 1967 had now been internalized by Israel—
that is, transformed into a continuous confrontation with a resentful Pales-
tinian population, living under occupation within Israel’s post-1967
borders. Many Israelis were persuaded by the intifada that continuing occu-
pation was not tenable and that the Palestinians were indeed a people,
whose national movement had to find some political expression if there was
to be a peaceful accommodation between the two sides (Kelman, 1997c).
By the end of the 1980s, there was a strong interest on all sides in finding
a peaceful accommodation and an increasing recognition that some version
The Role of National Identity in Conflict Resolution
189


of a two-state solution would provide the best formula for a broadly accept-
able historic compromise. The political obstacles to such a solution, how-
ever, remained severe. A number of strategic and micropolitical considera-
tions—traceable, in particular, to the end of the Cold War and the
aftermath of the Gulf War—eventually brought the leaderships on both
sides to the negotiating table at the Madrid Conference in 1991 and the
subsequent talks in Washington. These talks, however, never developed
momentum. It was only after Prime Minister Yitzhak Rabin came into
power in Israel in 1992, at the head of a government led by the Labor Party,
and gradually (and reluctantly) concluded that Israel would have to deal di-
rectly with the PLO leadership in order to make progress in the negotia-
tions, that a breakthrough was finally achieved. This breakthrough oc-
curred in the secret Oslo talks, which culminated in the exchange of letters
of mutual recognition between the PLO and the State of Israel (which, in
my view, was the most significant achievement of the Oslo process) and the
Declaration of Principles signed in Washington in September 1993 (see
Kelman, 1997c, for further details).
The Oslo breakthrough occurred because a number of long-term and
short-term interests of the parties—to which I alluded briefly—persuaded
them of the necessity of reaching an agreement. “But a significant factor
contributing to the breakthrough was the conclusion, on both sides, that
negotiations were not only necessary but also possible—that they could
yield an acceptable agreement without jeopardizing their national existence.
This sense of possibility evolved out of interactions between the two sides
that produced the individuals, the ideas, and the political atmosphere re-
quired for negotiations” (Kelman, 1997d, p. 213). A variety of unofficial
contacts between the two sides, including the workshops to be described in
this chapter, contributed to creating this sense of possibility and the climate
conducive to negotiations (see Kelman, 1995).
Though the Oslo agreement was a genuine breakthrough in the effort to
resolve the Israeli-Palestinian conflict, the process of negotiating an interim
agreement and initiating final-status negotiations has faced numerous ob-
stacles and experienced dangerous setbacks, all reminding us of the in-
tractability of conflicts between identity groups, such as the Israeli-Palestin-
ian conflict. Significant changes in the conflict have taken place with the
establishment of a Palestinian Authority (PA) in the West Bank and Gaza
Strip, which constitutes the scaffolding for a Palestinian state. But the path
to a negotiated agreement has been obstructed by such events as the He-
bron massacre of Palestinians at prayer, the assassination of Yitzhak Rabin,
the suicide bombings against Israeli civilians, the continuation of the Israeli
settlement process in the West Bank, the deteriorating economy and quality
of life in the West Bank and Gaza, the corrupt and autocratic leadership of
the PA, and the three-year rule of Prime Minister Netanyahu in Israel.
The election of a new Israeli government in May of 1999 brought new
hope that serious negotiations of the final-status issues will be pursued. But
190
The Role of Social Identity in Reducing Intergroup Conflict


the negotiations failed and, in late September 2000, a new intifada broke
out, producing a violent Israeli response and a serious deterioration of the
peace process. Nevertheless I feel fairly confident that the negotiations will
eventually succeed in establishing a Palestinian state. I am less confident
that this new state will be truly independent, secure, and viable, and that its
establishment will form the basis of a long-term peace and a cooperative,
mutually enhancing relationship with Israel, conducive to ultimate recon-
ciliation between the two peoples. How the identity issues that divide the
two peoples are addressed in the negotiations and the public debates that
surround them will play a major role in determining the quality and dura-
bility of the solution achieved at the negotiating table.
The Struggle over National Identity
Although national identity is carried by the individual members of a na-
tional group and can thus be studied as a property of individuals, the pre-
sent chapter refers to it as a collective phenomenon—as a property of the
group. “Insofar as a group of people have come to see themselves as consti-
tuting a unique, identifiable entity, with a claim to continuity over time, to
unity across geographical distance, and to the right to various forms of self-
expression, we can say that they have acquired a sense of national identity.
National identity is the group’s definition of itself as a group—its concep-
tion of its enduring characteristics and basic values; its strengths and weak-
nesses; its hopes and fears; its reputation and conditions of existence; its in-
stitutions and traditions; and its past history, current purposes, and future
prospects” (Kelman, 1997a, p. 171; see also Kelman, 1998b).
In the Israeli-Palestinian and other such conflicts, the threat to collective
identity is a core issue in the conflict, which is integrally related to the
struggle over territory and resources. Both peoples and their national
movements claim the same territory, and each seeks ownership of that terri-
tory and control over its resources as the basis of an independent state that
gives political expression to its national identity. The integrity of this col-
lective identity is critical to each group for several reasons. First, the in-
tegrity of the national identity is an end in itself, in that the identity serves
as a source of distinctiveness, unity, and continuity for the group and of a
sense of belongingness for its members. Second, the national identity con-
stitutes the ultimate justification of the group’s claim to ownership of the
land and control of its resources. And third, the national identity provides a
focus for developing and maintaining the group’s distinctive culture, reli-
gion, and way of life. The collective identity of each group is bolstered by a
national narrative—an account of the group’s origins, its history, and its re-
lationship to the land—that explains and supports its sense of distinctive-
ness, its positive self-image, and the justice of its claims and grievances.
In conflicts such as that between Israelis and Palestinians, in which the
The Role of National Identity in Conflict Resolution
191


two sides live in the same space and claim ownership of the same territory, it
is not only the actions of the other but also the identity and the very exis-
tence of the other that are a threat to the group’s own identity. The other’s
identity and its associated narrative challenge the group’s claims to owner-
ship—at least to exclusive ownership—of the land and its resources. The
other’s presence in the same space, particularly if it is accompanied by 
demands for a share of the power and for recognition of the other culture,
religion, and/or language, is perceived as a threat to the integrity and cohe-
siveness of the group’s society and its way of life.
These dynamics lead to a view of the conflict as a zero-sum struggle, not
only around territory but also around identity (Kelman, 1987). Acknowl-
edging the other’s identity becomes tantamount to jeopardizing the iden-
tity—and indeed the national existence—of one’s own group. Each side
“holds the view that only one can be a nation: Either we are or they are.
They can acquire national identity and rights only at the expense of our
identity and rights” (Kelman, 1987, p. 354). Thus, over the course of the 
Israeli-Palestinian conflict, there has been a systematic tendency on each
side to deny the other’s identity as a people, the authenticity of the other’s
links to the land, the legitimacy of the other’s claims to national rights, and
the very existence of the other as a national group (Kelman, 1978, 1982).
Negation of the other’s identity and of the narrative in which it is embedded
becomes so important to the conflict that it is incorporated in the identity
that each group constructs for itself and in the narrative that it presents to
the world (Kelman, 1999).
The contrasting Israeli and Palestinian narratives about the creation of
the State of Israel in 1948 both rely on the negation of the other to bolster
the justice of their own cause. For Israelis, the creation of Israel represented
a rightful return of the Jewish people to its ancestral homeland. Establish-
ment of a Jewish state in Palestine did not, in their eyes, constitute an injus-
tice to the Arabs who resided there, because Palestinian Arabs were not a
people, distinct from the Arab inhabitants of surrounding countries, and
never exercised sovereignty in Palestine. Moreover, in the Israeli narrative,
the responsibility for the Palestinian refugee problem and the suffering of
the Palestinian Arab population rests with their own aggressive and incom-
petent leadership, which rejected all compromise and initiated violent at-
tacks in the effort to block the establishment of Israel. For Palestinians, by
contrast, the creation of Israel represented an act of usurpation by Euro-
pean settlers, who forcefully displaced the indigenous population and de-
stroyed their society, their property, and their way of life. In the Palestinian
narrative, Jews are a religious group, not a nation entitled to its own state,
and Zionism is a form of settler colonialism that imposed itself on a region
in which it has no roots. Each identity gains some of its strength and legiti-
macy from negating and delegitimizing the other.
Identities that rest in part on negation of the other inevitably take on 
an exclusivist and monolithic character (Kelman, 1997b). In the Israeli-
192
The Role of Social Identity in Reducing Intergroup Conflict


Palestinian conflict, such exclusivist and monolithic definitions of identity
have begun to give way in recent years. For significant segments of the two
population, however—and in some respects even for large majorities—such
definitions still prevail. 
In the Israeli-Palestinian case, a defining element of each group’s identity
is its relationship to the land and its history. Insofar as this relationship is
exclusive—that is, insofar as the group’s identity rests on the view that the
land and its history belong to it alone and that the other’s claims on them as
part of its own identity are illegitimate and inauthentic—there is little room
for conflict resolution. Conflict resolution becomes an option when the
parties accept the possibility that certain elements of identity may be shared
with the other, acknowledging that the other also has a profound attach-
ment to the land, anchored in authentic historical ties to it. Israelis and
Palestinians have been gradually moving toward acceptance of shared ele-
ments of identity as they have been searching for a political formula for
sharing the land. It has proved more difficult for the two sides, so far, to ac-
cept Jerusalem as a shared element of the two identities and to develop a
political formula to reflect that view.
Identities that rest on negation of the other also take on a monolithic
character; that is, all dimensions of the group’s identity—such as ethnicity,
religion, and language—tend to be viewed as highly correlated. The ide-
ology calls for complete correspondence between ethnic boundaries, politi-
cal boundaries, boundaries of emotional attachment, and boundaries of in-
tensive interaction. Self and other are, in principle, completely separated
along all of these lines. The Israeli-Palestinian conflict, as well as other pro-
tracted conflicts, particularly between identity groups living in close prox-
imity within a small space (such as Northern Ireland or Cyprus), might be
more amenable to resolution if there were some degree of disaggregation of
the monolithic identity, based on distinctions between different types of
boundaries. Such distinctions would allow for the development of a tran-
scendent identity—not in place of the particular ethnonational identities, but
alongside of them. In the Israeli-Palestinian case, a transcendent identity
could be fostered by separating the concept of the state, as a sovereign po-
litical entity, from that of the country, as a geographical entity. This distinc-
tion would allow the two communities to treat the entire country (Eretz
Yisrael or Palestine) as an object of common sentimental attachment and as
the framework for common instrumental pursuits (in such areas as develop-
ment and use of water resources, environmental protection, public health,
and tourism), while living in and identifying with separate political states
within that country.
The zero-sum view of identity and the mutual denial of the other’s iden-
tity that I have described create serious obstacles to conflict resolution. All
issues tend to become existential—matters of life and death for each side.
Compromise solutions that involve sharing of the land or agreeing on dif-
ferent boundaries for different purposes are likely to threaten exclusivist
The Role of National Identity in Conflict Resolution
193


and monolithic identities. The demonized other is not trusted to negotiate
in good faith and respect agreements. In short, when acceptance of the
other’s national rights and recognition of the other’s national identity are
seen as relinquishing the group’s own rights and jeopardizing its own iden-
tity, distributive solutions based on compromise are hard to achieve. Even if
the parties agree to make certain compromises in response to reality de-
mands and external pressures, these compromises are unlikely to lead to
durable changes in the relationship between the conflicting groups, con-
ducive to stable peace, mutually enhancing interaction, and ultimate recon-
ciliation. Lasting change requires mutual adjustments in collective identity.
Identity Changes
The stubborn resistance to change in collective identities is widely recog-
nized and taken for granted. Yet identities have to change, at least tacitly, if
protracted identity conflicts are to be settled and, certainly, if they are to be
resolved in a way that transforms the relationship and opens the way to rec-
onciliation. South Africa provides perhaps the best illustration of an arena
of intense, protracted conflict in which fundamental identity changes paved
the way to resolution and reconciliation, although it also illustrates the dif-
ficulties in changing the worldviews and the structural realities that became
entrenched in the apartheid era.
Despite their undeniable rigidities, identities are potentially changeable
(and indeed negotiable) for two reasons: First, unlike territory and resources,
they are not inherently zero-sum; though they are perceived and debated as
such in intense conflicts, it is in fact not the case that A’s identity can be rec-
ognized and expressed only if B’s identity is denied and suppressed. 
If the two identities are to become compatible, however, they have to be
redefined. And this points to the second reason for the potential change-
ability of group identities: they can be redefined because they are to a large
extent constructed. T
o view national identity as a social construction does
not imply that it is manufactured out of nothing. There may be cases in
which one can properly speak of an imagined past, invented to buttress a
newly formed identity (cf. Anderson, 1983). Generally, however, the social
construction of an identity draws on a variety of authentic elements held in
common within a group: a common history, language, or religion; common
customs, cultural expressions, experiences, values, grievances, aspirations
(Kelman 1997b). Typically, the social construction of an identity involves a
dual process of discovery (or rediscovery) and creation of such common ele-
ments (Kelman, 1997a). The social construction of the identity implies a
degree of arbitrariness and flexibility in the way the identity is composed
(which elements are admitted into it and which omitted from it), and in
what its boundaries are (who is included and who is excluded). These
choices depend on the opportunities and necessities perceived by the elites
194
The Role of Social Identity in Reducing Intergroup Conflict


that are engaged in mobilizing ethnonational consciousness for their politi-
cal, economic, or religious purposes (Kelman, 1997b). Serbs and Croats, for
example, share a common language and culture, but differ in religion and
historical experiences. Political leaders have at times focused on the simi-
larities in the effort to shape them into a single nation; at other times they
have magnified the differences to define them as separate—and mutually
antagonistic—nations.
Thus, although national identities are generally constructed out of real
experiences, these experiences can be ordered in different ways, resulting in
different boundaries and priorities. As a consequence, they can be—and
typically are—deconstructed and reconstructed. “In fact, the reconstruction
of identity is a regular, ongoing process in the life of any national group.
Identities are commonly reconstructed, sometimes gradually and some-
times radically, as historical circumstances change, crises emerge, opportu-
nities present themselves, or new elites come to the fore” (Kelman, 1997b,
p. 338). Clearly, therefore, there is room for maneuver in a group’s self-
definition, particularly with respect to the definition of group boundaries
and the priorities among different elements of the group’s identity.
Changes in identity over the course of a protracted conflict come about
through a combination of changed perceptions of the necessity and the possi-
bility of resolving a conflict that has become increasingly costly to the par-
ties. The mounting costs and dwindling prospects of governing Algeria for
the French, of maintaining apartheid for white South Africans, of the occu-
pation of Palestinian territories for Israelis, and of the armed struggle for
the Palestinians created the necessity for changes in identity: Algeria as an in-
tegral part of France, South Africa under exclusive white control, Israel
within the borders of Greater Israel, and Palestinian repossession of the en-
tire homeland were assigned lower priority in the national identities of
these groups as it became clear to a majority that these aspirations could not
be realized at an acceptable cost. 
What made it possible to change these priorities was often the discovery
that accommodation of the other’s identity need not destroy the core of the
group’s own identity, and that a compromise solution to the conflict was
therefore negotiable. This kind of learning can take place in the course of
official or unofficial interactions between the groups or their members, in-
cluding the problem-solving workshops that my colleagues and I have con-
ducted: In the course of Israeli-Palestinian workshops, for example, partici-
pants have learned to differentiate their image of the enemy by discovering
that there are potential negotiating partners on the other side, that there is
a distinction between the other’s ideological dreams and operational pro-
grams, and that the other has positive goals beyond destruction of their
group (Kelman, 1987). They were enabled to enter into the enemy’s per-
spective, thus discovering the historical sources of the other’s claims and
grievances, the depth of the other’s fears, and the authenticity of the other’s
sense of peoplehood. They began to visualize a different future, discovering
The Role of National Identity in Conflict Resolution
195


possibilities for mutually beneficial coexistence and cooperation. As such
experiences multiply, and as the learnings produced by them are infused
into the two political cultures, each group may gradually change its identity
by eliminating the negation of the other’s identity as an element of its own
identity and perhaps even admitting the possibility of a partnership as a new
element of its own identity.
Negotiating Identity
The changes I have described are often the result of an explicit or implicit
process of negotiating identity. At its core, national identity is clearly non-
negotiable; indeed, the very idea of negotiating identity sounds like an oxy-
moron. National identity is a collective psychological conception, which
cannot be dictated or prescribed by outsiders. A group of people who define
themselves as a nation cannot be told that they have no right to do so be-
cause their self-definition does not conform to some set of theoretical, 
juridical, or historical criteria for doing so, or because their nationhood is
inconvenient to others. Nor does it make sense to tell them how to draw the
boundaries of the group: whom to include and whom to exclude. People are
a nation if they perceive themselves as such and are prepared to invest 
energy and make sacrifices in terms of that perception (Kelman, 1978).
Neither Palestinians nor Israelis will give up the core of their identity: their
sense of peoplehood, their attachment to the land, their conviction about
the historical authenticity of their links to that land, their commitment to
their national culture, language, and way of life. Nor will they give up the
national narrative that substantiates the justice of their cause. 
But there are many elements that can be added to or subtracted from an
identity without jeopardizing its core. In fact, changes in less central ele-
ments of the identity are often advocated precisely in order to protect the
core of the identity. It was on that basis that the majority of Israelis and
Palestinians came to accept territorial compromise—that is, a shrinking of
the territorial dimension of their identity—as the best available option for
maintaining their national identity. The Peace Now movement in Israel, for
example, advocated withdrawal from the occupied territories largely on the
grounds that this was the only way in which Israel could maintain its char-
acter as both a Jewish state and a democratic state. Yehoshafat Harkabi
(1986), a former chief of Israeli military intelligence and a prophetic voice
in the debate about Israeli-Palestinian peace, explicitly advocated a “smaller
Israel”—a “Zionism of quality” rather than a “Zionism of acreage.” He 
argued that Israel had to choose between withdrawing from the West 
Bank and making way for a Palestinian state there, or annexing the West
Bank with the consequence that Israel would eventually become a Palestinian
state. On the Palestinian side, the territorial dimension of the Palestinian
identity has gradually changed as the movement reflected on its realistic 
196
The Role of Social Identity in Reducing Intergroup Conflict


options. The thinking of the PLO evolved from advocacy of a Palestinian
Arab state in the whole of Palestine, to a secular democratic state, and even-
tually to a Palestinian state alongside of Israel comprising the West Bank
and Gaza (cf. Muslih, 1990). Significant segments of both societies still re-
ject territorial compromise on religious or ideological grounds and link
their national identity to possession of the land in its entirety. But the Pales-
tinian and Israeli mainstreams have by now come to terms with a national
identity that finds its political expression in only part of the land, as evi-
denced by the opinion polls that are now conducted on a regular basis in
both societies.
Such changes in elements of identity are a legitimate subject for “nego-
tiation” between groups whose identities clash, because the identity that
one group chooses for itself has significant implications for the rights, in-
terests, and identity of the other. Whenever one group translates the self-
definition of its nationhood into action—“by making territorial claims, by
demanding an independent state, by seeking to redraw borders, by declar-
ing who is included in the national identity and who is excluded from it, or
even by selecting a name for itself” (Kelman, 1997b, p. 337)—the other is
inevitably affected. Each group, therefore, has a legitimate concern about
the way the other defines itself, the way it formulates its national identity. It
is not surprising, then, that identity issues play an important role in the for-
mal and informal processes of pre-negotiation and negotiation—that is, in
the efforts to open a way to the negotiating table and to reach agreement
around the table.
T
o some extent, identity issues are part of the subject matter of the offi-
cial negotiations. I have already referred to the territorial dimension of
identity: insofar as Israelis and Palestinians are negotiating on the basis of a
“land for peace” formula, they are accepting territorial limits to their na-
tional identities, which have, after all, been historically linked to the whole
of the land. Similarly, the mutual recognition between Israel and the PLO,
as expressed in the exchange of letters between Arafat and Rabin—which I
have described above as the most important breakthrough of the Oslo
agreement —can be viewed as a product of the negotiation of identity: an
act of acceptance and legitimization of the other who in the past had been
defined as the antithesis to the self.
Although redefined identities are thus promulgated around the official
negotiating table, the “negotiation” of identity is primarily an informal, un-
official process in which members of the conflicting parties explore and 
invent ways of accommodating their group identities to one another. The
purpose of negotiation in this looser sense of the term is not to produce 
political agreements, but to develop joint understandings and formulations
that can help pave the road to political agreements at the official level. Im-
plicitly and explicitly, this kind of “negotiation” has been a central focus for
problem-solving workshops between Israelis and Palestinians that my col-
leagues and I have conducted over the past quarter-century.
The Role of National Identity in Conflict Resolution
197


Problem-Solving Workshops
Problem-solving workshops are the central instrument of interactive problem
solving, an unofficial, third-party approach to the resolution of international
and intercommunal conflicts, derived from the pioneering work of John
Burton (1969, 1979, 1984) and anchored in social-psychological principles
(see Kelman, 1972, 1979, 1992, 1998a). A workshop is a specially con-
structed, private space in which politically involved and often politically in-
fluential (but generally unofficial) members of conflicting communities can
interact in a nonbinding, confidential way. The microprocess of the work-
shop provides them the opportunity to penetrate each other’s perspective;
to explore both sides’ needs, fears, priorities, and constraints; and to engage
in joint thinking about solutions to the conflict that would be responsive to
the fundamental concerns of both sides.
Let me describe a typical “one-time” workshop for Israeli and Palestin-
ian participants. By “one-time” I mean that this particular group of Israelis
and Palestinians convenes only for this one occasion. Some of the indi-
viduals may have participated in more than one such workshop, and the 
one-time workshops that we have held over the years have had a cumulative
effect within the two societies. But, until 1990, we made no effort to recon-
vene the same group of participants for another occasion. The workshops
take place under academic auspices and are facilitated by a panel of social
scientists knowledgeable about international conflict, group process, and
the Middle East. The workshops usually begin with two pre-workshop 
sessions, about four hours in length, during which the third party meets
separately with each of the two parties. The workshop itself typically lasts
about two and a half days, often scheduled over an extended weekend. The
participants include three to six Israelis and an equal number of Palestin-
ians, plus a third party of three or more members.
The Israeli and Palestinian participants have included parliamentarians,
leaders and activists of political parties or political movements, journalists,
editors, directors of think tanks, and politically involved academicians—that
is, scholars who not only publish academic papers but who also write for
newspapers and appear in the media, who serve as advisors to political lead-
ers, and some of whom move back and forth between government and
academia. Some of our participants have been former diplomats, officials,
or military officers, and many were later to become negotiators, ambas-
sadors, cabinet ministers, parliamentarians, and leading figures in the media
and research organizations. We look for people who are within the main-
stream of their societies and close to the center of the political spectrum. At
the same time, they have to be people who are at least willing to explore the
possibility of a negotiated solution and to sit down as equals with members
of the other party.
The central ground rule of problem-solving workshops is the principle
of privacy and confidentiality. In the early days of our work, confidentiality
198
The Role of Social Identity in Reducing Intergroup Conflict


was particularly important for the protection of our participants, because
the mere fact that they were meeting with the enemy was controversial and
exposed them to political and even physical risks. Confidentiality is equally
important, however, for the protection of the process that we are trying to
promote in workshops. We are trying to encourage the participants to talk
and listen to each other, rather than focus on their constituencies, on third
parties, or on the record. We want them to think out loud, to experiment
with ideas, to explore different options, without having to worry about how
others would react if their words in the group were quoted outside. We
want them to engage in a type of interaction that is generally not feasible
among parties engaged in a bitter conflict—a type of interaction that, in-
deed, deviates from the conflict norms that usually govern their behavior:
an interaction that is analytic rather than polemical, one in which the par-
ties seek to explore each other’s perspective and gain insight into the causes
and dynamics of the conflict; an interaction that is problem-solving rather
than adversarial, one in which the parties sidestep the usual attempt to allo-
cate blame and, instead, take the conflict as a shared problem that requires
joint effort to find a mutually satisfactory solution.
The agenda of a problem-solving workshop is designed to allow this kind
of interaction to unfold. The core agenda of a one-time workshop has four
components. First, each side is asked to discuss its central concerns in the
conflict—the fundamental needs that would have to be addressed and the
existential fears that would have to be allayed if a solution is to be satisfac-
tory to it. The parties are asked not to debate the issues raised, although
they may ask for clarification of what the other says. The purpose is for
each side to gain an adequate understanding of the other’s needs, fears, and
concerns, from the perspective of the other. Once they have demonstrated
that they understand the other’s needs to a significant degree, we move to
the second phase of the agenda: joint thinking about possible solutions.
What participants are asked to do in this phase is to develop, through an in-
teractive process, ideas about the overall shape of a solution for the conflict
as a whole, or perhaps, a particular issue in the conflict, that would address
the needs and fears of both sides. They are given the difficult assignment of
thinking of solutions that would meet not only their own side’s needs, but
the needs of both sides.
Once the participants have developed some common ground in this
process of joint thinking, we turn to the third phase of the workshop: dis-
cussion of the political and psychological constraints within the two soci-
eties that would create barriers to carrying out the ideas for solution that
have been developed in the group. We deliberately leave the discussion of
constraints to the third phase, so that it does not hamper the creative
process of jointly generating new ideas. Finally, depending on how much
progress has been made and how much time is left, we ask the parties to en-
gage in another round of joint thinking—this time about ways of overcom-
ing the constraints that have been presented. The participants are asked to
The Role of National Identity in Conflict Resolution
199


come up with ideas about what their governments, their societies, and they
themselves might do—separately or jointly—that would help to overcome
the barriers to negotiating mutually satisfactory solutions to the conflict.
The third party in our model enacts a strictly facilitative role. It does not
propose solutions, nor does it participate in the substantive discussions. Its
task is to create the conditions that allow ideas for resolving the conflict to
emerge out of the interaction between the parties themselves. The facilita-
tion of the third party, however, is an important part of the process. The
third party sets the ground rules and monitors adherence to them; it helps
to keep the discussion moving in constructive directions, tries to stimulate
movement, and intervenes as relevant with questions, observations, and
even challenges. It also serves as a repository of trust for parties who, by de-
finition, do not trust each other: They feel safe to come to the workshop
because they trust the third party to maintain confidentiality and to protect
their interests.
Workshops have a dual purpose: to produce changes, in the form of new
insights and ideas, in the individual participants; and to transfer these
changes into the political process at the levels of both public opinion and
decision making. These two purposes may at times create contradictory re-
quirements—most notably in the selection of participants (Kelman, 1992).
T
o balance these contradictory requirements, we look for participants who
are not officials, but politically influential. They are thus more free to en-
gage in the process but, at the same time, because of their positions within
their societies, any new ideas that they develop in the course of a workshop
can have an impact on the thinking of decision makers and the society at
large.
The Israeli-Palestinian workshops we carried out until 1990 were all
one-time events designed to create a climate conducive to movement to the
negotiating table. In 1990, we organized our first continuing workshop, in
which a group of high-level Israelis and Palestinians met periodically over a
three-year period (Rouhana & Kelman, 1994). In 1994, Nadim Rouhana
and I convened a Joint Working Group on Israeli-Palestinian Relations,
which is still in progress.2 For the first time in our program, this group has
been engaged in producing jointly authored concept papers on some of the
final-status issues in the Israeli-Palestinian negotiations and on the future
relationship between the two societies. One of three papers that the group
has published so far—a paper on the problem of Palestinian refugees and
the right of return (Alpher & Shikaki, et al., 1999)—provides the starting
point for my detailed illustration of the negotiation of identity, to be pre-
sented in the next section.
Much of the discussion in our workshops, from the beginning and until
this day, has focused, in effect, on a process of negotiating collective identi-
ties. In our experience, such a process can be productive only if it is based
on mutual respect for the core of the other’s identity and on the principle of
reciprocity. Each side must know that the other does not seek to undermine
200
The Role of Social Identity in Reducing Intergroup Conflict


its group identity, and each must take care not to undermine the other’s
identity. And each must know that the risks it takes in acknowledging the
other’s claims, rights, and authenticity will be reciprocated by the other’s ac-
knowledgment of its claims, rights, and authenticity.
Starting from the understanding that neither side is prepared to negoti-
ate the core of its identity—its peoplehood, its relationship to the land, the
basic justice of its cause—or the general lines of its national narrative, there
remain various elements of each group’s identity that can be “negotiated” in
the interest of mutual accommodation. Let me illustrate some of the possi-
ble changes in identity that can be and have been discussed in problem-
solving workshops and similar encounters and that have, over time, begun
to penetrate the Israeli and Palestinian political cultures.
1. Many members of both communities have become able to remove the
negation of the other’s identity as an integral part of their own identity.
Though the other may still be seen as an obstacle to achieving one’s own
national goals, the other is not as often seen as the antithesis of one’s own
identity whose demise is a condition for one’s own survival. Thus, many 
Israelis have come to accept the reality of Palestinian peoplehood, particu-
larly after observing Palestinians’ readiness to make sacrifices for their na-
tional cause during the intifada and Palestinians’ celebration of the signing
of the Oslo agreement in September 1993. Interestingly, Israelis saw paral-
lels between these events and their own struggle for statehood and celebra-
tion at attaining it—a significant degree of identification with the other
whose existence had previously been denied. Many Palestinians, on their
part, now recognize the right of Israelis to their state, on the grounds that
the state has existed for over half a century and that its dismantlement
would create a new injustice to the generations that were born into it. Very
few Palestinians, on the other hand, are prepared to acknowledge the his-
torical links of Jews to the land, which might be seen as justification for the
establishment of the Jewish state in the first place.
2. We have seen signs of softening of the exclusiveness of group identity,
which allows for the recognition that—despite the validity of one’s own
claims—the other too has valid claims. The recognition of elements of
identity shared with the other, such as a common identification with the
land, opens the way to political solutions based on sharing territory and re-
sources. In a recent workshop, for example, mainstream Israelis and Pales-
tinians were able to agree—much to everyone’s surprise, including their
own—on a formula for sharing Jerusalem: a united city containing the capi-
tals of both states. Public opinion polls on both sides also suggest greater
willingness to share Jerusalem than had been widely assumed (Segal, 1999).
3. Workshop participants have experimented with disaggregating the
monolithic nature of their identities, recognizing that there are different
boundaries of group identity (such as ethnic boundaries, political bound-
aries, boundaries of sentimental attachment) that do not necessarily coin-
cide. This recognition opens the way to the development of transcendent
The Role of National Identity in Conflict Resolution
201


identities, which might allow the two peoples to maintain a common at-
tachment to the country while “owning” only part of that country as their
political state. The concept of a “united country with divided sovereignties”
was discussed in one of our workshops in the early 1980s. In a more recent
workshop, the idea of establishing different kinds of boundaries was ex-
plored in the attempt to find solutions to the problem of Israeli settlements
in the areas in which Palestinians hope to establish their state.
4. Workshop discussions can help to identify outdated elements of
group identity, which refer to maximalist goals and dreams of glory or self-
aggrandizing images that have no current political relevance but poison the
climate for conflict resolution. Examples here might be Palestinian refer-
ences to the armed struggle as the way to eliminate the Zionist entity, or 
Israeli references to the Zionist project of making the desert bloom. Work-
shops have often sensitized participants to words and images that humiliate
and frighten the other and could be discarded with minimal cost to group
identity.
5. In the course of workshop discussions, participants may decide to re-
order the priorities within their national identities, such that certain ele-
ments (e.g., territorial ambitions), which may not have been given up but
have become too costly to pursue, are relegated to low priority and thus be-
come available for negotiated compromise. Thus, over time, Palestinians
(in our workshops and in the larger society) decided to give priority to end-
ing occupation and establishing a Palestinian state over recovering the lost
land in its entirety. Israelis gave priority to maintaining the Jewish character
of Israel over controlling the whole of the land.
6. Finally, workshop participants may “negotiate” changes in national
narratives that accommodate the other’s view of history as much as possible,
such as accepting a share of the responsibility for the course of the conflict.
The above-mentioned concept paper of our current Joint Working Group
on Israeli-Palestinian Relations on “The Palestinian refugee problem and
the right of return” (Alpher & Shikaki, et al., 1999) provides a good illustra-
tion of such an effort to negotiate identity. I shall conclude this chapter with
a more detailed discussion of this concept paper, trying to show what made
it possible to achieve common ground in this effort, what still remains to be
done, and possible directions for achieving further progress.
The Problem of Palestinian Refugees
From the point of view of the struggle over national identity, the problem
of Palestinian refugees is probably the most important and the most diffi-
cult issue to resolve in the Israeli-Palestinian conflict. It may ultimately be
more important and more difficult than the issue of Jerusalem, because it
bears directly on the contrasting views of the justice or injustice of the es-
tablishment of Israel presented by the two national narratives.
202
The Role of Social Identity in Reducing Intergroup Conflict


The problem of Palestinian refugees goes back to the 1948 war that 
followed the declaration of an independent State of Israel. In the course 
of that war, somewhere between 600,000 and 760,000 Palestinian Arabs—
about half of the Arab population of mandatory Palestine—were displaced
from their homes in what became the State of Israel (Morris, 1987, 
pp. 297–298). They were turned into refugees and most of them and/or
their descendants remain refugees to this day. The precise number of Pales-
tinian refugees today is in dispute. The figure given by UNRWA—the
United Nations Relief and Works Agency, which is charged with the wel-
fare of Palestinian refugees—is almost three and a half million (as of Janu-
ary 1998). This number includes Palestinians who registered as refugees
with UNRWA after losing their homes in 1948, as well as their descen-
dants. The majority of this population lives in refugee camps in Gaza, the
West Bank, Jordan, Lebanon, and Syria. Palestinian refugees in Lebanon
(some 350,000) and Syria (some 300,000) by and large remain stateless, as
do many Palestinians living elsewhere in the Arab world. 
Palestinians and Israelis have differed sharply in their account of the
Palestinian exodus in 1948. According to the Palestinian narrative, “the
Arab refugees were forcibly expelled by Jewish forces or left in a panic flight
to escape massacre and . . . they were helped on their way by occasional
massacres, committed by Jewish forces to keep them running” (Alpher &
Shikaki, et al., 1999, p. 173). According to some Israeli accounts, the Arab
population was urged by its own leaders to leave their homes in anticipation
of an early return after the defeat of the Jewish forces by the Arab armies.
The more recent writings of Israeli “revisionist historians” (e.g., Morris,
1987) support the claim that there were indeed some systematic efforts by
Jewish forces to expel the Arab population at various times and places. No
doubt, a combination of circumstances created refugees in a war that was
fought in the midst of the country’s towns and villages.
How the refugee problem is ultimately resolved raises major practical
and symbolic issues for both sides that reach to the core of their respective
identities. By practical issues, I refer to those relating to the implementation
of any agreement on the refugees—that is, to the concrete steps that will be
taken to deal with the plight of the current refugee population. By symbolic
issues, I refer to those relating to the principle of the right of return of the
refugees to their original homes.
At the practical level, Palestinians are concerned with normalizing the
status of the refugees to the satisfaction of the individual Palestinians in-
volved and of the entire Palestinian community. Normalization of the status
of individual refugees could take any one or combination of the following
forms: return to Israel (perhaps their original homes, where that is still fea-
sible), return to the Palestinian state, financial compensation, resettlement
in a country outside of Palestine, integration with citizenship in the state in
which they are now living, or Palestinian citizenship and permanent status
in the state in which they are now living. The options offered to the
The Role of National Identity in Conflict Resolution
203


refugees must be such as to satisfy the entire Palestinian community that
the final peace agreement was not achieved at the expense of any part of
that community.
At the symbolic level, Palestinians are asking for an Israeli acknowledg-
ment of the right of return as a matter of principle, even if they are prepared
to compromise on the implementation of that right. From the Palestinian
perspective, Israeli acknowledgment of the right of return presupposes ad-
mission of direct moral responsibility for the plight of the refugees. Such an
acknowledgment and admission would confirm the Palestinian narrative
about the consequences for the Palestinian people of the Zionist enterprise,
of Jewish settlement in Palestine, and of the establishment of Israel. In 
confirming the Palestinian narrative, such an acknowledgment would also
recognize the Palestinian national rights that are anchored in this narrative.
There seem to be two interrelated, but separable, reasons for Palestinians’
emphasis on this symbolic issue. Psychologically, they need an Israeli ac-
knowledgment that Palestinians have a right to return to the homes from
which they were wrongfully displaced, in order to satisfy their sense of justice
and be able to let go of the conflict. Politically, they need this acknowledg-
ment to support the validity of their claims (such as claims for compensation)
in the bargaining that lies ahead.
For Israelis, too, major practical and symbolic issues are at stake in the
resolution of the refugee problem. At the practical level, Israelis are pri-
marily concerned with the impact of a return of Palestinians in large num-
bers to Israel proper (i.e., to Israel within its pre-1967 borders). A large-
scale return of refugees (beyond a symbolic number of up to 100,000)
would be disruptive to the society because it would raise many divisive
questions about the ownership of homes, villages, and other properties that
have changed hands or been destroyed after the 1948 events. T
o what pre-
cise destination would refugees return whose former houses are now in-
habited by Jewish families or whose homes or villages no longer exist? Fur-
thermore, and most centrally, Israelis are concerned about the impact of a
large-scale return of Palestinians to Israel on the demographic balance and
the Jewish character of the state. In short, they see a large-scale return of
refugees to Israel proper as creating an existential threat to Israel by under-
mining the Jewish majority and the viability and stability of the state. There
is also concern among Israelis about a mass return of refugees to the future
Palestinian state because of the impact on the stability and economy of that
state, but this concern is shared by the Palestinians themselves, who are at
least as mindful as Israelis about the absorptive capacity of the Palestinian
state.
At the symbolic level, Israelis have serious concerns about acknowledg-
ing the principle of a Palestinian right of return, even if Palestinians agree
that this right will not be implemented through a large-scale return of
refugees to Israel proper. Acknowledgment of that principle is particularly
troubling, from the Israeli perspective, if it entails admission of moral re-
204
The Role of Social Identity in Reducing Intergroup Conflict


sponsibility for the plight of the refugees. Such an acknowledgment and ad-
mission would undermine the Israeli narrative about the rightful return of
the Jewish people to its ancestral homeland, the establishment of the State
of Israel, the Arab rejection of partition and attack on the new state, and the
resulting refugee problem. Again, there seem to be two reasons for Israelis’
stance on this symbolic issue and, in particular, for their reluctance to ac-
cept direct moral responsibility for the plight of the refugees. Psychologi-
cally, they are not willing to accept a national self-image that negates the
righteousness of their cause and implies that Israel was “born in sin.” Po-
litically, they are concerned that accepting the principle of the right of 
return and moral responsibility for the refugee problem would keep the 
issue open, even after a peace agreement has been signed, and make Israel
vulnerable to future claims. 
Our Joint Working Group on Israeli-Palestinian Relations discussed, in
considerable detail, the divergent narratives of the two communities; the
fundamental needs, fears, and concerns of each community that would have
to be addressed if a solution of the problem of Palestinian refugees is to be
acceptable to each; and the general principles that must underlie such a so-
lution. The joint concept paper that finally emerged from the group’s dis-
cussion (Alpher & Shikaki, et al., 1999) starts with a review of the two nar-
ratives, of the basic needs of each side, and of the principles that must
govern a solution. It then proceeds to present four possible solutions to the
refugee/right of return issue. Solutions 1 and 2 reflect the traditional posi-
tions—conceivably, the opening negotiating positions—of the Palestinian
and Israeli sides, respectively. Solutions 3 and 4 are compromise solutions,
from a Palestinian and an Israeli standpoint, respectively. These two solu-
tions represent serious efforts, on the part of each side—based on intensive
discussions within the group—to come up with a compromise that would
accommodate the concerns of the other side without jeopardizing its own
vital interests. The paper discusses the advantages and disadvantages of
each solution, highlights the areas of agreement between solutions 3 and 4
as well as the remaining gaps between the two sides, and points to possible
ways in which these gaps might be bridged.
Solutions 3 and 4 represent a considerable narrowing of the gap between
the two sides when compared to the opening positions of solutions 1 and 2.
Still, there remains a significant gap between the two positions, particularly
with respect to the symbolic issues that divide the two sides. Solution 3 calls
for an Israeli acknowledgment of the moral right of Palestinians to return
to their homes based on Israel’s acceptance of responsibility for creating the
refugee problem (although it recognizes that the actual exercise of the right
of return would be limited). Solution 4, the Israeli compromise proposal,
acknowledges that Israel shares practical (but not moral) responsibility for
the plight of the refugees, and accepts the right of return to the Palestinian
state, but not to Israel proper (although it agrees to the return of a limited
number of refugees to Israel). 
The Role of National Identity in Conflict Resolution
205


The third party tried very hard to encourage the participants to bridge
the remaining gaps between solutions 3 and 4 and try to come up with a
joint solution 5—a compromise of the compromises, as it were. We did not
succeed in this effort, however. In part, further movement toward a joint
position on the symbolic issues was blocked by the great difficulty of the as-
signment. We were dealing here with issues that reached to the core of each
group’s identity and associated narrative. For Israelis to acknowledge any-
thing more than shared practical responsibility for the refugee problem and
for Palestinians to accept anything less than an Israeli acknowledgment of
moral responsibility would undermine their respective narratives, and most
members of the group were not prepared to go that far. Furthermore, at
least some of the participants may have preferred to leave the final bridging
of the gap between the two sides to the official negotiators. Since that final
bridging would inevitably entail large concessions on one or both sides,
they may have felt that it could best be done in the context of an overall
agreement, where concessions on the refugee issue could be made with ap-
propriate trade-offs on other issues.
Even though the Joint Working Group did not come up with a joint so-
lution to the refugee problem, the concept paper that it produced repre-
sents a significant contribution to the final negotiation of this issue. The
publication of the paper is in itself significant, in that it is the joint product
of a group of politically mainstream and influential Israelis and Palestinians,
who worked on it over the course of two years and agreed to make it public.
It is also significant that the Israeli and Palestinian authors of the paper ex-
plicitly premise their discussion on the assumption of a two-state solution,
advocating the establishment of a viable, independent Palestinian state. In
the paper, each group recognizes the other’s narrative, without abandoning
its own. The different options for resolving the refugee problem are evalu-
ated in terms of their success in addressing the fundamental needs and fears
of each side. What emerged from this effort is a considerable narrowing of
the gap between the two sides, even though the paper does not entirely
bridge that gap. Solutions 3 and 4 represent substantial movement from the
initial positions reflected in solutions 1 and 2—movement based on each
side’s sincere attempt to accommodate the basic concerns of the other.
With respect to the practical issues—that is, the actual implementation
of the right of return—the gap between the two sides was substantially nar-
rowed in the Working Group’s paper. “Solutions 3 and 4 concur that, at the
practical level, the core of the solution involves four components: a ‘return,’
however defined, of a limited number of refugees to Israel proper; return of
a larger number to the Palestinian state; permanent absorption in host
countries (notably Jordan); and compensation. Both sides agree that a suc-
cessfully negotiated solution should close the file on the refugee issue”
(Alpher & Shikaki, et al., 1999, p. 184). The two solutions disagree about
some important details. For example, the Israelis frame the limited return
of refugees to Israel proper as part of a program of family reunification,
206
The Role of Social Identity in Reducing Intergroup Conflict


whereas the Palestinians frame it as part of the exercise of the right of re-
turn. With regard to compensation, solution 3 calls for both collective and
individual compensation for the refugees’ losses, whereas solution 4 calls
for collective compensation only. Furthermore, solution 4 stipulates a simi-
lar mechanism of collective compensation for Jewish refugees from Arab
countries (albeit without an operational link to the compensation of Pales-
tinian refugees), whereas solution 3 eschews any such linkage. These differ-
ences clearly require further negotiation. But what is important is that the
two solutions agree on the crucial practical issues of implementing the right
of return: Palestinians concede that there will be no large-scale return of
refugees to Israel proper; and Israelis accept the refugees’ right of return 
to the Palestinian state, agreeing to leave it up to Palestinian authorities to
control the inflow of refugees in accord with the state’s absorptive capacity.
With respect to the symbolic issues—those relating to the principle 
of the right of return—the concept paper represents significant movement
toward a mutually acceptable formulation, although there is still a wide gap
between the two positions. Solution 4, the Israeli compromise statement,
acknowledges that Israel “shares practical (but not moral) responsibility, to-
gether with the other parties to the process that culminated in the 1948 war,
for the plight and suffering of the refugees, and that rectification of that
plight by all parties is a central goal of the Arab-Israeli peace process”
(Alpher & Shikaki, et al., 1999, p. 181). This statement comes closer than
past statements by mainstream Israelis to acknowledging the suffering of
the Palestinian refugees and accepting a share of the responsibility for it,
and thus beginning to address the Palestinians’ sense of justice. The move-
ment reflected in this statement was generated by an intensive process of
negotiating identity within the Working Group. Though it represents sig-
nificant movement, this acknowledgment is not sufficient to meet the Pales-
tinians’ call for full Israeli acknowledgment of the principle of the right of
return, based on acceptance of direct moral responsibility for the plight of
the refugees. Solutions 3 and 4, thus, still remain far apart at the symbolic
level. This remaining gap is understandable in that the two solutions,
though they represent serious efforts at compromise, have their point of de-
parture in two sharply clashing national narratives. The question is whether
it might be possible to bridge that remaining gap through a further process
of negotiating identity.
Closing the gap at the symbolic level is clearly difficult because of the
sharp clash in national narratives that gives rise to it. Neither party can be
expected to give up the core of its narrative in the course of the current ne-
gotiations; it is likely to take several generations before such changes can
occur. But it is possible, as part of the process of negotiation and reconcilia-
tion, for the parties to make changes in how they formulate their narratives,
where they place their emphasis, how they present the other within their
own narratives, and what they require by way of acknowledgment or valida-
tion of their narratives. Such changes at the edges of the narratives can po-
The Role of National Identity in Conflict Resolution
207


tentially be “negotiated,” in the sense of being explored and jointly formu-
lated by the parties in a spirit of mutual accommodation.
Thus, in the case of the right of return, we can assume that Palestinians
will not abandon their conviction that they have a moral right to return to
the homes from which they feel they were wrongfully displaced by the 
Israelis, and that Israelis will not admit guilt for the plight of the refugees.
The question is whether it is possible to jointly formulate an Israeli ac-
knowledgment that would be substantially responsive to what Palestinians
need to hear about the right of return without requiring either side to aban-
don these core elements of their narratives. In the terms of my summary,
toward the beginning of this section, of the symbolic issues that the refugee
problem raises for the two sides, the question would be: Can the parties 
formulate an Israeli acknowledgment that would provide significant
(though probably not total) satisfaction of Palestinians’ sense of justice and
validation of their claims for restitution, without requiring Israelis to negate
the righteousness of their cause and making them vulnerable to future
claims?
Let me offer three possible approaches to answering this question, which
can be seen as three steps in the negotiation of identity that might further
narrow the gap between the two sides manifested in solutions 3 and 4.
First, it may be possible to redefine the conception of responsibility in a way
that would not imply guilt and thus be acceptable to the Israeli public, and
yet be viewed by Palestinians as confirming their narrative, validating their
claims, and satisfying their sense of justice. One possibility here is for Israel
to accept a degree of causal responsibility without necessarily accepting
guilt. An Israeli acknowledgment might, for example, describe the refugee
problem as a tragic and unjust outcome of a historical process that arose
with the establishment of the State of Israel. It might go on to say that,
while there are disagreements about the precise historical events and the
distribution of blame, Israel acknowledges that it was an active participant
in this historical process and, indeed, a beneficiary of it, and thus bears part
of the responsibility for these events. The statement might conclude with
an offer of compensation as an expression of that responsibility. A statement
of this kind clearly implies causal responsibility without explicitly acknowl-
edging guilt. It thus has the potential, if properly fine tuned, of an acknowl-
edgment consistent with both the Palestinian and the Israeli narratives.
Second, it may be possible to find a mutually acceptable moral basis for the
right of return, which satisfies Palestinians’ need for an Israeli acknowledg-
ment of the right of return as a moral principle without linking that moral
right to the refugees’ original homes in Israel. One approach here would be
to anchor the moral basis for the right of return in return to Palestine: 
Israelis might acknowledge the Palestinians’ moral right to return to Pales-
tine, based on the proposition that both peoples have legitimate claims to
the land and that sharing the land (by establishing two states within it) 
is, therefore, a just solution to the conflict. Although the two sides differ in
208
The Role of Social Identity in Reducing Intergroup Conflict


their views of the justice and injustice of the events of 1948, the two-state
solution allows them to develop a shared view of justice for the future. 
Israeli acknowledgment of the refugees’ moral right to return to Palestine,
while not fully addressing Palestinians’ sense of justice, would do so to a sig-
nificant degree. It would also strengthen the validity of Palestinian claims
by confirming (in a context of reciprocity) Palestinian national identity and
rootedness in the land. For Israelis, anchoring the moral basis for the
refugees’ right of return in return to Palestine, rather than to their original
homes, would greatly reduce their concerns about future claims, even
though it may not entirely eliminate them.
Third, in conjunction with the two approaches just discussed and possi-
ble other approaches to narrowing the gap between the two sides at the
symbolic level, it may be possible for each side to find ways to affirm the
identity of the other. Such affirmations, perhaps in the form of acknowledg-
ments that the other has integral links to the land and national rights, could
compensate each side for the concessions they make to resolve the refugee
issue: the Israelis for the risks they take in accepting causal responsibility for
the plight of the refugees, and the Palestinians for accepting something less
than the right of the refugees to return to their original homes. Affirmation
of the other’s identity can take place only in the context of a process of ne-
gotiating identity, based on the principle of reciprocity, in which each side
learns to uncouple negation of the other from affirmation of the self. Each
can affirm the other’s identity to the extent that it feels assured that it is not
thereby jeopardizing its own identity and its own claims.
The three approaches to narrowing the remaining gap between the Is-
raeli and Palestinian positions in our Working Group’s paper on the prob-
lem of refugees all represent efforts to negotiate identity. As such, they
search for ways whereby each group can accommodate the narrative of the
other without undermining its own narrative. In such a process, each group
is encouraged to affirm or protect the identity of the other if, in turn, its
own identity is protected or affirmed. Each may be willing to grant some-
thing more to the other than it was originally prepared to offer, or receive
something less from the other than it originally demanded, in return for a
substantial concession by the other at the symbolic level.
Conclusion
In the Israeli-Palestinian conflict, as in other protracted ethnic conflicts, the
ever-present disputes over territory, resources, and political control are ex-
acerbated by perceived threats to national identity and national existence
that underlie the actions and reactions of the opposing communities.
Threats to identity and existence create obstacles to the settlement of con-
flicts, even when both parties have concluded that a compromise agreement
is in their best interest. Moreover, even when specific issues in conflict are
The Role of National Identity in Conflict Resolution
209


settled and political agreements signed—often with the mediation of pow-
erful third parties—these agreements may not lead to stable peace, fruitful
cooperation, and ultimate reconciliation between the two parties, unless
they have formed a new relationship based on mutual respect for their na-
tional identities.
The experience of my colleagues and myself in problem-solving work-
shops on the conflicts in Israel/Palestine, Cyprus, Sri Lanka, and Northern
Ireland, and experiences with similar efforts in other parts of the world,
support the central argument of this chapter about the role of national
identity in intercommunal conflict and conflict resolution. A central feature
of such deep-rooted, protracted conflicts is the struggle over national iden-
tity between the two parties, which leads them to perceive their conflict in
zero-sum terms, not only with respect to territory and resources, but also
with respect to national identity and national existence. Long-term resolu-
tion of such conflicts and reconciliation between the former enemies re-
quires changes in the groups’ national identities; in particular, they require
a redefinition of each group’s identity so that affirmation of its own identity
is no longer predicated on negation of the identity of the other. Such iden-
tity changes are possible, provided they leave the core of each group’s iden-
tity and national narrative—its sense of peoplehood, its attachment to the
land, its commitment to the national language, culture, and way of life—in-
tact. Furthermore, such changes must and can be “negotiated” between the
two groups. It is possible for the groups to accommodate their national
identities and the surrounding narratives to one another, as long as each 
respects and acknowledges the other’s peoplehood and thus reassures the
other that the core of its identity will not be jeopardized by changes in 
the periphery. Although some negotiation of identity takes place around the
negotiating table, it is primarily a process that must engage the entire body
politic on each side, at all levels and through different media. Problem-
solving workshops provide a specially constructed arena for engaging in this
process and jointly developing new insights and formulations that can be
injected into the public debate and can penetrate the political culture on
both sides.
A central lesson from our experience is that national identity, though
very much part of the problem in ethnic conflicts, can also become part of
the solution. The way we talk about our identity affects the way we think
about it and ultimately the way we act on it. In groups that are caught up in
protracted conflict, identity depends on the conflict and is shaped by the
conflict: Many elements of identity are constructed as vehicles for pursuing
the conflict. It should be possible, within limits, to reconstruct these ele-
ments as vehicles for peace and reconciliation. What is needed is an invest-
ment of identity in conflict resolution and in a new relationship with the
former enemy. Development of such a new, transcendent identity confronts
many obstacles. It cannot bypass the political process of negotiating a mu-
tually acceptable agreement, nor can it be allowed to threaten or undermine
210
The Role of Social Identity in Reducing Intergroup Conflict


the particularistic identity of each group. But within these constraints, the
potential for reconstructing the national identities of former enemies in the
service of peace and reconciliation exists and needs to be nurtured.
Notes
1. The work reported here is carried out under the auspices of the Program on
International Conflict Analysis and Resolution (PICAR), which I direct at the
Weatherhead Center for International Affairs, Harvard University. I am grateful to
the Center for providing a home for PICAR, to the William and Flora Hewlett
Foundation for providing financial support for PICAR over a number of years; and
to Donna Hicks, Nadim Rouhana, and my other PICAR colleagues.
2. I am grateful to the organizations that have provided financial support for the
Joint Working Group over the years: the Nathan Cummings Foundation, the
Carnegie Corporation, the Ford Foundation, the Charles R. Bronfman Foundation,
the U.S. Information Agency, and the William and Flora Hewlett Foundation, as
well as the Renner Institut in Vienna and the Weatherhead Center for International
Affairs.
References
Alpher, J., & Shikaki, K., with the participation of the additional members of the
Joint Working Group on Israeli-Palestinian Relations (1999). Concept paper:
The Palestinian refugee problem and the right of return. Middle East Policy, 6 (3),
167-189. (Originally published as Weatherhead Center for International Affairs
Working Paper No. 98-7. Cambridge, MA: Harvard University, 1998.)
Anderson, B. (1983). Imagined communities: Reflections on the origins and spread of na-
tionalism. London: Verso. 
Burton, J. W. (1969). Conflict and communication: The use of controlled communication in
international relations. London: Macmillan.
Burton, J. W. (1979). Deviance, terrorism and war: The process of solving unsolved social
and political problems. New York: St. Martin’s Press.
Burton, J. W. (1984). Global conflict: The domestic sources of international crisis.
Brighton, England: Wheatsheaf.
Gerner, D. J. (1991). One land, two peoples: The conflict over Palestine. Boulder, CO:
Westview Press.
Halpern, B. (1969). The idea of a Jewish state (2nd ed.). Cambridge, MA: Harvard
University Press.
Harkabi, Y. (1986). Israel’s fateful hour. New York: Harper & Row.
Hertzberg, A. (1973). The Zionist idea. New York: Atheneum.
Kelman, H. C. (1972). The problem-solving workshop in conflict resolution. In
R.L. Merritt (Ed.), Communication in international politics (pp. 168–204). Urbana:
University of Illinois Press.
Kelman, H. C. (1978). Israelis and Palestinians: Psychological prerequisites for mu-
tual acceptance. International Security, 3, 162–186. 
Kelman, H.C. (1979). An interactional approach to conflict resolution and its appli-
cation to Israeli-Palestinian relations. International Interactions, 6, 99–122.
The Role of National Identity in Conflict Resolution
211


Kelman, H. C. (1982). Creating the conditions for Israeli-Palestinian negotiations.
Journal of Conflict Resolution, 26, 39–75.
Kelman, H. C. (1987). The political psychology of the Israeli-Palestinian conflict:
How can we overcome the barriers to a negotiated solution? Political Psychology,
8, 347–363.
Kelman, H. C. (1988, Spring). The Palestinianization of the Arab-Israeli conflict.
The Jerusalem Quarterly, 46, 3–15.
Kelman, H.C. (1992). Informal mediation by the scholar/practitioner. In J. Bercov-
itch & J. Rubin (Eds.), Mediation in international relations: Multiple approaches to
conflict management (pp. 64–96). New York: St. Martin’s Press.
Kelman, H. C. (1995). Contributions of an unofficial conflict resolution effort to the
Israeli-Palestinian breakthrough. Negotiation Journal, 11, 19–27.
Kelman, H. C. (1997a). Nationalism, patriotism, and national identity: Social-
psychological dimensions. In D. Bar-T
al & E. Staub (Eds.), Patriotism in the lives
of individuals and nations (pp. 165–189). Chicago: Nelson-Hall. 
Kelman, H. C. (1997b). Negotiating national identity and self-determination in 
ethnic conflicts: The choice between pluralism and ethnic cleansing. Negotiation
Journal, 13, 327–340.
Kelman, H. C. (1997c). Some determinants of the Oslo breakthrough. International
Negotiation, 2, 183–194.
Kelman, H. C. (1997d). Group processes in the resolution of international conflicts:
Experiences from the Israeli-Palestinian case. American Psychologist, 52, 212–220.
Kelman, H. C. (1998a). Social-psychological contributions to peacemaking and
peacebuilding in the Middle East. Applied Psychology: An International Review,
47(1), 5–28.
Kelman, H. C. (1998b). The place of ethnic identity in the development of personal
identity: A challenge for the Jewish family. In P. Y. Medding (Ed.), Coping with
life and death: Jewish families in the twentieth century (Studies in Contemporary Jewry
XIV) (pp. 3–26). Oxford: Oxford University Press.
Kelman, H. C. (1999). The interdependence of Israeli and Palestinian identities:
The role of the other in existential conflicts. Journal of Social Issues, 55(3),
581–600.
Khalidi, R. (1997). Palestinian identity: The construction of a modern national conscious-
ness. New York: Columbia University Press.
Mendelsohn, E. (1989). A compassionate peace: A future for Palestine, Israel, and the
Middle East (rev. ed.). New York: Farrar, Straus & Giroux.
Morris, B. (1987). The birth of the Palestinian refugee problem, 1947-1949. Cambridge:
Cambridge University Press.
Muslih, M. (1988). The origins of Palestinian nationalism. New York: Columbia Uni-
versity Press.
Muslih, M. (1990). T
owards coexistence: An analysis of the resolutions of the Pales-
tine National Council. Journal of Palestine Studies, 19 (4), 3–29. 
Rouhana, N. N., & Kelman, H. C. (1994). Promoting joint thinking in international
conflicts: An Israeli-Palestinian continuing workshop. Journal of Social Issues, 50
(1), 157–178.
Segal, J. M. (1999). Defining Jerusalem. Middle East Insight, 14 (1), 27–28, 51–54.
T
essler, M. (1994). A history of the Israeli-Palestinian conflict. Bloomington and 
Indianapolis: Indiana University Press.
212
The Role of Social Identity in Reducing Intergroup Conflict


Richard D. Ashmore
9
Lee Jussim
David Wilder
Jessica Heppen
Conclusion
Toward a Social Identity Framework 
for Intergroup Conflict
We began this volume with two goals: (1) to review and organize social sci-
entific knowledge, across several disciplines, about social identity and inter-
group conflict; (2) to set the stage for future work by highlighting the theo-
ries, concepts, and methods with the most promise. These goals provide the
basic structure for this concluding chapter. First, we take a look back: we
summarize the chapters, distill the major points, and identify areas of agree-
ment and disagreement. Second, we take a look forward: again drawing
from the preceding chapters, we present a framework that highlights the
ideas that we feel should be the highest priority as we and other social and
behavioral scientists attempt to construct a more comprehensive and coher-
ent social identity perspective on intergroup conflict.
A Look Back 
As noted in the introductory chapter, the authors were given considerable
freedom in how they carried out their task, and this shows in the types of
contributions they have produced. Both chapters in Part I are conceptual
analyses; however, Brewer presents a general theory, whereas Eriksen uses
conflicts in three societies to illustrate his key points. In Part II, the chap-
ters by Citrin, Wong, and Duff and Sidanius and Petrocik report empirical
investigations, both based on survey data. In Part III, White presents a de-
tailed analysis of the group-based violence in Northern Ireland, while Staub
describes the place of self and identity within a general framework for un-
derstanding genocide. In the single chapter in Part IV, Kelman delineates
213


the ways in which national identity is involved in problem-solving work-
shops designed to help ameliorate the Israeli-Palestinian conflict. We turn
now to a more in depth review of these chapters.
Part I: The Contribution of Individuals’ Identities and the 
Collective Identities of Social Groups to Intergroup Conflict
Although both Brewer and Eriksen present conceptual analyses, there are
clear differences between these contributions. Brewer details a very tight
theory addressing a very specific question, When does ingroup love become
outgroup hate? This theory is stated in the most general and abstract form
and is similar to the grand hypothetico-deductive theories of learning in the
1940s (e.g., Hull, 1943) and to social psychological models of the 1950s
(e.g., Festinger’s theories of communication [1950], social comparison
[1954], and cognitive dissonance [1957]). And, as is true of this latter group
of conceptual frameworks, it is individualistic in orientation, focusing on
the thoughts and feelings of individuals, though admittedly in a complex 
social environment. Further, the model seeks support and testing primarily
in laboratory experiments. 
Brewer begins with Sumner’s notion of ethnocentrism: social groups ex-
ist in a state of war with one another, and, as a consequence, ingroup love
and outgroup hate are functionally interconnected sentiments necessary for
group survival. Brewer, however, suggests that Sumner’s theory can be de-
composed into four separate propositions: (1) social categorization—indi-
viduals and groups partition the social world into categories, and this cate-
gorization involves ingroup assimilation and outgroup contrast, which
make social categories more distinct in individual and collective representa-
tions than in reality; (2) ingroup positivity—given that self is a member of
some groups and that self is positively valued, there is a tendency to view
one’s group (the ingroup) as good; (3) intergroup comparison—individuals
compare their group to other groups, which can lead to judgments, not just
that we are good but also that we are better than other groups; (4) outgroup
hostility—the preceding steps, together with certain additional conditions, lead
people to hate an outgroup and actively work to harm it. Brewer suggests
that these four propositions are not highly correlated, as Sumner argued,
but instead constitute a progression of possible relations between ingroup
orientations and feelings about, and behavior toward, the outgroup.
Brewer further proposes that social categorization and ingroup positivity
are likely universal. She also draws an important implication from the latter
principle. Much discrimination is driven by ingroup positivity and is, thus,
favoritism for one’s own group, rather than discrimination against an out-
group. She presents laboratory experimental evidence in favor of this pro-
posal. She also notes that an important task for future research will be to
develop ways of disentangling pro-ingroup from anti-outgroup discrimina-
tion in real-world settings.
214
Conclusion


This raises the question that Brewer began with, When does ingroup
love (and “discrimination for us”) turn to outgroup hate (and “discrimina-
tion against them”)? Brewer argues that it is not “realistic, objective con-
flict” that provides the significant impetus. Rather, she suggests that the an-
swer lies within her optimal distinctiveness theory of social identity, which
she links to evolutionary theory. Her argument is that humans evolved, not
as isolated individuals or even in families, but instead as members of larger
social groups. T
o allow the individual to learn from and cooperate with oth-
ers, humans had to trust other people. However, to trust all others would be
dangerous. Thus, human beings developed the propensity to trust all others
within their group (“depersonalized trust”). When groups encounter one an-
other, each operating according to this same orientation gives rise to uni-
versal ingroup-outgroup stereotypes: “we” are close knit and cooperative;
“they” are clannish and treacherous. 
The final ingredient in the progression to outgroup hate is provided by
emotions. When members of different groups interact, they feel uncom-
fortable with one another and misattribute this negative emotional state to
“They are___________.” Under conditions of limited contact and little 
social mixing, the dominant emotions are likely to be contempt (“They are
inferior and repugnant”), and this should motivate distancing (Karen 
Horney’s “moving away from” [Ashmore, 1970]). However, under condi-
tions of close contact and possible integration, people will likely feel 
invaded, and this will give rise to anger and fear (“They are violent and dan-
gerous”). According to Brewer, it is these social conditions and associated
feelings and attributions that turn ingroup love into outgroup hate and 
produce group-based harm-doing (Horney’s “moving against”). 
In contrast to Brewer’s development of a general theory of group identi-
fication and intergroup conflict with most evidence provided by laboratory
research, Eriksen is concerned with identity politics, especially how social,
especially ethnic, identities are manipulated by political elites, and experi-
enced and enacted by everyday citizens, in real-world ethnic conflicts. Ac-
cordingly, his data are not from controlled experiments, but rather are case
studies of three actual “ethnic conflicts,” though it is clear from his analysis
that he would not accept this label. Eriksen builds his case, not by a step-by-
step elaboration of a small set of ordered propositions as Brewer does, but
instead by first describing the current position in cultural anthropology
concerning ethnicity and its implications for intergroup conflict. This posi-
tion essentially argues that ethnicity is not primordial (a social category de-
fined by shared biological descent or ancestry and thus a grouping that is
fundamental and enduring for both the group itself and individual group
members) and not centrally about what a group of people share. Instead,
ethnicity is best seen as a set of socially constructed group and interpersonal
performances that communicate to self and others the boundaries between
groups, which are not fixed entities but rather collectives in flux. T
o capture
this view, many anthropologists have turned from studying ethnicity per se
T
oward a Social Identity Framework for Intergroup Conflict
215


to seeking to understand social identities. One implication of this position is
that group identity is conceptualized as both conscious and strategic—indi-
viduals are aware of their multiple social identities and deploy them situa-
tionally in order to achieve personal and group goals. 
Although Eriksen generally accepts this basic position, he notes that it
leaves the person and his or her everyday experiences out of the ethnicity-
identity-intergroup conflict equation. He proposes that, in addition to con-
ceiving ethnicity as about social construction, boundaries, and strategic en-
actment, it is necessary to consider the personal identities of individuals and
how these can become fused with social identities. More specifically with
regard to intergroup conflict, he suggests that, when the lived experience of
individual group members is consonant with the key elements of a social
identity currently proffered by authorities, it sets the stage for destructive
intergroup conflict fought in the name of that group identity. Thus, for ex-
ample, although Bosnians and Serbs in Bosnia shared many cultural ele-
ments and in multiple ways were quite similar to one another, they had
lived separate lives in the private sphere, with little intermarriage and few
cross-group close friendships. This situation of separate private spheres al-
lowed group myths and pejorative outgroup beliefs to be part of the every-
day experience of members of each group. Thus, when leaders espoused the
separate and conflicting religiously based social identities of Bosnian and
Serb, the citizenry had personal selves and everyday experiences that made
these appeals resonate. 
Although they take two very different approaches and seem to offer in-
commensurate, if not incompatible, frameworks for understanding inter-
group conflict, the Part I contributions are actually much closer than might
appear at first glance. Brewer starts with the individual and does a nice job
of identifying and ordering the cognitive and affective factors that set the
stage for ingroup attraction and, in turn, outgroup hostility. She concludes
by describing factors that moderate the model she has constructed. These
moderators are all societal variables (e.g., social structural complexity), 
and it is precisely such factors that Eriksen elaborates in his contribution.
Eriksen draws the circle complete by his insistence that only when socially
propagated group identities fit with the lived experience of individuals and
their personal sense of self will the stage be set for ethnically based violence.
In complementary ways, then, both Brewer and Eriksen note the impor-
tance of individual and social factors in the ways in which social identity is
implicated in intergroup hostility.
The foregoing should not be construed as meaning that Brewer and
Eriksen are in total agreement about the personal and societal underpin-
nings of group-based harm doing. The clearest point of friction concerns
the role of objective conflict between groups as a precipitator of hostile re-
lations. Although he never uses the term, Eriksen at several points presents
ideas consistent with the realistic group conflict (RGC) view of intergroup
conflict (Ashmore & Del Boca, 1976; Boulding, 1962). According to this
216
Conclusion


perspective, groups come into conflict over limited material resources—
land, energy sources such as oil, and the like. Whereas Eriksen identifies
objective competition as crucial to conflict development and proliferation,
Brewer sees such “realistic” competition as much less important. In fact, she
might be read as suggesting that tangible conflicts come only after the stage
is set by the elements offered in her model.
The second major disagreement between Brewer and Eriksen is equally
instructive. Brewer’s presentation is all about the here and now, about indi-
viduals’ thoughts and feelings about me, us, them, in the present moment.
Eriksen, on the other hand, grounds his analysis in history. For each of his
three examples, he describes what happened in the past, not just as stage
setting but also as crucial to understanding why the groups are fighting
now. This difference reflects, in part, two contrasting disciplinary para-
digms. Brewer is a psychological social psychologist, and present-day prac-
titioners of this social science continue to be much influenced by Kurt
Lewin’s proscription against “historical” or “genetic” accounts: do not ex-
plain a person’s current behavior by recourse to what happened to that per-
son in her childhood; instead account for action on the basis of the contem-
poraneous field of forces acting on the person. Although Lewin’s ahistorical
approach referred to the life history of the individual, many of his students
who, in turn, shaped today’s social psychologists, took Lewin’s approach
more generally to mean that the important determinants of action are cur-
rent social factors. As a consequence, not only has the life history of the in-
dividual been left to developmental psychologists, but the past of social
groups has been left for historians. In contrast, social and cultural anthro-
pologists, viewing culture as, in part, something that one generation passes
to the next, have been much more inclined to see a group’s history as im-
portant to understanding its present organization and functioning—Thus,
Eriksen’s descriptions of current conflicts feature historical information. 
One important next step will be to bring these perspectives together.
One way, which will be elaborated below, is to extend the notion of life
story and personal narrative into the analysis of social identity and group-
based conflict. The basic notion is that, as individuals and groups, we make
sense of who we are now, in part, by looking back and creating stories—
about self, about our ethnic group, about our nation. And it is these stories,
as part of personal and group identities, that may be crucial in setting the
stage for violence and oppression done in the name of us vs. them.1
Part II: The Contribution of Ethnic and National Identities to 
Political Conflict in the United States
The two chapters in the second part follow very well from Eriksen’s contri-
bution. Methodologically, he suggests that it is useful to look closely at spe-
cific intergroup conflicts; substantively, ethnic conflict within nation-states
is a major issue around the globe and is not limited to Third World or de-
T
oward a Social Identity Framework for Intergroup Conflict
217


veloping countries. In Part II, the focus is on patterns of ethnic divergence
and convergence in just one nation, the United States, arguably the most
economically advanced country in the world. Further, only one type of con-
flict is considered: political contentions among ethnic groups.
Citrin, Wong, and Duff’s central concern is American national identity:
What is it? Are there ethnic group differences in beliefs about American
identity? How do individuals prioritize ethnic and national identity? How
do ethnic and national identity relate to policy preferences? Citrin and his
co-workers set the stage for their empirical research by carefully laying out
their notion of social identity, in this case national identity. Following T
ajfel
(1978) they conceive of national identity as involving both “identification
as,” or self-categorization (e.g., “I am an American”) and “identification
with,” or affective attachment to the nation. Regarding the latter, they fur-
ther distinguish patriotism, or love for one’s country (“Being American is
very important to me”) from chauvinism, or feelings of national superiority
and degrading of other countries (“If only other countries were like
America the world would be a much better place”). 
Citrin et al. suggest that they go beyond T
ajfel (a point we return to be-
low in considering the multiple dimensions of social identity) by adding a
third component to national identity—normative content. The use of the
word “normative” is deliberate and does not mean average. Instead, norma-
tive conceptions are assertions about how society should be organized, how
people should behave to make society work well, and how society should
treat groups and individuals in the nation. As political scientists, Citrin et al.
bring an important theme to the volume—normative as pre- and pro-scrip-
tive, not just descriptive. This reflects the subfield of political theory and
the issue of how political systems should be constructed. Given the empha-
sis on ethnicity and national identity, Citrin and his colleagues stress three
different normative views of how ethnic and national commitments should
be related in multiethnic states: liberal, in which national membership de-
pends on sharing a set of ideals (e.g., respect for the individual, freedom)
and not shared ethnicity (hence, it is often termed civic, not ethnic, nation-
alism); nativism, in which American identity is equated with one ethnic
group, white Anglo Saxons, and thus, to be an American means accepting
the values and other characteristics of this dominant ethnic group; multicul-
turalism, in which ethnicity is prized above national identity.
Citrin et al. also preface their data reporting by introducing symbolic
politics theory (Sears, 1993), which stresses that political orientations are
socialized in childhood and maintained by adult engagements (e.g., work).
According to the authors, national identity is widely socialized in the
United States, and, thus, there should be more ethnic consensus than con-
flict concerning national identity. 
Citrin et al. address their questions about national identity and predic-
tions from symbolic politics theory using survey data collected from both a
national sample and Los Angeles County respondents. Concerning their
218
Conclusion


first question, What is American identity?, they found considerable agree-
ment that, in regard to political issues, most Americans define self as “just
an American” (self-categorization) and feel considerable pride in and patri-
otism for America, but not high levels of chauvinism. Also, there was con-
siderable support for nativistic beliefs (though the pattern was uneven) and
even more support for assimilation notions. 
In answer to their second question, Are there ethnic group differences in
views on American identity?, Citrin et al. concluded that there are few dif-
ferences in defining American identity except on issues pertaining to ethnic
groups per se (e.g., how the nation treats minority groups). A most intrigu-
ing finding was that, on average, African Americans were more nativist than
whites. This was especially true for endorsing being a Christian as a crite-
rion for being “a true American.” As the authors note, this suggests yet an-
other important social identity, religion. 
Turning to the third question, How do individuals prioritize ethnic and
national identity?, Citrin and colleagues found that, when given the option,
about 50 percent of African Americans chose “both ______ [my ethnicity]
and American,” about one-third chose ethnicity as their primary identity,
and about one-sixth chose “just American.” Hispanics and Asians responded
similarly, but were somewhat less likely to select “just American” and some-
what more likely to choose “both _______ [my ethnicity] and American”
and ethnic identity. For whites, on the other hand, “just American” was far
and away the predominate response. 
The final question was, How do ethnic and national identity relate to
policy preferences? As predicted by symbolic politics theory, race (given
small sample sizes, only whites and blacks were compared) did not predict
views on multicultural policies or international affairs (military spending or
protectionism), whereas national identity did. On race-related policies, race
did predict policy preferences. However, when ethnic identification was
added to the predictive equation, felt closeness to one’s ethnic group did not
enhance support for race-based policies for blacks, but for whites felt close-
ness increased disapproval of programs that benefited minority groups.
That it was among whites that identification with one’s group related to po-
litical issues regarding race may seem paradoxical to some given that the
practice of identity politics is most often associated with minority groups.
The Citrin et al. contribution highlights three points. First, social identity
is multidimensional. Their chapter considers self-categorization, affective
attachment, and normative content. Below we identify additional possible
facets of social identification. Second, a particularly important aspect of ex-
periencing ethnicity, though often neglected, is the “shoulds” attached to
one’s group. These are related to the notion of normative content, but they
focus, not on broad issues of what it means, for example, to be a true Ameri-
can, but on the very specific group norms for how “one of us” should act—
the food we should eat, the way we should talk, and the like—and the pres-
sures an individual feels from members of her or his own group to abide by
T
oward a Social Identity Framework for Intergroup Conflict
219


these “rules.” T
o address this issue, Contrada et al. (in press) have recently
developed a self-report instrument to assess individual differences in own-
group pressure to behave in specific ways. Third, as Citrin and his colleagues
note, social identities are not necessarily political identities. This raises the
crucial question, How do people get from social definition to action? 
In the second chapter of Part II, Sidanius and Petrocik take up much the
same general topic as Citrin and his collaborators; the relationships be-
tween ethnicity and national attachment in the United States. And like 
Citrin et al., Sidanius and Petrocik identify three perspectives on this issue:
melting pot (each ethnic group contributes something to an emergent 
national American identity, and, in the process, members of these ethnic
groups lose their own sense of ethnicity and become identified with, and 
attached to, the nation); pluralism (individuals and groups retain their
uniqueness, but all agree to be tolerant and to also identify with the nation);
group dominance model (in all societies, groups are arranged in a hierarchy
of prestige and power, and this means that groups not only maintain their
distinctive identity, but also that, for those at the top, ethnocentrism and
patriotism are positively linked, but for those at the bottom this correlation
is negative). Again in parallel with the preceding chapter, Sidanius and
Petrocik assess with survey data the predictions they derive from each per-
spective. These data come from five different samples, from university stu-
dents, to Los Angeles residents, to a national sample. 
Sidanius and Petrocik address three specific issues: the salience of ethnic
identity; the relation between ethnicity and patriotism; the relationship be-
tween patriotism (attachment to the nation) and ethnocentrism (by which
they mean a negative attitude toward American ethnic groups relative to
one’s own group). In regard to the first issue, Sidanius and Petrocik con-
clude that ethnic identity is salient for Americans and that this is especially
true of minority-group Americans. With respect to the second issue, Sida-
nius and Petrocik find that white Americans are highest in patriotism and
African Americans lowest. Concerning the third issue, Sidanius and Petro-
cik report that, for white Americans, patriotism and ethnocentrism are of-
ten positively correlated, but the correlation for minority Americans is gen-
erally negative. 
Of the many excellent points in the Sidanius and Petrocik contribution,
we underline just one, their group dominance model. Individuals develop
their intergroup attitudes and ethnic groups come into conflict within soci-
ety, and all known societies contain group dominance structures: Some so-
cial categories are privileged and others are disadvantaged. Thus, the soci-
etal context for intergroup relations is never neutral or even-handed, and
this is true even in economically advanced nations with strong democratic
traditions. This is a particularly important reminder at this time in the
United States, where we surmise that a large portion of the dominant white
populace believes that the civil rights and women’s movements have some-
how leveled the playing field for all Americans and now it is time to treat
220
Conclusion


everyone as individuals and not as group members. Without denying the
social changes that have occurred over the past four decades, it is simply not
true that American society and culture treat all people and groups as equal. 
The chapters by Citrin et al. and Sidanius and Petrocik present a para-
dox: both address the same basic issue (the interplay between ethnic and na-
tional identity in the United States), they deal with two similar folk models
(though they use somewhat different labels—liberal [Citrin et al.] and plu-
ralism [Sidanius & Petrocik]; nativist [Citrin et al.] and Americanization
and melting pot [Sidanius & Petrocik]; see also Fredrickson [1999] for a
historical review of models of American ethnic group relations), they both
draw on similar previous studies (especially that by de la Garza, Falcon, &
Garcia, 1996), and they take the same basic methodological approach (sur-
vey data; drawing in one case from the same survey). But they reach quite
different conclusions: Citrin and co-workers see a small ethnic divide in the
United States and little tension between ethnic and national identity,
whereas Sidanius and Petrocik perceive a large ethnic divide and consider-
able conflict between ethnic and national attachment. 
There are three factors that make this paradox more apparent than real.
First, they use different items to assess ethnic/national self-categorization.
Second, the two chapters are not really directed at precisely the same set of
phenomena. Citrin et al. predict policy preferences from ethnicity and eth-
nic identity, whereas the primary outcome variable of interest to Sidanius
and Petrocik is the correlation between national attraction (especially pa-
triotism) and prejudice against U.S. ethnic groups other than one’s own
group. Third, and we believe most important, in reporting results, Citrin
and colleagues concentrate on absolute levels of attachment to the United
States (which are generally high) and play down group differences (which,
for some measures, are substantial), whereas Sidanius and Petrocik stress
group differences and deemphasize absolute levels of national attachment
by ethnic group members. 
The Citrin et al. and Sidanius and Petrocik chapters underline four im-
portant points. First, how one asks people to report their group self-catego-
rization is crucial. Some items implicitly pull for “just American,” whereas
others pull for ethnicity and nationality self-codings either instead of, or in
addition to, labeling self as “American.” Our recommendation is that sur-
veys concerned with ethnicity and national identity should use the 1995
LACSS item (described by Citrin et al.) that offers the respondent the fol-
lowing choices: “just an American,” “_______ [whichever national or ethnic
group the person sees self as a member of],” “both American and ______
[self-categorized ethnic group].” This seems to us to provide the most neu-
tral way of getting at how respondents define and categorize self in terms of
a superordinate national identity (“American”) and more particular ethnic
self-codings. Although the results reported by Citrin et al. are from a small
and local sample, they suggest that the modal white American sees self as
“just American,” while minority Americans define self in three very distinc-
T
oward a Social Identity Framework for Intergroup Conflict
221


tive ways: about half select a dual identity, about one in three categorizes
self in terms of ethnicity only, and about one in six regards self as just
American.
Second, the Citrin et al. and Sidanius and Petrocik chapters remind us
that it is easy to slip from ethnic identity, or identification, to ethnicity, or
simple ethnic group differences. Although both chapters report analyses
dealing with individual differences in group identification (especially self-
labeling), they also report analyses in terms of ethnic group differences.
However, simple ethnic group differences do not necessarily mean differ-
ences as a function of ethnic or national identity. In addition, such an ap-
proach obscures the wide variation within any social category in the content
and degree of identification with the group. For example, what other social
identity, intergroup attitude, and public policy preference differences are
there among minority group Americans who self-label as “just American,”
“American and ____,” and minority group label only?
Third, ethnic groups are not monolithic. Obviously there is variation by
age, socioeconomic status, and a variety of other variables that Citrin et al.
and Sidanius and Petrocik partial out in some analyses. At the same time,
some of these partialed-out variables constitute important social identifica-
tions. For example, there is considerable variation in the African-American
(and white) community in terms of religious commitment, especially defin-
ing self as a “fundamentalist Christian.” 
Fourth, just as ethnic and national identity are multifaceted, so too is 
national attachment. Both chapters note that there is a positive form of na-
tional attachment (patriotism) and a more virulent brand (chauvinism/
nationalism—basically “We are the best in the world . . . and others
should emulate us”), as well as issues of willingness to leave the country and
beliefs about what makes someone a true American. 
The foregoing summary suggests that the Citrin et al. and Sidanius and
Petrocik chapters have not resolved the question of how ethnicity and eth-
nic identity relate to national identity and attachment. They have, however,
provided some important clues. In general, there is fairly high attachment
to the United States in all groups, but there are also substantial differences
between groups, especially on issues pertinent to ethnicity. More impor-
tant, the two chapters begin to point us toward the necessary next genera-
tion of research. This will need to use multiple measures of ethnic or na-
tional self-categorization, especially items that do not tilt toward or away
from ethnic (or national) identification, and multiple measures of national
attachment. Thus, future work will be directed at two important questions:
How do the various facets of ethnic and national identity (self-categori-
zation, felt closeness, and the like) relate to the various components of 
national identity? How, in turn, do these sometimes competing and some-
times complementary forms of self-identification relate to variables perti-
nent to intergroup conflict (e.g., attitudes toward other groups, political
policies and programs)? 
222
Conclusion


Part III: The Contribution of Social Identity to Violent Intergroup Conflict
White continues the focus of Part II to look intensively at one conflict, but
the site of his analysis is Northern Ireland, not the United States, and the
type of conflict is deadly violence, not political strife. White begins his
chapter in a very dramatic fashion by presenting a table indicating who has
killed whom in the conflict. This table not only helps him make some very
important points about the conflict but also reminds the reader that the ul-
timate outcome of identity-based disputes is that people may kill one an-
other. Of the many excellent points made by White, we highlight three: (1)
the past importantly shapes the contemporary conflict; (2) the violence in
Northern Ireland is three- not two-sided; (3) identity is involved in the con-
flict in multiple ways. 
First, the current fighting is rooted in a long history, and that past not
only shapes the present conflict but also sets parameters for how the hos-
tility might be reduced. White describes how the violence in Northern Ire-
land began with the Norman/English invasion of Ireland over 700 years
ago, how religion was introduced into the conflict in the 16th century with
the Reformation, how a Protestant ascendancy was created during the 17th
century, how Northern Ireland was created in the early part of this century,
and how the current protracted violent confrontation began in the 1960s. 
This history is important to the current conflict in three major ways.
1. It underlines the importance of losing for group definition. On multi-
ple occasions over seven centuries, the Irish have lost battles and wars to the
English. These repeated defeats have been incorporated into the group nar-
rative of the Republicans of Northern Ireland. This parallels the history
and national narrative of Serbs as described by Eriksen (this volume; see
also Staub this volume). This 700-year record also suggests why the Irish
Republican Army may be so reluctant to “decommission,” which, as this
chapter is being written (spring 2000), may derail the recently agreed to
peace accord. The Irish have lost many times to the English; they can not
accept another “defeat,” which is how many perceive decommissioning,
even if the current agreement is widely supported by Protestants and
Catholics.
2. The Protestants in Ireland were imported by the British and took
over land and status from defeated Irish. They have always been in a mi-
nority in Ireland as a whole (though a majority in Northern Ireland), and in
White’s words, have a “siege mentality.” It is likely that being a numerical
minority with political and economic majority status (as whites in Southern
Africa and the Tutsi in Rwanda until recently) also has important implica-
tions for social identity. Groups that have superior power but are numeri-
cally in the minority (power majority, numerical minority) likely stereotype
the other group (power minority, numerical majority) as, at the same time,
both immature and childlike (to justify their inferior position) and poten-
tially violent in a primitive way (to make sense of past uprisings and fears of
T
oward a Social Identity Framework for Intergroup Conflict
223


what might happen if “they” won). This is much the same set of images that
characterized U.S. whites’ views of blacks during the slavery period (Boskin,
1970). In complementary fashion, the self-image of power majority yet nu-
merical minority groups features traits such as grown up and mature (re-
sponsible, productive, and the like) and civilized.
3. As White notes, a long history of fighting an external oppressor also
sets the stage for role identities in this struggle, a point to which we return
below. Here we simply note that often roles are hard to give up, thereby
making intergroup conflict difficult to reduce.
Second, what looks to an outsider as a simple two-sided religious con-
flict—Northern Ireland Catholics versus Protestants—is, upon more care-
ful analysis, really a multisided political conflict with a religious gloss.
White shows that there are three key contestants—Protestant paramili-
taries (who are just a portion of the Protestant community); Irish Republi-
cans (most IRA members are Catholic, but not all Catholics are affiliated
with the IRA, and some IRA members are Protestant); and the state, espe-
cially British authorities. At the end of his analysis, White notes that there is
actually a fourth important party, the people and government of the Repub-
lic of Ireland, and, with the current efforts at peace, the role of that party
will likely be increasing. The United States has also been involved in the re-
cent efforts to resolve the conflict, and thus there is a fifth “side.” The cru-
cial point for the present volume is that apparently “us against them” con-
flicts are seldom just two-sided; and the various parties are likely to have
different “identity shares” in the conflict, which leads to the next major
point we wish to underscore about White’s chapter.
Third, identity is implicated in multiple and diverse ways in the conflict.
T
o uncover how identity enters into the political violence, White focuses on
accounts given by the various participants—quite simply how do various
participants describe the conflict and their role in it? White begins his
analysis by using Thoits and Virshup’s (1997) distinction of “we” from “me”
social identities. The former are collective social identifications (member of
some social category [e.g., nation, ethnic group]), whereas the latter are role
identifications (occupant of a particular role [e.g., parental, occupational]).
White argues that for Republicans their involvement reflects both a collec-
tive social identity (“we”) and role identity (“me”). Importantly, the collec-
tive self-definition is not me as Irish Catholic versus Irish Protestants, but
me as Irish Nationalist versus the British. The role identity is a set of enact-
ments that the person has learned from either growing up in a Republican
home or as part of joining a Republican organization. White illustrates a
Republican role identity by describing the hunger strikes of the early 1980s.
He notes that hunger strikes have a long history in Republican efforts to get
the British out of Ireland, and this means not only that these are part of the
“Republican prisoner” role but also that there are specific men who have
gone on hunger strikes (and often died) and who now serve as role models
(see also Aretxaga, 1993). 
224
Conclusion


White’s suggestion that role identities can be important parts of involve-
ment in intergroup conflict is a significant addition to role identity models
(McCall & Simmons, 1966; Stryker,1987). Just as the normal everyday ac-
tivities of society require roles, so, too, do protracted conflicts. What might
these conflict roles be? Obviously, there are the roles of active fighter (sol-
dier) and prisoner. But in addition, “armies” require that other functions be
fulfilled. One important function is intelligence, or the collecting of infor-
mation to assist in waging the fight. White describes a British soldier who
slept outside a bakery; it is possible that the owner of the bakery could have
provided intelligence reports to either the Republican or Protestant para-
militaries and, if so, would have taken on a conflict-related role identity. 
In addition, when an army is an insurgent one, other roles are necessary,
among them, fund-raiser, politician, and civilian support. In one of the Re-
publican narratives reported by White, the respondent describes his grand-
mother, of whom he said, “She never wavered.” She may also have supplied
a safe house to paramilitaries and fulfilled additional functions in support of
others with active fighting roles, and, in so doing, her role of “supporter”
would have been active and not passive. 
According to White, the Protestant Loyalists and British actors in
Northern Ireland construe their involvement solely in “we” social identity
terms. For the Protestant paramilitaries, the conflict is “us against them,”
where “We are British Protestants” and “They are Irish Catholics.” For the
English, White argues that the conflict is “them versus them,” a religious
conflict among the Irish. 
White acknowledges possible role identities for Protestants and British,
but we believe that he underplays them. Although in the earliest years of the
current “troubles,” the Loyalists may have been relatively unorganized with
little sense of history, certainly as the conflict evolved it is probable that
Protestant paramilitary groups became relatively well organized and devel-
oped systems of roles. Role identities are even more probable for the British
or English first-line participants in Northern Ireland, the military person-
nel. It is likely that many English soldiers construed their involvement there
as “me as a soldier doing a soldier’s job,” and not as “we British.”
White’s primary data are verbal accounts, and gathering such narratives
in the midst of an armed conflict is very difficult. This difficulty raises some
questions about White’s conclusions. As he notes, the verbalizations he re-
ports are not factual histories or descriptions, but instead they are stories
that a person has told to someone else. As with any story, the content of the
account depends, in part, on who is talking to whom for what purpose. This
raises the most general question: Are the narratives given by the three
groups of protagonists (Republicans, Protestant paramilitaries, English sol-
diers and elites) comparable? This can be partitioned into two more specific
queries: Who gave the accounts? T
o whom were the accounts given and
how did the respondent think the accounts would be used? Regarding the
first or “who?” question, White acknowledges that the three groups were
T
oward a Social Identity Framework for Intergroup Conflict
225


not equated: many of the Republicans were Sinn Fein members; Protestant
paramilitary accounts were from former paramilitaries; the English voices
were from soldiers at the start of the current conflict (late 1960s) and politi-
cal leaders. Regarding the “to whom and how used?” question, the Republi-
cans were all interviewed by White himself, and it is likely that they were
assured confidentiality and told that the conversations were for research
purposes. The Protestant paramilitary stories and statements by English ac-
tors were taken from secondary sources, most from memoirs and remem-
brances or from official statements. The question of how the respondents
thought the accounts would be used is less certain, but at least some had
clear public relations and political implications. As a consequence, these
statements may have been tilted toward a “we” social identity perspective
and away from explaining the individual’s roles in the conflict. 
While White presents a detailed analysis of just one violent conflict,
Staub describes a general framework for understanding genocide that is
based on, and applied to, several different murderous group conflicts. Basi-
cally, this model indicates that mass killing and genocide are the end result
of a set of personal and social processes that are set in motion by one of two
general types of instigators: difficult life circumstances (e.g., hyperinflation)
and intergroup conflict. These instigators threaten basic needs, including
those related to identity. This sets the stage for a spiral toward large-scale
systematic violence as follows: the psychological elevation of the ingroup
(and generally a derogation of the outgroup), which facilitates initial harm-
doing; this, in turn, produces altered societal institutions and individual
identities. These lead to the final step in the model, mass killing and geno-
cide. T
o this causal sequence, Staub adds two moderators. The first, culture,
involves a variety of societal-level factors (e.g., history of devaluation of 
an outgroup); the second, bystanders, indicates individuals both within 
and outside the group who are not (at least initially) involved in the harm-
doing.
Among the many insights provided by Staub’s model, we highlight five.
First, both group and individual identity play important roles in intergroup
violence. Regarding the former, Staub suggests that ethnic groups and na-
tions vary in their “group self-concept,” and this shared notion of “us” can
lead some groups toward mass killing and genocide while directing others
away from such harm-doing. In particular, Staub highlights as a cultural
factor predisposing toward systematic intergroup violence, “unhealed vic-
timization,” the belief that “we” have been hurt coupled with no opportu-
nity to personally and socially come to terms with this wound. 
Second, and closely related to the first point, a personal and societal
sense of vulnerability (“They have killed our people before, and they could
attack us at any time”) can not only predispose a group toward violence but
also make preemptive violence more likely. As Eriksen noted regarding
Serbs in the former Yugoslavia and White with regard to the Irish Republi-
cans, these groups had histories of losing, which justified, in the partici-
226
Conclusion


pants’ minds, acts of “getting them before they get us.” This type of mind-
set helps sustain these conflicts 
Third, regarding the through-time set of factors leading to intergroup
violence, most important for the present volume is that these involve
changes in individual and group identities. Those in a society who engage
in initial violence are more receptive to self-definitions that make their
harm-doing not only acceptable but, in some cases, also morally right; those
who at the outset are bystanders can come to see killing or support for
killing as part of how “I ought to behave” and alter their self-conception 
accordingly. 
Fourth, in terms of culture moderators, it is possible to identify societal-
level identity types that are more likely to participate in mass killing, and
these are not neatly divided along Eastern versus Western modal self-
concept configurations. In fact, as Staub notes, both the autonomous indi-
vidual self more common in Western industrialized societies and the more
social and connected self-configuration that is more common in Eastern
and less industrialized nations have their own vulnerabilities to genocide.
The challenge for the next generation of researchers is to assess these types
and test Staub’s suggestions. (See Norenzayan et al. [1999] and Lyengar et
al. [1999] for additional ideas about different types of cultures and the im-
plications for intergroup strife.) 
Fifth, third parties play significant roles both in creating conditions ripe
for large scale group violence and as a major resource in limiting or stop-
ping genocide. Staub describes how Belgians built on existing Hutu-Tutsi
dominance systems and added the notion of “blood” to help make the soci-
ety more manageable for them. Years later this made the violence more
likely. 
Part IV: The Role of Social Identity in Reducing Intergroup Conflict
Staub’s emphasis on the importance of bystanders, especially external by-
standers, as crucial to preventing the escalation to mass killing and genocide
leads very nicely into the final part, which focuses on reducing conflict. The
chapter in this section, by Kelman, describes problem-solving workshops as
a conflict-reduction method that features third parties. Kelman begins his
chapter by describing the history of the Israeli-Palestinian conflict. Just as
Eriksen and White suggest that the past is important to understanding why
and how intergroup harm-doing develops, Kelman proposes that efforts to
reduce conflicts must take into consideration the historical foundations of
current strife. Next, Kelman lays out his notion of national identity. He ar-
gues that, “Although national identity is carried by the individual members
of a national group and can thus be studied as a property of individuals, the
present chapter refers to it as a collective phenomenon—as a property of
the group.” In the simplest terms, national identity is the shared and con-
sensual definition of us as a group and nation; this includes our claims to
T
oward a Social Identity Framework for Intergroup Conflict
227


land and continuity over time, our beliefs about who we are, and what
makes us distinctive. Kelman suggests that national identity is accompanied
by a “national narrative—an account of the group’s origins, its history, and
the relationship to the land.” Thus, for Kelman, social identity as a nation
involves both a self-definition (national identity) and group story (national
narrative). 
In seeking to understand how national identity might underlie conflict,
Kelman argues that in protracted intergroup fights, especially in which
both sides have claim to the same land, the groups’ national identities can
become exclusive (e.g., “This land belongs to us and can belong to no one
else”) and monolithic (all the multiple dimensions of nationhood [religion,
ethnicity, history, and the like] become seen as highly intercorrelated). Both
of these properties make conflict resolution more difficult, identity exclu-
sivity by making sharing impossible by definition, and identity unidimen-
sionality by making us all the same and reducing any possibility that in
some ways they are like us. One final aspect of protracted conflicts such 
as that between the Israelis and Palestinians is that the process of fighting
can become part of the national identities—“We are the people who fight
them.” This echoes Staub’s observation that engaging in intergroup vio-
lence changes individuals. Kelman proposes that intergroup conflict can
change whole nations and become part of the shared collective self-defini-
tion and the consensual national narrative.
Kelman then describes how identity change, though difficult, is still pos-
sible. Here he makes two very significant points. First, although people and
nations see their identities as fixed, they are, in fact, not. Identities, like all
belief systems, are not immutable essences, but rather are socially con-
structed and, as such, are modifiable. Second, unlike land and other physical
objects, identities are not zero-sum; the definition of “us” need not include
the defeat or devaluation of “them”.
How might these possible changes in national identity be achieved? Kel-
man describes the problem-solving workshops that he and his colleagues
have been conducting for many years in the Middle East. These feature
third parties who serve both to moderate the discussions and to provide an
unbiased resource for both sides. Another key aspect of the workshops is
that the participants are influential people in the Israeli and Palestinian
communities (e.g., business leaders) but not formal political leaders. Thus,
they have “clout” but do not need to represent a political position. 
The workshops attempt to find ways to bridge the gaps between the two
sides. Kelman highlights how this involves altering national identities. For
example, workshop participants have considered what aspects of their na-
tional identity might be outdated and thus might be psychologically dis-
carded. Kelman concludes his contribution with a detailed description of
how one workshop grappled with a crucial issue in the Israeli-Palestinian
conflict: what to do with, and for, the Palestinian refugees. A key feature of
his analysis is that this question has both material (what Kelman terms
228
Conclusion


“practical”) and symbolic facets. On the material or objective side, how can
Israel and Palestine accommodate the large number of refugees in other
countries? On the symbolic side, how can each side resolve the refugee is-
sue in such a way that it does not do violence to its national identity and
group narrative?
A Look Forward: A Social Identity Framework 
for Intergroup Conflict
As the preceding “look back” indicates, our authors have done an excellent
job of addressing the first goal of this volume: to survey what the social and
behavioral sciences have learned about the ways in which social identity and
intergroup harm-doing are interconnected. They have also succeeded in
addressing our second goal in putting together this book: to identify the
needed next steps in moving toward a more complete account of social
identity and intergroup conflict. In this section we pull together and orga-
nize the next-generation conceptual and methodological tools identified by
our contributors in the form of a social identity framework for intergroup
conflict that is schematically depicted in figure 9.1. It is important to em-
phasize that this is not a tight deductive model such as proposed by Brewer;
instead, we offer a heuristic scheme that identifies important constructs and
suggests how they might be interconnected. We present this framework as
follows. First, we provide a brief overview of the scheme. Second, the ad-
vantages of phrasing the framework in terms of a general threat or coping
paradigm are described. Third, the details and implications of the frame-
work are presented.
Overview of Framework 
The figure depicts three general sets of constructs—threat and appraisal;
destructive coping: intergroup harm-doing; and social identity—and identi-
fies these as mutually influencing one another. The arrow from “threat and
appraisal” to “destructive coping: intergroup harm-doing” marks this as a
general threat/coping framework as used in psychological analyses of stress
(Lazarus & Folkman, 1984). When members of a group perceive actual or
possible harm to themselves and other ingroup members (threat), they seek
to makes sense of this (appraisal), and if the other group is somehow judged
as responsible for the threat, they may cope with the appraised threat de-
structively by harming that group and its members. 
The return arrow from “destructive coping” to “threat” indicates that in-
tergroup harm-doing can escalate threat and appraisal: when one group in-
flicts harm on another, it raises the possibility of retaliation (i.e., threat).
The figure indicates that threat and appraisal and harm-doing are each con-
nected to social identity, and also that social identity moderates the link of
T
oward a Social Identity Framework for Intergroup Conflict
229


threat and appraisal to intergroup violence. Social identity consists of both
group (socially shared) and individual (personal) representations about own
and other group. Group identities that feature own group as victim and
other group as oppressor or aggressor heighten objective threat and ap-
praisal of danger; the reverse path indicates that being threatened and see-
ing another group as responsible is likely to influence individual and group
definition (e.g., living in a state of war, declared or not, with another group
will help create social and personal identities congruent with this state). 
Similarly, the “intergroup harm-doing” to “social identity” arrow indi-
cates that groups that engage in outgroup violence will have their personal
and group self-definitions altered, again to bring them into a psychologi-
cally consistent relation (e.g., “we are the people who are making the world
safe/pure/ . . . by subjugating/killing . . . them”). Finally, “social iden-
tity” moderates the link between “threat and appraisal” and destructive cop-
ing.” Groups that have a socially shared sense of “us” as victim and “them”
as dangerous are more likely to cope with threat via intergroup violence
than groups that do not have such a view of “us” and “them.”
The figure is obviously simplified in several respects. First, the frame-
work is from the perspective of only one group in relation to just one other
group. As White’s chapter in particular informs us, this is seldom the case in
real-world conflicts, and the framework, if it is to prove useful, will need to
be expanded to include additional groups beyond “us” and “them.” Second,
there are many other instigators to intergroup violence than threat. Often
one group aggresses against another for material gain because the group’s
leaders believe that they can get away with it. For example, the people of
West Africa in the 17th through 19th centuries posed no threat to Ameri-
cans; slaves were introduced into the United States simply because Euro-
230
Conclusion
figure 9.1 Social Identity Framework for Intergroup Conflict
&


peans and other slave traders had more power than Africans, and American
plantation owners wanted cheap labor. Third, threat, even if it is linked
clearly to some outgroup, does not always lead to violence. Fourth, there
are many inputs into intergroup conflict in addition to threat and appraisal
and social identity. At the societal level, type of economic system and politi-
cal structures are important, and, at the individual level, personality and so-
cial attitudes are likely significant contributors to group-based harm-doing.
Fifth, the figure suggests a static view of the links among the variables de-
picted. This is not the case. As Staub’s chapter makes very clear, there is a
through time path to mass killing and genocide. Thus, our figure stands for
many causal paths that unfold through time. For example, a group might
experience a large influx of immigrants to their country as an economic
threat (“They will take our jobs”); this may, in turn, lead to discriminatory
laws and outbreaks of violence against that group; if continued over time,
these actions will likely come to be prominent parts of views of us and them;
whether this would escalate to widespread and systematic harm-doing such
as genocide would depend on threat and appraisal and social identity ac-
cording to figure 9.1. 
Advantages of Threat/Coping Framework 
There are four major advantages to using a general threat/coping paradigm.
First, as has been demonstrated at the individual level, unpredictable and
uncontrollable threats are experienced as particularly stressful, and such
threats often lead people to seek the comfort and security of the ingroup
(Worchel, 1999). For most people, the world today is characterized by un-
predictable and uncontrollable forces and events—modernization in many
developing countries and globalization of the economy for all nations (Erik-
sen, this volume), rapid technological change, and often unplanned social
change. Thus, the model suggests that intergroup strife will be more, not
less, likely as we enter the 21st century, even though much of the globe is
experiencing considerable material progress. 
Second, as stress and health workers have demonstrated, it is fruitful to
distinguish problem- from emotion-focused coping. In the former, the per-
son, upon encountering a situation that signals current or future harm, at-
tempts to deal directly with the source of threat (e.g., upon learning that she
has high blood pressure, which it is explained to her increases the risk 
of health problems, the person changes her diet to reduce salt and fat).
Emotion-focused coping, on the other hand, involves attempts to reduce
the fear and anxiety that accompany threats. Thus, using the same example,
the newly diagnosed hypertensive individual could drink alcohol to avoid
thinking about the mysterious and scary condition about which she has just
learned. As applied to intergroup conflict, the 1929 Stock Market Crash
ushered in a worldwide depression. In the United States, one response was
the set of political measures termed the New Deal, which among other
T
oward a Social Identity Framework for Intergroup Conflict
231


things sought to provide assistance for the economically afflicted. This
problem-focused coping can be contrasted with the rise of Nazism in Ger-
many at about the same time. Hitler’s programs sought more to alleviate
feelings of despair and national impotence following the loss of World War
I and the economic and political turmoil of the Weimar years, and these led
to the “Final Solution” and the Holocaust.
Third, there are clear individual differences in styles of coping, and the
preceding chapters suggest that there are likely reliable societal and group
differences in styles of coping as well. Thus, for example, Staub notes that
societies differ in degree of authority orientation. It is likely that those high
in authority orientation will more likely accept violence-against-outgroup
“solutions” for threats offered by leaders even if they contradict societal
norms about harm-doing.
Fourth, and very important for the present analysis, there is evidence
that self and identity are implicated in the threat-coping process (Contrada
& Ashmore, 1999). Thus, there is reason to believe that an identity-inter-
group conflict model based on a threat-coping framework may be fruitful.
Framework Details and Implications
Destructive Coping: Intergroup Harm-Doing.
As depicted in figure 9.1, our
model seeks to account for intergroup harm-doing, by which we mean
harm by members of one group to those of another group based on group
membership. As suggested in the figure, we see intergroup harm-doing as
multidimensional. Intergroup violence can involve physical and/or psycho-
logical hurting. For example, during the slavery and Jim Crow eras in the
United States, African Americans were subjected to both physical violence
(e.g., beatings, lynchings, poor housing) and psychological violence (e.g.,
societal stereotypes of innate inferiority and racial etiquette that made in-
terracial contacts demeaning for blacks). T
oday the balance is much shifted
toward psychological harm-doing (persisting cultural stereotypes), though
physical violence against blacks has not been eliminated. 
Another important dimension is the degree to which harm-doing is sys-
tematized. At one extreme, members of one group engage in local and
spontaneous acts of violence against outgroup individuals. At the other ex-
treme, there are elaborate societal systems for hurting the other group. In
Nazi Germany, there was an organized set of institutions for carrying out
the “Final Solution.” It is not just open violence that can be highly system-
atized. The slavery system in the United States, the caste system in India,
and apartheid in South Africa all were undergirded by complex sets of 
social institutions (legal, educational, and the like) that oppressed certain
groups. 
There are two other important dimensions of intergroup harm-doing
not depicted in the figure: (1) extent—how widely across the other group is
the violence extended? and (2) duration—how long has the harm-doing
232
Conclusion


been going on? As both increase so too does the strength of the links of in-
tergroup harm-doing to social identity and threat. 
Threat and Appraisal. According to our model, a significant contributor to
intergroup harm-doing is threat and appraisal. Threat and appraisal are
highly related but separable concepts. Just as work on social problems has
profited from a distinction between actual conditions in the social environ-
ment and the public “seeing” and labeling of these as problematic (Thomp-
son & Fine, 1999), we too feel it is useful to distinguish the objective threat
and the individual and collective recognition and interpretation (or ap-
praisal) of this condition (see also Northrup, 1989). Thus, for example, dur-
ing the 1860s many Chinese immigrated to the United States to work on
the transcontinental railroad. The societally projected image of Chinese
during this period was relatively benign. However, when the railroad was
completed and many Chinese became competitors for jobs, unions and
other groups promoted public representations (in newspapers and the like)
of the Chinese as dangerous, and these came to be shared widely enough
that local discrimination occurred and national immigration restrictions
were implemented. Thus, it was not the sheer numbers of Chinese that
mattered or even the economic competition with whites, but the shared so-
cial representation of Chinese as dangerous. 
Threat, like harm-doing, is viewed as multidimensional. Figure 9.1 iden-
tifies three important dimensions: material, symbolic, link to other group.
Material and symbolic threat are often treated as opposite ends of a single
continuum. We believe, however, that it is best to consider these as poten-
tially independent dimensions. As noted above, material threat is the core of
realistic group conflict (RGC) approaches to intergroup conflict (Ashmore
& Del Boca, 1976). These are often contrasted with “symbolic politics” ap-
proaches (see Citrin et al., this volume; Sears et al., 1999). In symbolic poli-
tics (SP) models, groups come into “unrealistic” non-self-interested conflict
over symbolic issues. For example, Whose language should be spoken in
this country? Whose history should be taught in our schools? Our view is
that the conflict between the RGC and SP positions is more apparent than
real. Material things, especially land, often acquire symbolic meaning. And
symbols may have significant material implications. Continuing the same
examples, deciding “whose language” will, in turn, partially shape which
group’s members can make the better living economically. Whose history is
taught will legitimate and buttress the economic and political status of some
groups and undermine that of others. Thus, even symbolic conflicts have
important material implications.
In his contribution to this volume, Kelman stresses that the Israeli-Pales-
tinian conflict has both material, or “practical,” and symbolic facets, and
both of these need to be addressed in attempts to reduce the conflict. The
other chapters in this book tend to emphasize either material or symbolic
threat. As noted above, Brewer’s primary argument is in terms of symbolic
T
oward a Social Identity Framework for Intergroup Conflict
233


threat, whereas Eriksen stresses material conflicts. Citrin and his colleagues
were guided by symbolic politics theory; Sidanius and Petrocik’s group
dominance model is, at its core, a materialist framework—group dominance
in the service of material outcomes. 
The third dimension of threat depicted in the figure is link to the other
group. Some threats directly and clearly involve “them.” Some examples:
Chinese competing with whites in the United States after the completion of
the transcontinental railroad, Hutus and Tutsi attacking one another in
Rwanda (Staub, this volume), and laws in India designed to set aside jobs for
the lower castes (Eriksen, this volume). At the same time, intergroup harm-
doing can be triggered by threats that do not objectively arise from the
other group. In Germany in the 1920s, the hyperinflation and other eco-
nomic problems did not arise because of the actions of Jews. Hitler identi-
fied this group as a cause, however, and this appraisal was widely accepted
given the longstanding societal representations of Jews as sneaky and dan-
gerous (Goldhagen, 1996). Examples of types of threats that do not directly
involve the outgroup that yet can start a cycle of intergroup harm-doing are
economic instability and hardship (poverty, globalization that fractures 
traditional economic practices and procedures, inflation), political insta-
bility and corruption (changing governments, lack of societal agreement on
political processes, widespread graft by public officials), and rapid social
change (large-scale movements from rural to urban areas, changes in family
roles, and other social changes associated with “modernization”). 
In sum, our conceptualization of threat underlines two major points.
First, threats can be about tangible things (material threat) and about intan-
gibles as represented in valued symbols (symbolic threat). Further, a given
threat can have both a material and symbolic component. Second, conflict
between groups can begin with factors that are not objectively related to the
groups at all. Quite simply, some nongroup-related threats can initiate the
set of processes that lead to intergroup violence. In these cases, leaders di-
rect appraisal to “them” as the cause and away from the economic, political,
and social factors, leading individuals to see themselves and their groups as
in danger. In such cases, groups cope with the emotions aroused by the
threat rather than with the threat itself.
There are two other important dimensions of threat that are not de-
picted in the figure, and these parallel those for intergroup harm-doing: (1)
extent—how widely across the ingroup is the threat experienced? and (2)
duration—how long has the threat been in existence? As both increase, so
too do the links of threat to intergroup harm-doing. 
Appraisal is necessary for threat to lead to collective group-based vio-
lence. Although the figure suggests that this is a static phenomenon, it is
not. Instead, appraisal unfolds over time. T
o capture this idea we have iden-
tified both primary and secondary appraisal. Primary appraisal is an initial
making sense of the threat: What is the problem? What or who is responsi-
ble? Thus, for example, when a society with many foreign workers (as is
234
Conclusion


true of many nations) encounters an economic downturn, it might be ex-
plained relatively spontaneously as “them taking jobs from us,” and this
could lead, without much thought or organization, to isolated attacks on
the foreign workers.
However, for violence to be systematic and widespread requires mobiliz-
ing a larger segment of society and providing more organized paths to ac-
tion. This involves secondary appraisal guided by both group and individual
social identities as indicated in the figure. Regarding the former, group
identities must first be invoked to make sense of the threat. As Brewer,
Eriksen, and Staub note, leaders are crucial in this process (see also
Worchel, 1999). At least initially, these leaders are not likely to be national
political leaders but instead proponents of “fringe” groups, labor leaders, or
local religious figures. And, in an early coping stage, the efforts they direct,
though more organized than a spontaneous attack or a “race riot,” are still
relatively reflexive.
But what makes individual citizens ready to accept a group-based (“them
against us”) explanation for their troubles? Eriksen’s chapter points us in the
right direction when he notes that the shared group social identities prof-
fered by leaders must be consonant with the social and personal identities of
individuals and also fit with their daily experience and enactment of these
selves. This melding of a group identity with the identities of individuals
can result in mobilization of substantial numbers of individuals for action,
in this case a second stage of harm-doing.
As actions against the outgroup develop and widen, the ingroup engages
in secondary appraisal where the “guilt” of the outgroup is solidified in so-
cietally projected images, and harm-doing is explained as a necessary or
perhaps even a noble activity.
If mobilization is to be enduring, individuals will have to accept some
personal costs. This raises the crucial question: Under what circumstances
will individuals and groups be willing to endure personal hardship to either
inflict harm or to experience harm? Stated somewhat differently, what are
the factors that determine whether threat leads to mobilizing individuals
and the group to undertake action? As figure 9.1 suggests, we believe that
part of the answer lies in the content and structure of both group and indi-
vidual social identity. With regard to the group narrative, a national or eth-
nic story that features being a victim will increase the likelihood of mobi-
lization, and this will be true if there is also the sense of group vulnerability
(Eriksen, White, Staub, this volume). At the individual level, we suggest
that the relative lack of valued personal selves makes a society susceptible to
group-based mobilization. Thus, for example, in a nation in which men and
women are not able to successfully enact work and family roles because of
poverty, discrimination, and the like, they will be more likely to be suscepti-
ble to social identities that involve outgroup violence or support others 
doing harm.
An important implication of this line of reasoning is that protracted con-
T
oward a Social Identity Framework for Intergroup Conflict
235


flicts create social identities, at both the group and individual level, that
make the progression to widespread and systematic violence more likely.
Harming others based on group membership or standing by while others
do this changes people. Cognitive consistency models suggest that people
and groups would develop personal and shared reasons that justify this 
behavior. These have included belief systems specifying that they are not
human (and, thus, norms about hurting others do not apply), are inferior
humans (and, thus, need to be taken care of as slaves), or are powerful and
dangerous (and, thus, must be killed before they destroy us). These ratio-
nalizations become part of transformed group and individual identities 
as depicted by the arrow from intergroup harm-doing to social identity in
figure 9.1. Given the focus of this volume on social identity, we turn now to
a more detailed analysis of the group and individual identity portion of the
scheme.
Social Identity. As indicated at the top of figure 9.1, and suggested at several
places above, social identity is here viewed as a significant input to the
threat-intergroup harm-doing relationship. In the proposed framework, we
distinguish group from individual social identity. 
• Group Social Identity. Group social identities are group-level phe-
nomena. This means that both the referent and locus of the identity is 
the group. By referent we mean what the identity is about, the group—
America/Americans; Serbia/Serbians, and so on. The content of the iden-
tity is the set of things that are thought to “define,”—for example, America
and Americans, including geography (e.g., 50 states, mostly between
Canada and Mexico), politics (e.g., democracy, two-party system), econom-
ics (e. g., capitalism), morals and values (e.g., caring, freedom), history (e.g.,
revolution, civil war), and the like. By locus of the identity, we mean that
the group is also the level of representation of the identity. Such representa-
tion takes both psychological and socio-structural form. Regarding the for-
mer, labeled “Social Representations” in the figure, the group identity is a
set of beliefs that are widely accepted in the society, even though many indi-
viduals do not know or endorse all of the beliefs (Thompson & Fine, 1999).
The socio-structural form of group identity, “Societally Projected Images”
in figure 9.1, is manifested in two primary and intertwined ways: (1) it is
woven into the political, economic, and social structure of the society, often
in implicit ways (e.g., institutional racism in the United States [ Jones,
1997]); (2) it is conveyed by the societal mechanisms of socialization such as
schools and the media (e.g., cultural racism in the United States [ Jones,
1997, Worchel, 1999]). 
In addition to the rather static group social identity notion (“We are
_____”), we propose that there is a group-level analog to life story and per-
sonal narrative at the individual level (McAdams, 1997; Gergen & Gergen,
1988). That is, just as individuals develop stories that specify not only what
“I am like now” but also “how I got here” and “where I will likely go in the
236
Conclusion


future,” we propose that groups develop narratives that tell the story of
their group. This proposed construct is explicit in only one of the foregoing
chapters; Kelman makes national narrative a central part of his analysis, but
he draws no link to life story or self-narrative at the individual level. The
notion of group narrative, however, is implicit in several other contribu-
tions. For each of the three case studies of intergroup conflict described by
Eriksen, a group narrative was propagated by political leaders who identi-
fied “us” as noble and deserving and “them” as dangerous (e.g., in Bosnia
the Serbian national narrative featured past defeats at the hands of powerful
outsiders, especially the Turks, and the local Muslim community was de-
fined as representing external invaders even though this definition was inac-
curate). In describing the role identities of Republicans, White emphasized
the importance of a family history of involvement in the cause and the fact
that this history was often passed down from one generation to the next.
We suggest that, as one generation of Republicans socialized the next, they
not only told the story of “our family’s” involvement but also told the story
of “We Republicans,” the narrative of this group. 
There are four additional important points about group narrative:
1. Group stories can lie relatively dormant for periods of time. For
example, in Yugoslavia during the Tito period, the divergent group
narratives of the Serb, Croat, and Muslim communities were not in
the foreground of societally projected images or social representa-
tions of the citizenry.
2. At certain times and under certain conditions, there are forces
pushing for demonstrating and featuring a national narrative.
Threat, in the form of societal crises or outgroup actions, can
bring group stories to prominence and political leaders can use
such narratives (Brewer, Eriksen, this volume).
3. Group narratives can change. As Kelman noted, history is not
fixed, and, thus, groups can alter their story (Rothman, 1997).
4. Group narratives often contain folk theories of ethnicity, and these
generally do not match scientific understanding of this concept. As
Eriksen notes, most current-day anthropologists reject the notion
that ethnicity is primordial, and yet it is precisely this version of
ethnicity that activists and politicians use in mobilizing people to
action on the basis of ethnicity.
• Individual Social Identity. At the individual level, identity components
are generally partitioned into social identities and personal identities (Ash-
more & Jussim, 1997b), and we follow that convention here even though
drawing the line between personal and social is not always easy (Deaux,
1992). Social identities are those self definitions that identify self in terms
of, or in relation to, other people and groups (e.g., “I am a teacher” which
specifies a role that I and other teachers enact in relation to other people
who fill the role of student; “I am an American,” which specifies that I am 
T
oward a Social Identity Framework for Intergroup Conflict
237


a member of a social category along with others). Personal identities are
self-definitions that identify me as a unique individual (e.g., “I am obedi-
ent”). Personal identity also includes self-narrative (an organized story
about me, again as an individual). 
Given the thrust of this volume, we feature social identity here, though
we briefly describe below how individual-level personal identity might be
important to intergroup conflict. We follow Thoits and Virshup (1997) in
partitioning individual-level social identities into “we’s” (i.e., self as a mem-
ber of a collective group [e.g., “I am an African American”]) and “me’s” (i.e.,
self as fulfilling a role [e.g., “I am a mother”]). 
1. Individual social identity: “we selves.” In the application of social identity
to intergroup conflict, “we selves” are obviously important and have been
the most often studied. Thus, in this volume, all authors have been con-
cerned with how individuals define self in relation to an ethnic or national
group. And, they have done this in a variety of ways: self categorization in
terms of an ethnic label such as black or white, felt closeness to one’s group,
beliefs about one’s group and another group, and the like. This diversity 
of approaches raises an important question for the next generation of work
on social identity and conflict: what are the components of “we” social 
identities? 
Most researchers regard self-definition in terms of collective social
groups as multidimensional, but there is far from consensus on what these
dimensions are. T
o illustrate, table 9.1 lists the distinctions and components
that three research groups have proposed for ethnic identity, racial identity,
and cultural identity, perhaps the most often studied types of “we” social
identities. On the same line are listed dimensions that are the same or
highly similar. It can be seen that there are five areas of partial agreement
on what are the major dimensions of “we” social identity. T
wo of the three
articles included: (1) salience/centrality; (2) evaluation and feelings/private
regard; (3) values and attitudes/ideology; (4) interest and knowledge/famil-
iarity; (5) behaviors and practices/behavior and language.
At the same time, no dimension is mentioned by all three authors. And,
there are three components listed by just one author: (1) self-identification,
(2) commitment, (3) public regard. De la Garza et al. and Sellers and his as-
sociates may have omitted self-identification because they were proposing
scales to measure discrete dimensions of a specific social identification, and,
in so doing, they may have taken for granted that the first step in social
identification is placing self in a social category. At the same time, it is likely
that a comprehensive assessment of “we” social identity will require atten-
tion to self-identification, including such issues as preferred ethnic label and
the possibility of self-subtyping. Just as perceivers discern subtypes within
age (Brewer, Dull, & Lui, 1981), gender (Clifton, McGrath, & Wick, 1976),
and racial categories (Devine & Baker, 1991), it is likely that people may de-
fine self in terms of such subtypes (e.g., Eriksen noted that Yugoslavians liv-
ing in Belgrade and other urban areas had very different views of being a
238
Conclusion


Serb, Muslim, or Croat than those in rural areas, and this may involve sub-
typing). Regarding commitment, most would agree that something akin to
the Eriksen/Marcia notion of the developmental path that the person took
to her or his current sense of group identification is a potentially important
component of racial/ethnic/cultural identity. Cross’s (1995) work is a re-
lated but different approach to social identity as a developmental construct.
Finally, even though only Sellers et al. explicitly include a public regard di-
mension as part of social identity, it is likely that others would regard this as
a potentially important dimension.
T
able 9.1 illustrates one other complication in conceiving “we” social
identities: to what extent must the “other” group be included within the
concept of collective social self-definition? De la Garza and associates are
very clear that Latino cultural identity includes familiarity with and lan-
guage skills relevant to both the ingroup and the “other” group. Should as-
pects of identification in terms of “them” routinely be included in assess-
ments of social identities?
As is apparent, table 9.1 suggests a large number of possible components
of “we” social identities. As researchers grapple with which are necessary to
include, it is important to consider construct validity. In the simplest terms,
where does collective social identity end and related variables (causes, cor-
T
oward a Social Identity Framework for Intergroup Conflict
239
table 9.1 Dimensions of “We” Social Identity Proposed by Three Research
Groups
AUTHOR(S)
Phinney (1995)
de la Garza et al. (1995)
Sellers et al. (1998) 
CONSTRUCT
Ethnic identity
Latino(a) cultural identity
African-American 
racial identity
DIMENSIONS
Self-identification
—
—
Salience
—
Centrality
Evaluation of
—
Private regard
and
feelings about
group
—
—
Public regard
—
Values and attitudes
Ideology
Interest in
Familiarity with
—
and
Latino(a) culture
knowledge of
American culture
group
Ethnic behaviors 
Behavior 
—
and practices
and
language
Spanish proficiency
English proficiency
Commitment
—
—


relates, and consequences) begin? For example, is “interest in and knowl-
edge about” one’s ethnic group part of one’s social identity, or is it best seen
as an outcome predictable from identification? This question becomes even
more difficult regarding behavior: are ethnic behaviors part of social iden-
tity or an outcome that can be used to assess the construct validity of social
identity measures? 
One of the most important messages of the present volume is that even
conflicts that appear to be simple “us against them” on the basis of one so-
cial identity distinction (e.g., religion in Northern Ireland) also involve
multiple other social identities. These other social identifications are not
just regarding ethnic group and nation, which have been the focus of the
foregoing chapters. For example, in Part II of the volume addressing ethnic
and national identity in the United States, there was evidence that gender,
socioeconomic status, and religion were associated with ethnic and national
identifications and with the intergroup attitudes and policy positions under
investigation. Thus, as highly social creatures, people have multiple identi-
fications with collectives, and these interrelate with one another in complex
ways to shape orientations toward other groups and toward political issues.
2. Individual social identity: “me selves.” As White’s chapter reminds us,
how people relate to their own and other groups depends not just on collec-
tive self-definitions (“we’s”), but also on thinking of self in terms of social
roles (“me’s”). As noted above, role identities pertinent to group conflict are
especially likely to arise and be important in protracted conflicts where
many “supporting” roles are needed to carry on the fight. Also as noted
above, there is a second major way in which role social identities are impor-
tant to understanding intergroup conflict: to the extent that societal condi-
tions allow individuals to have and successfully fulfill valued role identities
(e.g., husband and father), it will be more difficult to recruit them for doing
violence against others. On the other hand, when poverty, oppression,
modernization, and other social forces make it difficult for large numbers of
people to fulfill social roles, it makes it easier for them to be drawn to social
identities that provide feelings of positive evaluation and potency even if
they also explicitly or implicitly involve doing harm to others. 
In addition, the relation between social role organization and group
boundaries is likely to have important implications for group-based vio-
lence. Eriksen’s observation that Bosnians and Serbs had separate private
spheres means that they did not have role identities (husband-wife, in law,
and the like) across social category lines. Social roles organized largely along
in-group lines may make it easier to recruit people to intergroup harm-do-
ing. Further, it is likely that friendships, too, were predominantly within
group. We suggest that cross-group friendship is likely to reduce inter-
group violence (Ashmore, 1970; Pettigrew, 1998). If I see myself as Richard
(Bosnian) friend of Mel (Serb) and we have extensive first-hand experience
with one another, it will be harder to mobilize either of us for “us against
them” identity conflict.
240
Conclusion


Personal Identity.
Although social identities are most pertinent to inter-
group conflict, we suggest that personal identifications (self-construals that
identify me as a unique individual) are also likely important. We offer two
illustrative examples. First, although Staub treated “authority orientation”
as a societal-level construct, it is likely to have an individual-level represen-
tation also, with substantial individual differences even within a society high
(or low) in this orientation. T
o the extent that I define myself as obedient to
authority, I will be a more likely recruit for group-based violence fomented
by political leaders. Second, as noted above, we believe that self narratives
are important to identity at both the group and individual level. White’s
chapter suggests that self stories are most likely to be conflict-related in so-
cieties characterized by protracted conflict. White notes that many current-
day Republicans come from families with a long commitment to the cause
of freeing Ireland from England. Further, in these families, that tradition is
a central part of the stories that parents and grandparents tell to children,
and, in turn, this history of one’s family’s involvement becomes part of the
individual’s personal story. Thus, just as Kelman noted that the group nar-
rative can feature “We are the people who fight them,” so, too, at the indi-
vidual level, a central part of an individual’s narrative can be that “My family
and I are people who fight them.”
Whose Social and Personal Identities? It is useful to distinguish three general
types of actors in intergroup conflict situations:
1. Elites—individuals who have political, economic, or information
power. As Brewer, Eriksen, Staub, and Kelman note, elites play a
pivotal role in social strife, often in propagating “us versus them”
social identities to mobilize citizens (Worchel, 1999).
2. Agents of violence—people who actively and directly carry out the
harm-doing (e.g., in White’s chapter, the Protestant and Republi-
can paramilitaries and English soldiers).
3. “Ordinary _____s” (as per Goldhagen’s [1996] “ordinary Ger-
mans”)—general citizens or average group members. Whatever
the merits of Goldhagen’s arguments, he has highlighted a crucial
point, How do ordinary citizens become participants in organized
violence between groups? Staub suggests that bystanders who “do
nothing” while organized violence is done will be touched, possi-
bly even taking on the identity of the hurting party.
Caveats
We reiterate that the proposed framework is not a tight deductive theory,
but instead a heuristic scheme intended to highlight the possible roles of
identity in intergroup conflict. Thus, it is necessary to note two significant
caveats.
T
oward a Social Identity Framework for Intergroup Conflict
241


One of the most important caveats about the model is what is left out.
Many factors that contribute to intergroup violence are not depicted. As
noted above, these omissions are made to underscore the importance of so-
cial identity. At the same time, as with any framework, our scheme hides as
well as highlights. Even some factors that we feel are quite important (e.g.,
the role of political, economic, and religious leaders in promoting identity-
based conflict) are only indirectly represented in the figure (e.g., as soci-
etally projected images under group social identity).
A second caveat concerns context. As depicted in figure 9.1, we have es-
sentially frozen in time a conflict between two groups and also wrenched
that conflict out of its regional and global context. As is clear in the chapters
by Eriksen, Sidanius and Petrocik, White, and Kelman, however, many eth-
nic and national conflicts unfold over a relatively long time period, and, as
noted above, early events profoundly influence the current situation
(Fisher, 1990; Northrup, 1989; Rothman, 1997). 
White’s chapter highlights the “space/place” context caveat for our
model: In the case of Northern Ireland, England/Britain is an “outside”
(not “them” or “us” if one focuses just on the two groups in Northern Ire-
land) yet potent factor (and if White is correct, the “them” party for Repub-
licans). Further, White notes but does not emphasize other important out-
siders—the Republic of Ireland and the United States. The differing
responses of the world community to the conflict in Yugoslavia and Rwanda
suggest the importance of regional context, together with associated strate-
gic interests and belief systems (e.g., racism). The fighting in Kosovo, on
the edge of Europe, led to a military campaign by NATO, whereas the
genocide in Rwanda resulted in action from the world community that was
“too little too late” (New York Times, December 16, 1999).
Identity and Conflict Reduction
Given that we anticipate a future volume of the Rutgers Series on Self and
Social Identity completely devoted to reducing intergroup conflict, here we
discuss just two issues important to defusing intergroup conflict: social psy-
chological mechanisms and types of interventions.
Regarding social psychological processes, several of the preceding chap-
ters noted that getting people to accept higher order, superordinate identi-
ties might be useful in ameliorating intergroup strife (Eriksen, Staub, Kel-
man). Brewer, too, raised this possibility, but also noted that according to
her optimal distinctiveness model and the present extension and elabora-
tion of it, overarching identity units could create conflict under some con-
ditions. The possible utility of higher order identities for conflict reduction
is the direct focus of Gaertner and his colleagues. In multiple experiments,
they have demonstrated that intergroup bias is significantly reduced when
242
Conclusion


members of different groups adopt a superordinate or common group iden-
tity (e.g., Gaertner, Dovidio, Nier, Ward, & Banker, 1998)
Perhaps the furthest extension of the concept of superordinate categories
is that of “global culture.” The goal of global culture is to unite all people
on the basis of characteristics shared as humans, thereby changing group
boundaries so members of currently different groups see themselves as be-
longing to the same group (Worchel, 1999). However, it is important to
note that global culture violates the basic need of identity uniqueness as
specified by Brewer’s optimal distinctiveness model (see Brewer, this vol-
ume). It also requires that individuals deny all previous group identities
(e.g., national, ethnic) in favor of the global identity. Thus, although the
notion of global culture builds on research on higher order identities, it
conflicts with evidence that the benefits of superordinate identities occur
even when subordinate identities are also strong. These findings suggest
that intergroup conflict reduction does not require that people give up their
ethnic identity or other relatively specific collective social identities (Gaert-
ner et al., 1998; Citrin et al., this volume).
Although the notion of higher order identities is appealing and Gaertner
and his co-workers have provided evidence for such an approach, we would
note that the power of transcendent identifications is not just inside the
head. Rather, we believe that it is action based on shared goals that is im-
portant. This was Sherif’s (1958) basic insight when he introduced the idea
of “superordinate goals”—goals that members of both groups desired but
that could not be achieved by either group on its own; as a consequence the
groups had to work together. Ashmore’s (1970) shared coping model took
Sherif’s logic a step further by proposing that shared goals induced coop-
erative action and that such behavior promoted cross-group positive affec-
tive ties, especially friendships, and these emotional ties promoted stereo-
type destruction, dissonance reduction, and the like that, in turn, produced
prejudice reduction (see also Pettigrew, 1998). The importance of cross-
group friendships as a mediator of intergroup contact fits with Eriksen’s
emphasis on lived experience. Just as the separate private spheres of Bosni-
ans and Serbs provided fertile ground for religiously based hostility, so, too,
could changes that promote cross-group friendship help in reducing con-
flict and making citizens of Bosnia more resistant to identity politics appeals
from leaders.
The second issue related to reducing intergroup strife to be addressed is
implementation of conflict reduction strategies. Here we discuss two ap-
proaches: educational programs designed to promote learning about other
groups, and third-party interventions. Many school-based programs have
been designed to foster intergroup learning (Aboud & Levy, 1999; Gurin,
1999). Bilingual and multicultural educational programs and materials have
been shown to have positive effects on intergroup relations among students
(Bigler, 1999; Genesee & Gandara, 1999; Stephan & Finlay, 1999). The
T
oward a Social Identity Framework for Intergroup Conflict
243


goal of these programs is to create an environment that allows groups to
maintain their cultural identity and, at the same time, accept other groups
(Fisher, 1990). Through these programs, individuals may see that they are
less different from members of other groups than they thought (Stephan &
Finlay, 1999). The danger, of course, is that individuals may find learning
about their group highly involving and satisfying, while ignoring or deni-
grating lessons about other groups. If so, such curricula might unintention-
ally magnify group boundaries and increase intergroup friction.
Another approach to conflict reduction, designed for groups engaged in
active, often violent conflict, is third-party intervention (e.g., Worchel,
1999; Rothman, 1997). Although the positive role of third-parties in reduc-
ing conflict is most obvious in Kelman’s chapter in the present volume, it is
also clear in the contribution by Staub. Fisher (1990) suggests that it is im-
portant to distinguish different types of third-party interventions, and pre-
sents a taxonomy of four types of third-party interventions that differ in
terms of whether or not the third-party focuses on the subjective (including
perceptions and evaluations of the conflict) and objective (including con-
flicts over both scarce resources [material] and over values [symbolic]) as-
pects of the conflict. Third-party consultation, the type performed by Kel-
man and his colleagues, is seen as the most desirable because it involves a
clear emphasis on both the objective and subjective aspects of the conflict
(Fisher, 1990). In contrast, there are numerous examples of conflict inter-
ventions carried out by other nations or aggregations of nations that focus
only on objective facets of the conflict. Worchel (1999) suggests that such
third-party interventions can sometimes save ethnic groups from persecu-
tion and destruction, but that they rarely lead to long-term peace. It is pos-
sible that these interventions are often unsuccessful, in part, because they
fail to adequately address the identity issues of the conflict. 
The chapters by Kelman and Staub raise an important question regard-
ing third-party intervention: When should third parties intervene? At the
roundtable discussion at the conference associated with this volume there
was clear agreement that early intervention was desirable, though Citrin
pointed out that there is little political incentive for leaders to intervene un-
til major conflict has emerged with clear evidence of harm-doing. Further,
Rothman (1997) notes that negotiating too early in a conflict may instead
escalate identity conflicts.
Summary and Suggestions
In this concluding chapter we both summarized the preceding chapters and
offered a conceptual framework that might guide future research.
Part I of the volume was, in part, concerned with the issue of the locus of
social identity—does social identification reside in the individual or in soci-
ety? Brewer and Eriksen did not resolve this starting question, and, in fact,
244
Conclusion


these authors answered that social identity is both “in” society and “in” the
individual. Thus, social identity, as with self and identity more generally, is
both personal and social (Ashmore & Jussim, 1997b). Brewer and Eriksen
identified the individual and societal building blocks of social identity, espe-
cially as they are linked to conflict between groups.
In Part II we asked the contributors to address one type of conflict (po-
litical strife) in one society (the United States). Not only did they share this
charge, the authors also used a similar method, survey research. Yet given
this shared topic and empirical approach, Citrin et al. and Sidanius and
Petrocik reached contrasting conclusions, the former seeing generally low
levels of ethnically based political conflict and the latter concluding that 
the ethnic divide is wide and deep. As noted above, the difference in con-
clusions is more apparent than real, reflecting, in part, methodological and
interpretational differences. The Part II authors leave us with many im-
portant ideas, including the multifaceted nature of ethnic and national 
identification.
In Part III, White and Staub addressed violent conflict, the first in an
analysis of Northern Ireland and the latter with a general model of the fac-
tors contributing to mass killing and genocide. These two chapters dovetail
well in highlighting the importance of history and action for understanding
the links between identity and conflict. White detailed the history of the 
violence in Northern Ireland and showed how the distant past was influen-
tial in shaping the most recent protracted fighting beginning in the late
1960s and is still much alive as the contestants seek to build on, and extend,
the 1998 peace accord. Staub underlined that identities are not just in 
the head but, instead, both influence action and are shaped by action (and
inaction). 
Kelman, in the single Part IV chapter, described the problem-solving
workshop approach that he and his colleagues have developed to ameliorate
protracted conflicts. In so doing he underlined the central role that national
identity plays in maintaining such conflict, and, at the same time, he offered
the very optimistic message that national identity as a social construction
can be altered.
The second part of this concluding chapter was devoted to a “look
ahead.” We presented a social identity framework for intergroup strife.
This scheme, depicted in figure 9.1, identifies three major constructs: threat
and appraisal, destructive coping: intergroup harm-doing, and social iden-
tity. It is proposed that these are mutually causally linked and further that
social identity moderates the threat and appraisal to intergroup harm-doing
link.
We offer three general suggestions for theorists and researchers under-
taking the next generation of work on social identity and intergroup con-
flict. First, we believe it will yield dividends if intergroup relations re-
searchers looked to the concepts and methods of the general field of “self
and social identity.” Except for White’s use of role identity models, all the
T
oward a Social Identity Framework for Intergroup Conflict
245


chapters in this book were guided by social identity perspectives specifically
intended for understanding intergroup relations (e.g., T
ajfel, 1981; Turner,
Hogg, Oakes, Reicher, & Blackwell, 1987). There is, however, a broad in-
terdisciplinary field concerned more generally with the issues of self and so-
cial identity (Ashmore & Jussim, 1997a). We urge scientists interested in in-
tergroup conflict to look more systematically at this field. We illustrated
what we have in mind above in linking “national narrative” as described by
Kelman to the individual-level concept “life story” (McAdams, 1997) and
personal narrative (Gergen & Gergen, 1988).
Second, we suggest that those interested in the interface between iden-
tity and group conflict consider a broader range of conflicts than here ad-
dressed. Because we anticipate a future volume directed at identity and gen-
der, we did not include female-male relations in this volume. Yet gender
and other forms of intergroup conflict are likely to share some features with
ethnic and national conflict but also to have some unique facets.
Third and finally, we hope that the general framework we have proposed
will generate both empirical work and conceptual debate. Our broader
hope is that the heuristic scheme, and the volume as a whole, will spur oth-
ers to develop a more comprehensive and coherent social identity paradigm
for intergroup conflict
Note
1. History is also important for knowing the “objective” ways in which present
day material realities and symbols are constructed. For example, the Rwandan geno-
cide was made more likely by the distant and recent past of Rwanda: Several cen-
turies ago, the indigenous Hutu were subdued by Tutsi invaders (Smith, 1998), who
were physically different (e.g., taller); the Tutsis implemented a feudal system that
subjugated the numerically superior Hutus and propagated a value system that ex-
alted Tutsi physical attributes. Early in the twentieth century, colonial Belgians used
these economic and value systems to help maintain order and their control over
both Hutus and Tutsis. In addition, they added the notion of “blood” to explain
group differences. These historical “facts” shape today’s material and symbolic 
realities that guide action.
References
Aboud, F. E., & Levy, S. R. (1999). Are we ready to translate research into pro-
grams? Journal of Social Issues, 55, 621–625.
Aretxaga, B. (1993). Striking with hunger: Cultural meanings of political violence in
Northern Ireland. In K. B. Warren (Ed.), The violence within: Cultural and political
opposition in divided nations (pp. 219–253). Boulder, CO: Westview Press.
Ashmore, R. D. (1970). Prejudice: Causes and cures. In B. E. Collins (Ed.), Social
psychology: Social influence, attitude change, group processes, and prejudice. Reading,
MA: Addison-Wesley.
246
Conclusion


Ashmore, R. D., & Del Boca, F. K. (1976). Psychological approaches to understand-
ing intergroup conflict. In P. A. Katz (Ed.), Towards the elimination of racism
(pp. 73–123). New York: Pergamon Press.
Ashmore, R. D., & Jussim, L. (Eds.) (1997a). Self and identity: Fundamental issues
(Vol. 1). New York: Oxford University Press.
Ashmore, R. D., & Jussim, L. (1997b). T
oward a second century of the scientific
analysis of self and identity. In R. D. Ashmore & L. Jussim (Eds.), Self and iden-
tity: Fundamental issues (Vol. 1, pp. 3-19). New York: Oxford University Press.
Bigler, R. S. (1999). The use of multicultural curricula and materials to counter
racism. Journal of Social Issues, 55, 687–705.
Boskin, J. S. (1970). Sambo. In G. Nash and R. Weiss (Eds.), The great fear: Race in
the mind of America. New York: Holt, Rinehart and Winston.
Boulding, K. E. (1962). Conflict and defense: A general theory. New York: Harper.
Brewer, M., Dull, V. & Lui, L. (1981). Perceptions of the elderly: Stereotypes as
prototypes. Journal of Personality and Social Psychology, 41, 656–670.
Clifton, A. K., McGrath, D., & Wick, B. (1976). Stereotypes of woman: A single
category? Sex Roles, 2, 135–148.
Contrada, R. J., & Ashmore, R. D. (Eds.) (1999). Self, social identity, and physical
health (Vol. 2, pp. 3–20). New York: Oxford University Press.
Contrada, R. J., Ashmore, R. D., Gary, M. L., Coups, E., Egeth, J. D., Sewell, A.,
Ewell, K., & Goyal, T. (in press). Measures of ethnicity-related stress: Psycho-
metric properties, ethnic group differences, and associations with well-being.
Journal of Applied Social Psychology.
Cross, W. E. (1995). The psychology of Nigresence: Revising the Cross Model. In 
J. G. Ponterotto, J. M. Casas, L. A. Suzuki, and C. M. Alexander (Eds.), 
Handbook of multicultural counseling (pp. 93–122). Thousand Oaks, CA: Sage 
Publications.
Deaux, K. (1992). Personalizing identity and socializing self. In G. Breakwell (Ed.),
Social psychology of identity and the self-concept (pp. 9–33). London: Academic Press.
de la Garza, M. F-O., Newcomb, M. D., & Myers, H. F. (1995). A multidimensional
measure of cultural identity for Latino and Latina adolescents. In A. M. Padilla
(Ed.), Hispanic psychology: Critical issues in theory and research (pp. 26–42). Thou-
sand Oaks, CA: Sage Publications.
de la Garza, R. O., Falcon, A., & Garcia, C. (1996). Will the real Americans please
stand up: Anglo and Mexican an-American support of core American political
values. American Journal of Political Science, 40, 335–351.
Devine, P. G., & Baker, S. M. (1991). Measurement of racial stereotype subtyping.
Personality and Social Psychology Bulletin, 17, 44–50
Festinger, L. (1950). Informal social communication. Psychological Review, 57, 271–
282.
Festinger, L. (1954). A theory of social comparison process. Human Relations, 7,
117–140.
Festinger, L. (1957). A theory of cognitive dissonance. Stanford: Stanford University
Press.
Fisher, R. J. (1990). The social psychology of intergroup and international conflict resolu-
tion. New York: Springer-Verlag.
Fredrickson, G. M. (1999). Models of American ethnic relations: A historical per-
spective. In D. A. Prentice & D. T. Miller (Eds.), Cultural divides: Understanding
and overcoming group conflict (pp. 23–34). New York: Russell Sage.
T
oward a Social Identity Framework for Intergroup Conflict
247


Gaertner, S. L., Dovidio, J. F., Nier, J. A., Ward, C. M., & Banker, B. S. (1999).
Across cultural divides: The value of superordinate identity. In D. A. Prentice &
D. T. Miller (Eds.), Cultural divides: Understanding and overcoming group conflict.
New York: Russell Sage.
Genesee, F., & Gandara, P. (1999). Bilingual education programs: A cross-national
perspective. Journal of Social Issues, 55, 665–685.
Gergen, K. J., & Gergen, M. M. (1988). Narrative and the self as relationship. In 
L. Berkowitz (Ed.), Advances in experimental social psychology (Vol. 21, pp. 17–56).
New York: Academic Press.
Goldhagen, D. (1996). Hitler’s willing executioners: Ordinary Germans and the holo-
caust. New York: Alfred A. Knopf.
Gurin, P., Peng, T., Lopez, G., & Nagda, B. A. (1999). Context, identity, and inter-
group relations. In D. A. Prentice & D. T. Miller (Eds.), Cultural divides: Under-
standing and overcoming group conflict (pp. 133–170). New York: Russell Sage.
Hull, C. L. (1943). Principles of behavior. New York: Appleton-Century-Crofts.
Iyengar, S. S., Lepper, M. R., & Ross, L. (1999). Independence from whom? Inter-
dependence with whom? Cultural perspectives on ingroups versus outgroups. In
D. A. Prentice & D. T. Miller (Eds.), Cultural divides: Understanding and overcom-
ing group conflict (pp. 273–301). New York: Russell Sage.
Jones, J. (1997). Prejudice and racism (second edition). New York: McGraw-Hill.
Lazarus, R. S., & Folkman, S. (1984). Stress, appraisal, and coping. New York:
Springer.
McAdams, R. C. (1997). The case for unity in the (post)modern self: A modest pro-
posal. In R. D. Ashmore & L. Jussim (Eds.), Self and identity: Fundamental issues
(Vol. 1, pp. 46–78). New York: Oxford University Press.
McCall, G. J., & Simmons, J. L. (1966). Identities and interactions. New York: Free
Press. 
New York Times, 16 December 1999.
Norenzayan, A., Choi, I., & Nisbett, R. E. (1999). Eastern and Western perceptions
of causality for social behavior: Lay theories about personalities and situations. In
D. A. Prentice & D. T. Miller (Eds.), Cultural divides: Understanding and overcom-
ing group conflict (pp. 239–272). New York: Russell Sage.
Northrup, T. A. (1989). The dynamic of identity in personal and social conflict. In
L. Kriesberg, T. A. Northrup, & S. J. Thorson (Eds.), Intractable conflicts and their
transformation (pp. 55–82). Syracuse, NY: Syracuse University Press.
Pettigrew, T. F. (1998). Intergroup contact theory. Annual Review of Psychology, 49,
65–85. 
Phinney, J. S. (1995). Ethnic identity and self-esteem: A review and integration. 
In A. M. Padilla (Ed.), Hispanic psychology: Critical issues in theory and research
(pp. 57–70). Thousand Oaks, CA: Sage Publications.
Rothman, J. (1997). Resolving identity-based conflict in nations, organizations, and com-
munities. San Francisco, CA: Jossey-Bass.
Rosenberg, S. (1997). Multiplicity of selves. In R. D. Ashmore & L. Jussim (Eds.),
Self and identity: Fundamental issues (Vol. 1, pp. 23–45). New York: Oxford Uni-
versity Press.
Sears, D. O. (1993). Symbolic politics. In S. Iyengar and W. J. McGuire (Eds.), 
Explorations in political psychology (pp. 113–149). Durham: Duke University Press.
Sears, D. O., Citrin, J., Cheleden, S., & Van Laar, C. (1999). Cultural diversity and
multicultural politics: Is ethnic Balkanization psychologically inevitable? In 
248
Conclusion


D. A. Prentice & D. T. Miller (Eds.), Cultural divides: Understanding and overcom-
ing group conflict (pp. 35–79). New York: Russell Sage.
Sellers, R. M., Smith, M. A., Shelton, J. N., Rowley, S. A. J., & Chavous, T. M.
(1998). Multidimensional model of racial identity: A reconceptualization of
African American racial identity. Personality and Social Psychology Review, 2, 18–39.
Sherif, M. (1958). Superordinate goals in the reduction of intergroup conflict.
American Journal of Sociology, 63, 349-363
Smith, D. N. (1998). The psychocultural roots of genocide: Legitimacy and crisis in
Rwanda. American Psychologist, 53, 743–753.
Stephan, W.G., & Finlay, K. (1999). The role of empathy in improving intergroup
relations. Journal of Social Issues, 55, 729–743.
Stryker, S. (1987). Identity theory: Developments and extensions. In K. Yardley &
T. Honess (Eds.), Self and identity: Psychosocial perspectives (pp. 89–103). New York:
Wiley. 
T
ajfel, H. (1978). Differentiation between groups: Studies in the social psychology of inter-
group relations. London: Academic Press.
T
ajfel, H. (1981). Human groups and social categories: Studies in social psychology. Cam-
bridge: Cambridge University Press 
Thoits, P. A., & Virshup, L. K. (1997). Me’s and we’s: Forms and functions of social
identities. In R. D. Ashmore & L. Jussim (Eds.), Self and identity: Fundamental 
issues (Vol. 1, pp. 106–136). New York: Oxford University Press.
Thompson, L., & Fine, G. A. (1999). Socially shared cognition, affect and behavior:
A review and integration. Personality and Social Psychology Review, 3, 278–302.
Turner, J. C., Hogg, M. A., Oakes, P. J., Reicher, S. D., & Blackwell, M. S. (1987).
Rediscovering the social group: A self-categorization theory. Oxford: Basil Blackwell.
Worchel, S. (1999). Written in blood: Ethnic identity and the struggle for human har-
mony. New York: Worth Publishers.
T
oward a Social Identity Framework for Intergroup Conflict
249


This page intentionally left blank 


Aborigines, 53
abstract communities, 60–61, 62
accentuation principle, 30
acculturation, 110–11
Act of Union (1801; Great Britain;
1801), 137, 144
affect, 72, 74–75, 81
affirmative action, 8, 93, 96
African Americans, 11, 76, 78, 107,
126n.2
disengagement by, 8
ethnic group identification, 86–87,
89, 96, 117, 219
immigration and, 104
oppression of, 103, 125
patriotism and, 11, 83, 84–85, 95,
110, 119–24, 220
policy preference issues, 93–94, 219
pride in America, 83, 84
racism toward, 23
religion and, 85, 219, 222
slavery and, 224, 230–31, 232
agents of violence, intergroup conflict
role, 241
aggression
defensive, 162, 171
intergroup, 26–28, 35
See also harm-doing; violence
Albanians (ethnic), in Yugoslavia, 49
Algeria, 195
Alien Act (1798; U.S.), 104
aliens, resident. See resident-aliens
altruism, 29
Americanization, of immigrants, 76,
105, 221
ancestry
American perspectives, 76
shared, 43, 215
anger, 33, 215
Anglo-Irish Treaty (1921), 137
antagonism
group-based, 65
ideologies of, 170–71, 172, 176
toward outgroups, 10, 17, 26–28, 32,
34, 36
anthropology, 7, 12
of identity, 43–47, 215
political, 62–66
anti-Semitism, 170, 234
apartheid, 44, 194, 195, 232
appraisal, 229–31, 233–36, 245
Arab-Israeli conflict, 188–189
Camp David accord (1978), 189
Palestinianization of, 189–91
251
Index of Subjects


Arab-Israeli conflict (continued)
See also Isreal-Palestinian conflict;
Palestinians
Arab-Israeli war (1967), 189
Arab League, 189
Arab nationalism, 188, 188
Arabs, 133, 155, 188–89, 192, 203. See
also Palestinians
Argentina, 178
Armenia, 172
Armenians, 8, 11, 171–72
Asian Americans, 11, 78, 86–87
Chinese nineteenth-century immi-
grants, 233, 234
ethnic group identification, 117, 219
multiculturalism and, 89
national loyalty and, 102–3
patriotism and, 97, 119–21, 123–24
pride in America, 83
assimilation
American perspectives, 72, 76–77,
82–85, 92, 93, 105, 219
intragroup, 22, 28, 29, 214
of self, 27
See also melting pot
atrocities, 8. See also ethnic cleansing;
genocide; mass killing
attachment
ingroup, 17, 18, 20, 28, 29, 34, 35
psychological, 74, 81, 219
authority orientation, 232
genocide, mass killing, and, 162, 163,
167–68, 173, 177, 180
personal identity and, 241
autonomous identities, 166, 167, 227
avoidance, 32–33
Azerbaijan, 172
Bahai, 179
Balkanization, 72, 76, 97
basic needs, 160–64, 168, 180, 181, 226
Belgium, 170, 227, 246n.1
better world ideologies, 165
Bharatiya Janata Party (BJP), 54, 55
bias, 7
ingroup, 17, 18, 20, 24–27, 34, 36
BJP
. See Bharatiya Janata Party
black Americans. See African Americans
blame, 33
blind patriotism, 75, 167, 178
Bosnia-Herzegovina, 52, 53, 56–59, 237
ethnic cleansing, 101
intergroup conflict in, 11, 49–50, 64,
187, 216
need for healing, 171, 172, 175–76,
177
Bosnians (people), 6, 8, 10, 49–50, 216,
240, 243
British mandate (Palestine), 188
Burundi, 59
bystanders, to violence, 162, 163,
173–74, 179–80, 181, 226, 227,
241
Cambodia, 11, 165
Camp David accord (1978), 189
Canada, 58–59
caring, inclusive, 178, 179
caste system, India, 55–56, 57, 61, 232,
234
categorization. See self-categorization;
social categorization
Catholics. See Roman Catholics
chauvinism
American views on, 81–85, 92, 93, 95,
219, 222
patriotism vs., 74–75, 218
children, values development in, 178
Chinese immigrants, American nine-
teenth-century hostility toward,
233, 234
Christians
African American, 85, 219, 222
anti-Semitism and, 170
in Fiji, 51
in India, 54
in Yugoslavia, 49–50
citizenship, 77
ethnic definitions of, 101
Civil Rights Act (1964; U.S.), 77
civil rights movements
in Northern Ireland, 138
in United States, 105, 220
civil society, 177–78
class, 48, 62
class-based social movements, 63
cognitive balance, 22
cognitive dissonance, 214
252
Index of Subjects


collective identities, 56–61, 66, 144
Israeli-Palestinian conflict and, 191,
194, 200–201
self and, 59, 139
Yugoslavia and, 50
See also group identity; social identity
collectivist cultures, 34, 166
communal differences, national identity
and, 101–25, 220–21
communication, theories of, 214
communism, 49, 165
communities
abstract, 60–61, 62
imagined, 45, 60, 65
comparison
intergroup, 19, 23–26, 27, 32, 214
modernization as trigger for, 57
social, 24, 34, 214
compensatory identities, 168–69
competition, 31, 65
for resources, 24, 28, 56, 60, 217
social, 24–26, 35
confidentiality, workshops and, 198–99,
200
conflict, 7, 8, 12–13. See also intergroup
conflict
conflict resolution
national identity and, 187–211
workshops on, 7, 12, 187, 190,
195–202, 210, 214, 227, 228, 245
conflicts of interest, 28, 34
connected identity, 166–67, 181, 227
conquest, 3, 8, 27
constructive patriotism, 167, 178
constructivism, 44, 47, 53, 62, 66, 215
consumerism, in India, 55
contempt, 18, 19, 32–33, 215
content of identity, 72, 74, 75–76, 164,
218, 219
cooperation, 23, 28–30, 33–34, 35, 36
coping
destructive, 229–30, 232–33, 245
problem– vs. emotion–focused,
231–32
shared, 243
styles of, 232
corruption, intergroup harm-doing and,
234
courage, moral, 178–79
criminalization (British policy), 151–
53
Croatia, 49
Croats, 48–50, 59, 133, 155, 177, 195,
237
Crusades, 11
cultural flows, 45–46
cultural identity, 238–39, 244
culture, 66
ethnicity and, 43–46, 215
Fiji, 52
global, 243
as intergroup conflict factor, 34, 46
Yugoslavian issues, 48–51
See also ethnic groups; genocide; mass
killings
Cyprus, 187, 193, 210
Dalits (Untouchables caste), 55, 57
Declaration of Principles (PLO and Is-
rael; 1993), 190
defensive aggression, 162, 171
democratic societies, genocide and,
177–78
destructive coping, 229–30, 232–33,
245
devaluation of outgroups, 161–63,
170–73, 176, 177, 226
differentiation, 20–22, 29–33, 35, 36,
170
discrimination, 7, 10, 84
apartheid policy, 44, 194, 195, 232
against outgroups, 18, 20, 23, 25,
26–27, 35, 36, 214–15
See also prejudice
disgust, 32
distinctiveness, 34, 36
ethnic, 72
of Israeli-Palestinian conflict issues,
191
optimal, 10, 21–22, 29–30, 32–33,
215, 242, 243
positive, 25, 26, 27
distrust, outgroup, 30–36
diversity
American identity and, 71, 76–77,
82–83, 103–6
as conflict source, 102, 106–7
immigration as source of, 92
Index of Subjects
253


dominance
group, 106–9, 119, 124, 125, 220–21,
234
social, 11, 103, 106–10, 122, 124, 125
Easter Rising (1916; Ireland), 137, 138
economic instability, intergroup harm-
doing and, 163, 234
education, 8, 61
about conflict reduction, 243–44
egalitarianism
in India, 55–56, 57
in United States, 95
Egypt, 188, 189
Egyptian-Israeli peace treaty (1979;
Camp David accord), 189
elites, intergroup conflict role of, 241
embedded identity, 166–67, 168, 178
emotion-focused coping, 231
emotions, 32–33, 215. See also specific
emotions
empathy, 23, 171, 175, 176
enemies
identification of, 161, 162, 170–71
image differentiation, 195–96
engagement of experiences, 175
England. See Great Britain
equality, social, 57
Eritrea, 59–60
essentialism, 44, 45, 46, 65
Ethiopia, 59–60
ethnic cleansing, 4, 11, 27, 33, 49, 65,
101
ethnic conflict, 59–60. See also inter-
group conflict
ethnic grievances, 114, 118
ethnic groups, 3, 7, 45
minority group identification with,
11, 86–87, 89, 95–96, 117, 219,
221–22
multiculturalism and, 77
reciprocal attitudes among, 18–19
in United States, 71, 72, 76, 77–78,
86, 95–96, 103–8, 218–22
See also specific ethnic groups
ethnic identity, 3, 5, 6, 8–11, 72
components of, 238–39
intergroup conflict and, 42–66, 215
multiculturalism and, 77
in United States, 5, 10–11, 78, 86–87,
89, 93–96, 108–9, 116–18, 124,
219–22
See also social identity
ethnicity, 42, 62, 75, 215–16
culture and, 43–46, 215
group narratives and, 237
Marxist view, 48
monolithic identity and, 193
symbolic politics theory and, 78, 219
U.S. perspectives on, 76, 77, 86, 109,
119–22, 218–22
ethnic pride, 72, 77
ethnocentrism, 11, 29, 30, 44
American perspectives on, 109, 110,
117, 122–24, 220
generalized, 112, 113, 117
group orientations, 19–28
hierarchical, 112–13, 122–23
Sumner’s theory of, 17–19, 214
superiority feelings and, 168, 169
European Community, 172
exclusionary patriotism, 107–8, 123–24,
125
exclusivist identities, 192–94, 201, 228
experience. See personal experience
exploitation, 3, 10
Fatah, 189
favoritism
ingroup, 17, 23–26, 34, 72, 75, 78,
214
outgroup, 25
fear, 32–33, 215
Fiji, 6, 10, 42–43, 51–53, 56–58, 60, 61,
63, 64
Fijians, 6, 10, 51–53, 56–58, 64
fission of groups, 64, 65
France, 169, 172, 195
friendship, cross-group, 240, 243
functional theory, 18
fusion of groups, 64, 65, 87
Gaza Strip, 188, 189, 190, 197, 203
GEI. See generalized ethnocentrism
gender, intergroup conflict and, 246
gender-based identity, 5, 166
gender-based social movements, 63
gender relations, in Fiji, 52
254
Index of Subjects


generalized ethnocentrism, 112, 113,
117
General Social Survey (1996), 79–86,
92, 115–16, 126n.2
genocide, 3, 6–8, 10, 11, 213, 226, 231,
245
definition of, 159
ethnic cleansing as, 4, 11, 27, 33, 49,
65, 101
Holocaust as, 11, 172, 173–74, 232
individual and group identities in, 12,
159–81
origins of, 160–63
prevention of, 175–80, 181
Genocide Convention (U.N.), 159
Germany, 8, 169, 234
Holocaust and, 11, 172, 173–74, 
232
Nazism and, 165, 173–74, 179, 232
Weimar Republic and, 178, 232
World War II and, 172, 173–74, 179
Yugoslavia and, 48
global culture, 243
globalization, 43, 66, 71, 231, 234
goals
common, 33–34, 243
superordinate, 9, 33, 243
Government of Ireland Act (1920;
Great Britain), 137
Great Britain
conflict in Northern Ireland and, 12,
134–38, 140–41, 144–45, 149–52,
154, 155–56nn.1–3, 180, 223–26,
242
Fiji and, 51, 52
mandate over Palestine, 155, 188
group-based hierarchies, 107–8, 220
group dominance model, 106–9, 119,
124, 125, 220–21, 234
group identity, 3–5, 22, 74, 216
changes to, 194, 201–2, 210–11, 227,
228
genocide, mass killing, and, 159–65,
170–71, 178, 180–81, 226
in social identity framework, 230,
235, 236–37
See also ingroups; outgroups; social
identity
group integration, 63, 65
group narratives, 210, 228, 235, 236–37,
241, 246
Israeli and Palestinian, 191–92, 202–9
See also myths, national
groups
integration into, 63, 65
ostracism from, 179
self-concepts of, 164, 168–70, 171,
174, 180, 226
GSS. See General Social Survey
Guatemala, 159
Gulf War, 190
Guyana, 53
Gypsies, 49, 173
harm, infliction of, 26–27, 214, 215, 
226
harm-doing, intergroup, 3, 10, 229–36,
245
hate, 17, 18, 25, 27, 32–34, 36, 37n.1,
214–15
healing, victimization and, 162, 163,
171–72, 175–77
HEI. See hierarchical ethnocentrism
hierarchical ethnocentrism, 112–13,
122–23
hierarchies, group-based, 107–8, 220
Hindus, 6
in Fiji, 51
in India, 10, 54–61, 65
hindutva movement, 10, 54–58, 61, 65
Hispanic Americans, 11, 78, 125
ethnic group identification, 86–87,
89, 117, 219
national loyalty and, 102–3
patriotism and, 83–84, 85, 97, 109,
111, 119–21, 123–24
pride in America, 83, 84
See also Mexican Americans
history, role of, 217, 223, 227, 245,
246n.1
Holocaust, 11, 172, 173–74, 232
Jewish Palestinian homeland and, 188
Home Rule, Ireland and, 137, 149
hostility, outgroup, 17–19, 27, 28,
33–36, 37n.1, 75, 78, 97, 214, 216
human welfare, caring about, 178–79
Hungarians, in Yugoslavia, 49
hunger strikes, 151–52, 224
Index of Subjects
255


Hutus, 59, 133, 155, 161–62, 170–71,
174, 234, 246n.1
identification, 238
ingroup, 17–36, 214–15
with vs. as, 72, 73–74, 218
identity, 12–13, 245
anthropology of, 43–47, 215
autonomous, 166, 167, 227
collective. See collective identity
compensatory, 168–69
connected, 166–67, 181, 227
definitions of, 5
embedded, 166–67, 168, 178
ethnic. See ethnic identity
exclusivist, 192–94, 201, 228
gender-based, 5, 166
genocide, mass killing, and, 160,
161–81, 213
group-based, 3, 4, 5, 139. See also
group identity
individual. See self
insecure, 167
as intergroup conflict factor, 3, 4, 7–9
monolithic, 192–94, 201, 228
national. See national identity
negation of, 192–93, 196, 201, 210
negotiation of, 196–97, 200–202,
207–10
racial, 11, 238–39
role, 135, 138–40, 144, 146–48,
151–54, 156n.6, 224–25, 237, 240,
245
secure, 167, 181
social. See social identity
social construction of, 194–95
superordinate, 242–43
in threat-coping process, 232
transcendent, 193, 201–2, 210–11,
243
identity politics, 42–47, 53, 58–66, 215,
243
in United States, 4, 94–95, 97, 219
ideological asymmetry, 124–25, 126n.4
ideologies
of antagonism, 170–71, 172, 176
of better world, 165
genocide, mass killing, and, 161–62,
165–66, 170–71, 173–74, 178, 181
monolithic identity and, 193
nationalistic, 165
imagined communities, 45, 60, 65
immigration, 231
to Palestine, 188
United States and, 8, 71, 76–77, 92,
96, 97, 104–5, 110, 233
Immigration and Nationality Act (1965;
U.S.), 77
inclusion, need for, 21–22, 29–30, 32,
33, 35, 36
inclusive caring, 178, 179
India, 6, 42–43, 60, 63
caste system, 55–56, 57, 61, 232, 234
hindutva movement, 10, 54–58, 61,
65
Muslim population, 4, 54–58, 60, 61
nationalist movement, 74
Indians (Fiji people), 6, 10, 51–53,
56–58, 64
individual-group discontinuity effect,
31
individual identity. See individual social
identity; self
individualism, U.S. perspectives, 109,
111, 126n.1
individualist cultures, 34, 166
individual social identity, 230, 235, 236,
237–40
ingroup positivity, 18–19, 20–23, 32, 36,
214
ingroups, 10, 17–36, 214–16
comparison and competition issues,
23–26, 34
ethnocentrism and, 17–19, 29, 214
favoritism toward, 17, 23–26, 34, 72,
75, 78, 214
formation of, 17, 19, 28–35
optimal distinctiveness theory, 21–22,
29–30, 32–33, 215
outgroup antagonism by, 26–28, 36
positivity toward, 18–19, 20–23, 32,
36, 214
scapegoating by, 161, 162, 164, 165,
180
social categorization and, 20, 214
injured social identities, 12
injustice, 57–58, 161
insecure identities, 167
256
Index of Subjects


instigators of violence, 160–63, 173,
180, 181, 226, 230
Institute for Social Research (University
of Michigan), 114
Institute for Social Science Research
(UCLA), 79
instrumentalism, 44, 45, 46–47, 53, 62
integration, group, 63, 65
interactive problem solving, 198
interdependence, 33, 34
cooperative, 35, 36
obligatory, 29
intergroup comparison, 19, 23–26, 27,
32, 214
intergroup conflict
definition of, 6–7
diversity as source of, 102
ethnic and national identity as issues
in, 42–66, 215, 222
gender and, 246
identity’s role in, 3, 4, 7–9
ingroup identification and, 17–36,
214–15
Israeli-Palestinian, 12, 187–211,
227–28
in multiethnic states, 101–2
multilevel analysis of, 7, 9
in Northern Ireland, 133–55, 223–26
realistic group conflict view of, 8,
216–17, 233
reduction of, 5, 7, 9, 12, 36, 242–44
self ’s role in, 3, 4
social identity and, 4, 5, 7, 8–12,
213–46
social identity framework for, 213,
229–42, 245
as violence instigator, 160–61, 180,
226
intergroup contrast, 22, 214
intergroup harm-doing, social identity
framework and, 229–36, 245
intergroup relations, 6, 11, 17, 18,
35–36
conflict-reduction strategies, 243–44
social categorization’s role in, 19–20
social identification and, 22–23
intermarriage, 57
in Fiji, 52
in India, 56
in United States, 105
in Yugoslavia, 50
international conflicts, 3, 187, 198. See
also war; specific conflicts and wars
interventions, as conflict-reduction
strategy, 243, 244
intifada, 189, 191, 201
intolerance, 32
intragroup assimilation, 22, 28, 29, 
214
invasion, feelings of, 33
IRA. See Irish Republican Army
Iran, 179
Ireland, 136–38, 149, 155–56nn.1, 3,
223, 224, 242
home rule, 137–38
See also Northern Ireland
Irish Free State, 137–38
Irish Nationalists, 135, 137, 138, 141,
142, 144, 152, 224
Irish National Liberation Army, 155n.2
Irish Republican Army, 12, 133,
137–38, 140, 145, 151, 152,
155nn.1, 2, 223, 224
Irish Republicans, 134–37, 140–42,
144–49, 151–53, 223–26, 237
paramilitaries, 11, 133–35, 137–38,
140, 148–49, 151, 154,
155–56nn.2, 3
self-narratives, 241
Islam. See Muslims
Israel, 133, 179, 187–211, 227–29, 233
Camp David accord (1978), 189
creation of, 188, 189
Declaration of Principles (1993), 
190
national identity concerns, 191–97,
200–201, 207–9, 214
overview of Palestinian conflict,
188–91
problem-solving workshop participa-
tion, 7, 12, 198–202, 210
Israeli-Palestinian conflict, 188–91
Arab states and, 188–89
Palestinian refugee issue, 188–89,
192, 202–9, 228–29
Palestinian uprising, 189, 191, 201
peace moves, 189–90
See also Arab-Israeli conflict
Index of Subjects
257


Jews, 8, 177, 179
anti-Semitism and, 170, 234
Holocaust and, 11, 172, 173–74, 232
in United States, 11, 104
Zionism and, 188, 192, 196, 202, 204
See also Israel
Joint Working Group on Israeli-Pales-
tinian Relations, 200, 202, 205–7
Jordan, 188, 203
justice, 161
Palestinian sense of, 204, 207, 208,
209
Khmer Rouge, 165
kinship
ethnicity vs., 57
identity politics and, 58, 60, 61, 64
Yugoslavian issues, 48, 50, 51
Korean War, 166
Kosovars, 8
Kosovo, 8–9, 48, 57, 58, 101, 155, 242
LACSS. See Los Angeles County Social
Surveys
language
conflict over, 8
in Fiji, 51, 52, 53
in India, 54
monolithic identity and, 193
in United States, 72, 82
in Yugoslavia, 48
Latino National Political Survey, 83
Latinos. See Hispanic Americans; Mexi-
can Americans
leaders, role of, 175, 177, 235, 237, 242
Lebanon, 203
liberalism, 76, 77, 78, 95, 218, 221
liberalization, in India, 55
life conditions, 160, 161, 163, 166, 168,
180, 226
literary studies, postcolonial movement
and, 44
Los Angeles County Social Surveys
(1994, 1995), 79, 83, 86–87, 221
Los Angeles County Social Surveys
(1997, 1998), 113–14
love
of country, 74–75
ingroup, 17, 18, 34, 214–15
Loyalists, Irish, 141, 142–44, 225
loyalty
identity politics and, 58, 64
ingroup, 28, 29–30, 33, 34
national, 71, 75, 101–3
pluralism and, 106
Macedonia, 49
Madrid Conference (1991), 190
Mandal commission (India), 56
Maoris, 53
marriage. See intermarriage
Marshall plan, 172
Marxism, 48
mass killing, 3, 6–8, 10, 11, 231
definition of, 159
individual and group identities in, 12,
159–81, 226, 245
origins of, 160–63
prevention of, 175–80, 181
See also genocide
material threat, 233–34
Mauritius, 53
media, 55, 61, 236
melting pot, America as, 11, 92, 103,
104–9, 116–19, 122, 124, 220, 221
Mexican Americans, 83–84, 109–11,
125
minimal group paradigm, 18
minority groups, 8
ethnic identifications by, 11, 86–87,
89, 95–96, 117, 219, 221–22
multiculturalism and, 77, 82
state-sanctioned violence against, 3,
11
subtle racism toward, 23
in United States, 71, 72, 77, 82–87,
89, 95, 96, 107, 219–22
See also specific minority groups
misattribution, 32, 215
modernity, 44–45, 47, 65, 66
modernization, 45, 234, 240
as conflict factor, 10, 56–57, 231
in Fiji, 53
monolithic identity, 192–94, 201, 228
monolithic societies, 162, 163, 177
Montenegro, 49
moral courage, 178–79
moral superiority, 10, 32–33, 34
258
Index of Subjects


Mughals, 55, 57
multiculturalism, 42, 44
United States and, 71, 76–78, 81–83,
89, 92–93, 96, 103, 218
multiethnic states, 75, 76, 78, 87, 218
communal and national identity in,
101–25
multiple voices, 44, 45–46
murder. See genocide; mass killing
Muslims
in Fiji, 51
in India, 4, 6, 54–58, 60, 61
in Yugoslavia, 49, 50, 52, 58, 59, 237
See also Arabs
myths, national, 6, 10
identity politics and, 58, 61
Yugoslavia and, 50, 216, 237
See also group narratives
narratives. See group narratives; self-
narratives
National Election Study (1992; 1996),
114–15
national identity, 3, 5, 6, 8–11, 245
American, 71–97, 101–25, 218–22
changes in, 187, 194–96, 201–2,
210–11, 228
definition of, 191
intergroup conflict and, 42–66
Israeli-Palestinian conflict issues and,
12, 187–211, 214, 227–28
negotiation of, 196–97, 200–202,
207–10
Northern Ireland hostilities and, 136
See also myths, national; social iden-
tity
nationalism, 10, 42, 44–46, 65, 115
Anderson’s (Benedict) view, 60
Arab, 188
civic vs. ethnic, 75
Indian hindutva movement and, 54,
55
as integrating ideology, 87, 89
patriotism vs., 74–75
Quebecois, 58–59
nationalistic ideologies, 165
nationalist movements, 8, 74
Nationalists, Irish, 135, 137, 138, 141,
142, 144, 152, 224
nationality, 48, 75
U.S. perspectives on, 76, 77, 86, 95,
105
National Opinion Research Center, 79
nations, 3, 7
definition of, 74
ethnic groups and, 44–45
loyalty to, 71, 75, 101–3
minority groups in. See minority
groups
See also specific countries
Native Americans, 76
nativism, 76–77, 82–85, 92, 218, 219,
221
NATO (North Atlantic Treaty Organi-
zation), 242
Nazism, 165, 173–74, 179, 232
Ndembu, 58
needs, basic, 160–64, 168, 180, 181, 226
negation of identity, 192–93, 196, 201,
210
negotiation of identity, 196–97,
200–202, 207–10
NES. See National Election Study
New Deal, 175, 231–32
Northern Ireland, 7, 133–55, 187, 193,
210, 213, 242, 245
activist accounts concerning, 140–51,
225–26
bystanders to violence in, 179–80
historical background of hostilities in,
136–38
hunger strikes in, 151–52, 224
political violence in, 133–55, 155n.1,
223–26
social and role identities in, 11–12,
135, 138–40, 143–44, 146–48,
151–54, 156nn.5, 6, 223–26, 237
Norway, 58
Nuer, 61, 63
Nuremberg trials, 172
obligatory interdependence, 29
one-time workshops, 198–200
oppression
of African Americans, 103, 125
institutionalized, 6
optimal distinctiveness, 10, 21–22,
29–30, 32–33, 215, 242, 243
Index of Subjects
259


“ordinary” people, intergroup conflict
role of, 241
Oslo agreement (1993), 190, 197, 201
ostracism, 179
outgroups, 10, 17–36, 214–16
American perspectives on, 73
antagonism toward, 10, 17, 26–28,
32, 34, 36
devaluation of, 161–63, 170–73, 176,
177, 226
distrust of, 30–36
ethnocentrism and, 18, 19, 214
favoritism toward, 25
hatred of, 17, 18, 25, 27, 32–33, 34,
36, 37n.1
hostility toward, 17–19, 27, 28,
33–36, 37n.1, 75, 78, 97, 214, 216
intergroup comparison and, 23–26,
27, 32
scapegoating of, 161, 162, 164, 165,
180
social categorization and, 20, 214
subtle racism toward, 23
PA. See Palestinian Authority
Pakistan, 55
Palestine, 155, 188–90, 192, 197, 203,
208–9
Palestine Liberation Organization, 189,
190, 197
Palestinian Authority, 190
Palestinians, 179, 187–211, 227–29, 233
Declaration of Principles (1993), 190
national identity concerns, 191–97,
200–201, 207–9, 214
Oslo agreement (1993), 190, 197, 201
overview of Israeli conflict with,
188–91
problem-solving workshop participa-
tion, 7, 12, 198–202, 210
refugee issue, 188–89, 192, 200,
202–9, 228–29
See also Arab-Israeli conflict; Israeli-
Palestinian conflict
passivity toward victims, 173–74, 179
patriotism, 115
blind, 75, 167, 178
chauvinism vs., 74–75, 218
constructive, 167, 178
exclusionary, 107–8, 123–24, 125
in United States, 10–11, 78, 81,
82–85, 92, 95, 97, 109–11, 119–24,
126nn.2–4, 219–22
Peace Now movement (Israel), 196
personal experience, 47
identity politics and, 58–61, 65, 66
personal identity. See self
PLO. See Palestine Liberation Organi-
zation
pluralism
in American society, 11, 103, 105–9,
118–19, 122, 124, 220, 221
genocide and, 177
pogroms, 27
policy issues, U.S., 73, 78, 87–96, 219,
221
immigration as, 76–77, 96, 104, 105
impact of racism on, 77
political anthropology, 62–66
political groups, 3, 72, 139, 159
political instability
genocide and mass killing origins
and, 163
intergroup harm-doing and, 234
political science, 7, 12
political violence, in Northern Ireland,
133–55, 155n.1, 223–26
politicide, 159
politicization, 34
politics
American ethnic group concerns
about, 117–18
of identity, 4, 42–47, 53, 58–66,
94–95, 97, 215, 219, 243
of power, 34
positive distinctiveness, 25, 26, 27
positive self-concept, 169
postcolonialism, 44
power politics, 34
power structure, 10
prejudice, 7, 8, 10, 221, 243
against outgroups, 17, 23, 25, 36,
37n.1
See also discrimination
pride
in America, 10, 78, 79, 81, 83, 84, 95,
219
ethnic, 72, 77
260
Index of Subjects


primary appraisal, 234–35
primordialism, 44, 45, 76, 215, 237
prisoners of war, 166
problem-focused coping, 231–32
problem-solving workshops, 7, 12, 187,
190, 195–202, 210, 214, 227, 228,
245
Protestants (in Northern Ireland),
133–38, 140, 141, 151, 153, 154,
155–56n.3, 223–26
Loyalists, 141, 142–44, 225
paramilitaries, 11, 12, 133–35, 137,
138, 140–44, 149, 151–54,
155–56nn.1–3, 224–26
Unionists, 137, 141–42
Protestants (in Rwanda), 178
Provisional Irish Republican Army. See
Irish Republican Army
Provisional Sinn Féin. See Sinn Féin
psychology, 12
punishment, 175
Quebecois nationalism, 58–59
race, policy preferences and, 93–95, 96,
219
racial identity, 11, 238–39
racism, 23, 77
Rashtriya Swayamsevak Sangh, 54
realistic group conflict, 8, 216–17, 233
reconciliation
identity change as factor in, 194,
210–11
need for, 171, 172, 175–77, 181
refugees, Palestinian, 188–89, 192, 200,
202–9, 228–29
reification, 45, 46
relational identities, 166
religious groups, 7
in Fiji, 51, 52
in India, 10, 54–56, 57, 60
monolithic identity and, 193
Northern Ireland hostilities and,
11–12, 133–38, 140–44, 149, 151,
153, 154, 155–56nn.3, 4, 223–25
in United States, 71, 72, 104, 219,
222
in Yugoslavia, 48, 49–50
See also specific religious groups
Republicanism, 136–37, 140. See also
Irish Republicans
resident-aliens, in United States, 11,
119–22
resources
competition over, 24, 28, 56, 60, 217
as Israeli-Palestinian conflict issues,
187, 191–92, 201
revenge, 57–58, 176
RGC. See realistic group conflict
rhetoric, political, 58
right of return, Palestinian refugees
and, 203–9
RIR. See Royal Irish Regiment
ritual, 58, 60
role identity, 135, 138–40, 144, 146–48,
151–54, 156n.6, 224–25, 237, 240,
245
Roman Catholics
in Northern Ireland, 133, 135–38,
140–44, 153, 154, 155–56nn.3, 4,
180, 223, 224
in Rwanda, 178
in United States, 104
Romans, 11
Royal Irish Regiment, 135
Royal Ulster Constabulary, 134–35,
137, 151–53, 155n.2
RSS. See Rashtriya Swayamsevak Sangh
RUC. See Royal Ulster Constabulary
Rwanda, 11, 59, 101, 174, 223, 234,
246n.1
church leaders’ actions in, 178
conflict instigators, 161–62
ideologies of antagonism and, 170–71
need for healing in, 171–72, 175–77
world community and, 242
See also Hutus; Tutsis
Sami, 58
scapegoating, 33, 161, 162, 164, 165,
180
SDLP
. See Social, Democratic, and
Labour Party
secondary appraisal, 234–35
secure identities, 167, 181
security forces (Northern Ireland),
133–35, 138, 140, 151, 152, 154,
155n.2
Index of Subjects
261


segmentary organization, 64, 65
self, 3, 4, 215–16, 245–46
collective identity and, 59, 139
definitions of, 5
genocide, mass killing, and, 12, 159,
160, 162–67, 169, 171, 174, 176,
179, 180–81, 213, 226
identity politics and, 47, 65
intergroup comparisons and, 23–24
monolithic identity and, 193
in optimal distinctiveness theory, 22,
30, 32
role identities and, 139–40
social identification and, 21, 87, 238
in social identity framework, 230,
235, 236, 237–41
in threat-coping process, 232
self-categorization, 3, 10, 20, 222
Americans and, 81, 86–87, 219, 221
social identity and, 72, 73–74, 139,
218, 238
self-concept, group, 164, 168–70, 171,
174, 180, 226
self-esteem, 21, 22, 25, 72, 169
self-evaluation, 24, 25
self-narratives, 238, 241, 246
self-worth, 21, 26, 176
Serbia, 49
Serbs, 6, 59, 61–62, 223, 226
Bosnia and, 10, 177, 216, 243
Croats and, 133, 155, 195
group narrative by, 58, 216, 237
Kosovo and, 8–9, 57, 58
role identities of, 240
Yugoslavia and, 48–50
shared coping, 243
Simmel’s Rule, 63, 65
Sinhalese-T
amil conflict, 58
Sinn Féin, 137, 138, 140, 152, 226
situational identification, 45–46
Six-Day War (1967), 189
slavery, 3, 8, 10
in Fiji, 51
in United States, 224, 230–31, 232
Slavs, 49–50
Slovenes, 48–49
Slovenia, 48, 49
social categorization, 10, 19–20, 29, 33,
35, 214
social change, intergroup harm-doing
and, 234
social class. See class
social comparison, 24, 34, 214
social competition, 24–26, 35
Social, Democratic, and Labour Party,
143
social dominance theory, 11, 103,
106–10, 122, 124, 125
social equality, 57
social identification, ingroup positivity
and, 18–19, 20–23, 32, 36, 214
social identity, 3–13, 47, 213–46, 228
Americans’ ethnicity and, 116–18,
124
class and, 48
communal differences and, 102,
108–9
definition of, 5–6, 21, 138–39
dimensions of, 73–74, 218, 219, 238
genocide, mass killing, and, 160, 162,
164–65, 171, 178–79
injured, 12
intergroup conflict framework and,
213, 229–42, 245
intergroup conflict reduction and, 5,
7, 9, 12, 36, 242–44
intergroup conflict role of, 3, 4, 7–9
multilevel analysis of, 7, 9
multiple aspects of, 178, 240
Northern Ireland hostilities and,
11–12, 138–40, 143–44, 146, 148,
154, 156n.5, 223–26
optimal distinctiveness and, 10,
21–22, 29–30, 32–33, 215, 242, 
243
social categorization’s role in, 19–
20
types of, 71
as violence factor, 5, 11–12, 133, 135,
154, 160, 162
See also group identity
social identity theory, 3, 10, 21, 27, 164
U.S. national identity and, 72–75,
77–78, 87, 97, 124
socialization, 31, 78, 84, 86, 218, 236
social movements
class-based, 63
gender-based, 63
262
Index of Subjects


social psychology, 7, 10, 17
conflict-reduction mechanisms,
242–43
social representations, 236, 237
social structure, 34–35
social welfare. See human welfare
societally projected images, 236, 237
sociology, 7, 12
Somalia, 59
South Africa, 187, 223
apartheid policy, 44, 194, 195, 232
Sri Lanka, 187, 210
status of groups, 24, 25, 35
stereotypes, 7, 8, 10, 223–24, 232, 243
categorization and, 20
outgroup, 30–31, 215
Sudan, 59
suffering
acknowledgment of, 171, 175
desensitization to, 173–74
images of past, 57–58, 61
superiority
beliefs of, 168–69
moral, 10, 32–33, 34
superordinate category identity, 36
superordinate goals, 9, 33, 243
superordinate identities, 242–43
symbolic politics theory, 73, 78–79, 84,
87, 95, 218, 219, 233, 234
symbolic threat, 233–34
symbolism, political, 58
Syria, 189, 203
T
aig (term), 142, 156n.4
T
amils, 58
territory, Israeli-Palestinian conflict
over, 187, 191–93, 196–97, 201,
202
third-party interventions, 243, 244. See
also problem-solving workshops
Thirty Years’ War (1618–1648), 169
threat, 28, 32–34, 35, 36
feelings of, 162, 163, 165, 171, 209,
237
material vs. symbolic, 233–34
in social identity framework, 229–36,
245
threat-coping frameworks, 229, 231–32
tolerance, 77, 95, 177
traditionalism, 55
transcendent identity, 193, 201–2,
210–11, 243
tribal societies, 64
Trinidad and T
obago, 53
trust, 31, 36, 171, 175, 176
ingroup, 10, 28–30, 34, 35, 215
outgroup distrust, 30–36
Turkey, 171–72
Turks, 8, 9, 11, 57, 172
Tutsis, 59, 133, 155, 161–62, 170–71,
178, 223, 234, 246n.1
Ulster Defence Association (UDA),
138, 142, 143
Ulster Defence Regiment (UDR), 135,
153
Ulster Democratic Party (UDP), 153
Ulster Freedom Fighters (UFF), 138,
142, 143, 152, 155n.2
Ulsterization (British policy), 151, 
152
Ulster Volunteer Force (UVF), 137,
138, 142, 143, 152, 155n.2
U.N. See United Nations
Unionists, Irish, 137, 141–42
United Irishmen, 136–37
United Kingdom. See Great Britain
United Nations, 179
Genocide Convention, 159
Palestine partition plan, 188
United Nations Relief and Works
Agency, 203
United States, 5–8, 10–11, 155, 217–22,
240, 245
American identity conceptions in,
75–78, 81, 84, 92, 218–19
Americanization of immigrants in,
76, 105, 221
chauvinistic views in, 81–85, 92, 93,
95, 219, 222
communal and national identity in,
101–25, 220–21
communal conflict in, 102, 106–8
historical perspectives on ethnic di-
versity in, 103–6, 221
identity politics in, 4, 94–95, 97, 219
immigration issues in, 8, 71, 76–77,
92, 96, 97, 104–5, 110, 233
Index of Subjects
263


United States (continued)
multiculturalism in, 71, 76–78,
81–83, 89, 92–93, 96, 103, 218
national identity’s meaning in, 71–97,
218–20
Northern Ireland conflict and, 224,
242
patriotism in, 10–11, 78, 81, 82–85,
92, 95, 97, 109–11, 119–24,
126nn.2–4, 219–22
policy issues in, 73, 77, 78, 87–96,
104, 105, 219, 221
pride in country in, 10, 78, 79, 81, 83,
84, 95, 219
slavery in, 224, 230–31, 232
unit relationships, 22
universal stereotypes, 30–31, 215
University of California, Los Angeles
(UCLA), Institute for Science 
Social Research, 79
University of Michigan, Institute for
Social Research, 114
UNRWA. See United Nations Relief
and Works Agency
Untouchables (Indian caste), 55, 57
UVF. See Ulster Volunteer Force
values, genocide, mass killing, and, 162,
177, 178–79, 180
victimization, 162, 163, 171–72,
175–77, 226
Vietnam, 165, 172
violence, 8, 216
agents of, 241
bystanders to, 162, 163, 173–74,
179–80, 181, 226, 227, 241
instigators of, 160–63, 173, 180, 181,
226, 230
in Northern Ireland, 133–55, 155n.1,
223–26, 245
as revenge for past, 57–58
social identity’s role in, 5, 11–12, 133,
135, 154, 160, 162
state-sanctioned, 3, 11
See also aggression; ethnic cleansing;
genocide; harm-doing; mass
killing; war
vulnerability, sense of, 163, 168, 172,
176, 226
war, 3, 6, 10, 27
in Ireland, 136, 137, 223
in Middle East, 188–89
in Rwanda, 161
Sumner’s theory of, 18
in Yugoslavia, 47–51
See also international conflicts; specific
wars
weakness, sense of, 168
Weimar Republic, 178, 232
West Bank (Middle East), 188, 189,
190, 196, 197, 203
women
identity and, 166
in United States, 76
women’s movement, 220
workshops, problem-solving, 7, 12, 187,
190, 195–202, 210, 214, 227, 228,
245
World War I, 169, 188
World War II, 10, 172, 173–74, 177,
179. See also Holocaust
Yugoslavia, 42–43, 47–52, 57, 61, 63,
77, 237
ethnic cleansing policy, 4, 49, 101
world community and, 242
See also Bosnia-Herzegovina; Kosovo;
Serbs
Zambia, 58
zero-sum perspective, 194, 210
discriminatory behavior and, 26
in Israeli-Palestinian conflict, 12,
187, 192, 228
social competition and, 24–25
Zionism, 188, 192, 196, 202, 204
264
Index of Subjects


Aboud, F. E., 243
Abrams, D., 21
Aiken, L. S., 123
Alba, R. D., 86, 105, 111
Alexander, king of Yugoslavia, 48
Allport, G. W., 7, 20, 177
Alpher, J., 200, 202, 203, 205, 206, 207
Anastasio, P. A., 9, 23
Anderson, Benedict, 45, 60, 62, 194
Andrews, Irene, 143
Appiah, K. A., 77
Arafat, Yasser, 189, 197
Aretxaga, B., 224
Arthur, M., 150
Ashe, Thomas, 151, 152
Ashmore, Richard D., 3–14, 71, 213–49
Bachman, B. A., 9, 23
Baker, S. M., 238
Banaji, M. R., 21
Banker, B. S., 97, 243
Barber, B., 71
Baron, R. A., 23
Bar-tal, D., 75, 115, 164
Barth, F., 43, 46, 62
Bauman, Zygmunt, 42, 66
Baumeister, R. F., 5, 168
Beike, D., 23–24
Bell, J. B., 138, 151, 152
Bettencourt, B. A., 26
Bigelow, K. R., 179
Bigler, R. S., 243
Billig, M., 18
Blackwell, M. S., 73, 139, 246
Blanz, M., 27
Blumer, H., 107
Boskin, J. S., 224
Boulding, K. E., 216
Bourhis, R., 25
Brewer, Marilynn B., 5, 7, 10, 17–41,
75, 87, 170, 213, 214–17, 229,
233–34, 235, 237, 238, 241, 242,
243, 244–45
Brimelow, P., 76
Bringa, T., 50
Brown, R., 17, 18, 25
Bruce, Steve, 135, 141, 142, 152, 153
Bundy, R., 18
Burton, John, 198
Cairns, E., 133, 180
Calderwood, Alex, 143
Campbell, D. T., 18, 25, 30, 31
Caporael, L. R., 29
265
Index of Names


Chapman, M., 62
Charles I, king of England, 149
Cheleden, S., 77
Chirot, D., 3, 71
Churchill, Winston, 149–50
Citrin, Jack, 5, 6–7, 10, 71–100, 109,
213, 218–22, 233, 234, 243, 244,
245
Clifton, A. K., 238
Cohen, Abner, 44, 62
Cohen, A. P., 47
Connor, W., 74, 75, 87
Contrada, R. J., 220, 232
Cornell, S., 50, 51
Craig, G. A., 169
Crocker, J., 21
Cromwell, Oliver, 136, 149
Cronin, S., 136, 137, 144
Cross, W. E., 239
Cusack, J., 153
Dahl, R. A., 106
Danziger, K., 5
Darby, J., 180
Davis, J., 79
Deaux, K., 237
de la Garza, R. O., 72, 83, 87, 103, 106,
109, 110, 111, 221, 238, 239
Del Boca, F. K., 8, 216, 233
Deschamps, J-C., 20
des Forges, A., 162, 171, 174, 178
Devine, P. G., 238
de Weerd, M., 139
Diehl, M., 18
Dietz, M., 74
Doise, W., 20
Dorr, N., 26
Dovidio, J. F., 9, 22, 23, 97, 243
Drake, C. J. M., 135
Duff, Brian, 5, 10, 71–100, 213, 218
Dull, V., 238
Eidheim, H., 58
Elliott, S., 138, 151, 152
Epstein, A. L., 58
Epstein, S., 174
Erickson, G., 21
Eriksen, Thomas Hylland, 5, 6, 7, 10,
42–68, 171, 213, 214–17, 223, 226,
227, 231, 234, 235, 237, 238–39,
240, 241, 242, 243, 244–45
Evans-Pritchard, E. E., 61, 63, 64
Falcon, A., 72, 83, 103, 221
Fang, C. Y., 125
Fanon, Frantz, 44
Farley, R., 71
Farrell, M., 152
Featherstone, M., 43
Federico, C., 125
Feeney, B., 133
Feierabend, I. K., 115
Fein, H., 165
Feshbach, S., 74, 75, 81, 103, 112, 115
Festinger, L., 24, 214
Fincham, F., 26–27
Fine, G. A., 233, 236
Finlay, K., 243, 244
Fisher, R. J., 242, 244
Fiske, S. T., 31
Flackes, W. D., 138, 151, 152
Flament, C., 18
Fogelman, E., 174
Folkman, S., 229
Fordham, S., 8
Ford, T. E., 23
Fortes, M., 64
Fraser, A., 136, 149
Fraser, M. R., 140, 144
Frederico, C. M., 78, 165
Fredrickson, G. M., 221
Freedman, S. G., 105
Frey, D., 23
Fried, A., 74
Friedman, J., 42
Froystad, K., 61
Funk, C. L., 28, 95
Furnivall, J. S., 50
Gaertner, S. L., 9, 23, 97, 242–43
Gandara, P., 243
Gandhi, Rajiv, 54
Garcia, C., 72, 83, 103, 221
Gaughan, Michael, 152
Geertz, Clifford, 44
Gellner, Ernest, 44–45, 46, 61, 62
Genesee, F., 243
George V, king of England, 149
266
Index of Names


Gergen, K. J., 236, 246
Gergen, M. M., 236, 246
Gerner, D. J., 188
Gerstle, G., 77
Giddens, A., 42, 47
Glazer, N., 77, 97
Gleason, P., 76
Gluckman, Max, 63
Goldberg, D. T., 77
Goldhagen, D., 3, 234, 241
Gordon, Milton, 77, 105
Goring, Herman, 173
Gourevich, P., 167, 174, 178
Greeley, A. M., 105
Green, D. P., 81
Greenfeld, L., 3, 71, 75
Greenwald, A. G., 21
Grey, Edward, 150
Gubin, A., 177
Gurin, G., 74
Gurin, P., 74, 243
Gurr, T. R., 3, 101, 159
Gurtman, M. B., 22
Haas, E. B., 72
Hagengimana, A., 177
Halpern, B., 188
Handelman, D., 58
Handler, R., 59
Hanna, B. A., 30
Hannerz, U., 43
Hanning, P., 27
Hansen, T. B., 56
Harff, B., 101, 159
Harkabi, Yehoshafat, 196
Hartmann, D., 50, 51
Hartz, L., 78
Haslam, S. A., 7, 139
Heider, F., 22
Heppen, Jessica, 213–49
Herman, J., 171, 172, 175
Hertzberg, A., 188
Hewstone, M., 21, 26–27
Higham, J., 76
Hilberg, R., 173
Himmler, Heinrich, 173
Hinkle, S., 18, 25
Hirt, E. R., 21
Hitler, Adolf, 55, 58, 165, 232, 234
Hobsbawm, E., 62
Hochschild, J. L., 84
Hofstetter, C. R., 115, 125
Hogg, M. A., 20, 21, 73, 139, 246
Holland, D., 4, 5, 46, 47
Hollinger, D. A., 71, 77, 89
Horney, Karen, 215
Horowitz, D. L., 56, 78, 97, 101
Hull, C. L., 214
Huntington, S., 48, 53
Huntington, S. P., 76
Hyndman, M., 145
Ignatieff, Michael, 48, 65
Insko, C. A., 31
Jackman, M., 107
James II, king of England, 136
Janoff-Bulman, R., 171
Jarymowitz, M., 170
Jaspars, J., 26–27
Jenkins, R., 47
Johnson, G., 23
Jones, J., 3, 236
Jones, James, 96
Jussim, Lee, 3–14, 71, 213–49
Kaistaniemi, L., 169
Kapferer, B., 58
Kateb, G., 77
Kaukiasnen, A., 169
Keegan, J., 3
Kelley, H. H., 32
Kelley, R., 104
Kelly, J. D., 51
Kelman, Herbert C., 3, 5, 6, 7, 12, 75,
170, 177, 187–212, 213–14,
227–29, 233, 237, 241, 242, 244,
245, 246
Kelters, S., 133
Kennedy, C., 21
Keough, K. A., 171, 175, 176
Khalidi, R., 188
Kincaid, E., 166
Kinder, D. R., 83, 84
Klandermans, B., 139
Kleppner, P. J., 104
Klicperova-Baker, M., 115
Konner, M., 174
Index of Names
267


Kosterman, R., 74, 75, 81, 112, 115
Kramer, R. M., 30
Kreisberg, L., 3
Kressel, N. J., 167
Laclau, E., 65
Lagerspetz, K. M. J., 169
Lash, S., 43
Lau, R., 78
Lavine, H., 167
Lazarus, R. S., 229
LeVine, R. A., 30
Levin, S., 26, 78, 103, 125, 165
Levy, S. R., 243
Lewin, Kurt, 217
Lifton, R. J., 173
Lijphart, A., 101, 102, 103
Lind, E. A., 89
Lind, M., 76
Lipset, S. M., 78, 103
Lui, L., 238
Lyengar, S. S., 227
Lyons, F. S. L., 137, 149
Major, B., 21
Malanchuk, O., 74
Malle, B. F., 107
Manchester, William, 149–50
Manzi, J., 25, 26
Marcia, 239
Martin, P., 82
Maudling, Reginald, 149
McAdams, R. C., 236, 246
McCall, G. J., 139, 225
McCann, L. I., 171, 172, 175
McCaughey, Sean, 152
McClintock, C., 26
McDonald, H., 153
McDonald, M., 62
McGarry, J., 136
McGarty, C., 139
McGrath, D., 238
McKee, Billy, 151, 152, 153
McKittrick, D., 133, 134
McMichael, Gary, 153
McMichael, John, 153
Meertens, R. W., 23
Mendelsohn, E., 188
Merelman, R. M., 82
Meyer, G., 20
Miller, A. H., 74
Miller, D., 72, 74, 77, 89
Miller, D. T., 87
Morris, B., 203
Moxon-Browne, Edward, 141–42, 144
Moynihan, D. P., 77, 97
Mullen, B., 25
Mummendey, A., 25–26, 27
Muslih, M., 188, 197
Muste, C., 72, 77
Myers, D. G., 6, 164
Myrdal, G., 96
Neidenthal, P. M., 23–24
Nelson, S., 142
Netanyahu, Benjamin, 190
Nier, J. A., 97, 243
Norenzayan, A., 227
Northrup, T. A., 3, 233, 242
Novak, M., 105
Oakes, P. J., 7, 20, 73, 139, 246
O’Connell, A., 27
Ogbu, J. U., 8
Ohbuchi, K., 89
Okin, S. M., 77
O’Leary, B., 136
Oliner, D., 174, 179
Oliner, S. B., 174, 179
O’Malley, P., 150
Opotow, S., 173
Orwell, George, 75
Otten, S., 26, 27
Parenti, M., 105
Pearlman, L. A., 171, 172, 174, 175,
176, 177
Pennebaker, J. W., 171, 175, 176
Perdue, C. W., 22
Petrocik, John R., 5, 6–7, 11, 101–29,
165, 213, 220–22, 234, 242, 245
Pettigrew, T. F., 23, 177, 240, 243
Platow, M., 27
Polsby, N. W., 106
Portes, A., 82
Powell, C. L., 106
Powell, G. B., Jr., 102
Pratto, F., 103, 107, 125
268
Index of Names


Premdas, R., 51
Prentice, D. A., 87
Prentice-Dunn, S., 23
Prunier, G., 162, 171
Rabin, Yitzhak, 190, 197
Rabinowitz, J. L., 78, 125, 165
Ranger, T., 62
Raz, J., 72, 77
Reese, H., 170
Reich, R. B., 71
Reicher, S., 20, 73, 139, 246
Reingold, B., 72, 81
Rex, J., 44
Roediger, D. R., 86
Rogers, R. W., 23
Rokkan, S., 103
Roosens, E. E., 45
Roosevelt, Franklin, 175
Rose, Richard, 141, 144
Rothman, J., 237, 242, 244
Rouhana, Nadim, 200
Rubin, M., 21
Rummel, R. J., 177
Ruscher, J., 31
Rust, M. C., 9, 23
Sachdev, I., 25
Sadat, Anwar, 189
Said, Edward, 44
Salins, P. D., 105
Salmivalli, C., 169
Sampson, E. E., 166
Sanders, L. M., 83
Sands, Bobby, 152
Savarkar, Veer, 55
Schaar, J. H., 74, 75
Schatz, R., 167
Schlesinger, A. M., Jr., 44, 72, 77, 106
Schöpflin, G., 48
Schopler, J., 31
Schreiber, H., 25–26
Schwartz, S. H., 26
Sears, D. O., 28, 73, 77, 78, 79, 83, 84,
87, 95, 218, 233
Segal, J. M., 201
Sellers, R. M., 238, 239
Shave, R., 27
Shaw, J. S., 25, 26
Sherif, M., 7, 9, 33, 243
Shikaki, K., 200, 202, 203, 205, 206,
207
Shore, C., 62
Sidanius, Jim, 5, 6–7, 11, 25, 26, 78, 85,
101–29, 165, 213, 220–22, 234,
242, 245
Simmel, Georg, 63
Simmons, J. L., 139, 225
Simon, B., 25
Skerry, Peter, 125
Smith, Anthony D., 44, 45, 62
Smith, C., 25
Smith, E. R., 32
Smith, N. S., 167
Smith, P., 21
Smith, Rogers, 76
Smith, T., 79
Smith, William, 143
Soeters, J. L., 177
Stahelski, A. J., 32
Stallworth, L. M., 107
Stangor, C., 23
Staub, Ervin, 3, 5, 7, 12, 74, 75, 81, 115,
159–84, 213, 223, 226, 227, 228,
231, 232, 234, 235, 241, 242, 244,
245
Stephan, C., 32
Stephan, W. [G.], 32, 243, 244
Stevenson, J., 142, 143
Strathern, M., 43
Streich, G., 82
Struch, N., 26
Stryker, S., 139, 225
Sugawara, I., 89
Sullivan, L. A., 23
Sullivan, J., 74
Sumner, W. G., 17–19, 214
Sutton, M., 134, 155n.1
T
ajfel, H., 3, 4, 5, 7, 10, 17, 18, 20, 21,
26, 31, 72, 74, 87, 117, 124, 164,
218, 246
T
akagi, E., 29
T
ambiah, S. J., 47
T
amir, 87
T
aylor, C., 74
T
aylor, Peter, 153
T
ec, N., 174, 179
Index of Names
269


T
esser, A., 25
T
essler, M., 188
Thatcher, M., 135
Thoits, P. A., 3, 4, 72, 73, 139, 140, 224,
238
Thompson, L., 233, 236
Thornton, C., 133
Thorson, S. J., 3
Tilly, C., 155n.1
Tinker, H., 51
Tito, Josip Broz, 48, 237
T
onkin, E., 62
Triandis, H. C., 34, 166
Turner, J. C., 3, 7, 10, 17, 18, 20, 21, 24,
72, 73, 117, 124, 139, 164, 246
Turner, P
., 21
Turner, V., 58, 60
Turton, D., 47
T
yler, R. B., 22
T
yler, T., 89
Urry, J., 43
van der Veer, P., 55
van Laar, C., 77
Virshup, L. K., 3, 4, 72, 73, 139, 140,
224, 238
Walch, T., 105
Ward, C. M., 97, 243
Warner, R. S., 71
Waters, M., 86
Weber, J. G., 21, 23
Weber, Max, 75
West, S. G., 123
Wetherell, M., 20
White, John, 142–43
White, Robert W., 5, 7, 11–12, 133–
58, 180, 213, 223, 224–27, 
230, 235, 237, 240, 241, 242, 
245
White, T. F., 135, 155–56n.3
Whyte, J., 133, 134
Wick, B., 238
Wilder, David, 3–14, 213–46
Wilkes, 20
William of Orange, 136
Williksen-Bakker, S., 52
Wilson, Paddy, 143
Wolfe, A., 72
Wolfinger, R. E., 105
Wong, Cara, 5, 10, 71–100, 213, 218
Wood, J. V., 24
Worchel, S., 231, 235, 236, 241, 243,
244
Young, I. M., 72, 77
Zaller, J., 79, 87
Zillman, D., 21
270
Index of Names

List o
f
 Tables and Figures 
List o
f
 Contributors 
Pre
f
ace 
Contents 
Part I 
Representations of Self and Group 
Regarding the Relationship Between Social Identity and 
Personal Identity 
Jean-Claude Deschamps and Thierry Devos 
VII 
IX 
XVll 
2 
Social Representations in Personal Identity 
1 3  
Willem Doise 
3 
The Personal-Social Interplay: Social-Cognitive Prospects on 
Identity and Self-Others Comparison 
24 
Carmencita Serino 
4 
Self 
-We-Others Schemata and Social Identifications 
44 
Maria Jarymowicz 
Part II Establishing Group Identity 
53 
5 
A Developmental View of the Search for Group Identity 
53 
Stephen Worchel 
6 
The Role of Prototypicality in Group Influence and Cohesion: 
Contextual Variation in the Graded Structure of Social 
Categories 
75 
Penelope Oakes, S. Alexander Haslam and John C. Turner 
7 
Group Beliefs as an Expression of Social Identity 
Daniel Bar-Tal 
8 
Social Identity and National Identity 
Jose Miguel Salazar 
Part III Identity and GrouplIntergroup Process 
9 
Social Identity and Intragroup Differentiation as Subjective 
93 
1 14 
124 
Social Control 
124 
Jose M. Marques, Dario Paez and Dominic Abrams 


VI 
Social identity: international perspectives 
10 
Categorization and Social Influence 
142 
Juan Antonio Perez and Gabriel Mugny 
11 
Social Identity and Interpersonal Relationships 
154 
Miguel Moya 
12 
Social Identity and Aspects of Social Creativity: Shifting to 
New Dimensions of Intergroup Comparison 
166 
Steve Hinkle, Laurie A. Taylor, Lee Fox-Cardamone and 
Pamela G. Ely 
13 
Mere Categorization and Competition: A Closer Look at 
Social Identity Theory and the Discontinuity Effect 
180 
Stephen M. Drigotas, Chester Insko and John Schopler 
Part IV 
The Impact of Culture on Identity and Categorization 
199 
14 
Discrimination and Beliefs on Discrimination in Individualists 
and Collectivists 
199 
J. Francisco Morales, Mercedes Lopez-Saez and Laura Vega 
15 
Constructing Social Identity: The Role of Status, Collective 
Values, Collective Self-Esteem, Perception and Social 
Behaviour 
211 
Dario Paez, Cristina Martinez-Taboada, Juan Jose Arrospide, 
Patricia Insua and Sabino Ayestaran 
Re
f
erences 
Index 
230 
255 


List of Tables and Figures 
Tables 
4.1 The social self as a structural basis of different forms of 
social identifications 
4S 
S.1 Perceived differences in ingroup and outgroup similarity 
70 
S.2 Minority influences in groups over time 
72 
1l.l Meta-contrast ratios in self-perception and the four stimuli 
used by Moya and Peplau 
162 
13.1 Max. reI. as a function of matrix type (PDR versus PDC), 
trials and groups versus individuals, experiment 2 
190 
14.1 Summary of some differences between individualists and 
collectivists 
202 
14.2 Relation between identification with the goup and intergroup 
differentiation as a function of individualism/collectivism and 
autonomous/relational orientation 
203 
IS.1 Perceived ingroup and outgroup heterogeneity 
216 
IS.2 Perception of ingroup and outgroup by salience versus 
non-salience of social comparison 
217 
IS.3 Perceived social desirability of attributes for engineers and 
psychologists 
218 
15.4 Attributes for engineers and psychologists 
220 
IS.5 Factorial analysis of the 'most important social identity 
scale' (Garza & Herringer, 1986) 
224 
IS.6 Differences between engineers and psychologists as a 
function of their degree of collectivism 
227 
Figures 
1.1 Personal and social dimensions in psychosocial identity 
11 
3.1 Mean scores (x 100) of each content category elicited in self/ 
ingroup versus self/outgroup similarity judgements ( N  = 121) 
39 
4.1 Selfˮwe schemata distinctness and the degree of individuals' 
involvement with outgroups and dissimilar others: the 
theoretical model 
SO 
4.2 Recall of children's problems 
SI 
4.3 Motivation to solve children's problems 
51 
4.4 Readiness to host different nationalities 
SI 


Vlll 
Social identity: international perspectives 
4.5 Tolerance of German minority activity 
5 1  
4.6 Positivity of evaluation of Slovak and Irish paintings 
52 
4.7 Positivity of evaluation of Japanese painting 
52 
5.1 Stages of group development in existing groups 
58 
5.2 Nature of desired interaction with outgroup as a function of 
group life 
68 
6. 1 A categorial hierarchy of vegetation 
78 
6.2 A categorial hierarchy of Europeans 
79 
9. 1 Manipulation of group uniformity and group members' 
status: distribution of ingroup and outgroup members' 
responses 
1 34 
9.2 Evaluations of attractive and unattractive ingroup and 
outgroup members as a function of group uniformity 
1 35 
9.3 Example of ranking of characters A-F used to build modal 
and deviate members 
1 36 
9.4 Effects of norm awareness on judgements of modal and 
deviate ingroup and outgroup members 
1 37 
9.5 Effects of ingroup and outgroup accountability on 
judgements of modal and deviate ingroup and outgroup 
members 
1 38 
9.6 A two-step process of inter- and intragroup judgement 
140 
1 1 . 1  Interpersonal and intergroup demands according to 
Stephenson (1984) 
1 58 
1 1 .2 Example of a hypothetical identity taken from Deaux's 
formulation and Moya's data 
161 
12. 1 
Ingroup favouritism on different comparison dimensions as a 
function of outgroup status 
1 73 
12.2 Ingroup favouritism on different comparison dimensions as a 
function of outgroup status 
1 75 
1 3. 1  A Prisoner's Dilemma Game matrix (PDG) 
1 88 
1 3.2 A Prisoner's Dilemma max. reI. matrix (PDR) and a 
Prisoner's Dilemma confounded matrix (PDC) 
1 89 
1 3.3 A Prisoner's Dilemma Game alternative matrix (PDG-Alt.) 
195 


List of Contributors 
Dominic Abrams is Professor of Social Psychology and Director of the 
Centre of the Study of Group Processes at the University of Kent, UK. His 
main research interests are intergroup conflict and differentiation, social 
identity processes and social self-regulation. He is co-author and co-editor 
(together with M.A. Hogg) of four books: Social Identifications (London: 
Routledge, 1988); Social Identity Theory (Hemel Hempstead: Harvester 
Wheatsheaf, 1990); Group Motivation (Hemel Hempstead: Harvester 
Wheatsheaf, 1993); and Social Identity and Social Cognition (Oxford: 
Blackwell, forthcoming, 1998). 
Juan Jose Arrospide is currently Assistant Professor at the Department of 
Social Psychology, School of Psychology, Basque Country University, 
Spain. He has done research on the topic of social identity, group dynamics 
and social representions. He is co-author (together with S. Ayestaran and 
C. Martinez-Taboada) of a chapter on 'New Techniques for the Study of 
Groups' ('Nuevas Tecnicas en el estudio de los grupos') included in 
Methods and Techniques o
f
 Research in Social Psychology ( Mhodos y 
Tecnicas de Investigacion en Psicologia Social, Madrid: Eudema, 1994). 
Sabino Ayestaran is currently Professor of Social Psychology at the 
Department of Social Psychology, School of Psychology, Basque Country 
University, Spain. His research interests focus on social identity and group 
dynamics, with special attention to intra- and intergroup conflict. He is the 
editor of a volume on The Group as a Social Construction (El Grupo como 
Construccion Social, Barcelona: Plural, 1996). 
Daniel Bar-Tal is Professor of Psychology at the School of Education, Tel­
Aviv University, Israel. His research interests focus on political and social 
psychology, studying societal beliefs regarding conflict, delegitimization, 
security, patriotism and siege mentality. He is the author of Group Belie
f
s
 
(New York: Springer Verlag, 1990) and Societal Belie
f
s
 o
f
 Ethos (London: 
Sage, in press) and co-editor of Social Psychology o
f
 Knowledge (with W. 
Stroebe, A.W. Kruglanski and M. Hewstone; New York: Springer Verlag, 
1988), Stereotyping and Pre
judice (with C. Grauman, A.W. Kruglanski and 
W. Stroebe; New York: Springer Verlag, 1989), Patriotism in the Lives o
f
 
Individuals and Nations (with E. Staub; Chicago: Nelson-Hall, 1997) and 
Concerned with Security (with D. Jacobson and A. Klieman; Greenwich, 
CT: JAI Press, in press). 


x 
Social identity: international pers
pectives 
Jean-Claude Deschamps, born in France, is currently Professor of Social 
Psychology at the School of Social Sciences, Lausanne University, 
Switzerland. He has studied social identity and intergroup relations, the 
processes of causal attribution and social representations. He is co-author 
(with W. Doise) of a chapter on 'Crossed Categories Membership and 
Intergroup Relations', included in H. Tajfel (ed.), Differen
Social Groups (Cambridge: Cambridge University Press, 1978) and (with A. 
Clemence) of a volume on Everyday Explanation (L'ex
plication quotidienne, 
Friburg: Delval, 1987). 
Thierry Devos is Assistant Professor at the School of Social Sciences, 
Lausanne University, Switzerland. He has studied social identity, causal 
attributions and social representations. He is co-author (with J.-c. 
Deschamps and L. Comby) of 'Assymmetries in Judgment of Ingroup 
and Outgroup Variability', European Review o
f
 Social Psychology (1996). 
Willem Doise, born in Belgium, is Professor of Social Psychology at the 
School of Psychology, University of Geneva, Switzerland. His research has 
focused on the link between mind and society, specifically social represen­
tations and the role of socio-cognitive conflict in mental development. He 
has published several books, among them Groups and Individuals 
(Cambridge: Cambridge University Press, 1978) and Levels o
f
 Explanation 
in Social Psychology (Cambridge: Cambridge University Press, 1986). 
Stephen M. Drigotas is currently Assistant Professor of Psychology at the 
Southern Methodist University, Dallas, Texas, USA. His recent publications 
include: 'On the Peculiarities of Loyalty' (together with G.A. Whitney and 
C.E. Rusbult), Personality and Social Psychology Bulletin (1995), 'The Role 
of Identifiability in the Reduction of the Interindividual-Intergroup Dis­
continuity Effect' (together with J. Schopler et al.), Journal o
f
 Personality 
and Social Psychology (1995), and 'Willingness to Sacrifice in Close Rela­
tionships' (together with P.A.M. Van Lange et al.), Journal o
f
 Personality 
and Social Psychology (in press). 
Pamela G. Ely is Assistant Professor and Chair of Psychology at Saint 
Andrews College in Laurinburg, North Carolina, USA. She also directs the 
Human Services Program of the Saint Andrews at Sandhills program. Her 
research focuses on the relationship between individualism/collectivism and 
group and intergroup processes, including group-oriented instructional 
methods. Her most recent publication is 'Recognizing Group Diversity: 
Individualist-Collectivist and Autonomous-Relational Social Orientations 
and Their Implications for Intergroup Processes' (with R.J. Brown et al.), 
British Journal o
f
 Social Psychology (1992). 
Lee Fox-Cardamone is Assistant Professor of Psychology at Kent State 
University - Stark Campus in Canton, Ohio, USA. Her research interests 
include political activism, intergroup processes and group decision-making. 
Her most recent publications are 'Grassroots Political Actions as an 


List o
f
 contributors 
Xl 
Intergroup Phenomenon' (with S. Hinkel et al.), Journal o
f
 Social Issues 
(1996), and 'Recognizing Group Diversity: Individualist-Collectivist and 
Autonomous-Relational Social Orientations and Their Implications for 
Intergroup Processes' (with R.J. Brown et al.), British Journal o
f
 Social 
Psychology (1992). 
S. Alexander Haslam is a Senior Lecturer in Social Psychology at the 
Australian National University. His main research interests are in the social 
psychology of stereotyping and social judgement with particular emphasis on 
the role which group memberships play in those processes. This work 
addresses issues of 
social categorization, social identity and prejudice, dealing 
specifically with topics of stereotype consensus, perceived group homo­
geneity, illusory correlation, prejudice and extremism. His recent publica­
tions include Stereotyping and Social Reality (with P.J. Oakes and J.C. 
Turner; Oxford: Blackwell, 1994), The Social Psychology o
f
 Stereotyping and 
Group Lif
e
 (with R. Spears, P.J. Oakes and N. Ellemers; Oxford: Blackwell, 
1997), The Message o
f
 Social Psychology (with C. McGarty; Oxford: 
Blackwell, 1997), and Doing Psychology (with C. McGarty; London: Sage, in 
press). 
Steve Hinkle is an Associate Professor of Psychology at Miami University 
in Oxford, Ohio, USA. His main research interests are in motivational 
aspects of social identity theory, the relationship between individualism/ 
collectivism and aspects of group and intergroup processes, and the psy­
chology of political action. Recent publications include 'Grassroots 
Political Action as an Intergroup Phenomenon' (with L. Fox-Cardamone 
et al.), Journal o
f
 Social Issues (1996) and 'Recognizing Group Diversity: 
Individualist-Collectivist and Autonomous-Relational Social Orientations 
and Their Implications for Intergroup Processes' (with R.J. Brown et al.), 
British Journal o
f Social Psychology (1992). 
Chester Insko is Professor at the Department of Psychology, University of 
North Carolina, USA. Research interests are intergroup relations, inter­
personal attraction and attitude change. Recent publications include 
'Memory for and Experience of Differential Competitive Behavior of 
Individuals and Groups' (with M.J. Pemberton and J. Schopler), Journal 
o
f
 Personality and Social Psychology (1996), and 'Differential Distrust of 
Groups and Individuals' (with J. Schopler) and 'Entitativity, Loss of Per­
sonal Control, and Evolution' (with J. Schopler and C. Sedikides), both 
included in C. Sedikides et al. (eds), Intergroup Cognition and Intergroup 
Behavior (Hillsdale, NJ: Erlbaum, in press). 
Patricia Insua is Assistant Professor of Language and Thinking in the 
Department of Basic Psychology, School of Psychology, University of the 
Basque Country. She is co-author, with D. Paez and J. Marques, of a chapter 
on 'Social Cognition' in the book edited by J.F. Morales et al. entitled Social 
Psychology (Psicologia Social) (Madrid: McGraw Hill, 1994). Her main 


xu 
Social identity: international perspectives 
research interests in social psychology are social representations, attitudes, 
AIDS and drug-addiction processes, and they also embrace general 
psychological processes of language, cognition and emotional problems. 
Maria Jarymowicz has been Professor of Psychology at the School of 
Psychology of the University of Warsaw and Head of the Laboratory 
'Studies on Personal and Social Identity' since 198 1 .  Her main research 
interests focus on ego and self in social perception, self-we-others schemata 
distinctiveness and personal/social identity differentation - in relation to 
egocentric/non-egocentric perception, the individualism-collectivism rela­
tionship and the affective basis of social stereotypes. Her most representative 
publications (both as editor) are To Know Self 
- to Understand Others (Delft: 
Eburon, 1993) and Beyond Egocentrism in Self- and Social Perception (Poza 
EgocentrycznQ PerspektywQ Widzenia Siebie i Swiata, Warsaw: Wydaw­
nictwo Instytutu Psychologii PAN, 1994). 
Mercedes Lopez-Saez is Assistant Professor of Social Psychology at the 
Department of Social Psychology of the Spanish Open University at 
Madrid (UNED). Her main research interests include gender stereotypes, 
social influence and persuasion. Recent publications include 'Cultural and 
Individual Processes Involved in Gender Stereotyping: An Empirical 
Approximation to Career Election' ('Procesos Culturales e Individuales 
Implicados en la Estereotipia de Genero: Una Aproximacion Empirica a la 
Eleccion de Carrera'), Revista de Psicologia Social (1994), 'Gender Stereo­
types in the Spanish Population: Looking Toward the Future' (with 1.F. 
Morales), in L. Amancio and C. Nogueira (eds), Gender, Management and 
Science (Braga: Institute of Education and Psychology, 1995), and The 
Election o
f a Typically Feminine or Masculine Career f
rom a Psychosocial 
Perspective (La Eleccion de una Carrera Tipicamente Femenina 0 Masculina 
Desde la Perspectiva Psicosocial), Madrid: Ministry of Science and 
Education, 1995). 
Jose M. Marques is Professor of Social Psychology at the University of 
Porto, Portugal. Research interests centre mainly on the field of intergroup 
relations, specifically the role of intergroup processes in generating group 
cohesiveness and enhancement of social identification. Recent publications 
include 'The Black Sheep Effect: Social Categorization, Rejection of 
Ingroup Deviates, and Perception of Ingroup Variability' (with D. Paez), 
European Review o
f
 Social Psychology (1994), 'The Representation of 
Emotions in Groups: The Relative Impact of Social Norms, Positive­
Negative Asymmetry and Familiarity on the Perception of Emotions' (with 
D. Paez and A.I. Vergara), European Journal o
f
 Social Psychology (1996), 
and 'Collective Memory Processes Associated with Traumatic Experiences: 
Social Sharing, Emotional Climate and the Transgenerational Transmission 
of Information in the Case of the Portuguese Colonial War' (with D. Paez 
and A.F. Serra), in 1. Pennebaker et al. (eds), Social Memory and Collective 
Trauma (Hillsdale, Nl: Erlbaum, 1997). 


List o
f
 contributors 
X 
111 
Cristina Martinez-Taboada is Assistant Professor at the Department of 
Social Psychology, School of Psychology, Basque Country University, 
Spain. He has done research on the topic of social identity, group dynamics 
and social representions. He is co-author (with S. Ayestarin and J.J. 
Arrospide) of a chapter on 'New Techniques for the Study of Groups' 
('Nuevas Tecnicas en el Estudio de los Grupos') included in Methods and 
Techniques o
f
 Research in Social Psychology ( Mhodos y Tecnicas de 
Investigaci6n en Psicologia Social, Madrid: Eudema, 1994). 
J. Francisco Morales is Professor of Social Psychology at the Department 
of Social Psychology, of the Spanish Open University at Madrid (UNED). 
He has done research on leadership, social identity and its relation to 
individualism-collectivism. Recent publications include 'Three Portraits of 
Social Psychology in Europe', Proceedings o
f
 the Tenth Meeting o
f
 Cheiron 
( 1992), 'Leadership in Two Types of Health-Care Organizations' (with F. 
Molero), in J.M. Peiro et al. (eds), Work and Organizational Psychology: 
European Contributions o
f
 the Nineties (London: Taylor & Francis, 1995), 
Treatise o
f
 Social Psychology I: Basic Processes ( Tratado de Psicologia 
Social I: Procesos Basicos (with M. Moya); Madrid: Sintesis, 1997) and 
Social Psychology ( Psicologia Social) (with M. Moya et al.); Madrid: 
McGraw-Hill, 1 994). 
Miguel Moya is Professor of Social Psychology at the School of Psychology 
of the University of Granada, Spain. His main research interests are 
gender, interpersonal and intergroup relations, prejudice and stigma. 
Recent publications include Treatise o
f
 Social Psychology I: Basic Processes 
( Tratado de Psicologia Social I: Procesos Basicos), (with J.F. Morales; 
Madrid: Sintesis, 1 997), Psychology and Gender (Psicologia y Genero) (with 
M.e. Gomez; Granada: Universidad de Granada, 1997), and Social 
Psychology (Psicologia Social), (with J.F. Morales et al.; Madrid: 
McGraw-Hill, 1994). 
Gabriel Mugny is Professor of Social Psychology at the Faculty of Psy­
chology and Educational Sciences, University of Geneva, Switzerland. His 
main research interests are social influence, intergroup relations, social 
identity, cognitive development and social representations. Recent 
publications include Psychologie Sociale et Developpement Cognitif (with 
W. Doise; Paris: Armand Colin, 1997), La Psychologie Sociale, Vol. i: 
Relations Humaines, Groupes et In
fluence Sociale (with D. Oberle and J.L. 
Beauvois; Grenoble: Presses Universitaires de Grenoble, 1995), The Social 
Psychology o
f Minority Influence (with J.A. Perez; Cambridge: Cambridge 
University Press, 1991) and In
fluences Sociales (with J.A. Perez et al.; 
Neuchatel: Delachaux et Niestle, 1993). 
Penelope Oakes born in England, is currently Senior Lecturer in Social 
Psychology at the Australian National University in Canberra, Australia. 
She has published widely on stereotyping and group processes, including 


XlV 
Social identity: international pers
pectives 
Stereotyping and Social Reality (with A. Haslam and J.C Turner; Oxford: 
Blackwell, 1 994). She is also co-editor of The Social Psychology o
f
 Stereo­
typing and Group Lif
e (with R. Spears et al.; Oxford: Blackwell, 1997). 
Dario Paez, born in Chile, is currently Professor of Social Psychology at 
the Department of Social Psychology, School of Psychology, Basque 
Country University, Spain. He has done research mainly on social identity, 
emotions and health, collective memory and social representations. He has 
published 'Inhibition, Alexythimia and Health' (with N. Basabe et al.) in J. 
Pennebaker (ed.), Emotion, Disclosure and Health (Washington: APA, 
1995), and Collective Memory o
f
 Political Events (with J. Pennebaker and 
B. Rime; Mahwah: Erlbaum, 1997). 
Juan Antonio Perez is Professor of Social Psychology at the University of 
Valencia, Spain. His main current research interests are in minority 
influence and racism against gypsies. Recent publications include Influences 
Sociales (with G. Mugny et al.; Neuchatel: De1achaux et Niestle, 1993) and 
'The Conflict Elaboration Theory of Social Influence', (with G. Mugny), in 
E. Witte and J. Davis (eds), Understanding Group Behavior, Vo!' 2 
(Hillsdale, NJ: Erlbaum, 1990). 
Jose Miguel Salazar is Professor of Psychology at the Central University of 
Venezuela, where he pioneered social psychology studies in the late 1950s. 
Currently he is Editor of the Revista Interamericana de Psicologial 
Interamerican Journal o
f
 Psychology. His main interests centre on issues of 
nationalism, national identity and supra-nationalism, and he has done 
research on the relationships of nationalistic attitudes to behaviour, Latin 
American identity and value changes, and, more recently, the construction 
of national and quasi-national identities. He has published Psychological 
Foundations o
f
 Nationalism (Bases Psicol6gicas del Nacionalismo, Mexico: 
Trillas, 1987) and numerous articles and chapters on the subject. 
John Schopler is Professor of Social Psychology, Department of 
Psychology, University of North Carolina, Chapel Hill, USA. His main 
research interests are in intergroup relations, social power, interdependence 
theory and conflict resolution. His recent publications are 'The 
Discontinuity Effect in Interpersonal and Intergroup Relations: Generality 
and Mediation' (with CA. Insko), European Review o
f
 Social Psychology 
(1992), 'The Role of Communication in Interindividual-Intergroup 
Discontinuity' (with CA. Insko et al.), Journal o
f
 Con
flict Resolution 
(1993), 'Individual-Group Discontinuity: Further Evidence for Mediation 
by Fear and Greed' (with CA. Insko et al.), Personality and Social 
Psychology Bulletin (1992), 'The Survival of a Cooperative Tradition in the 
Intergroup Discontinuity Context' (with CA. Insko et al.), Motivation and 
Emotion ( 1994), 'The Role of Identifiability in the Reduction of 
Interindividual-Intergroup Discontinuity' (with CA. Insko et al.), Journal 
o
f
 European Social Psychology (1995), 'Memory for and Experience of 
Differential Competitive Behavior of Individuals and Groups' (with M.J. 


List o
f
 contributors 
xv 
Pemberton and CA. Insko), Journal o
f
 Personality and Social Psychology 
(1996), and Intergroup Cognition and Intergroup Behavior (with C 
Sedikides and CA. Insko; Hillsdale, NJ: Erlbaum, in press). 
Carmencita Serino is Lecturer in Social Psychology at the Department of 
Psychology of the University of Bari, Italy. Her main research topics are 
social categorization, intergroup relations, similarity judgement, self-others 
comparison and social representations. Her recent publications include 
Asymmetry Phenomena in the Interpersonal Comparison (with L. Arcuri; 
Naples: Liguori, 1992), 'Information Processing and Self-Others Compar­
ison: Social Cognitive Extensions of a Non-Metrical Approach to Simi­
larity', in M.C Hurtig et al. (eds), Studies on the Self and Social Cognition 
(New York: World Scientific Publishing, 1993), 'Mental Illness in Psycho­
social Research: Processes of Explanations and Implicit Personality 
Theories' ('La Malattia Mentale Nella Ricerca Psico-Sociale: Processi di 
Spiegazione e Teorie Implicite della Personalita'), in G. Bellelli (ed.), The 
Other Illness: How Society Thinks about Mental Illness (L'Altra Malattia: 
Comme la Societa Pensa la Malattia Mentale, Naples: Liguori, 1994), 
Singular Worlds: Socio-Cognitive Perspectives on Identity ( Mondi al 
Singolare: Prospective Socio-Cognitive Sull'identita) (with J.C Deschamps; 
Naples: Liguori, 1996), and 'Similarities and Differences in the Social 
Representations of the Handicapped' (,Somiglianza e Differenza nella 
Rappresentazione Sociale dell'Handicap', in G. Petrillo (ed.), Social Psy­
chology o
f
 Health ( Psicologia Soziale della Salute, Naples: Liguori, 1997). 
Laurie A. Taylor is the Manager of Community Evaluation Processes at the 
United Way of the Greater Dayton Area in Dayton, Ohio, USA. Her 
research interests include group decision-making, intergroup perceptions 
and identity formation, conflict management and applied outcome 
evaluation. She has most recently published 'Recognizing Group Diversity: 
Individualist-Collectivist and Autonomous-Relational Social Orientations 
and Their Implications for Intergroup Processes' (with R.J. Brown et al.), 
British Journal o
f
 Social Psychology (1992). 
John C. Turner, born in England, is Professor of Psychology and Head of 
Division, Division of Psychology, Australian National University. His 
research interests are in social identity, intergroup relations, group 
processes and social cognition, particularly from the perspective of self­
categorization theory. He collaborated with Henri Tajfel at Bristol in the 
1970s in the development of social identity theory and is the originator of 
self-categorization theory. Recent publications include Stereotyping and 
Social Reality (with P.J. Oakes and S.A. Haslam; Oxford: Blackwell, 1994), 
'Social Identity, Interdependence and the Social Group: A Reply to Rabbie 
et aI.' (with R.Y. Bourhis), in W.P. Robinson (ed.), Social Groups and 
Identity (Oxford: Butterworth Heinemann, 1996), and 'The Socially Struc­
tured Mind' (with P.J. Oakes), in CA. McGarty and S.A. Haslam (eds), 
The Message o
f
 Social Psychology (Oxford: Blackwell, 1997). 


XVI 
Social identity: international pers
pectives 
Laura Vega has done graduate studies in Social Psychology at the Depart­
ment of Social Psychology of the Spanish Open University at Madrid 
(UNED). She is currently working for her doctoral dissertation and doing 
research on the relations between individualism and identity differentiation. 
She is co-author of a paper on 'Individualism, Collectivism and Social 
Identity' (,Individualismo, Colectivismo e Identidad Socia!'), included in a 
monographic volume on 'Social Identity' published by the Revista de 
Psicologia Social ( 1992). 
Stephen Worchel is Dean of the College of Arts and Sciences and Professor 
of Psychology at the University of Southern Maine. His research areas 
include intergroup relations, group dynamics, conflict resolution and cross­
cultural psychology. Recent publications related to social identity include 
the following: 'The View from Below: Intergroup Relations from the 
Perspective of the Disadvantaged Group' (with H. Rothgerber), Journal o
f
 
Personality and Social Psychology ( 1997); 'A Dynamic Approach to Group 
Dynamics', in P. Nye and A. Brower (eds), What's Social about Social 
Cognition? (Newbury Park: Sage, 1996); 'Changing the Stereotype of 
Stereotypes' (with H. Rothgerber), in R. Spears et al. (eds), The Social 
Psychology o
f
 Stereotyping and Group Lif
e (Oxford: Blackwell, 1997); and 
Con
flict between People and Groups (with J. Simpson; Chicago: Nelson­
Hall, 1993). 


Preface 
This volume deals with the issue of identity, the concern over which is far 
from being something new. In fact, it can be traced to classical Antiquity. 
Identity is present in the famous 'Know Thyself' precept, inscribed on the 
frontispiece of the Delphi Temple. It is also the target of many of the 
famous Socratic questions and the stimulus for Plato's well-known reflec­
tions. It can be considered, without exaggeration, an issue which has 
occupied humankind since the beginning of civilization. 
It also focuses on categorization or, more explicitly, on the analysis of 
how we perceive other people (stereotypes) and how we decide to behave 
towards them (prejudice, discrimination), and, in general, on the under­
standing of our readiness to come to grips with the social environment. All 
these problems have a long tradition, but they cannot be taken as old­
fashioned or as belonging to the past, since they are pervasive in our 
current, so-called 'civilized' societies. 
Admittedly, it would be illusory to try to find in the Socratic-Platonic 
analysis of the Delphic formula the point of departure of a current 
psychosocial analysis of the personal and social self. Only in the last 20 
years or so have social psychologists provided such an analysis, framed in 
the recent, but already well-known theories of social identity, and social 
and self-categorization. It is the purpose of this volume to offer an updated 
and balanced view of these fairly new developments, in the expectation that 
it will foster a better understanding of some of the more urgent and central 
problems of our time. 
The latest elaborations of the theories and the most recent research 
inspired by them are especially attended to. A detailed summary of them 
would exceed the length usually allowed to any preface. However, we can, 
at least, highlight three of their main contributions, closely linked to three 
salient features of the theories mentioned. 
In this respect, it should be stressed that one of the most important assets 
of the social identity and self-categorization theories is the building of an 
original theoretical corpus which has provided social psychology with new 
theoretical insights. 
Undoubtedly, before Tajfel's formulation of social identity theory, there 
were several psychosocial approaches to the study of identity. However, 
perhaps due to their low theoretical coherence, they did not succeed in 
making social identity a crucial concept of the discipline of social psy­
chology. The same can be said of self-categorization theory. 


XVlll 
Social identity: international perspectives 
This theoretical corpus was not without its appropriate empirical content, 
since the theoretical concepts and propositions were elaborated through an 
intense dialogue with certain psychosocial phenomena demanding explana­
tion, such as intergroup discrimination, the effect of stability and legitimacy 
of status differences, the processes of social comparison and of group 
competition, among others. In providing such an explanation, both theories 
became more and more influential and began to play a leading role within 
social psychology. 
Going still a step further, they widened their scope to accommodate 
other psychosocial phenomena. Let's take the example of the group con­
cept. Tajfel himself had suggested the need for its reformulation. However, 
this was only done by Turner several years later. The covariation hypoth­
esis, formulated by Deschamps, and linking intra-
and intergroup 
differentiation, is just another example of the success of the theories in 
their efforts to cover other psychosocial phenomena. Besides the group 
concept and the covariation hypothesis, the following processes, among 
others, are worth mentioning: influence, group polarization, collective 
behaviour, the impact of minorities, prejudice, stereotypes and leadership. 
This speaks in favour of the fruitfulness of both theories. 
These three features correspond approximately to the stages of develop­
ment of the theories, but, at the same time, all three have been, and still 
are, open and subject to changes and modifications. The fact that there is a 
mutual influence between them cannot be forgotten either. Thus, the 
intensive study of intergroup discrimination has helped to reformulate the 
early concepts, throwing light on the functions of comparison dimensions. 
By the same token, the extension of the theory to the group concept has 
allowed Turner to replace the early interperson-intergroup continuum by a 
new, more accurate interperson-group one. 
In recent years, the sustained vitality of the theories has allowed them to 
gain new empirical content with the corresponding increase in the amount 
of explanations of phenomena of interest for social psychoiogy. As a result, 
they have become both more complex internally and more sophisticated in 
their theoretical formulation. The contributors to this volume, among them 
some of the original formulators of both theories, have tried to give them 
even more impulse. They raise new issues, examine critically some crucial 
points, and explore new areas of research. Their work will serve to expand 
the theories and, we hope, to persuade others to join the effort. 
The overall aim of this volume is, then, to demonstrate the breadth of 
scope of issues related to social identity. Indeed, one of the most unique 
features of the area is that it has found a warm welcome in a variety of 
camps. As such, we begin the volume with chapters addressing identity and 
the self. We move from here to contributions that deal with identity in the 
group and intergroup context. And the final contributions venture into the 
realm of cultural influences on social identity. 
But the breadth of social identity is not only demonstrated by its con­
tribution across the social spectrum. It is also a theory that has captured 


Pre
f
ace 
xix 
the imagination of investigators in diverse lands with different cultural 
perspectives. In an effort to represent this geographical scope, the chapters 
in this volume represent a mosaic from five continents (Asia, Europe, 
North America, South America, Australia) and seven countries (Israel, 
Australia, Spain, Great Britain, the United States, Poland, Venezuela, 
Switzerland). Although the specific focus of each chapter represents a 
combination of the investigator's interest and his or her cultural founda­
tion, the social identity framework provides a common language that 
eclipses geographical, political and cultural boundaries. 
We'd like to extend our appreciation to our colleagues who have com­
mented on specific chapters. Special thanks is given to Dawna Coutant, 
John Turner and William Webb. Jennifer Ford, Kim Cozzi and Jon Iuzzini 
provided invaluable editorial assistance in preparing the manuscript for 
publication. And above all, we extend our appreciation to Ziyad Marar 
and Lucy Robinson, who were brave enough to accept such an ambitious 
project and whose expertise and gentle guidance moved it from the idea 
stage to book form. 




PART I 
REPRESENTATIONS OF 
SELF AND GROUP 
1 
Regarding the Relationship Between 
Social Identity and Personal Identity 
lean-Claude Deschamps and Thierry Devos 
The problems to be discussed in this chapter can be considered the central 
issue regarding both identity and a general perspective of social psychology; 
the theme of this chapter describes the relationship between two concepts 
called social identity and personal identity. 
Few concepts are as polysemic as identity. The idea of identity appears in 
the first texts of Antiquity (the famous 'Know Thyself' of Delphi) and is 
used in both everyday language and scholarly discourses; it is also the 
concern of many scientific disciplines. There are many synonyms or near 
synonyms for identity (one can mention at random: oneself, t, the self, we, 
self-perception, self-image, self-representation or self-awareness, the ego, 
and so on). The self can refer to a familiar subject, but also to a social 
agent's action. It can also refer to active processes and mechanisms that rule 
behaviour. Therefore, our remarks are limited to specific aspects of identity. 
Social identity and personal identity 
Identity is a central concept in social psychology, probably because it is one 
of the main concerns of that field. As Codol wrote (1979, p. 424), that 
major concern - which has been the subject of very old debates regarding 
philosophy and religion as well as ideology and the humanities - simply 
refers to the conflict of affirmation and individual necessity versus affir­
mation and collective necessity, the search for personal identity versus the 
search for collective identity, what constitutes individual difference versus 
what constitutes similarity to others (same things constitute both), social 


2 
Social identity: international perspectives 
visibility versus conformity, in short the conflict of the individual versus the 
group. And that concern, which can take many forms, refers to the 
relationship between the concept of the individual and the collective, often 
viewed as conflicting. This relationship is central in social psychology. In 
other words, it is an opposition between the diverse and the homogeneous, 
or, as Robert Pages notes, it is the recurrent theme of the same and the 
other. 
Studies about self-concept and identity also deal with the opposition 
between the individual and the social; and that opposition is codified 
according to the distinction made between personal identity and social 
identity. This distinction was already made - although not explicitly - at 
the beginning of modern psychological thinking. At the end of the last 
century, William James ( 1890) wrote about the distinction between the I 
and the me and came up with the idea of duality in self-representation. 
Nowadays, this distinction continues to have a great impact on studies on 
identity. Mead ( 1934) goes deeper into the question with his idea that the 
self consists simultaneously of a sociological component (the me) - which 
would only be an internalization of the social function - and a more 
personal component (the I); the developments of that definition of the self 
have gradually led to the distinction between social identity and personal 
identity. Of course, everyone agrees that identity is both personal - in the 
sense that it is 'situated within' a person - and social, inasmuch as its 
processes of individual formation are social, but the individual is indeed 
characterized by two poles - one is psychological and the other sociological 
- that have no connection to each other. At first, the individual/society 
dichotomy seemed outmoded, particularly with Mead's idea of conversation 
between the I and the me, according to which the individual is a continuous 
creation of society, and society an unremitting creation of individuals. But 
we will see that the two poles, the individual and the social, tend to be 
considered opposite. 
Let us first make clear the concepts of social identity and personal 
identity. These concepts are based on the idea that every individual is 
characterized by social features which show his or her membership of a 
group or a category, on the one hand, and by personal features or indi­
vidual characteristics which are more specific, more idiosyncratic, on the 
other. 
The former features define the social identity of a person. Belonging to a 
group or to a given social category is the most important. Social identity is 
codified as the part of the self which refers to cognitions ensuing from 
social ecological positions (Sarbin & Allen, 1968). Those who have similar 
positions and common backgrounds have similar social identities, therefore 
social identity does indeed refer to the similarity pole. However, the feeling 
of belonging to a group and the phenomena of identification are only 
possible in connection with groups or categories one does not belong to. 
Thus social identity refers to the fact that the individual perceives him- or 
herself as similar to others of the same background (the we), but social 


Social identity and personal identity 
3 
identity also refers to a difference, to a specificity of that we in connection 
with members of other groups or categories (the them). We have then a 
double motion which combines ingroup similarities and intergroup or 
categorial differentiation. The stronger the identification with a group, 
the more significant the differentiation of that group from other groups 
will be. 
More personal features or specific character attributes of each individual 
refer to personal identity and the idea that each individual is a unique 
combination of features which make him different from others, and which 
explains his uniqueness and the fact that he is specific. Personal identity is 
not well defined. However, it indicates how an individual is aware of his 
difference with respect to others. That feeling can only be experienced in 
relation to others and personal identity refers to the fact that the individual 
perceives himself as identical to himself; in other words he is the same in 
time and in space, but that is also what specifies him and marks him out 
from others. Personal identity is what makes you similar to yourself and 
different from others. 
Of course, one could debate this postulate regarding self-permanence as a 
relatively stable and durable concept held by everyone (see, for instance, 
Goffman [1956] with his descriptions of the individual who plays different 
roles for different audiences and Gergen [1965, 1982] with his idea of 
fluidity of the self). It is therefore necessary to speak of identities in the 
plural, considering that every social agent - whether individual or collective 
- can actualize, mobilize or produce identities according to the context. 
This also leads us to the question of the historical universality of this feeling 
of identity. But that does not change fundamentally the problem that 
concerns us. The fact that the feeling of identity is socially built into every 
individual and modulated in every social position and that it is an 
ideological construction does not mean that it is not real and efficient; it is a 
guide for individuals' actions and it is essential for the functioning of our 
society. 
Now, we must see how the articulation, the organization, the synthesis of 
personal and social aspects of identity are conceptualized. Social identity 
refers to a feeling of similarity to (some) others; personal identity refers to a 
feeling of difference in relation to the same others. This distinction between 
personal identity and social identity is indeed only an example of duality 
between the individual and the collective or difference and similarity. On 
the basis of generally accepted definitions of what is conventionally called 
social identity and personal identity, we have to deal with the conflict 
between the individual and the social. This lack of differentiation on a 
certain level - in other words social identity - and this differentiation -
personal identity - are considered most of the time as two opposite poles 
between which human behaviour oscillates. This combination of similarities 
and differences is problematic. It is hard to conceive how one can feel very 
similar and very different at the same time; it seems as though one has to 
choose to be one or the other. 


4 
Social identity: international perspectives 
Social identity theory 
The studies carried out by the 'Bristol School' emphasize the link between 
social identity and personal identity. In order to understand how the con­
cept of identity is theorized and used in that perspective, first of all a certain 
number of developments regarding the process of categorization must be 
addressed. 
Categorization and social categorization 
Categorization refers to psychological processes which tend to organize the 
environment into categories or groups of persons, objects, events (or groups 
of some of their characteristics) according to their similarities, their equi­
valences concerning their actions, their intentions or behaviour (Tajfel, 
1972a, p. 272). Categorization divides the environment into groups of sub­
jects which are or seem to be similar according to certain criteria; these 
subjects are or seem to be different from other subjects according to the same 
criteria. Categorization plays a specific part in structuring the environment 
by systematizing (dividing and organizing) and thus by simplifying it. 
One of the major effects of categorization and the simplification it implies 
is that it emphasizes the differences between categories (contrast or cogni­
tive differentiation effect) and the similarities within the same category 
(assimilation or cognitive stereotype effect). Each element of the category is 
characterized by the features shared by all the subjects of that category. 
Thus, a category is a group of elements which have in common one or 
several features. Cognitive apprehension is a simpler apprehension of 
reality. The emphasis on similarities and differences is one of the aspects of 
that simplification. Dealing with a psychological process, there is no reason 
at first sight to think that this process does not also play a role in the 
perception of the others, who are characterized by the fact that they belong 
to social categories. The fact that the elements referred to are persons and 
that we deal with social categories does not matter; the process does not 
change in that perspective. Actually, a social category is principally a 
cognitive category, where individuals who are supposed to share one or 
several features are grouped together. Stereotypes are - on a social level -
the expression of the attribution of features shared by different members of 
a group without taking into account the interindividual differences. The 
classification used as a basis for the stereotypes minimizes the differences in 
features among the members of a social group. As Doise noted in 1976, 
categorization is then a psychological process that simplifies the perception 
of physical and social worlds and the way the individual organizes the 
subjective perception of his or her environment. 
But apart from the homology between the perception of physical environ­
ment and social environment, social categorization - unlike physical stimuli 
categorization - implies that subjects are themselves inside a system of 
categories. They are subjects and objects of categorization at the same time. 


Social identity and personal identity 
5 
Social categorization minimizes the in-category differences and exaggerates 
the differences between categories. In addition, these differences are used for 
evaluation. When people hold the representation of a dichotomous universe 
where you can belong only to one category, they end up having a dis­
criminating attitude towards the members of other categories. 
Social categorization and identity 
How can discrimination between groups be explained? Unlike in Sherif's 
theory of the objective conflict of interest between groups (see, for instance, 
1966), competition is not a sufficient explanation for discrimination between 
groups. For Tajfel (l 972a) the idea of belonging to two different groups 
leads to discrimination in favour of the group one belongs to. But then, one 
must refer to concepts other than competition to explain the phenomenon 
of discrimination. This is when the concept of identity is introduced. It is 
important to recall the main points of social identity theory (see in 
particular Tajfel & Turner, 1979). 
According to Tajfel ( l972a), social identity is conceptualized as being 
connected to the individual's knowledge of belonging to a certain social 
group and to the emotional and evaluative signification that results from 
this group membership. Thus, it is through their belonging to different 
groups that individuals acquire a social identity defining their specific posi­
tions in society. But belonging to a given group contributes to the devel­
opment of a positive social identity only if the characteristics of that group 
can be compared favourably to other groups. In comparing the groups, 
individuals tend to establish a difference which is in favour of their own 
group. 
From 1 974 on, Tajfel adds an a priori distinction between two opposite 
poles in social behaviour. At one extreme (which can probably not be found 
in such a 'pure' form in 'real life'), we find interactions between two or more 
individuals. These individuals are totally determined by their interpersonal 
relations and by their individual characteristics and not affected at all by the 
different groups or social categories they belong to. At the other extreme, 
there are interactions between at least two individuals (or groups of 
individuals) which are totally determined by their respective membership of 
different groups or social categories and not affected at all by inter­
individual relationships among the concerned persons. Here again, there are 
few chances of encountering such 'pure' forms of this extreme in 'real' social 
situations (Tajfel & Turner, 1979). 
The two poles refer to interpersonal attitudes on the one hand, and to 
intergroup behaviour on the other. As Turner (198Ja) noted, personal 
identity corresponds to interpersonal behaviour - which means differ­
entiation between the self and the others; and social identity corresponds to 
intergroup behaviour - which means differentiation between groups or 
between 'we' and 'them'. 


6 
Social identity: international perspectives 
When identification with a group is emphasized, one changes from the 
interpersonal pole to the intergroup pole. The introduction of this distinc­
tion between interpersonal and intergroup phenomena is important for 
things are not going to be as simple as in Tajfel's early theory, according to 
which this distinction motivates a need for a positive difference in favour of 
one's own group compared to other groups, the creation of that difference 
and the emphasis on it. What really matters it the need for self-esteem, the 
'desire' for a positive self-evaluation, the need for individual positivity. The 
basic postulate is that 'individuals tend to maintain or increase their self­
esteem; they try to reach a positive self-concept' (Tajfel & Turner, 1979, 
p. 34). Thus, the individual tries to preserve or reach a positive self-image, 
and if she cannot evaluate herself according to the interpersonal behaviour 
pole - which lies in evaluating oneself directly by comparing oneself to the 
others - she can satisfy her desire for positive self-evaluation through social 
competition between groups (which must be considered as a groups com­
parison) with this tendency to introduce a positive difference in favour of 
one's own group in relation with other groups. 
In fact, social identity theory is founded on two bases: one is cognitive 
and the other is motivational. As mentioned before, it is cognitive insofar as 
the categorization process leads the subjects to overestimate intergroup 
differences and to underestimate ingroup differences. It is motivational 
insofar as what motivates discrimination is the need for self-esteem - or 
self-respect, according to Emler and Hopkins (1990). 
Turner (1975a, 1975b), who belongs to that school of thought, concludes 
that subjects act in terms of differentiation between groups only when this 
differentiation is the only means for them to achieve a positive self­
evaluation, a positive identity. If they can act in terms of a differentiation 
between self and others, they do not resort to the intergroup 'bias'. This only 
occurs when identification with the group provides for a positive identity. In 
other words, categorization on its own would not be enough to observe a 
differentiation between groups; the tendency to establish a positive differ­
entiation between self and others underlies the social categorization process. 
To put it differently, if the individual accedes to a positive identity through 
his identification with a group, he establishes differentiation between groups 
but no longer tends to establish differentiation with the other members of his 
group. If the individual is able to differentiate himself from the others 
directly - acceding in this way to a positive self-evaluation - he no longer 
establishes differentiations with other groups which he knows. 
The intention here is not to develop the limits of these studies or to 
question aspects of social identity theory (see Deschamps, 1984, 1991, for 
more details). However, it must be underlined that on the basis of this 
asserted dichotomy between interindividual and intergroup behaviour, one 
can say that the stronger social identity is, the less important is personal 
identity, and the more prominent a personal identity is, the less the 
individual needs a social identity, since social identity and personal identity 
satisfy the same need for a positive self-image. According to the social 


Social identity and personal identity 
7 
identity model, we can predict that when the salience of belonging to a 
group increases, every subject identifies him- or herself more and more with 
the in-group and that will reduce differentiation between the self and the 
ingroup while exacerbating differentiation between groups. In other words, 
when identification with a group increases, one changes from the inter­
personal pole to the intergroup pole, and vice versa. This dichotomy solves 
the problem of differences and similarities since only one pole is possible, 
switching from one to the other. This is indeed what Turner refers to when 
he introduces the concept of 'depersonalization'. For him, the factors which 
intensify the salience of the ingroup/outgroup categorization tend to 
increase identification (similarity, equivalence, interchangeability) between 
the self and the other members of the group (and differentiations with 
members of another group). Consequently, these factors depersonalize the 
self according to stereotypes determining the group. Depersonalization 
refers to the 'auto-stereotypy' process through which individuals come to 
consider themselves above all as interchangeable models of a category 
rather than as unique individuals who are distinct from the others (Turner, 
1987, p. 50). As we can see, both analyses - the one based on difference and 
the one based on similarity - depend on each other negatively. The dicho­
tomy postulated between interindividual and intergroup behaviour, which 
actually refers to the duality of self-awareness and the opposition between 
personal identity and social identity solves - apparently - the problem of 
similarities and differences while involving two mutually exclusive poles. 
Self-de
finition and its diff
erent levels 
Turner ( 1987) tries to explain the opposition between the psychological 
aspect of identity - which refers to an individual's unity with its constel­
lation of specific features - and the sociological aspect of identity - which 
refers to the idea of group membership and similarities - by distinguishing 
three levels in self-definition with his self-categorization theory: 
(a) a supra-order level where the self is defined as a human being: it refers 
to a human identity based on comparisons between the species (simi­
larity with the human race, differences with other forms of life); 
(b) an intermediate level of self-definition where the self is defined as a 
member of a group (ingroup similarities and intergroup differences): it 
refers to a social identity based on intergroup comparisons (in-species); 
(c) a subordinate level of self-definition where the self is defined as a 
unique being (differentiation between the self and the others inside a 
group); it refers to a personal identity based on interpersonal (in-group) 
compansons. 
However, when individuals situate themselves at a level of categorization, 
they obviously have to occlude the two other levels. As a matter of fact, 
there is a fundamental antagonism between the salience of one self-categorization 
level and the other levels: the salience of one of the categorization levels engenders 


8 
Social identity: international perspectives 
in-class similarity and interclass differentiation which reduces or inhibits the 
perception of in-class differences and interclass similarities on which inferior and 
superior levels of categorization are based. (Turner, 1987, p. 49) 
Turner adds that if, for simplification purposes, the supra-orderly categor­
ization level is ignored, the (social) intermediate level and the (personal) 
subordinate level are linked by a reverse relation. Obviously, the intro­
duction of these levels does not solve the opposition between the collective 
and the individual problem. Their relations remain negatively dependent 
(self-definition in terms of categorial membership - both on the supra­
orderly and the intermediate levels - versus individual specificity). 
In fact, as a whole these studies give the impression that it is difficult to 
get rid of this opposition between the individual and the collective, and that 
we are trapped with a dichotomous analysis according to which similarity 
and difference, social identity and personal identity, are two poles which 
depend on each other negatively. Having questioned the link between 
ingroup resemblance and intergroup divergence, we shall now present 
another series of studies in order to go beyond this opposition between 
individual and collective forms. 
The categorization model, like social identity theory or self-categorization 
theory, states that the same factors increase differences between groups and 
similarity or homogeneity inside the groups. However, this especially 
applies to outgroups and the way they are perceived and treated. The 
individual's own group would be perceived in a much more heterogeneous 
way. In fact, a certain number of studies carried out over the last 1 5  years 
illustrate the fact that the individual believes that his or her own group is 
more differentiated than other groups. This means that the group one 
belongs to can be perceived as heterogeneous by its members. This tendency 
to consider one's own group as relatively less homogeneous than a group 
one does not belong to (or the fact that a group tends to be considered more 
heterogeneous by its own members and not so by individuals who are not 
part of it) has been called the 'outgroup homogeneity effect'. This phenom­
enon has been explained in many ways (some explanations are based on 
motivation, in terms of beliefs, and refer to the self as a special category, 
recalling the way information about ingroup and outgroup is processed and 
codified in the memory; see Ostrom & Sedikides, 1992). But this is not 
essential here. What is important is that the concept according to which 
differentiation between groups necessarily means convergence within groups 
has been questioned and is no longer the rule. 
Covariation between interindividual and intergroup differentiation 
hypothesis 
A number of studies carried out in the 1970s already stated that emphasis 
on similarities inside a group does not necessarily mean emphasis on 


Social identity and personal identity 
9 
differences between groups (for instance, Deschamps, 1972- 1973; 
Deschamps & Doise, 1978). These observations, which were astonishing 
at a time when categorization effects were taken for granted, indicated 
(Deschamps, 1979) that the simple representation of belonging to two 
different groups could cause discrimination between groups in favour of 
one's own group but it would also engender differentiation between the self 
and the rest of one's group. When the representation of a dichotomized 
world is not prominent or relevant in a situation, there is no discrimination 
between groups - which is obvious. Moreover, in certain conditions, the 
differentiation between self and others could be less important than that 
between the self and the rest of the membership group when belonging to a 
group and dichotomy in groups are significant criteria in defining a 
situation. 
This has led us to a new perspective on similarities and differences: the 
intergroup and in-group differences covariation. Instead of considering 
interindividual and intergroup differences as two extremes of a continuum, 
and that they are in this way mutually (at least partly) exclusive, we assume 
that in certain conditions the stronger the identification with the group, the 
more important is interindividual differentiation within the group. Codol 
(for instance, 1975) underlines this issue with what he calls the phenomenon 
of 'superior conformity of the self': the more an individual conforms to the 
standards of a group (and the more she identifies herself with that group), 
the more she will tend to consider herself different from the other members 
of the group, believing that she corresponds to the standards better than the 
others. 
Here is the central concept of the interindividual and intergroup differ­
entiation model, in a few words. A general process of cognitive centrism 
appears when individuals are induced with the representation of a dichoto­
mized world, divided into two mutually exclusive categories. According to 
this representation, both ingroup favouritism or intergroup differentiation 
(which can be called sociocentrism) and autofavouritism or differentiation 
between self and others (which can be called egocentrism) would increase 
when categorization is emphasized. 
The aim of a certain number of our studies was to test this postulate 
(see Deschamps, 1982a, 1983- 1984, 1984; Deschamps & Lorenzi-Cioldi, 
198 1 ;  Deschamps & Volpato, 1984). The experiments carried out in these 
studies are similar to the ones used in Tajfel's minimal group paradigm and 
refer to the changes introduced into that paradigm by Turner in 1975. An 
outline of the surveys will suffice here (Deschamps, 1984). One of them was 
carried out in a girls' college. Using a sample of 88 girls, aged between 16 
and 20, half of the subjects were divided into two groups (categorization 
condition), apparently according to their preference for the works of two 
modern painters (in reality, it was an arbitrary way of splitting them). The 
other half was not divided into two groups and did not belong to any 
group in terms of their artistic preferences (non-categorization condition). 
In accordance with our hypothesis, it was predicted that differentiation 


10 
Social identity: international pers
pectives 
between self and others as well as intergroup differentiation would be more 
significant in the categorization condition than in the non-categorization 
condition. 
Concerning differentiation between groups, the results confirmed our 
expectations: discrimination in favour of the ingroup in the categorization 
condition and of course no discrimination in the non-categorization condi­
tion. Concerning differentiation between self and others, in the categoriza­
tion condition, the differentiation is much more significant than in the 
non-categorization condition. In other words, when individuals are simply 
induced to the representation of an environment including their group and 
another one (which leads of course to differentiation between groups), it 
increases significantly autofavouritism 'bias' or differentiation between self 
and others, compared to a non-categorization condition (without categor­
ization induction). Emphasizing intergroup differentiation in an experiment 
(in this example, by inducing or not the subjects with the representation of a 
dichotomized world) is sufficient to increase the differentiation subjects set 
between themselves and others. 
But more observations must be addressed. The contrary is also true. If 
differences between self and others within a group are emphasized in an 
experiment, differentiation between groups increases simultaneously. 
Another survey was carried out in a high school with 1 12 boys and girls 
aged between 14 and 1 5  (see Deschamps & Volpato, 1984). As in the 
categorization condition of the experiment mentioned above, the subjects 
were placed into two groups, apparently according to their preference for 
the works of a contemporary music composer. Half of the subjects, the ones 
in the individualization condition, expected that each individual would be 
given exactly the same remuneration, according to the minimal group 
paradigm. The other half, the ones in the fusion condition, expected that all 
members of a group would receive the same remuneration, corresponding to 
an average of points which all the subjects would give to the different 
members of that group. Our expectations proved to be right. In the fusion 
condition, differentiation between self and others and discrimination in 
favour of one's own group are less than in the individualization condition. 
When differentiation between self and others is emphasized - by focusing 
not on the groups as a whole, but on the fact that individuals are char­
acterized by their group memberships - differentiation between groups 
increases in the same way. 
In addition, the covariation model seems to apply especially to indi­
viduals belonging to privileged groups. Not only do differentiation between 
groups and ingroup homogeneity not necessarily go together, but the 
simultaneous variation of differentiation inside a group and between groups 
also depends on the relative status of the groups in question (Deschamps, 
1 982b; see also Lorenzi-Cioldi, 1988). These are the conclusions of other 
studies. For instance, it has been shown (Deschamps, 1977) that when boys 
had to evaluate themselves as well as other boys and girls for different 
assignments, they discriminated between boys and girls and between 


Social identity and personal identity 
SI: x--x 
PI: x---------------x 
SI: x---------------x 
PI: x---------------x 
" 
" 
" 
" 
"
SI: x-------x / 
/ 
/ 
/ 
/ 
Social identity 
PI: x-------x 
axis (SI) 
-----Compromise -----
SI: x --x 
PI: x --x 
/ 
,, W 
. i:- / 
,,<9
‚ 
(similarity) 
n
o / 
,,ƒ
-? 
""
i-
{!>
/ 
"
'.90,,.
 
,(
..,0/ 
'- -? 
'- °
0 ' 
/ 
,, œ  
/ 
,,'.9-
Key 
SI: x ---------------x 
PI: x--x 
x--x 
restricted range of identity 
(personal and/or social) 
x -------x 
medium range 
x---------------x 
expanded range of identity 
(personal and/or social) 
Figure 1 . 1  Personal and social dimensions in psychosocial identity 
1 1  
themselves and the other boys. Girls would discriminate between boys and 
girls, but unlike the boys they would not tend to discriminate between other 
members of their own category. 
In conclusion, the results of these studies briefly mentioned underline the 
relevance of the simultaneous variation of differentiation between groups 
and between self and others. However, it does not mean that under certain 
circumstances there is an opposition between personal identity and social 
identity. Obviously, relationships between the individual and the collective 
can be examined according to situations, cultures and societies. Never­
theless, one must at least consider the possibility of simultaneity between 
similarity and difference. Therefore, the postulated axis defining the inter­
individual relations and behaviour pole versus the intergroup relations and 
behaviour pole should be reconsidered. In that prospect, similarity and 
difference, social identity and personal identity must no longer be con­
sidered as two poles of the same continuum which are negatively dependent. 
These different elements must rather be considered as two disconnected 
dimensions that can be 'orthogonalized' in some way (see Figure 1 . 1). 
Consequently, models based on Tajfel's work and the covariation 
hypothesis can be explained at the same time. 


12 
Social identity: international pers
pectives 
Note 
This chapter was translated from the French by Mehr-Afarin Khosrowdad. It takes up and 
develops arguments which have been put forth in some of the previous texts and quotes several 
passages from an article published in 1991 by the first author in Cahiers Internationaux de 
Psychologie Sociale called 'Identites, Appartenances Sociales et DiITerenciations Individuelles'. 


2 
Social Representations in Personal 
Identity 
Willem Doise 
Introduction: Self as social representation? 
Social representations can be defined as organizing principles of symbolic 
relationships. They permit the positioning of the individual in relation with 
significant social objects. The self certainly is such an object embedded in a 
network of relationships. Given its central position in communicative 
systems, personal identity is not just an idiosyncratic entity, it is one of the 
most important organizing principles of symbolic relationships. 
In order to validate the basic assumption of personal identity as a social 
representation, existence of common knowledge about personal identity 
should be evidenced together with systematic positioning of individuals in 
the frame of this common knowledge and the anchoring of such positioning 
in other socio-psychological meaning systems. Indeed, important compo­
nents of all studies on social representations are the definition of a common 
field, the systematizing of individual positioning in that field, and the 
anchoring of these positionings in related fields of representations (Doise, 
Clemence & Lorenzi-Cioldi, 1993). 
Several investigators have already examined aspects of personal identity 
in the frame of social representation theory (see, for instance, Durand 
Delvigne, 1992; Palmonari, Carugati, Ricci-Bitti & Sarchielli, 1979; 
Ravaud, Beaufils & Paicheler, 1986; Tap, 1985; Zavalloni & Louis­
Guerin, 1984; Zeegers, 1988). Nevertheless, personal identity is more often 
studied in other theoretical frameworks and this may explain why books 
containing prototypical social representation studies do not mention 
investigations of personal identity (see, for instance, Abric, 1994; Doise & 
Palmonari, 1986; Duveen & Lloyd, 1990; Jodelet, 1989). A remarkable and 
pioneering exception is the study by Codol (1 984a; see Farr & Moscovici, 
1984). In this study Codol analysed the system of relationships connecting 
representations of self, others, group and task. 
Another possible reason for the paucity of research on personal identity as 
social representation is the limited definition of the concept of personal 
identity. Generally, identity is conceived of as a very peculiar set of opinions, 
judgements, evaluations, attitudes, manifested by a person towards him- or 


14 
Social identity: international perspectives 
herself. Although research from different theoretical traditions bears on the 
general processes governing identity development and construction (Damon 
& Hart, 1982; Erikson, 1977; L'Ecuyer, 198 1 ; Marcia, 1966) and on 
situational regularities enhancing saliency of identity aspects (Codol, 1990; 
Kuhn & McPartland, 1954; McGuire & McGuire, 1988), the study of 
individual identity as socially organized still remains marginal. 
Nevertheless, identity, as other social representations, can be studied as a 
cognitive organization, oriented by a meta-system of social regulations. In 
order to illustrate this thesis I shall successively show (a) that common 
knowledge on identity exists, (b) that organizing principles rule individual 
positioning in relation with this common knowledge frame, and (c) that 
specific social insertions and experiences modulate such positioning. In 
developing my position, I draw on selected evidence from studies on social 
representation theory and a reinterpretation of traditional research on social 
identity. My main objective is to convince the reader that studying identity 
as a social representation is feasible. 
Common knowledge in personal identity 
Personal identity is often considered as unique and singular. On the other 
hand, shared group membership involves common characteristics. Hence, 
individuals would be considered similar to the extent that they belong to the 
same or similar categories. In this sense group membership would result in 
homogeneity and individual differences would correspond to personal char­
acteristics which are not shared by other group members. Such considera­
tions fit with Turner's conception of a functional antagonism between 
the salience of different levels of self-categorization, the group level and the 
individual level: 'The salience of one level produces the intra-class simi­
larities and inter-class differences which reduce or inhibit the perception of 
the intra-class differences and inter-class similarities upon which lower and 
higher levels respectively are based' (Turner, 1987, p. 49). 
Recent research, however, suggests that the problem of similarity and 
difference is not so simple: commonness could also be studied at the indi­
vidual level and specificity at the group level. Let us consider some of the 
research evidence. 
Ravaud et al. (1986) investigated school classes attended by physically 
handicapped pupils as well as classes of 
non-handicapped students. Students 
in both types of classes were asked to describe on the same scales members 
of their class (themselves included), as well as handicapped pupils in general 
and non-handicapped pupils in general. Factorial correspondence analysis 
yielded very contrasting descriptions of the general group of handicapped 
and non-handicapped students. However, self-descriptions of handicapped 
and non-handicapped pupils were very similar. In addition, when the 
descriptions targeted specific handicapped or non-handicapped individuals, 
there was a high degree of similarity. In terms of social representations, 


Social representations in personal identity 
1 5  
differences are found at the category level, but they do not manifest 
themselves when individuals are asked to describe individual members of 
the categories. Commenting on these results, the authors contest the gener­
ality and importance of the so-called 'deductive effect' of social categ­
orization. If such an effect were strong, distinctive traits of categories as 
such should be attributed to single members of each category. 
This is not the case: 
It is a rather striking observation that stereotypes are not very important in 
structuring representations of individuals considered to be bearers of the stereo­
types. The abstract individual described by the valid pupils is totally opposed to 
the descriptions of the concrete individuals and the image that valid pupils 
construct of real handicapped ones and that appears in their answers seems not to 
be mediated through stereotypical representations. This is a strong challenge for 
assumptions that consider stereotypes to be a necessary tool, a measure applied to 
all members of a stereotyped group. (Ravaud et aI., 1986, p. 1 80) 
However, it should be noted that the same research reports results showing 
that real discrimination can coexist with a blurring of stereotypes. When the 
students were asked to choose others for a party at their home or as work 
partners, non-handicapped and handicapped pupils clearly prefer non­
handicapped companions. Such discrimination does not reveal itself in the 
concrete images pupils give of themselves and of each other. 
In the same line of thought, research carried out on adolescent identity by 
Palmonari and colleagues is relevant (Palmonari et aI., 1979; Palmonari, 
Pombeni & Kirchler, 1989). Their results show striking similarities between 
descriptions of self given by adolescents belonging to different formal and 
informal groups. These descriptions reflect their openness to the social 
world, their commitment, disengagement and dissatisfaction. 
Similar results were obtained by Doise and Lorenzi-Cioldi ( 1 99 1 )  study­
ing Swiss and second-generation immigrant pupils. These adolescents gave 
highly contrasting descriptions of Swiss in general and foreigners in general. 
However, when describing themselves and their friends, Swiss and immi­
grant pupils gave quite similar 'concrete' descriptions. Moreover, variations 
between subjects of the same category in self-descriptions are not more 
important than variations between them in the description of group 
stereotypes. To state it differently, self-images are as stereotypical as group 
images. 
The results of other investigations (Nakbi, 1990; Nakbi & Arnal­
Duchemin, 1987) support similar conclusions. These studies gathered not 
only self-images and other-images of male and female students, but also 
ideal images and self-images attributed to others. Multidimensional analyses 
of these data revealed a very similar structure of these representations for 
males and females. 
This research demonstrates a high degree of similarity between self­
definitions of members of different groups. General societal norms affect 
these definitions much more than do specific group memberships. This is 


16 
Social identity: international perspectives 
also a conclusion that can be drawn from research by Cowan and Hoffman 
(1986) and by Inoff, Halverson and Pizzigati (1983). These investigators 
examined the self-descriptions of children, finding striking similarities in the 
self-descriptions of boys and girls, despite the fact that conventional gender 
stereotypes are used in the description of others. It was assumed that these 
children did not apply gender stereotypes to themselves because they tend to 
ascribe favourable traits to themselves, independently of the gender conno­
tation these traits may convey. A similar motivation underlies the self­
descriptions of Genevan pupils: Swiss and immigrant students assigned 
positive traits to themselves that are part of the stereotypes of Swiss 
nationals and foreigners. This tendency also corresponds to the inclination 
of individuals to ascribe to themselves numerous and contrasted charac­
teristics (Sande, 1990). 
The content of self-representations varies from culture to culture and 
epoch to epoch. Commonality in self-definitions in the Netherlands was 
studied by Zeegers (1 988), who analysed the contents of 'lonely heart 
advertisements' in Dutch newspapers from 1945 to 1986. Three styles of 
self-presentation appeared successively in these personal ads. During a first 
phase starting in 1945, a majority of the ads contained traits such as serious 
(degelijk), tidy (net), educated (beschaa
f
d), that is, transsituational evalu­
ative characteristics. During a second phase, from 1965 on, personality 
traits adapted to more specific situations appeared, carrying information 
about lifestyle considered important for developing commitment to a 
relationship that should above all result in friendship. After 1975, the 
advertiser's description of his or her identity became still more specific, 
including more personal traits and declarations of interest in particular 
activities. 
It would be remiss not to mention that there are individual differences in 
the presentation of self-images of members in a group. Experimental results 
also show that it is easy to entice individuals to differentiate themselves 
from other individuals within as well as between groups using characteristics 
deemed to underscore one's originality. However, such differentiations often 
involve an accentuation of conformity to prevailing norms, the so-called 
'superior conformity of the self' (Codol, 1975). 
Normative regulations strongly intervene in self-presentations. More 
generally, linguistic systems offer definite dimensions for describing the self. 
The English language contains numerous terms for individuals to describe 
their personality. Allport and Odbert (1936) identified more than 4,000 such 
terms. Various techniques have been used to systematize this set of traits. 
By asking individuals to use samples of these adjectives to describe them­
selves or other persons, or to indicate similarities between the traits, 
investigators consistently demonstrated the existence of a common structure 
(Goldberg, 1 990; Peabody & Goldberg, 1989). This structure shall be 
described in the next section, but it is important to note here that according 
to various authors this structure corresponds to a common-sense psychology 
embedded in daily language, much more than to the results of strict 


Social representations in personal identity 
17 
scientific observation (Leyens, 1983; Semin, 1987). The important point in 
this discussion is that individuals use common knowledge systems to 
describe themselves. 
Identity as individual positioning 
The images that individuals construct of themselves, although highly similar, 
nonetheless vary from one person to another. In many studies five dimen­
sions are considered necessary and sufficient to describe these interindividual 
variations. According to Goldberg (1990): 'These "Big-Five" factors have 
traditionally been numbered and labeled as follows: (I) Surgency (or 
Extraversion), (II) Agreeableness, (III) Conscientiousness (or Dependabil­
ity), (IV) Emotional Stability (vs. Neuroticism), and (V) Culture. Alterna­
tively, Factor V has been interpreted as Intellect' (p. 1217). 
In fact these organizing principles of between-individual variation deal 
with the way people behave towards each other. A person is involved daily 
in multiple social relationships and the way s/he participates in these 
interactions is being constantly evaluated. If common language, well before 
psychological expert language, contains an abundant range of terms to 
qualify individuals, this variety primarily serves the purpose of differentiat­
ing between individuals according to their behaviour in all kinds of social 
interactions. 
It is impossible to estimate the number of people who have been asked, at 
least once, to describe themselves on a personality questionnaire. If these 
individuals were able to comply with such a request it is because they 
disposed of common references, of comparison frames allowing them to 
compare their own way of behaving with the way prototypical persons 
behave. Apparently, a stable organization of the description of interactional 
systems can be traced back in ancient historical documents. 
Adamopoulos (1982) examined the descriptions of social interactions 
between figures in epic texts of various historical epochs (Homer's Iliad; an 
Old English poem of a fight with a dragon: Beowulf, and a story of a young 
recruit's adventures during the American Civil War: The Red Badge o
f
 
Courage). In each text, dyadic interactions were retrieved and the type of 
interaction between each dyad was classified. Finally, a matrix was con­
structed for each text linking types of interactions and pairs of persons. For 
each matrix a principal component factor analysis was performed. Three 
common factors were obtained for the interpersonal structures in the three 
documents: affiliation, power (status) and formality (versus intimacy), 
although the latter factor appears less clearly in the most ancient text. 
Adamopoulos and Bontempo (1 986) repeated the procedure for The 
Odyssey and The Song o
f
 Roland, and were able to retrieve the same three 
factors in both documents. 
These results can be compared to those obtained by Wish (1976; Wish, 
Deutsch & Kaplan, 1976), who analysed similarity judgements between 25 


1 8  
Social identity: international perspectives 
social relationships (for instance: between husband and wife, between prison 
guard and inmate, between two enemies, between two political opponents). 
The main aim of the study was to identify the dimensions underlying the 
conceptions of these different social relationships. 
Four dimensions described the judgements of prototypical relation­
ships. Dimension I contrasted competitive versus cooperative relationships: 
divorced couple, personal enemies, political opponents versus good friends, 
husband-wife and team mates. Dimension 2 involved unequal versus equal 
relationships: guard-inmate, interviewer-interviewee, teacher-pupil versus 
team mates, political opponents, good friends, personal enemies and cousins. 
Dimension 3 concerned socio-emotional-oriented versus task-completion­
oriented relationships: parent-child, brother-sister, husband-wife versus 
interviewer-job applicant, seller-buyer and political opponents. Dimension 
4 included intense versus superficial relationships: parent-child, husband­
wife, good friends, psychotherapist-patient versus chance encounters, 
interviewer-interviewee and employee-employer. Individual differences in 
the use of these common dimensions were also investigated. 
These representations of social relationships offer important coordinates 
for individuals who have to position themselves in self-descriptions. In this 
respect McAdams (1985) proposes a theory of self as composed of internal 
personified and idealized images. This Imago theory postulates that identity 
'is a story complete with setting, scenes, characters, plot, and recurrent 
themes. Identity formation is the process of constructing a self-defining life 
story. The main characters in the story are imagoes' (McAdams, 1985, 
p. 127). References are classical authors such as Erikson (1977), James 
(1 892), Jung (1943), Klein (1948), but also Markus's (1984) ideas on 
'possible selves'. The symbolic relationships between these prototypical 
images remind us of Adamopoulos's analyses: 
The two major thematic lines around which the content of identity can be 
organized are agency (power/mastery/separation) and communion (intimacy, 
surrender, union; Bakan, 1966). Initial investigations of life stories of college 
students and men and women at midlife suggest that the power motive (Winter, 
1973) and the intimacy motive (McAdams, 1980) are two independent personality 
dispositions - both assessed via the Thematic Apperception Test or TAT - that 
serve as significant predictors of salience in life stories of thematic lines of agency 
and communion, respectively. (McAdams, 1985, p. 128) 
Furthermore, McAdams presented a taxonomy of 12 major deities of the 
Greek pantheon according to their high or low degree of 'agency' and 
'communion', a taxonomy that he also applied to the analysis of identities 
of 50 respondents on the basis of personality tests and of a life story 
interview. 
Individuals define themselves in relation to each other, referring to rep­
resentations of their participation in social interaction settings. The research 
tradition in which I locate my own work emphasizes the importance of 
group and category membership in self-definitions. However, we have 
already seen that individuals often do not assign for themselves the same 


Social representations in personal identity 
19 
characteristics they ascribe to their groups. Notwithstanding such findings, 
it remains true that differences exist between individuals and that some 
people more readily attribute typical group characteristics to themselves. 
This is especially true for gender characteristics. Each time individuals are 
invited to express their degree of agreement with the relevance of masculine 
and feminine traits to their self-descriptions, factors opposing both kind 
of traits can be found in the data (for research with French-language 
subjects, see Lorenzi-Cioldi, 1994; Tap, 1985). This effect certainly cannot 
be explained by mere chance that this organizing male-f
emale principle 
actualizes the opposition between 'agency' and 'communion' described by 
McAdams. In his taxonomy of the pantheon, deities with high 'agency' and 
low 'communion' were masculine (Zeus, Hermes and Ares) whereas deities 
with opposed characteristics were feminine (Demeter, Hera and Aphrodite). 
The organizing principles of identity regulate symbolic relationships 
between social agents and, therefore, correspond to the definition of social 
representations (Doise, 1986). But the fact that positioning in relation with 
each other actualizes dimensions of 'agency' and 'communion' does not 
preclude the importance of other dimensions. 
Verkuyten (1990) has analysed the identity components of a nationwide 
sample of Dutch and Turkish adolescents living in the Netherlands (N = 
2,710 Dutch and 237 Turkish). Some significant differences were found in 
self-descriptions between the Dutch and Turkish youngsters, but these 
differences explained no more than 1 % of the overall variance. More 
importantly, in order to analyse relationship between self-esteem and five 
components of the self-concept (ethnic identity, academic ability, sports, 
popularity, body image), a standardized multiple regression analysis with 
global self-esteem as dependent variable was used for respondents of Dutch 
and Turkish origin separately. The results were highly similar. For the 
Turkish and the Dutch adolescents, body image had the greatest impact on 
self-esteem and ethnic identity, while academic ability and sports had the 
lowest. The similarity of results for both samples suggests that general 
norms are very important in shaping self-esteem, and that general societal 
values transcend, in this case, cultural barriers, probably as a consequence 
of the homogenizing effect of common schooling. 
Makris-Botsaris and Robinson (1991) adapted for Greek youngsters 
scales elaborated by Harter (1985) for measuring different components of 
self-evaluation and perception in the USA. They compared factorial 
structures obtained in Greece and in the USA with youth aged II to 14. In 
both countries they found a highly similar five-factor structure of scholastic 
competence, social acceptance, athletic competence, physical appearance 
and behavioural conduct. Correlations between subscales corresponding to 
these factors and global self-worth were comparable in Greece and the 
USA, the correlation between physical appearance and self-worth always 
being the highest. 
Common knowledge reflects itself in the way individuals define themselves, 
others, and their relationships. Application to oneself of this knowledge is 


20 
Social identity: international perspectives 
highly normative and results in significant interindividual resemblances. 
Nevertheless, between-individual variations are equally important and they 
are organized along dimensions which are also consensually defined and 
which refer to characteristics of the multiple relationships in which indi­
viduals take part. 
Anchored identities 
The anchoring of social representations has mainly been studied in three 
different ways. First, the relationships with general beliefs and values (such 
as belief in a just world or egalitarianism) have been investigated. Such 
values and beliefs are considered general to the extent that they supposedly 
organize symbolic relationships in various domains. Social representations 
have also been studied as anchored in the views individuals develop in 
structuring their social environment, as for instance the representations they 
hold concerning relationships between social groups and categories, such as 
gender. Finally, a third way of studying anchoring investigates how group 
memberships or social positions held by individuals influence their represen­
tations, the general hypothesis being that shared social insertions lead to 
specific interactions and experiences that modulate social representations. 
The first way of studying anchoring mainly uses correlational methods in 
order to determine if links exist between adherence to given beliefs or 
opinions. Such a method was used in the studies reported at the end of the 
previous section showing that important aspects of youth's identity, namely 
self-esteem, were linked to perceptions of their physical appearance and less 
to evaluation of their popularity, academic or sportive achievement. 
Anchoring in the views on the social environment is directly studied in 
research on psychological androgyny linking personal sexual identity to the 
actualization of a gender scheme (see, for instance, Bem, 198 1 ,  1985). 
Indeed, what else is the gender scheme if not a 'naive' theory on the 
respective positions of males and females in society? 
More generally, representations held by individuals on the nature of 
social groups intervene in the structuring of personal identity. Lorenzi­
Cioldi (1988) studied this intervention experimentally. He distinguished two 
conceptions of the social group: the collection group and the aggregate 
group. Members of dominant groups in our society tend more to adhere to 
a conception of their group as a collection of individuals, each having their 
specificity. These individuals define themselves through their personal 
characteristics, apparently not linking themselves to their group. However, 
members of more dominated groups tend to consider themselves as less 
differentiated from each other, as more homogeneous, and they attribute 
more directly to themselves characteristics of the group as a whole. 
Lorenzi-Cioldi (1988, 1994) illustrated the complexity of the links between 
collective and aggregate groups, gender groups, within- and between-group 
differentiation on common and specific dimensions. His data supported the 


Social representations in personal identity 
21 
general hypothesis of a homology between the representations of groups 
which are apparently not linked. Groups of men, of original and singular 
personalities, of competitors, on the one hand, and groups of women, 
individuals with similar personalities and cooperators, on the other, are 
respectively associated with respect to identity dynamics of their members. 
Social representations of groups, membership groups as well as other 
groups with which one interacts, intervene in actualizing personal identity. 
As social interactions often involve groups occupying asymmetrical posi­
tions in a representational field, one is easily led to the conclusion that 
identities of interacting individuals differ. This was not the starting point of 
my theorizing on personal identity as a social representation. Initially, I 
insisted on the similarity between the identities of members of different 
groups, such as autochthonous and immigrants. I will return to this point 
after discussing the relevance of the third way of studying anchoring in 
research on identity as a social representation. 
This approach to studying anchoring is still more directly concerned with 
the problem of differences between groups. Results presented by Durand 
Delvigne (1 992) can be used to illustrate this form of anchoring. She 
analysed self-descriptions of men and women, of different ages and social 
conditions, and found important similarities for different groups of 
respondents in factorial structures of answers to a French version of the 
sexual role inventory. Notwithstanding this similarity, a striking difference 
also characterized the structures of men and women, the opposition between 
the feminine and masculine poles being much stronger in the factorial 
structures of responses obtained from women than in those of men. Durand 
Delvigne summarized these results in the following way: 
This result is to be considered in relation with a hypothesis prevailing in relevant 
literature and stating that women more than men are referred to their social 
identity. Social models would be available for defining them, whereas men would 
be more easily individuated. For women, gender category would be a necessary 
marker of identity (Hurtig and Pichevin, 1986; Lorenzi-Cioldi, 1988). (Durand 
Delvigne, 1992, p. 1 82) 
Hence, differences exist in the organizing principles of male and female 
identities and they result in different self-descriptions. This is particularly 
true when comparing the self-descriptions of males and females of high and 
low social status. Masculinity scores are high for individuals of higher social 
status, independent of their gender, and femininity scores are especially low 
for men with a higher status. Androgyny scores are high for all groups of 
higher status independent of gender. Considering these results together with 
many others that are accumulating in the area of gender studies, one cannot 
but subscribe to Durand Delvigne's thesis: 
Gender, as identity, is not inherent to a sex group, but refers to social status. 
Otherwise said, gender orientation of women is related to the opportunities they 


22 
Social identity: international perspectives 
do have or do not have to overcome the social definition of themselves. In this 
case, and related to identity, gender is not a dynamic set of attributes produced by 
sexual category membership. It is a social marker linked with the asymmetric 
position of dominant and dominated individuals. (Durand Delvigne, 1 992, p. 223) 
Gender opposition does not represent sex opposition. It is a social con­
struction of hierarchical relationships linked with sex group reports and 
modulating personal identity through social differentiation processes. 
Nevertheless, the existence of similarities of personal identities across 
group boundaries is real. This apparent similarity may correspond to 
within-group differences along lines that were not analysed in the studies 
reported previously. Another explanation for the repeated findings of 
between-group similarities in identity research is suggested by some findings 
of Lorenzi-Cioldi (1988) and Amancio (1989), who report that identity 
processes are differently actualized as a function of intergroup encounter 
situations and of available comparison dimensions. In line with such 
findings, one can expect that a request to describe oneself on scales without 
explicit gender connotation or without any intergroup comparison will 
induce descriptions that privilege the individual as abstracted from his or 
her group memberships. Otherwise stated, it is possible that questionnaire 
instructions convey social representations of the 'collection' type. Depriving 
the individual of his or her representations of membership in 'aggregate' 
groups focuses his or her attention on representations of self as an indi­
vidual in search of originality and distinction. But such a search itself is 
highly normative and results in rather similar identity descriptions. 
The predominance of such normative self-definitions should not prevent 
us from studying less dominant ones. Reicher and Emler (1986) investigated 
such identities in their work on the origin of adolescent delinquency. 
According to them, behaviours such as vandalism, theft or physical 
aggressions are more often admitted by early to mid-teens than by any other 
age-group. There is evidence that such self-report scores are related to the 
actual level of involvement in delinquent activities (Singh, 1979). More 
importantly, such self-reports are analysed by Reicher and Emler (1986) as 
part of a reputation-building strategy. At this age life chances are con­
sidered to be settled by one's educational success or failure, and a young 
person's identity is characterized by its relation to his/her school career. 
Negative attitudes towards school authority are not infrequent and are 
generalized to other forms of institutionalized authorities, such as police and 
legal authorities. This means that important aspects of the social contract, 
namely that authorities will protect individuals from anarchy and redress 
violations of rights, are no longer considered valid: 
Thus delinquency can be seen in part as a strategy for coping with the lack of 
institutional protection. In the first place, delinquency acts as redress. It resolves 
grievances through attacks upon enemies, upon the police force and upon school 
and public property. Second, it offers its own kind of protection through the kind 
of reputation it creates. (Reicher & Emler, 1986, pp. 30-31) 


Social representations in personal identity 
23 
A self-presentation and social reputation is actively promoted in order to 
signify to others that one is the kind of person who stands up for one's 
rights. Another objective of the social construction of such a representation 
is to obtain support from others sustaining the same oppositional identity 
(see also Emler & Reicher, 1 995). 
Conclusion: A definition of personal identity 
Personal identity can be considered to be a social representation, an organ­
izing principle of individual positioning in a field of symbolic relationships 
between individuals and groups. This positioning is related to pre-existing 
insertions in a network of social relationships. 
Not all social relationships in which we take part are equally present in 
our minds. Depending on situations, various relationships become salient 
and give rise to different identity dynamics. Many self-descriptions only 
actualize general societal norms inducing individuals to define themselves in 
relation to each other on consensual dimensions, such as physical appear­
ance for youth. Other situations actualize social representations more 
directly anchored in group and category memberships. Memberships inter­
vene more intensively in conditions that are now well studied (Fiske & 
Neuberg, 1 990; McGuire & McGuire, 1988; Pavelchak, 1 989). Studying 
personal identity as a social representation should contribute to a better 
understanding of identity processes. 


3 
The Personal-Social Interplay: 
Social-Cognitive Prospects on Identity 
and Self-Others Comparison 
Carmencita Serino 
Introduction: 'Personal' and 'social' in social identity theory 
In his analysis of social categorization, Tajfel describes intra-categorial 
assimilation and inter-categorial differentiation as two directly related and 
simultaneously occurring processes. The minimal group paradigm (for 
example, Billig & Tajfel, 1 973; Tajfel, Flament, Billig & Bundy, 1 971) is 
intended to show that these effects usually occur even when group member­
ship is an extremely abstract and irrelevant experience for the subjects. In 
most of these experiments, the assimilation/differentiation processes manifest 
themselves by means of some form of favouritism: the mere categorization 
into different groups would elicit subjects' 'behaviours' (that is, giving 
money, rewards, and so on) which are clearly intended to favour their own 
group and to enhance the distance between ingroup and outgroup. 
Indeed, Tajfel aimed at providing a purely cognitive basis for intergroup 
differentiation (Hewstone, Hantzi & Johnston, 1 991). On the other hand, he 
also pointed out the existence of a collective level of reality which cannot be 
reduced to the sum of merely individual phenomena: under certain circum­
stances, individuals may act essentially as members of a particular group, 
even independently of their strictly individual tendencies or characteristics. 
Social identity theory (SIT) was later developed by Tajfel and Turner 
along these lines (for example, Tajfel & Turner, 1 979; Turner, 1 978). 
According to Tajfel ( l978b), 'at least in our kinds of societies, an individual 
strives to achieve a satisfactory concept or image of himself'. This 
individual will show, then, a tendency either to confirm or to change his/her 
condition of group member, depending on whether a particular group can 
contribute (or not) to the positive aspects of his/her identity. In other words, 
when his/her need for self-positive evaluation is not complied with, an 
individual will tend to leave the group (unless some objective conditions or 
particular values and norms interfere with this tendency). 
We may notice that, despite the original orientation towards a purely 
cognitive approach, motivational factors are mostly put forward by the SI 
theorists. A true 'axiom related to the individuals' need for positive self-


The personal-social interplay 
25 
evaluation' (Deschamps, Clemence & Roux, 1983) might in fact underlie 
their assumptions. 
Indeed, the idea according to which a fundamental need for self-positivity 
is at the basis of social identity and group membership could be easily 
verified in short-lived groups (such as the 'minimal' ones), wherein joining a 
group is a rather extrinsic and fortuitous (if not totally exploitable) experi­
ence. By contrast, when real social groups are involved, individual and 
social identities appear to be much more intertwined. Indeed, belonging to 
particular social groups is in general a deep-rooted aspect of persons' 
identity, a real 'way of being', which also affects the ways in which all social 
knowledges are filtered and reconstructed. ' As we will argue below, 
belonging to some particular social groups may even be the necessary 
condition allowing subjects to develop a sharper and better defined self­
image, to perceive themselves as being quite specific and highly distinctive 
individuals. 
In this vein, Tajfel and Turner's approach to the personal-social inter­
play is worth further analysis. Their idea of 'continuum' fails adequately to 
represent the dynamic connection of individual and collective processes. The 
personal-social continuum seems to imply an essential, unsolved dichotomy 
between individual and collective processes, personal and social identity, 
self
/others and intergroup comparisons. 
Alternative views on the personal-social continuum 
The close interdependence of intra-psychic and social psychological pro­
cesses in identity formation is recognized by Zavalloni (Zavalloni, 1973, 
1983; Zavalloni & Louis-Guerin, 1984). Her research suggests: (a) that even 
negative stereotypes about the ingroup are possible: these do not weaken, 
however, the individual feeling of being a group member; (b) that, on the 
other hand, the strong identification with a group does not exclude the self/ 
others differentiation. The intra-categorial differences can be as marked as 
the inter-categorial ones. Individuals can make a fine-grained distinction 
between the characteristics of their own group and the degree to which these 
characteristics refer to themselves (Serino, 1988a; Zavalloni, 1 983). Thus, 
the conditions in which group membership is made salient do not lead, 
necessarily, to an intra-categorial assimilation process. 
On the whole, the idea of a necessary and positive relationship between 
intra-categorial assimilation and inter-categorial differentiation was not 
always completely confirmed (cf. Eiser, 1983; Tajfel & Wilkes, 1963). 
By contrast, some experiments (Deschamps, 1982a) suggest a different 
hypothesis. In these studies, the greater the social identification, the greater 
is also the self/others differentiation occurring inside the membership group. 
On the other hand, when the self/others differentiation is emphasized, the 
intergroup differentiation is also found to become stronger. In other words, 
intragroup and intergroup differentiations do not appear as being 


26 
Social identity: international pers
pectives 
necessarily located at the opposite poles of a continuum. Sometimes, they 
appear as being consistent, simultaneous judgements which can 'vary 
together'. This is usually referred to as the 'covariation hypothesis'. 
The 'covariation' model was developed in the frame of a greater attention 
to comparison processes occurring in 'natural' sets, where the different 
groups are often characterized by a different social status. Natural groups 
and their functioning already represented a matter of central concern in 
Tajfel's perspective (see for instance, Tajfel, 1 982b). Yet, in his research, 
prominent attention was given to the minimal experimental conditions of 
social categorization. 
It was mostly Doise's model of the 'categorial differentiation' which led 
to a systematic employment of natural groups (for example, males/females; 
high school students/apprentices) in the experiments (Doise, 1 976). This 
yielded a special focus on the asymmetrical functioning of categorial differ­
entiation, due to the place taken by different groups in our society. In this 
vein, Deschamps ( l982b) focused on the relationships between 'dominant' 
and 'dominated' social groups. According to him, subjects belonging either 
to a 'dominant' social group or to a 'dominated' one would show important 
differences in perceiving intergroup similarities and in 'favouring' behav­
iours, as well as in the ways of representing themselves. Social identity 
would vary as a function of the 'symbolic and material capital' available to 
different subjects. In particular, the members of a 'dominant' social group 
would show a greater tendency to perceive themselves as highly distinctive 
individuals endowed with unique idiosyncratic traits. The members of a 
'dominated' group, instead, would usually show a tendency to perceive 
themselves as somewhat undifferentiated components of a 'collective object' 
(Deschamps, 1 980; Lorenzi-Cioldi, 1 988). 
In a study already mentioned above, Zavalloni suggested that there may 
be deeply different identities, depending on the social status of the mem­
bership group. In her view, a meaningful example is offered by the asym­
metrical development of men's and women's identities. The two genders are 
seen as the 'most basic identity groups, not only because they were the first 
to be learned . . . but also because they have traditionally englobed all 
others' (Zavalloni, 1 983, p. 2 1 5), thus becoming the basis of a social 
hierarchy codified in the culture. The dynamic interplay of men's and 
women's identities implies a power relationship, analysed with reference to 
its most subtle and elusive aspects. Language itself, for instance, reveals that 
only 'male categories' are usually associated with the positive cultural 
connotations of the concept of humanity (Zavalloni, 1 983, p. 217). 
There is an essential convergence in these different studies. Although 
adopting different methods and starting points, they emphasize the close 
interdependence between comparison processes, self-representations and 
individuals' concrete positions in a hierarchical social context. They suggest 
that the different levels of self/others comparison (namely intraindividual, 
interindividual, intergroup levels) are closely articulated with each other 
(Doise, 1 988). 


The personal-social interplay 
27 
In this view, Deschamps's studies on 'covariation' and those on the 
comparison between 'dominant' and 'dominated' social groups should be 
considered as a whole. The covariation hypothesis suggests that, instead of 
being opposed along a continuum, the personal 'pole' and the social one 
can be seen as 'similar' (Doise, 1 988, p. 1 05). Yet, in our opinion, this 
'similarity' is not to be understood merely as a form of isomorphism, that is, 
as a 'structural homology' between interpersonal and intergroup com­
parisons. A deeper and more concrete connection between different levels 
and aspects of self/others comparison can be underlined. This rests on the 
fact that social conditions may affect not only intergroup comparisons but 
also the forms of individuals' membership, as well as the structure and 
contents of personal identity. The links between interpersonal and inter­
group comparisons can be further analysed along these lines. 
We will briefly consider now two main questions singled out in this 
section. These refer, respectively, (a) to the relationships between similarity 
and difference and (b) to the interplay of cognitive and social processes in 
social comparison. 
Similarity relationships and self/others comparison 
Some of the studies mentioned above suggest that there may be several 
ways of connecting similarity and difference with each other in social 
comparison. This point can be further highlighted by referring to Codol's 
experiments on the 'superior conformity of self' behaviour (Codol, 1975). 
This is also described as the primus inter pares (PIP) eff
ect: the subjects of 
Codol's experiments, set in a highly normative situation, tended to present 
themselves as conforming, better than the others, to the shard norms. 
Actually, they based their own self-distinctiveness on the claim for a 'simi­
larity', deemed to be a value in this case. Considering these effects, we may 
notice that an entangled co-occurrence of similarity and difference appears 
at the intra-categorial level. This being similar and different at one and the 
same time (Codol, 1 982) sheds light on the dynamic interplay (if not, on the 
very 'genetic' relationship: see, for instance, Luria, 1 976) between personal 
and social identity. It is only by starting from some adhesion to a group, to 
its norms and values (that is, by starting from a certain conformity) that it 
becomes possible to develop the very idea of one's own uniqueness, and to 
set out an image of oneself as a highly distinctive individual. 
Further insights into the complex structure of the self/others comparison 
can be derived from a particular theoretical approach: that singled out by 
Tversky (1977) in his analysis of similarity. Tversky goes beyond the 
traditional 'metrical' approach, wherein similarity and difference are 
usually described just as proximity/distance between points along a linear 
dimension. He opposes this view with a 'non-metrical' approach: in this 
case, similarity is understood as a global judgement (that is, a J
eature­
matching process), taking into account both common and distinctive 


28 
Social identity: international perspectives 
features of the objects to be compared. This author demonstrated that 
similarity judgements usually take a specific direction. In general, one of 
the two objects is in fact 'prominent' (that is, more 'salient' or 'proto­
typical' than the other). It then becomes the reference point (the 'model') 
which the other term (that is, the 'variant') is usually referred and 
compared to. By systematical variations of the comparison reference point, 
Tversky showed that perceived similarity is higher in this standard form of 
comparison than in the inverse condition (that is, when the model is judged 
in reference to the variant). 
Following these studies on the 'asymmetry phenomena', several authors 
attempted to verify whether (and in what conditions) the self is a major 
reference point in interpersonal comparison (Holyoak & Gordon, 1 983; 
Srull & Gaelik, 1 983). Codol (1984b, 1 987)2 suggested that interpersonal 
comparison too is characterized by a cognitive asymmetry, depending on 
the different 'salience' of the terms (self, other) under consideration. Codol's 
findings show that, in general, perceived similarity is higher when the others 
are compared to the self than in the inverse condition. In practice, it is not 
the same to state: 'I am similar to the others' as to state: 'the others are 
similar to me'. A person would tend to claim others' similarity ('the others 
are similar to me more than 1 am similar to the others') as well as hislher 
own difference ('I am different from the others more than the others are 
different from me'). 
Starting from these results, several meanings of similarity and difference 
can be pointed out (Serino, 1 988b): similarity may mean 'conformity', but 
also 'representativeness'; difference may mean 'specificity' but it may refer 
to 'deviance' as well. These different meanings are the result of a social 
construction and can be 'recoded' depending on the comparison direction 
(Serino, 1 989). 
A 'non-metrical approach' to similarity may yield important conse­
quences for the analysis of social categorization. One of these consequences 
could be the greater attention to the 'productive' aspect of categorization. In 
fact, a feature-matching process involves a number of questions concerning 
the ways in which similarity judgement is performed whenever a few simple 
(and generally a priori provided) dimensions are not available as yardsticks 
for comparison. The multidimensional character of comparison is then 
emphasized: in this light, similarity and difference may even appear as 
compatible and co-occurring judgements, rather than inversely related and 
opposed along a continuum. 
A 'metrical approach' to similarity? 
The way in which concepts are operationalized may affect the way in which 
particular questions are focused on. Thus, when considering social identity 
theory, we may observe that (a) by focusing, mainly, on 'favouritism', it 
appears as shifting to a motivational analysis of group membership, and 


The personal-social interplay 
29 
that (b) by considering personal and social identity as inversely related 
levels, the true meaning and the subjective relevance of an identity group for 
the individuals risk being somewhat neglected. 
Starting from Tversky's analysis, we might even observe that Tajfel's 
approach to similarity seems to lie on some 'metrical' assumptions. Right 
from his earliest studies in this field (Tajfel, 1 959; Tajfel & Wilkes, 1 963), 
some similarity/difference judgements were inf
erred, being based on the 
perception of some 'proximity' or 'distance' between points along dimen­
sions (such as weight or length) already provided by the experimenter. The 
authors were quite aware that this procedure needed to be further 
controlled. Wilkes and Tajfel (1966) could in fact demonstrate that there 
exists a connection between the 'distance of an actor from the series of 
stimuli and its categorization as the "same" or "different from" the series' 
(Tajfel, 1 98 1 ,  p. 92). Yet, these experiments cannot explain how similarity 
judgements are performed whenever comparison is not based on a few 
simple units or dimensions which are already available and/or directly 
provided to perceivers. 
Some general issues of research on social categorization might be traced 
back to this 'metrical' view on similarity. For instance: 
(a) special attention is paid to the 'deductive' face of categorization, that 
is, to its consequences: categories are understood as a condition, at the 
origin of the categorizing activity, more than as a product of this 
activity; 
(b) judgement is split up into as many similarity relationships as there are 
comparison dimensions; 
(c) similarity and difference appear as being two opposite and mutually 
excluding forms of relationship between the terms to be compared. 
The same characteristics are also shared by social identity theory, as IS 
highlighted by the following SIT assumptions: 
As a consequence of the outstanding categorization (whether 'inter­
personal' or 'intergroup'), subjects will act either as mere individuals or 
as group members. 
2 
A more direct or explicit involvement of the self in the comparison 
would change the whole situation: the previous categorization would be 
cancelled and a different one would be introduced. Thus, the existing 
opposition between two groups is replaced by a more general (oneself
/ 
another) opposition, wherein membership of either category ends up 
becoming somehow irrelevant. 
3 
When an intergroup categorization is activated, the typical accentuation 
of intra-categorial (that is, self/others) similarity will be observed. Con­
versely, an accentuation of oneself/another difference is likely to occur 
only when intergroup categorization is not salient. 


30 
Social identity: international perspectives 
As a result, two main issues arise in the frame of social identity theory. 
First of all the close interdependence between personal and social aspects of 
identity is often overlooked. Second, self
/ingroup and self
/outgroup 
conditions are not sufficiently differentiated from each other. These are 
just described as two, more or less similar, examples of 'interpersonal' 
comparison. 
To summarize, the continuum model, based on a necessary co-occurrence 
of intra-categorial assimilation and inter-categorial dif
f
erentiation, is not the 
only way of describing the relationships between personal and social 
identity. In our view, even the analysis of 'covariation' could be further 
developed. Not only can intra- and intergroup comparisons appear to be 
'similar' rather than 'opposite' poles (as in Deschamps's experiments show­
ing that intra- and intergroup similarities can 'go together'). As a matter of 
fact, a special form of 'covariation' also occurs when the self takes a 
prominent (prototypical) role, thus appearing, simultaneously, as being 
both similar to and dif
f
erent f
rom the other ingroup members. Here, too, 
similarity and difference are linked to each other: yet, in this case, they 'co­
vary' in a quite unexpected way. 
The relationships between similarity and difference are not totally 
explained by an 'all or none' approach (Deschamps & Serino, 1 996), and 
this may yield some consequences for identity research. 
Self-categorization theory 
More recent developments of Turner's analysis were aimed at dealing with 
some of the problems pointed out above. Although an essential continuity is 
affirmed between social identity theory and Turner's self-categorization 
theory, nonetheless some important novelties are also introduced. In par­
ticular: (a) the attention given to motivational factors is reduced; (b) the 
'productive' face of categorization is emphasized; (c) similarity and 
difference are more closely connected with each other. 
Of course, an exhaustive description of this theory exceeds the scope of 
our chapter. Thus, we will just illustrate, briefly, the three points stated 
above. 
Focus on cognitive processes 
Turner integrates the prototypicality theory with the analysis of social 
categorization. 
The prototype concept was singled out by Rosch (1973, 1 975, 1978) in the 
frame of her analysis of categories as 'fuzzy' sets. According to Rosch, 
class-inclusion can hardly be conceived as a rigid all-or-none process. In 
most cases, in fact, classification appears as a probabilistic process. More or 
less typical examples are gathered together in the same category; therefore, 


The personal-social interplay 
3 1  
categories take a hierarchical structure, being organized around a central 
(prototypical) element. Categories are organized, in turn, into taxonomies 
(that is, hierarchical categorial systems), with different levels of abstraction 
and inclusiveness (Rosch, 1 978). 
Turner draws on these ideas when developing his self-categorization 
theory. According to him, the self is (also) a cognitive structure, and self­
categorizations can be described as 'a part of a hierarchical system of 
classification' (Turner, 1 987, pp. 44-45). In this view, 'all social com­
parisons with others depend upon the categorization of the others as part of 
a self-category at some level of abstraction' (Turner, 1 987, p. 48). 
Conforming to Rosch, the self-categorization process gives way to a 
number of superordinate and subordinate self-categories, with increasing 
levels of abstraction and inclusiveness. Turner explains this point by singling 
out three different levels, describing 'one's human, social and personal 
identity respectively'. A very abstract and inclusive level of self­
categorization is the one opposing 'human beings' to other forms of life 
(inter-species comparison). At an intermediate level (intergroup compar­
ison), the self-concept describes the individual as a member of particular 
social groups. The most concrete and least inclusive level of categorization 
would be based on interpersonal comparisons, wherein a person is 
distinguished from the others by means of hislher particular personality, of 
hislher unique and strictly individual characteristics. 
The productive f
ace o
f
 categorization 
The 'productive' character of social categorization is put forward by 
focusing on those processes directed to select, among different levels of 
abstraction/inclusiveness, the more pertinent ones in the given comparison 
conditions. For this reason, the question of salience is particularly 
emphasized (Oakes, 1 987; Oakes & Turner, 1 986). According to Turner and 
his co-workers, the salience of a given categorization can hardly be 
described as the effect of an automatic perceptual prominence of certain 
stimuli. Instead, it depends on the interaction between accessibility and fit 
(see Bruner, 1 957). Accessibility deals with the degree of readiness with 
which certain pieces of information are activated, while fit 'ties 
categorization firmly to reality' (Oakes, Turner & Haslam, 1991) in that 
it refers to an adequate and veridical representation of social situation. In 
this view, the salience of a given categorization is highly dependent on 
context, as well as on perceivers' purposes and goals. 
Unlike Rosch, Turner suggests that the basic categorization level (that is, 
the level at which the best 'compromise' between information generality and 
concreteness is achieved) is a 'highly contextual variable rather than fixed' 
(Turner, 1987). 
Cantor and Mischel (1979) already emphasized that the relation­
ships between different levels of abstraction is flexible rather than rigid in 


32 
Social identity: international perspectives 
categorial systems concerning social reality. These authors too applied 
prototypicality theory to the analysis of persons' perception, focusing on 
context and perceivers' purposes as crucial aspects of classification. They 
pointed out a number of analogies, but also some important differences in 
the ways categorial judgements are performed in the social versus non-social 
domain (see also Holyoak & Gordon, 1 984; Wyer & Podeschi, 1 978). For 
instance, the different levels of abstraction seem to be 'interchangeable' 
(Cantor & Mischel, 1 979) in the social domain. Moreover, class-inclusion is 
mostly based on a bipolar opposition, that is, on the incompatibility 
between different categories, rather than on merely descriptive typologies 
(Cantor & Mischel, 1979, p. 33). These differences between social and non­
social categorial systems emphasize the evaluative character of classification 
in the social domain. Elsewhere (Serino, 1990) we addressed this question 
by referring to Billig's (1985) analysis of 'particularization'. In a well-known 
example, Billig describes the process by which one general category (for 
example, Christianity) is split up into two distinct sub-categories (black and 
white Christians). Notice that the new distinction does not describe only the 
shift from a wide, superordinate category to less abstract and inclusive ones. 
The difference between black and white christians, in fact, may also entail a 
basic opposition between a lay and primitive religiousness and a deep and 
mature one. From this viewpoint, particularization yields an evaluative 
discrimination allowing the new ingroup to take its place at the top of an 
extremely general hierarchy (that opposing 'true' religiousness to super­
stition). The reference to a value system entails both particularization and 
generalization, as two interrelated aspects of categorizing. The reference to 
a value system also explains why, although the level of abstraction is not 
immutable, the 'inclusiveness of the category remains stable' (Cantor & 
Mischel, 1 979, p. 1 6). In fact, even the introduction of a very abstract level 
of classification can be aimed at enhancing interpersonal or intergroup 
differentiation! 
Our discussion converges on Turner's position. According to him, 
categorization and particularization are two interdependent aspects of the 
same process (Turner, 1 988). Turner's description of 'normative fit' as an 
important component of salience also emphasizes, as we do, the social 
meaning of categorization. He argues, for instance, that categorizing some­
one as 'Catholic' is not only the result of a comparison with other categories 
(cf. Oakes et aI., 1 991); it also deals with behavioural and normative 
contents of that person's identity. 
Indeed, social categories are highly meaningful and personally relevant 
for individuals. This might account for 'people's tendencies to perceive 
themselves and others in terms of psychologically opposing forces' (Cantor 
& Mischel, 1 979) which seem to affect the judgements of prototypicality in 
the social domain. In this light, proto 
typicality itself might carry a par­
ticular meaning: when social categories are involved, most 'prototypical' 
features would be in fact those endowed with a strong (positive or negative) 
value. 


The personal-social inter
play 
33 
Similarity and dif
f
erence as interdependent judgements 
In self-categorization theory, similarity and difference are linked to each 
other in a more complex fashion. Difference always implies a 'higher level 
identity', so that no comparative relation in terms of difference could be 
assessed between two stimuli if they were not suitable to be included in 
the 'same category' at a higher level of abstraction (Turner, 1 987). Accord­
ing to Turner, categorization follows the meta-contrast principle. This 
entails an inverse relationship between intra- and inter-categorial com­
parisons.3 Therefore, it is consistent with the idea that the perception of 
intra-categorial similarities and inter-categorial differences is at the basis 
of categories formation and functioning. In this view, the salience of a given 
category and even the prototypicality of its prominent examples can be 
understood as a function of the 
meta-contrast perceived between intergroup and intragroup differences. As the 
differences perceived between individuals belonging to different (implicit) groups 
increases, compared to the differences between individuals belonging to the same 
(implicit) groups, then the more salient will the social categorization become . . .  
i.e. members will tend to be perceived less as individually different persons and 
more as unitary social groups. (Oakes et aI., 1991, pp. 126-127) 
Turner (1987, p. 49) claims there is a 'functional antagonism between the 
salience of one level of self-categorization and other levels', assuming that 
'there tends to be an inverse relationship between the salience of the 
personal and of the social levels'. In this sense, Turner clearly draws on 
Tajfel's idea of 'continuum'. 
Prototypes: similar and different 
The meta-contrast ratio also provides a simple operational description of 
the prototype concept. As Turner suggests, a given stimulus will tend to be 
perceived as more exemplary or representative of the category as a whole 
the higher the meta-contrast ratio: the higher the mean perceived difference 
between the target stimulus and the out-group members over the mean 
perceived difference between the stimulus and other ingroup members, the 
more prototypical the target stimulus. 
Yet, by referring to the meta-contrast, only the comparative aspect (the 
'horizontal' dimension) of prototypicality is emphasized. Being aware of 
this, Turner himself recently focused on some aspects of the prototype 
which cannot be described in this way. 
In our view, the prototype concept also contains the idea that 
similarity and difference 'go together'. A prototype in fact differs from 
the other items because it better represents an abstract image or an ideal 
of the common category; an ideal which allows that category to be 


34 
Social identity: international pers
pectives 
defined and distinguished in turn from the other categories. Similarity 
and difference, conformity and uniqueness, are tightly intertwined in the 
prototype concept. 
This close connection of seemingly opposite judgements can be observed 
also when referring to the self-concept. On the whole, Codol's studies 
mentioned above seem to suggest that the self may act as a sort of proto­
type (Hardoin & Codol, 1 984) in interpersonal perception. By varying the 
order of two different description tasks (self-description, description of 
another stimulus-person), Codol (1987) also shows that common and dis­
tinctive traits carry a different weight across different conditions. Subjects 
starting with the self-description task used the same self-descriptive traits for 
describing the others. By contrast, when the other-description came first, 
subjects tended to use quite different traits in the subsequent description of 
themselves. (Different results were obtained when the other person was the 
subject's 'best friend'.) According to Codol, these results would suggest that, 
in general, 'the others' are assumed to belong to the same category as the 
self, whilst the reverse is not true. 
More generally, there would be a dynamic interplay between similarity 
and classification. As Tversky (1977) argued, similarity can determine the 
way in which objects are classified, but it also depends on the adopted 
classification. 
While developing his analysis of salience, Tversky suggests that each 
pertinent feature is weighted according to both its 'intensity' and its 'diag­
nostic value' (this refers to the 'kind of classification' introduced by that 
feature). Of course, these two different components are assumed to be both 
present and tightly intertwined in any matching process. Nonetheless, we 
may easily imagine some conditions where either intensity or diagnosticity 
is, in turn, particularly emphasized. 
In some cases, when two objects are to be compared, they are already 
assumed to belong to the same category. Comparison is just aimed at 
singling out the reasons for an overall accepted similarity. When a common 
category is involved (even implicitly), comparison might give rise, mainly, 
to the (differential) intensity of the shared (categorial) features. 
In other conditions, similarity is the final result of a more inductive 
process. Here, judges are searching for an appropriate category including 
both objects (thus making the comparison possible): greater attention would 
be paid, then, to the simple presence/absence of several (either common or 
distinctive) features, whose 'diagnostic value' is foregrounded. 
When considering in this light Turner's analysis of the different self­
categorization levels, it seems that similarities and differences are judged 
mainly by referring to the diagnostic value of particular traits. Thus, at the 
interpersonal level, individuals are deemed to make comparisons only on 
the basis of their unique, exclusive personality traits. These are obviously 
opposed to other traits (which are not recognized as pertaining to them­
selves) in a kind of judgement giving rise, mainly, to the simple presence/ 
absence of the critical features. 


The personal-social interplay 
35 
However, we are dealing here with an abstract idea of the individual. 
Indeed, when self and others are perceived in terms of 'psychological 
opposing forces', little attention will be paid to common characteristics. 
Nonetheless, interpersonal comparisons are often performed in a frame of 
already accepted common categories, relevant to each member and to his/ 
her sense of personal identity. In these conditions, greater attention would 
be paid to some characteristics which are present, to a different degree, in 
the different (more or less representative) ingroup members. 
In this way, we can also point out an important difference between self
/ 
ingroup and self/outgroup comparisons: these are conditions where a 
different focus on either intensity or diagnostic value of the relevant features 
might be observed. 
Research on intergroup relations already demonstrated that the ingroup 
members are perceived as varying along specific dimensions more than the 
members of an outgroup are (Park & Rothbart, 1 982); actually, it seems 
that 'ingroupers' are compared in terms of different intensity of the same 
features, while 'outgroupers' are not. 
At the intragroup level, the self/others comparison might be based on the 
different intensity of common characteristics, more than on the different 
salience of features unique to each individual. Again, some aspects of the 
personal-social interplay seem to escape an explanation in terms of 'con­
tinuum', and we can imagine an interpersonal level of self-categorization 
where some shared social characteristics do not stop being salient and 
important. 
In most cases, a special form of particularization is likely to occur at the 
intragroup level: it would be based on dimensional discrimination about 
some crucial categorial characteristics which are shared, to a different 
extent, by oneself and by the other ingroup members. This process seems to 
have been described, for instance, in a study by Montei! ( 1992): this shows a 
particular form of 'individualization', based on the individual appropriation 
of the positive characteristics of the group. 
On the other hand, a generalization process is required when self and 
other are (perceived as being) quite different. 'Similarity' can be assessed, in 
this case, only by shifting to a very inclusive category (see also Turner, 
1 987). At this level, people may share, at least, the fact of being 'indi­
viduals', endowed with their own personality. Thus, personal traits, which 
are usually considered as being the most genuine and concrete identity 
issues, would be typically elicited in some very general and abstract inter­
personal conditions. 
Starting from the prototype concept, personal identity and intragroup 
relations appear as an amazing, intertwined connection of similarity and 
difference, unicity and conformity, assimilation and distinctiveness (Codol, 
1 982). It would be hard to reduce this complex interplay to the frame of a 
personal-social bipolar opposition. In this light, what about the 'functional 
antagonism' defended by Turner? Is it possible to reconsider the same idea 
in a diff
erent perspective? 


36 
Social identity: international perspectives 
'Functional antagonism': different levels of categorization or different 
modes of processing information? 
According to Turner, different contents become salient at different (intra­
group, intergroup) levels of categorization. His idea of the functional 
antagonism between personal and social levels of self-categorization also 
entails the opposition of personal versus social features. This dichotomy can 
hardly be overcome just be postulating, as Turner does, a 'midpoint' of the 
personal-social continuum 'where the self-perception is likely to be located 
much of the time' (Turner, 1 987). 
We would refer to a different form of functional antagonism, where 
different modes of processing information are applied to a substantially 
unitary social reality. Cognitive research on person perception has pointed 
out a fundamental opposition between categorial and personalized 
information, top-down and bottom-up ('stimulus-related') processes (for 
example, Asch & Zuckier, 1 984; Fiske & Pavelchak, 1 986; Pratto & Bargh, 
1991). 
Brewer (1988), for instance, proposed a 'dual model', wherein which route 
is chosen depends, mainly, on the perceiver's purposes and goals. Similarly to 
Turner, Brewer suggests that 'the level of abstraction at which new 
information is initially encoded is expected to be different for category-based 
and person-based processing'. It is worth emphasizing, however, that exactly 
the same information 'can be processed in a top-down (de-personalized) 
manner (resulting in category-based cognitions), or in a bottom-up (per­
sonalized) fashion, that results in person-based representations'. Thus, 
'personalization refers to a different kind of cognitive structure in which the 
individual becomes the basis for organization of all relevant information'. 
When the personalized view is activated, perceivers are dealing with 
salient information which can hardly be reduced to strictly individual 
properties. Personalization does not lead, necessarily, to focusing on 
particular contents or on isolated individuals. Self-involvement may also be 
linked to the increasing awareness of a number of complex interactions. 
In other words, personalization could give way to an extremely sensitive 
appraisal of all pertinent variables and of their concomitant variations. This 
was suggested by a recent study (Serino & Mazzara, 1 995), bearing on 
persons' perception in particular social contexts. Subjects (high school 
students) were presented with a slide showing either a class-room (C) or a 
fast-food restaurant (F) and an ambiguous figure entering the situation. 
This figure was described either as 'a person' (N) or as a 'handicapped 
person' (H). Among other tasks, subjects had to describe themselves and the 
stimulus-person by means of a list of 1 5  bipolar dimensions. The order of 
these descriptions was systematically varied: half the subjects started by the 
self-description, later describing the other person; the other half described 
the stimulus-person first, while the self-description was performed later. On 
the whole, three independent variables were then introduced: stimulus­
person (NIH); situation (C/F) and re
f
erence point (either 'oneself' or 'the 


The personal-social interplay 
37 
other person', assuming that the term described first also serves as a 'model' 
in the subsequent description). 
According to our results, the description of the other person was mostly 
based on categorial information (that is, whether that person was presented 
as being 'normal' or 'handicapped': F(1 5, 120) = 4.87; P = .001). The self­
description was performed in a quite different way. Different effects of the 
three independent variables are observed in reference to each particular self­
descriptive trait, so that, for instance, some aspects of the self-description 
seem to be affected by the situation, other aspects depend mainly on the 
re
f
erence point, while, in other cases, the interaction of all three independent 
variables yields significant effects.4 
In short, the self-description seemed to be more mutable, supple and 
'personalized'. The special nature of this 'personalization' is worth empha­
sizing, however. It does not lie only on the reference to personal traits 
independent of the context. Personalization might also spring from an 
extremely fine-grained analysis of all variables involved in a given social 
context. 
Asymmetries and prototypes in intergroup relations 
In the previous paragraphs we challenged the idea that personal identity 
refers, necessarily, to an isolated individual and to hislher unique person­
ality traits. However, the close interdependence of personal and social 
identity can also be met from a different perspective. 
As we have already suggested, being a member of particular groups may 
deeply affect the individuals' experience, their ideas about social reality, 
others and themselves. In an intergroup context, usually characterized by 
several status differences, even self-concept, self
/others comparison and 
personal identity could be defined in different ways, depending on indi­
viduals' and groups' position in the social hierarchy. Thus, a number of new 
questions arise: What structure does the self
/other comparison take when 
the 'others' are ingroup or outgroup members, respectively? Do the nature 
and the social status of these different 'others' affect comparison modes and 
direction? What are prototypes in an intergroup frame? 
By referring to the asymmetry phenomena described above, we may even 
wonder whether there is any link or overlapping between the objective, 
social asymmetry of intergroup relations and the cognitive asymmetry of 
perceived similarity. By linking the cognitive asymmetry of similarity 
judgement to the structure of concrete intergroup relations, the close inter­
dependence between mental and social processes could be better highlighted 
(see also Boltanski, 1982). 
In a study by Hurtig, Pichevin and Piolat (1 992) this connection is 
highlighted. These authors focused on sex and age as factors affecting 
asymmetry phenomena. Their study centred on similarity judgements 
between pictures presenting two people of different sex and age. Indeed 


38 
Social identity: international perspectives 
these two variables were considered as defining different social categories, 
hierarchically related to each other (younger/older; male/female). Conform­
ing to the hypothesis, a strong asymmetry effect was found, due to the sex 
of the ref
erence point person. When the male photo was the reference point, 
the similarity perceived between the two stimulus-persons was higher. The 
authors argued that the male figure showed a great 'salience' (or 'proto­
typicality' degree) than the female figure. Moreover, when same-sex pairs 
were compared, similarity was perceived as being lower in male pairs than 
in female pairs: that is, the more salient (prototypical) stimulus also 
appeared as fitting a highly articulated and best differentiated social 
category. 
In the same vein, I focused on the self/other asymmetry occurring when 
different (advantaged/disadvantaged) social groups are involved. After a 
pilot-study, the powerful asymmetry between Northern and Southern 
Italians was taken as an example of relationships between 'dominant' and 
'dominated' social groups. The Northerners were considered as the upper, 
dominant group. 
In the first of these studies (Serino, 1 988b), 358 subjects belonging to the 
two groups (Northern and Southern Italians) had to perform a comparison 
between themselves and the Northerners or between themselves and the 
Southerners, later providing, by free explanation, the reasons for their 
previous judgement. The comparison reference point (oneself or others) was 
systematically varied, as in other experiments on the asymmetry effect.5 A 
2 x 2 x 2 design was set up, taking as independent (between subjects) 
variables the kind o
f
 judgement (in terms of similarity or difference), its 
content (comparison with the ingroup or with the outgroup) and the 
comparison re
f
erence point (oneself or the others, respectively). 
According to our results: 
A certain asymmetry (that is, an independent effect of the reference 
point variable) was observed. 
2 
This effect was found mainly in comparisons involving the group of 
higher status (that is, the Northerners). 
3 
Self
/ingroup and self
/outgroup comparisons tended to take an opposite 
'direction'. In the self
/ingroup comparison, the reference point was the 
group rather than the self (for example, 'I am similar to the Northerners 
more than the Northerners are similar to me'). An opposite (although 
non-significant) tendency appeared in comparisons with the outgroupers. 
On the whole, the different contents elicited in self
/ingroup and self
/ 
outgroup comparisons show that, at the intragroup level, subjects are com­
pared not only to other individuals, but also to a general norm. Similarity 
with the 'ingroupers' is based on a large amount of references to 'member­
ship'. Figure 3.1 suggests that group-belongingness is dramatically empha­
sized in the self
/ingroup condition. Thus, on the one hand, we can stress the 
difference between self
/ingroup and self/outgroup conditions; on the other 
hand we can observe that it is not a simple comparison between individuals 


The personal-social interplay 
1 20 
1 00 
• Self/ingroup 
D Self/outgroup 
80 
60 
40 
20 
0 
TR 
CU 
HIS 
BHV 
SEC 
MBSH D/ND 
OTH 
Key: 
TR = personality traits; CU = culture and habits; HIS = history; 
BHV = behaviours; SEC 
= social and economic conditions; 
NC 
MBSH = group membership; D/ND = difference and/or non-difference; 
OTH = other categories; NC = non-classified 
Figure 3.1 
Mean scores ( x  100) o
f
 each content category elicited in 
self/ingroup versus seljloutgroup similarity judgements (N = 121) 
39 
(just focusing on strictly idiosyncratic personal traits) that we observe at the 
intra-categorial level. 
The above-mentioned study did not provide decisive results. It suggested, 
however, that subjects belonging to the dominant group (that is, Northern­
ers) were more sensitive to the structure of the comparison: perceived 
similarity and difference varied as a function of the reference point (either 
self or group) even at the intragroup level. By contrast, Southern subjects 
appeared mostly concerned with the comparison content: in this case, sig­
nificant effects were due, mainly, to the 'type of comparison' variable (that 
is, whether the Northerners or the Southerners were involved). 
On the other hand, the free answers pointed out two different ways of 
referring to a common category. The Northerners employed 'we' mostly in 
relation to their own group (we, the Northern Italians). By contrast, the 
Southerners referred to a more inclusive category, describing the set formed 
by both groups (we all, Italians from the North and the South). Thus, it 
seems that the 'dominant' subjects focused mainly on their own category, 
while the 'dominated' subjects centred, mostly, on the relationshi
ps between 
the two categories (Northerners/Southerners). In a sense, our Northern 
subjects (that is, the 'dominant' ones) would given an example of 
comparison 
conditions wherein similarity and difference are activated simultaneously at 
the intra-categorial level. Conversely our Southern ('dominated') subjects 
would set out a kind of comparison where similarity and difference are 


40 
Social identity: international perspectives 
activated simultaneously at the inter-categorial level. A strong tendency to 
overcome the Northerners-Southerners opposition goes hand in hand with 
the clear awareness of the difference between the two categories (a difference 
recoded as unjustified discrimination, and then rejected). 
Some of the above results were worth further investigation. To this end 
new research on the same topic (Serino, 1 996) was carried out. A group of 
III subjects, equally divided into Northerners and Southerners,6 had to 
assess the perceived similarity between themselves and the Northerners or 
between themselves and the Southerners. The 2 x 2 x 2 design considered 
as (between subjects) independent variables: 
(a) the subjects' origin (either from the North or from the South); 
(b) the comparison content (comparison with the Northerners or with the 
Southerners); 
(c) the comparison re
f
erence point (either 'oneself' or 'the others'). 
Let us examine some results: 
The comparison content yields significant effects: greater similarity is 
assessed in comparisons with the Northerners than with the Southerners 
[F(l ,  101) = 7.76; p = .01]. This happens for all subjects, irrespective of 
their origin. All subjects underline that the Northerners (that is, the 
group of higher status) are 'more similar' to themselves. 
2 
On the whole, asymmetry phenomena are more marked in comparisons 
with the Northerners (tS4 = 2.80; P 
= .05) than with the Southerners 
(tSI = - l .65; n.s. ). This tendency may suggest that a group of high status 
also entail a more prominent, highly differentiated cognitive category. 
3 
Again, the subjects from the North and those from the South seem to 
construct their self
/others comparisons in different ways. It is mainly the 
subjects belonging to the group of high status (Northerners) who grant a 
prominent role to the self, which is taken as the comparison reference 
point. By contrast, the Southern subjects show a greater tendency to 
judge their own similarity in reference with 'the others'. 
4 The self
/ingroup comparisons appear to be more 'asymmetrical', which 
might suggest that ingroup refers, in any case, to a more differentiated 
and complex category than that describing the outgroup. This is true for 
all subjects, but, once again, the comparison direction is not the same 
for Northern and Southern respondents. For the Northerners, the self/ 
ingroup similarity is higher when it is judged in reference to themselves 
(t24 = 3.182; P = .004). They appear, then, as particularly centred on 
their own 'self' inside the membership group. By contrast, the Southern 
subjects tend to focus on their own similarity, by taking the 
'Southerners' (rather than 'themselves') as the comparison reference 
point ( t26 = -3.38; P = .01). 
On the whole, the different direction of these comparisons may suggest 
that different meanings of similarity are entailed by Northern and Southern 
subjects. In fact, the former seem to emphasize their own representativeness, 


The personal-social inter
play 
41 
while the latter appear as mostly concerned with their own conf
ormity to an 
external model. Cognitive asymmetry, which is more marked in those 
comparisons involving the group of higher status (that is, the Northerners), 
may shed light on the close connection between mental and social aspects of 
companson. 
Conclusion 
In this chapter, we have focused on personal and social identity by referring 
to the directional character of similarity (Tversky, 1 977), to the role of 
prototypes in person perception (Cantor & Mischel, 1 979; Rosch, 1978), 
and to the different ways of processing social information (Brewer, 1 988). 
We argued that the close relationship between individual and collective 
processes, personal and social dimensions, is not always suitably represented 
by the continuum model described in social identity theory. A dichotomic 
view of intra- and inter-categorial relations, of personal and social identity, 
still appear as the basis of Tajfel and Turner's research. More recent 
developments of the same perspective (such as Turner's self-categorization 
theory) integrate the theory under several important aspects. Yet that 
'dualistic conceptualization of macro/micro 
social relations' (Condor, 1990) 
which is typically expressed by the notion of 'continuum' does not appear to 
be completely overcome. 
When considering self/others comparisons performed in a concrete social 
context, it seems that (a) even personal identity is socially defined, 
depending on individuals' and groups' place in a given social hierarchy, and 
that (b) even categorial information can be processed in a 'personalized' 
manner. From this viewpoint, personal and social processes, individual and 
group identities, show their close connections. 
In short, the personal-social interplay is worth deeper analysis. Con­
cerning the self/others comparison, several strategies can be observed, where 
similarity and difference are linked to each other in unexpected ways which 
are beyond the classical connection of intra-categorial assimilation and 
inter-categorial differentiation. We explained this point by referring to the 
prototype concept. Prototypes are - at one and the same time - highly 
'conforming' (to the category standards) and clearly 'distinctive' (in refer­
ence to more peripheral categorial elements). A special interdependence of 
similarity and difference is then required. We have suggested that, in the 
social comparison, a similar close intricacy of similarity and difference often 
occurs at the intragroup level. Here, personal identity could be based not 
only on one's unique personality traits, but also on the different intensity of 
common characteristics, expressed to a different extent by oneself and by 
the other ingroup members. Thus, in some cases, different levels of self­
categorization could be characterized by different modes of judging (for 
example, intensity versus diagnostic value), more than by the emphasis on 
different contents. 


42 
Social identity: international perspectives 
The prototype concept implies a probabilistic view on class-inclusion and 
a non-metrical approach to similarity. We suggested that this non-metrical 
approach may yield some consequences for the analysis of social 
comparison, and we reconsidered some aspects of Tajfel's theory in this 
light. 
In the same vein, we referred to some studies on the asymmetry 
phenomena in similarity judgement, showing that these effects may also 
occur in the social domain. We must admit that it is not yet so clear what 
asymmetry really means when comparison involves such meaningful social 
'objects' as the self or an identity group. Nonetheless, these phenomena 
often reveal that some sophisticated strategies are activated in interpersonal 
perception and that respondents are sensitive to the context and to the 
(social, cultural) consequences of their judgement. From this viewpoint, 
asymmetry may even serve as an indirect measure of a certain stereotype, 
which operates independently of respondents' intention (Greenwald & 
Banaji, 1995; Serino, 1 996). 
The typical procedure of asymmetry research (based on systematic 
variations of the comparison reference point) would point out the different 
meanings of similarity and difference, which are often at the core of the 
interactions and 'negotiations' between different social groups. To illustrate 
this point, we referred to a few studies where asymmetry effects have been 
observed in similarity judgements performed by subjects belonging to 
groups of different status. Social status seems to affect the meaning of 
comparison, the ways in which self
/others similarity is assessed and the role 
of the self. 
From this point of view, asymmetry research may contribute to developing 
a typical social-cognitive level of analysis, wherein individual, interpersonal 
and intergroup processes are closely interrelated. Social categorization and 
similarity judgements are typical 'social-cognitive' processes: cognitive 
research can provide fruitful suggestions in order to understand their 
functioning. Yet the close articulation between mental and social processes 
still ought to be further highlighted. 
Notes 
I .  In this light, we may consider the extremely rich literature on social representations (for a 
review, see Jode1et, 1989). Intergroup differences in the ways of representing some particular 
aspects of social lif
e (such as mental illness) were highlighted in a series of studies carried out 
simultaneously in Italy and Spain (see Ayestanin, 1985; 8ellelli, 1987, 1994). 
2. The substantial continuity of Codol's research (from the discovery of the PIP effect to the 
analysis of its asymmetry phenomena in interpersonal comparison) is discussed in Serino 
(1992). 
3. Hogg and McGarty (1990) suggested that the meta-contrast ratio is more appropriate in 
simple unidimensional contexts than in more complex and cross-cutting social frames. 
4. For this reason, the self-description would hardly be explained by means of just one 
variable. Thus, we observe a slightly significant interaction of the three independent variables 
(F(15, 125) 
= 1 .69; p = .06) when considering all traits together. However, a number of 


The personal-social interplay 
43 
significant effects are observed when considering each particular bipolar dimension. Thus, for 
instance, some self-descriptive traits (for example, calm/agitated) vary according to the situation 
(F( 
I ,  141) = 16.46; p = .0 I); the re
f
erence point affects the description of oneself as being serious 
rather than f
rivolous (F(I ,  141) = 6.01; p < .01); the judgement of oneself as being gloomy or 
cheer
f
ul depends on the interaction of all three independent variables (F(I, 141) = 10.47; P = 
.0 
I); and so forth. 
5. For instance, some subjects had to assess 'their own similarity to the Northerners 
(Southerners), rather than the 'Northerners' (Southerners') similarity to themselves'. According 
to the experimental conditions, all kinds of comparisons were perf
ormed, by different subjects. 
6. The data in the Northern region were collected by N. Cavazza, University of Bologna. 


4 
Self 
-We-Others Schemata and Social 
Identifications 
Maria Jarymowicz 
Searching for the social self: the self and prosocial attachment 
In order to examine the social self from a social psychological perspective, 
we must refer to traditions in other areas of psychology (Jarymowicz & 
Greenwald, 1 993), and distinguish between the ego and the self (Allport, 
1 961; James, 1 890). In social psychology, this distinction is revealed in the 
view of the self as a sub
ject (or self as a knower) and the self as an object 
o
f
 knowledge (Epstein, 1980; Greenwald & Pratkanis, 1 984). In other 
words, we can describe the self as an active agent or as an object to be 
understood. 
Different conceptualizations of the social self define how the individual 
is connected to other people in the social environment. There are two 
particularly important approaches that deserve attention. On the one hand, 
the social self can be described by the individual's social identifications: 
different forms of syntony with other people and states of prosocial ego 
involvement. These states are accompanied by feelings of attachment, 
sympathy, affiliation, devotion, respect for others' expectations, common 
goals, understanding and acceptance. On the other hand, the social self 
can be associated with people's ability to perceive and recognize 
themselves, in other words, the ability to create cognitive representations 
of one's person and self schemata. The content of the self-description 
combined with the representation of other(s) constitutes the social self and 
we schemata. 
The relationship between the ego and the self, or between the ego-identity 
and self-identity (Erikson, 1956), has not been carefully examined in 
psychology. The aim of this chapter is to elaborate on the relation between 
the self and the other(s) schemata, on the one side, and the states of social 
identification, on the other. We are particularly interested in the relationship 
between the distinctness of self-we-other(s) schemata and attitudes towards 
others, ingroup versus outgroup members. 
In order to understand the process of social identifications, we can begin 
by questioning the generality of the popular thesis that states that the self­
others similarity leads to prosocial ego involvement and that the perceived 


Sel
f
-we-others schemata 
45 
Table 4. 1 
The social self as a structural basis o
f
 dif
f
erent f
orms o
f
 social 
identifications 
T
he self structure: 
internal basis of social identifications 
'T
he experiential social self' 
affect and drives mechanisms: emotional 
empathy, security, affiliation, and other 
hedonistic motives 
2 
'T
he puhlic self' 
dependence on group's reinforcements, 
the we-as-a-group concept (e.g. 'We - a 
family') 
3 
'T
he collective self' 
cognitive representations of real social 
divisions and memberships, the we-as-a­
category concept ('We - workers, women, 
etc.) 
4 
'T
he conceptual social sel
f' 
cognitive categorization based on abstract 
criteria, the attributive-we concept ('We -
honest, tolerant, etc.) 
5 
'T
he autonomous social self' 
differentiation of two self subsystems: 'me 
as an individual' and 'me as a member of 
social groups and categories' 
Ego-involvement: 
states of social identifications 
syntonic feelings, seeking for contact, 
affiliation, sensual comfort 
conformity: respect for particular ingroup 
norms and goals 
feeling of ingroup (categorial) homogeneity, 
internalization of social norms, ingroup 
favouritism 
f
eeling of we-ness across (and independent 
from) social groups and categories 
recognition of diff
erent and common values/ 
goals, enjoying communality with 
autonomous others, groups and categories 
similarity to others is the main determinant of attachment to them. This 
thesis has support in theoretical and empirical work (Byrne, 1 969; 
Reykowski, 1979). However, relationships between self-other(s) schemata 
and ego involvement are far more complex than this work suggests (Brewer, 
199 1 ;  Jarymowicz, 1987, 1 994c; Snyder & Fromkin, 1980). In order to 
understand this relationship, it is necessary to analyse different forms of 
prosocial ego involvement, ranging from the simple drives to affiliate or 
conform (when prosocial behaviour leads to affect comfort) to exocentric 
ego involvement (when prosocial behaviour is focused on other's states and 
needs, regardless of one's own needs: Jarymowicz, I 992a; Karylowski, 
1982). 
Social identifications, expressed by syntony with others and prosocial ego 
involvement, are sometimes considered to be defining characteristics of the 
social self. This view, however, creates more problems than benefits because 
there are many different forms of social identifications based on different 
social structures (Jarymowicz, 1 992b, 1994a; see Table 4. 1 ), 
On the primary level ( 1 )  drives and affective mechanisms lead to non­
specific affiliation and seeking syntony with others. The social self cannot be 
explicitly articulated by the individual, it is more like 'I have experienced it' 
or 'I f
eel it'. Identifications are automatic and strong, their objects are 


46 
Social identity: international pers
pectives 
persons who can gratify primary needs. The feelings last as long as the 
relationship with particular others endures. 
Thanks to a number of social reinforcers that can be offered by groups, 
the process of socialization determines the respect and conformity with 
group expectations (level 2). The social self may be a family, partnerships or 
classmates: we as a group is f
ormed (Jarymowicz, 1993a, 1 994b). Identi­
fications refer to the concrete group, and terminate when the person no 
longer belongs to the group. 
The ability to focus on broader social categories gradually develops, and 
the social self moves beyond face-to-face contacts (level 3): we as a category 
is created (Jarymowicz, 1 994b) based on intra-categorial differentiations. It 
gives rise to syntonical identifications with category members and results in 
group favouritism. Referents of these identifications are stable only so long 
as the cognitive map of the world reflects social reality. 
The human mind can create different types of categorizations as well 
(Brown & Turner, 1 981), for example those based on abstract criteria. 
Therefore, the attributive we appears ('We, optimists, art lovers'). This 'we' 
is formed independently of, and sometimes cross-cuts, social groups and 
categories (level 4). The identifications may refer to other people, regardless 
of their social membership. They are modified as new, cognitive, abstract 
categories of understanding the world (for example, acquiring the concept 
of 'loyalty' enables one to discern between loyal and disloyal people). 
Social categorization and social comparison processes are not the only 
ways the social self is identified. Individuals may define themselves beyond 
comparisons with others or membership to particular categories. Thanks to 
intra 
personal differentiation and formation of personal evaluative 
standards, one's self can be recognized as someone realizing the pursuit 
of certain superior values. Thus social identifications do not refer to a 
category of people but to common goals (level 5). When values are per­
ceived as common, identifications are likely to last a long time. When goals 
are common, identifications last as long as they are relevant. 
The distinction, we suggest, reflects the sequence of developmental stages 
in which people can use new cognitive categories and evaluative standards 
(Golab & Reykowski, 1 985; Hoffman, 1 989; Kohlberg, 1976; Piaget, 1 975; 
Reykowski, 1 989; Ziller, 1964). The stages also determine new areas and 
objects of people's identifications (Tap, 1 979): from social groups or 
categories of own membership (levels 1 -3), to categories going beyond 
social divisions (level 4) or categorization as such (level 5). 
How can we define the relationship between the social self, social identity 
and social identifications? In our approach the social self is considered not 
as a subjective state of identification with other people, but as a subsystem 
of self-knowledge. Prototypical elements of the social self constitute indi­
viduals' social identity (and, consequently, determine its content). On the 
other hand, social syntony and prosocial ego involvement are considered as 
states of social identifications, states of ego which depend on self-we-others 
schemata and thus on social identity. 


Self
-we-others schemata 
47 
Self-we-others schemata distinctness as a basis of identity 
Self-specific traits as personal identity attributes 
Our definition of identity is based on two groups of premises. The first 
reflects the traditional way of identity conceptualization: identity is con­
sidered as a result of recognition of the self (including language-mediated 
effects of 'talks with the self'; see Mead, 1934). Identity components are 
only these elements of the self which are explicitly indicated by a subject 
answering the question 'Who am I?' 
The other group of identity-definitional criteria refer to empirical data. 
According to the well-known studies by Markus (Markus, 1977; Markus 
& Smith, 1981), traits described by participants as (a) very important 
and (b) highly characteristic of the self are processed differently (for 
example, they are processed quicker). Markus labelled them schematic 
traits. 
2 
Our own studies on social comparison processes (Jarymowicz, 1991a) 
showed that traits of the self perceived as most clearly distinguishing the 
self from the others were processed differently (for example, egotism was 
found only for these traits of the self which were the most highly 
distinct). Our definition of identity refers to self-descriptive traits, traits 
which are not only characteristic of the self (and can be also charac­
teristic of the other), but also specific - distinguishing the self from the 
other. Such traits of the self, which are characteristic and distinguishing 
at the same time, we have labelled identity traits. 
In our definition (Jarymowicz, 1991a, p. 693), personal identity is a 
subsystem of self-knowledge, which is constituted by the traits perceived as 
the most characteristic of and the most specific (distinct) to the self. 
Self
-we-others schemata distinctness as a basis o
f
 personal versus 
social identity 
According to Byrne's studies (Byrne, 1969), interpersonal similarity leads to 
attraction. On the other hand, studies on identity have shown that, in some 
conditions, self/we-others similarity may be aversive and distinctness 
attractive (Brewer, 1991; Codol, 1979; Jarymowicz, 1982, 1987; Snyder & 
Fromkin, 1980). Seeking for explanations of the incoherent data concerning 
self/we-others similarity/distinctness and attractiveness, we found that the 
perceived dissimilarity was not a good measure of self/we distinctness from 
the other. People's estimation of the degree of similarity or dissimilarity is 
determined not only by the ability to differentiate between own person or 
own group and other people, but also by motivation to perceive self or we 
as similar or not similar to others (Suls & Miller, 1977). In order to create a 
better measure, we focused on the real cognitive distinctness of self/we­
others schemata, instead of the perceived similarity (depending often on 
subjects' wishful thinking). 


48 
Social identity: international pers
pectives 
The new operationalization was based on three procedural requirements. 
Individuals' attention on the self, the we or the other has to be focused 
without inducing social comparison processes (which are likely to 
activate social norms and evaluation). 
2 As was shown by Holyoak and Gordon (1983), the self is the usual 
reference point in social perception, and usually has stronger impact on 
perception of the other than does other on the self (see Hardoin & 
Codol, 1 984) - to focus attention on the self, the we and the others, a 
reverse order has to be applied. 
3 
To focus the subject's attention on descriptive and not evaluative 
dimensions of a particular object, nouns instead of adjectives have to be 
used (as the latter may help induce affective or evaluative associations). 
The so-called Questionnaire o
f
 Social Perception consists of a list of 70 
nouns, enumerating dimensions such as elegance, intelligence, tolerance (all 
presented in a positive manner). It contains three parts, presented to 
participants subsequently. The first part of the Questionnaire concerns the 
category of others. Participants read the list of dimensions and mark those 
which 'are on their mind when they think about other people'. Afterwards, 
out of the already indicated nouns, they mark l O  'which are considered the 
most frequent'. The second part examines the 'we' category. This category 
is first activated by an additional operation: participants are requested 'Try 
to think about those people who are on your mind when you use the word 
WE' and describe the chosen 'we' in an informal way. Then they read the 
same list of 70 nouns, and choose 10 dimensions 'most frequently used 
when thinking about people considered We'. The third part examines the 
self category: subjects choose 10 dimensions 'most frequently considered 
when thinking about oneself'. 
The three tasks refer to schematic dimensions (labelled analogous to 
Markus, 1977) of three objects. The distinctness of schemata is defined, 
however, by prototype, specific, distinctive dimensions, that is, only those 
dimensions which subjects ascribed to one and only one object. Measure­
ments built upon such material are likely to reflect the distinctness of self­
others, we-others and self-we schemata. 
Assuming that social identity may be derived from intergroup (or rather 
we-others) differentiation, in our approach the content of social identity is 
determined by distinctive dimensions of the we related to the others, that is, 
those dimensions which are marked by a subject as referring to the we and 
not marked as referring to others in general. 
Further on, assuming that personal identity may be derived from intra­
group (self-other members) differentiation (or distinguishing the self from 
similar ones; see Festinger, 1954), the content of personal identity is deter­
mined by distinctive dimensions of the self related to the we, that is, those 
dimensions which refer to the self and not to the we). 
We assume that the level of the distinctness of the self (from the we) and 
the we (from the others) is related to the stage of the identities' development 


Self
-we-others schemata 
49 
and their status, that is, their role in the regulation of individuals' func­
tioning: the higher the level of the distinctness, the greater the impact on the 
individual's functioning. 
Finally we posit (and present relevant data; larymowicz, 1 991b, 1993b): 
(a) social identity is likely to prevail in persons with a high level of we­
other schemata distinctness and a low level of self-we schemata 
distinctness; 
(b) personal identity is likely to prevail in persons with a high level of self­
we schemata distinctness and a low level of we-others schemata 
distinctness. 
By 'prevail' we mean both the higher level of development of a particular 
scheme and its relatively greater impact on information processing, evalu­
ation processes or behaviour. It is particularly clear when the social self and 
personal self do not overlap. This prediction was empirically verified and is 
presented elsewhere (Jarymowicz, 1993b, I 994a). 
Self-we schemata distinctness as a basis of social identifications 
The distinctness o
f
 the self as a precondition o
f
 understanding and 
acceptance o
f
 outgroups and dissimilar others: a theoretical model 
A number of theoretical arguments and empirical results contribute to the 
well-known thesis: similarity of an other to the self and the membership of 
the same social category facilitate identifications, and result in ingroup 
favouritism. In contrast, we focus on identifications with dissimilar others 
who do not belong to the same social group or category and study empirical 
findings concerning the relationships between self 
-we-others schemata 
distinctness and intergroup attitudes. 
The following predictions are proposed (larymowicz, I 992a, I 994c); 
The strong syntonic identification with one's own group leads to ingroup 
favouritism and, in certain conditions, to outgroup discrimination. 
2 
The lower the level of self-we schemata distinctness, the stronger the 
identification with one's own group and ingroup favouritism, and 
consequently the lower the ability to decentrate and to recognize the 
outgroup's perspective, and the lower the possibility of accepting others' 
goals and values. 
3 
The higher the level of self-we schemata distinctness, the weaker the 
influence of a reference group on individuals' functioning (including the 
tendency to ingroup favouritism and outgroup discrimination). 
We propose a theoretical model of the relationship between self-we 
schemata distinctness and possible identifications with dissimilar others. 
Two groups of premises are considered. 


50 
Social identity: international pers
pectives 
Degree of 
others-oriented 
involvement 
High 
Low 
Levels of self-we 
'--_----_e_----+-
distinctness 
Low 
Medium 
High 
0>----<0
 
Ability to recognize others' (outgroup members') states 
• 
• 
Motivation for engagement 
Figure 4.1 
Self
-we schemata distinctness and the degree o
f
 individuals' 
involvement with outgroups and dissimilar others: a theoretical model 
(a) instrumental (cognitive) premises - an ability to decentration and to 
recognize others' perspectives; 
(b) motivational premises - a readiness to focus on others' perspectives, to 
think what is beneficial to them, and to be involved. 
The model is presented in Figure 4. 1 .  
According to the model: 
The low level of self-we schemata distinctness relates to lack of ability 
to distinguish perspectives of one's own group from those of other 
groups, and to recognize others' states, needs, goals - which is a 
necessary precondition of others-oriented involvement. 
2 The high level of self-we schemata distinctness relates: 
(a) to an ability to distinguish perspectives of one's own group from 
those of other groups, and to recognize others' states, needs, goals 
- which is a necessary but not sufficient precondition of others­
oriented involvement; and 
(b) to different levels of pre 
social motivation (which depends on factors 
other than self-we distinctness) - thus, degree of others-oriented 
involvement varies in different situations. 
This model has been confronted with empirical data. Below some results 
are presented. 


Self
-we-others schemata 
6 .5 
(5.7) 
-0 
,g 5.0 /;:21 
-
eo 
o 
u 
l 
QJ 
QJ 
l 
.0 
(/) 3.5 
E E 
::> 
QJ 
(3 1) 
z :o  2.0 
0 
Ci 
L 
Low 
Medium 
High 
Self-we distinctness 
Figure 4.2 Recall o
f
 children's 
problems (Szuster-Zbro
jewicz, 
1 990) 
80 
(78 1) 
-
(/) 75  5J 
o 
(/) 
QJ 
QJ 
QJ .<: 70 
m -o  
Ol eo  
QJ 
QJ 
o 
l 
65 (68 9) 
L 
Low 
Medium 
High 
Self-we distinctness 
Figure 4.4 
Readiness to host 
diff
erent nationalities ( Cylwik, 
1 992) 
The model and empirical data 
35 
-
c 
o 0 
ill
 
'';::;
 
QJ 
eo 
l .Π
Ol +-'  
QJ 
0 
o E 
25 
(24 3) 
': " 
L 
(5 m 
Low 
Medium 
High 
Self-we distinctness 
Figure 4.3 
Motivation to 
solve children's problems 
(Szuster-Zbro
jewicz, 1990) 
90 
QJ 
U 
C 
o U 70 
c O  
6 V 50 
E $ 
<i   
30 
(81 5) 
L
(35 7) 
Low 
Medium 
(37 2) 
High 
Self-we distinctness 
Figure 4.5 
Tolerance o
f
 
German minority activity 
(Pawluczuk, 1992) 
5 1  
Participants of experiments were focused on others - presented either as 
anonymous 'others' or as 'they' (people of different social or national 
membership). In all groups self 
-we-others schemata distinctness was 
measured, as well as one of the following manifestations of others-oriented 
involvement: 
(a) recall of children's problems, mentioned prior to the experiment 
(Figure 4.2); 
(b) motivation to solve children's problems (Figure 4.3); 
(c) readiness to host at home students of different nationalities (Figure 
4.4); 
(d) tolerance of the activity of the German minority in Poland (to preserve 
their national tradition) (Figure 4.5); 
(e) positivity of evaluation of a painting, the artist being presented as a 
foreigner (Slovak, Irish or Japanese) (Figures 4.6-4.7). 
As can be seen, presented data are consistent with the model. 


52 
Social identity: international perspectives 
High 31 
(30.7) 
High 36 
(35.1) 
-
29 
-
33 
o ˆ 
o ˆ 
Q) S 
Q) 'S; 
‡ :-e 27 
(28 2) 
‡ :-e 30 
0l
Cf)
 
0l Cf)  
Q) 0 
Q) 0 
0
0.
 25 
o 
0. 27 
(25 8) 
(28 3) 
Low 
--( 
Low 
--((27 3) 
Low 
Medium 
High 
Low 
Medium 
High 
Self-we distinctness 
Self-we distinctness 
Figure 4.6 
Positivity o
f
 
evaluation o
f
 Slovak and Irish 
paintings (Sobczyk, 1992) 
Figure 4.7 
Positivity o
f
 
evaluation of Japanese painting 
(Sokolowska, 1992) 
The follow-up study established that participants with a low level of self­
we schemata distinctness were least likely to favour outgroups (similar 
results were obtained by Jarymowicz, 1994c; Kobuszewska, 1 989; 
Szuster-Zbrojewicz, 1988). The only inconsistency found was the motiva­
tional pattern presented in Figure 4.3 (Szuster-Zbrojewicz, 1 990). When 
subjects had to declare more than just good will, persons with a low level 
of the schemata distinctness were least likely to favour strangers. 
2 
The results of participants with a high level of the schemata distinctness 
are less homogeneous: in some studies a rectilinear rather than a 
curvilinear relationship is found. 
Conclusion 
There are a number of conclusions that can be drawn from this discussion. 
Let us outline them briefly. 
Egocentrism or ingroup orientation IS related to universal human 
properties. 
2 
The differentiation of cognitive schemata is a necessary precondition to 
go beyond own egocentrism: to recognize different social perspectives 
and understand others. 
3 
Social self and social identity (if related to social groups or categories -
see Table 4. 1) induce syntony and ingroup involvement. 
4 The necessary precondition of coexistence with outgroup members is an 
ability to distinguish the self from the we schemata: a formation of 
social and personal identity is necessary to enable the individual to 
switch focus and shift attention from one perspective to another. 
In summary, then, and coming back to the opening sentences of the 
chapter, we conclude by stating that we have tried to reconcile different 
traditions in the study of self in psychology. We think that from now on 
they should not be considered as separate. 


PART II 
ESTABLISHING GROUP 
IDENTITY 
5 
A Developmental View of the Search for 
Group Identity 
Stephen Worchel 
I spent several years of my life on the East Coast of the United States. 
Coming from Texas, which had only two types of weather, hot or rainy, I 
was awed by the changing seasons on the East coast. Each year, fall 
followed summer and signalled the coming of winter. I could consult my 
calendar to find the exact time of the metamorphosis, but nature was a 
better, but less predictable, source of information. Some years the change of 
season would be abrupt, easily noticeable and sudden when a winter storm 
would arrive and take control of my world for months. Other years, the 
changes of season were gradual, silent, almost secret; a few leaves might 
drop from a tree to whisper the advent of fall, but the full message took 
weeks. This four-act play would be performed every year. 
The drama of changing seasons has a message that carries far beyond the 
woodlands. The Navaho Indians believe that human behaviour is a mirror 
that reflects the forces of nature. If we wish to understand the behaviour of 
humans, we should look first to understand the natural environment. 
Indeed, there may be many messages about group and intergroup behaviour 
in the winds of seasonal change. 
I begin with this retreat into nature because it so sharply contrasts with 
the approach that characterizes much of the study of group and intergroup 
behaviour. Typically, social psychology searches for e
f
f
ects and the causes 
of these effects. Our methodological microscopes focus on a single behav­
iour, at a single point of time. The context in which that behaviour occurs 
(the person's previous actions, expectations about the present situation, or 
anticipations of the future) is seen as a pesky problem and given the 
ignominious title of 'error variance'. We speak of effects such as conformity, 


54 
Social identity: international pers
pectives 
group polarization, social loafing, cooperation, social categorization and 
intergroup competition as if they were stones in a behavioural river that 
exist in isolation. 
I want to offer a somewhat different perspective in this chapter. To begin, 
I want to question the concept of e
f
f
ect. Our 'effects' are not effects in the 
sense that we characterize them. Rather they are processes that can only be 
understood in the context of the preceding events and behaviours, the 
present interpretations of the setting, and the individual's anticipation of 
future activities. These processes are intertwined with the flow of behaviour 
over time and can best be understood and predicted in this context. They 
are the stream itself, not concrete forms that divert, interrupt and stand 
alone in the behavioural stream. This is not a new view. Lewin found that 
the effectiveness of leadership styles could be described only when one style 
was examined in light of the leadership style that had preceded it (Lewin, 
Lippitt & White, 1 939). Schachter (1951) found that reactions to a deviate 
not only changed over time, they were influenced by changes in the deviate's 
behaviour. If the deviate accepted the group position, he was quickly 
embraced by other group members. Continued disagreement with the group 
resulted in a cessation of communication directed towards him and 
rejection. Sherif and his colleagues (Sherif, Harvey, White, Hood & Sherif, 
1 961) found that hostility between groups could be reduced only after a 
series of cooperative encounters. And McGrath and his colleagues (Kelly, 
1 988; McGrath, 1 988) have lamented the neglect of time in the study of 
social psychology. 
I want to build on the foundation laid by these early studies and present 
group and intergroup behaviour as a procession of events, each related to 
the other, and each taking on meaning from the context in which it occurs. 
Groups, I believe, develop through rather predictable stages (or seasons). 
The process of group development is not one with a distinct beginning, and 
often there is no clear end. In fact, many groups 'live on' long after all the 
original members have departed. Rather, the development is a cycle that 
repeats itself. Each stage has a predictable impact on the behaviour of 
group members and on the relationship between groups. Actions that are 
accepted or encouraged at one point in the process will be censored and 
rejected at other points. Understanding and predicting group and inter­
group behaviour requires that we understand the developmental process. 
Independence versus group membership: a life-long dilemma 
Before jumping directly into issues of group and intergroup behaviour, let 
me be true to my psychology heritage and focus on the individual for a 
moment. Humanistic psychology argues that nearly from the moment of 
birth we embark on a life-long journey to 'find' our identity (Maslow, 1 954; 
Rogers, 1 95 1). The journey is a strenuous one, but at the heart of it is 
determining our relationship with the groups to which we belong. Defining 


A developmental view o
f
 the search f
or identity 
55 
this relationship throws us into two raging and competing cross-currents. 
On one hand is our desire to be accepted and valued by others. Groups 
offer us support, safety, protection and a reason for existence. While groups 
are a safe harbour in many ways, they also exact their price. They demand 
obedience and conformity, they require us to place group needs above 
personal needs, they coerce us to give up some of our personal identity. 
It is in this latter demand where the counter-currents meet with such fury, 
because paralleling the desire for group membership is the desire to be 
recognized as an independent, unique individual. In order to achieve this 
goal, we must differentiate ourselves from the groups to which we belong. 
We must resist their seductive lure to abandon our independent self and 
accept the group identity. Our lives are characterized by the constant 
struggle of deciding how much of our personal identity we will sacrifice for 
the substitute group identity. 
Tajfel and his colleagues (Tajfel, I 972b, 1 978d; Tajfel & Turner, 1986) 
anticipated this discussion in their development of social identity theory. 
They implied that we actually hold two identities (more correctly, one 
identity with two opposite poles), a personal identity that includes our 
personal characteristics and a social identity derived from our membership 
in groups. Turner (1982) argued that the identity (personal or social) which 
we stress at any moment is a function of the context in which a particular 
interaction takes place. Intimate interactions, for example, enhance the 
salience of our personal identity, while interacting with people from other 
identifiable groups may make our social identity salient (Doise, Deschamps 
& Meyer, 1978). For example, being the only woman in a group of men 
makes salient the social identity, especially of the minority person, and 
influences the nature of the interaction that will take place. Social identity 
theory further suggests that we wish to maintain a positive image of 
ourselves, and in order to enhance our social identity we will advantage our 
ingroup in relation to outgroups. 
Social identity theory offers unique insight into a host of social and 
individual behaviours. I would like to push the social identity approach a 
step further and suggest that the desire to emphasize one identity or the 
other is influenced by more than the social context in which people find 
themselves. I want to argue that the individual's own context (history, 
interpretation of the present situation, and anticipation of the future events) 
plays at least as important a role as does the social situations in determining 
the salience of personal/social identity. Further, an individual's life can be 
characterized as alternating between placing an emphasis on group identity 
and emphasizing personal differentiation. I accept the social identity posi­
tion that the social situation influences which identity is salient. However, I 
suggest that it is the individual's own context that motivates him or her to 
seek social situations that emphasize one identity or the other. For example, 
descriptions of close relationships suggest that individuals often alternate 
between periods of close intimacy with a partner followed by efforts to 
'touch base' with larger groups of people (Argyle, 1 992; Oldenquist, 1 992; 


56 
Social identity: international perspectives 
Reis, Nezlek & Wheeler, 1 980). Partners in intimate relationships are often 
threatened when they perceive the other 'pulling away' and becoming 
interested in the activities of social groups. However, the natural tide of 
relationships is a cycle of emphasizing the intimate relationship and 
emphasizing the membership in other groups. A threat to one aspect of the 
identity can lead a person to focus on and attempt to repair that identity. 
Individuals who have suffered threats to their social being, such as being 
fired from a job or going through a divorce, seek social support networks to 
reaffirm their social identity (Rook, 1 987). On the other hand, when the 
threat is to the personal identity, such as the case of experiencing an 
embarrassing event, the desire to be alone and repair the personal self is 
dominant (Morris et aI., 1976). Alternating between one identity and the 
other occurs not only in brief episodes during life. It also characterizes the 
larger pattern of human existence; greatest emphasis on establishing 
personal identity occurs during early childhood and middle age, while social 
identity themes are often most evident in adolescence and in the elderly. 
Finally, none of our behaviour escapes the embrace of culture, so it is not 
surprising that something so central as identity is a product of culture. 
Numerous investigators (for example, Hofstede, 1 980; Triandis, 1 988) have 
observed that cultures place varying emphasis on personal and social 
identities. In collective cultures such as those found in Asia, Greece and the 
Middle East, the emphasis is on the group and the individual's social 
identity is stressed. On the other hand, more individualistic cultures such as 
the United States and some countries of Western Europe force the 
individual into centre-stage, and personal identity plays a critical role. It is 
no accident that humanistic psychology with its emphasis on differentiating 
the self from the social context is a Western phenomenon. 
The development of groups 
This has been a rather long preamble to a discussion of groups and 
intergroup behaviour. However, I wanted to make the point that even at the 
level of personal and interpersonal behaviour there is an ongoing process of 
identification that involves emphasizing themes of inclusion and differentia­
tion at different points of time. The theme that is emphasized is a function 
of dynamics occurring within the individual and between individuals, as 
well as events that occur within the person's physical and social environ­
ment. My aim is now to demonstrate that similar processes characterize 
groups and the relationship between groups. My interest in group develop­
ment was initially sparked because so much of the research and theory in 
group and intergroup behaviour was not dynamic in approach, focusing 
only on specific behaviours at a single period, and ignoring the history of 
the members and the groups. I was concerned that most research on group 
dynamics began with the group already established, either in the sense of a 
laboratory group of strangers who arrived for an experiment or in an 


A developmental view o
f
 the search f
or identity 
57 
established group in a natural setting. This approach overlooks the process 
of group formation, how individuals decide to join groups, and how these 
variables affect group behaviour. These are important issues that may have 
dramatic impact on all group and intergroup behaviours. Unfortunately, 
these issues are beyond the scope of this chapter. However, although I will 
not address them, I want to raise them for future consideration. 
My second concern with traditional group research is that group devel­
opment and change is rarely addressed. Specifically, after reviewing the 
literature I was left with many questions about the dynamics of groups: for 
example, how do groups develop and change over time, and how does this 
development affect intermember and intergroup behaviour? In order to 
examine these questions, I turned to empirical accounts in social science 
literature on group dynamics and popular archival accounts of groups. My 
search of the archival literature included descriptions of small groups 
(families, athletic teams, splinter church groups, gangs), large groups 
(business organizations, political parties, nations) and social movements 
(women's rights movements, the civil rights movement, ethnic identity 
movements). I included a wide variety of groups because I wanted to 
identify both similarities and differences in the process involved in these 
groups. I was also guided by a personal interest in the processes involved in 
groups gaining independence from larger groups. This interest moved me to 
consider the area of group formation, as well as development, and carried 
me into the fascinating literature on revolution, protests and terrorism. I 
have discussed group formation and revolution in earlier papers (Worchel, 
Coutant-Sassic & Grossman, 1992), and space does not allow me to deal 
with this topic here. For the present, then, I will focus more directly on 
group behaviour from the time a group is initially formed. 
Actually, there has been a long history of efforts to examine group 
development. The t-group movement beginning in the early 1 950s offered a 
marvellous opportunity to observe groups over time and chart their devel­
opment and change. Based on data from these and similar groups, investi­
gators suggested that groups do move through rather predictable stages. 
For example, Tuckman (1 965; Tuckman & Jensen, 1977) identified the 
stages of forming, storming, norming, performing and adjourning. While 
this earlier work focused on the group, Moreland and Levine (1 982, 1988) 
examined the stages that individual members go through as they enter and 
move through the group. These studies of group and member development 
are rich and interesting, but it is difficult to determine where they have 
impacted group research. Rather than representing a coordinated effort, the 
studies and observations have been undertaken in isolation from each other 
so that it is not possible to identify a unified theory of group development. 
In addition, these fragmented observations of groups do not readily lend 
themselves to developing specific hypotheses. This is not necessarily an 
indictment of the approach, because much of the work was designed to 
describe group development, not generate research, but it does help explain 
why this approach has not been incorporated into mainstream research 


58 
Social identity: international pers
pectives 
Group identity 
Group productivity 
Individuation 
Note: During the decay period numerous members may leave the original group to join 
other groups or form new groups. While the old group re-enters the group identity 
stage, new groups formed by the spinning off of defectors will enter into their own 
stage of group identity. 
Figure 5. 1 
Stages o
f
 group development in existing groups 
efforts. Finally, many of these observations dealt with rather specific groups 
(t-groups, therapy groups), and the applications were rather group-specific. 
Most of the groups involved had specific beginning and end points, and 
they were formed for specific purposes. Therefore, the work rarely dealt 
with issues of group formation or examined cycles in development. Even the 
important work of Moreland and Levine follows the individual member 
through one group; there is no discussion of how the experience of 
membership in one group impacts the member's choice of additional groups 
or behaviour in these additional groups. In this sense, most earlier models 
are linear. As we will see, my observations suggested that group develop­
ment is cyclical and repetitive. This cyclical nature of groups has important 
implications for explaining intragroup behaviour and the continuing rela­
tions between groups. 
My work on group development used this earlier work as a starting point, 
and it was designed to deal with the limitations I perceived in these earlier 
approaches. As a beginning, I was able to identify four stages or periods 
that occur in intact groups (assuming group formation as the beginning 
point). The stages represented themes that dominated group activity during 
specific time periods. The stages are reciprocal in that they affect the 
behaviour of group members and the relation between groups, and they are, 
in turn, affected by member behaviour and intergroup relations. Beginning 
with the formation of the groups, the stages are: identification, group pro­
ductivity, individuation and decay (see Figure 5. 1). 
Identification 
Once a group is formed, its initial focus is on establishing its identity. 
Efforts are directed at defining the boundaries of the group, its 'place' in 


A developmental view o
f
 the search f
or identity 
59 
relation to other groups (and history), and the relationship between mem­
bers. This focus has profound effects on member behaviour and intergroup 
relations. The group establishes a central dogma or theme for the group. 
Attitudes on these issues are often extreme, clear and uncompromising. 
Minority positions are not tolerated and minorities are quickly excluded 
from the group. There are strong pressures on members to conform. Groups 
often adopt a dress code or uniform, a special language and other symbols 
that identify the group and mark people as being members. In many cases, 
efforts are made to 'reach back' in history to establish the legitimacy of the 
group, recognize patrons and previous founders, and locate the marks that 
will distinguish the group and identify its 'rightful' members. For example, 
the 1980s were a time of rebirth of Maori people in New Zealand. The 
Maori became especially interested in their rightful place in New Zealand 
history, their language, their customs and dress. There was a renewed push 
for recognition of the Maori within New Zealand society, a return of lands 
to the tribes, and greater powers for self-governance. 
Leadership during identification is often centralized. Members may be 
asked to demonstrate their loyalty to the group by making sacrifices or 
undergoing difficult initiation rites. The boundaries of the group are closed 
to new members. Recall that the initial stages of the doomsday cult studied 
by Festinger, Riecken and Schachter ( 1956) was characterized by 
withdrawing from public focus and closing the group to new members. In 
many cases, groups in a period of identification will physically withdraw, 
demanding that members break contact with other groups to which they 
have belonged. 
Relations with outgroups become especially strained at this point. Group 
boundaries are defined by the relation of one group to other groups. As a 
result the group either seeks to isolate itself from other groups or it invites 
conflict and competition with outgroups. Either of these actions defines 
boundaries and has a bonding effect within a group. Outgroups are 
presented as dangerous, aggressive and sinister, thereby forcing members 
together to deal with this outside menace. Similarities within the group 
membership and differences between groups are emphasized. In some cases, 
demands are made on outgroups which cannot be met, precisely to 'prove' 
that the outgroups are evil and unfair. For example, in many cases where 
ethnic groups attempt to 're-establish' themselves and incite member 
awareness, demands are made for the return of land or other resources that 
were taken from the group at an earlier period of history. Even when it is 
deemed that the resources were illegally seized, the wholesale return of these 
properties would create havoc in the larger society, so that the wholesale 
return is either impossible or will involve a slow, complicated process. The 
reluctance to return this property or the slow process is then used to 'prove' 
the lack of good will in the outgroup. 
From the group's perspective, the aim of the identification stage is to 
focus group members on their social identity, as opposed to their personal 
identity. Interactions within the group and between ingroup and outgroup 


60 
Social identity: international perspectives 
members emphasize social identity. At the same time that the group seeks to 
strip away the individuating characteristics of its members, it attempts to 
differentiate its own characteristics as a group, separate and different from 
other groups. 
Group productivity 
While efforts during the identification stage increase the cohesion within the 
group, they often interfere with group productivity. Therefore, as identity 
becomes established, group members begin to examine the goals of the 
group, especially ones involving group productivity. Although the orien­
tation is still on the ingroup, the topic of concern is productivity rather than 
identification. This focus creates numerous changes in the group and inter­
group behaviour. Overall, the emotional climate cools from the passionate 
state that characterized the identification process. Members now take a 
more analytical approach. During identification, group members were most 
concerned with the similarities of the group members. As the concern turns 
to productivity, members begin to examine the differences between them­
selves, but only those differences that may affect the group's task-related 
activities. The search is for the specific skills that members may possess that 
will help the group towards its productivity goals. Although still empha­
sizing social identity, the door is now opened for members to focus on their 
personal identity as it relates to group productivity. The boundaries of the 
group become more permeable and new members may be recruited to 
perform specific tasks. In some cases, the group may search for highly 
skilled members to fulfil high esteemed roles, but in many cases the new 
members will have a lower status and be expected to perform less attractive 
jobs, until they have 'proven themselves' worthy of group membership. This 
is often the fate of new immigrants to a country. 
The group becomes less antagonistic towards the outgroups, but its 
relationship with them remains cautious. In some cases, comparison with 
outgroups is sought in order to determine how 'productive' the ingroup is. 
Leadership becomes more task-oriented and the emotional pitch of the 
group is reduced. Minorities can influence the group on task-related issues, 
but these minorities will still meet quick rejection and censure if they 
threaten group identity. Group failures are attributed to external sources, 
while group success is attributed to internal group factors. 
Individuation 
During the productivity stage, members began to differentiate between 
themselves based on their task skills. The self-focus expands and accelerates 
during individuation. Members begin to assess their contributions to and 
rewards from group membership. They demand equity based on their 
contributions to the group rather than equality based on simple group 
membership. Subgroups appear during this stage, based first on skills and 


A developmental view o
f
 the search f
or identity 
61 
roles developed in the drive for productivity, but later expanding to include 
subgroups based on attitudes, interests and skills that may be unrelated to 
group activity. Individuals become more focused on personal needs and 
characteristics. Interactions within the group become more selective, but 
more intimate. Minority voice is not only tolerated, but encouraged. Social 
loafing (Latane, Williams & Harkins, 1979), in the sense of working less 
hard for group goals than personal goals, becomes common. Leadership 
may be questioned and failures blamed on poor leadership. 
Relations with the outgroup take on a very different flavour compared to 
the earlier two stages. Individuals begin a process of social comparison with 
outgroup members, as well as with ingroup members. They seek infor­
mation about out 
groups to use as standards for comparing their own group, 
and their treatment within the group. There is a tendency to overestimate 
the quality of life in the outgroup and the success of the outgroup. But this 
view is not necessarily used as a threat to the ingroup identity. Rather, it is 
presented as evidence that the ingroup is not meeting individual needs, and 
that changes are needed within the group. Rules, roles and norms of the 
ingroup are questioned, and there is a demand for greater personal freedom. 
Individual members begin negotiating with the ingroup to improve their 
personal standing. 
Decay 
From a group identity standpoint, the shift in focus to personal identity 
among group members poses a threat to group identity. There may be 
efforts to suppress this move towards individuality, but the group has 
become fragmented and the threat of rejection is less feared by members. 
The group becomes less important in the individual's self-concept. Indi­
vidual members actively explore defecting to other groups, or subgroups 
seek to break away from the group and form their new group. Marginal 
members who are not central to the group's identity are the first to leave. 
Later in this stage, more central members who have 'saleable' skills leave 
the group. These losses are viewed with alarm by the remaining members, 
who use these losses to demand change within the group. Although the call 
for change may result from a recognizable problem, members are often 
unclear about the specific change they desire. The result is that initial 
attempts to respond to these demands are often unorganized and ineffective. 
The easiest change is to replace the leader(s). There may be rapid succession 
of leaders as much is demanded from them, but members are unwilling to 
invest much power in leadership positions. Bitterness is expressed as 
members feel that they were lured into the group on the basis of false 
promises and expectations. Anticonformity as a way of demonstrating 
personal independence and displeasure will be common. 
Outgroups perceive the vulnerability of the group and efforts may be 
launched to encourage defection. Alarm over these incursions may be 
expressed, but the group is unable to mount a concerted defence. On a broad 


62 
Social identity: international pers
pectives 
scale, the decay of one group often creates conflict and hostility between 
outgroups as each attempts to profit from the condition. 
The cycle 
The stage of decay has a dramatic impact on the group. Many members 
leave the group and a negative emotional state predominates the group. But 
as the most disaffected members leave, the remaining members begin 
turning their attention back to the group. In the chaos, a strong leader often 
emerges, offering the remaining members the hope of re-establishing their 
group. The defections are seen as a group cleansing. Anger and hostility is 
expressed towards the defectors, and the group takes some solace in stories 
of disasters that befell these defectors. Attention is turned to rebuilding and 
redefining the group. The phoenix rises from the ashes. The group begins 
the identification stage once again. Often there is a surge of energy that 
comes from rebirth. The cleansing process may reach new heights as the 
group looks internally to define traitors or marginal members who have 
remained. These people are expelled, often with public ceremony. The 
fundamentalism that characterizes identification process grips the group. 
The reborn group begins differentiating itself from other outgroups, 
especially those that have defectors or profited form the demise of the 
group. Members' social identity is emphasized over personal identity. The 
cycle begins anew. 
The issue of inevitability 
This brief summary of the process of development raises as many questions 
as it answers. One of the first questions is whether this progression of stages 
is inevitable and must proceed uninterrupted in this order. The answer to 
this question is 'no'. Although this cycle may be the 'natural' flow of events, 
it is very clear that the process can be interrupted or short-circuited by 
numerous intended and unintended events. 
At any point during the process, events may occur that force the group to 
focus on specific issues and drive it to a particular stage. The most common 
situation is a return to group identification concerns. A clear or perceived 
threat from an outgroup, regardless of when it occurs in the development 
process, often makes group identity issues salient. An outgroup may make a 
bold effort to capture resources (land, material, status) held by a group. For 
example, it is not uncommon for employees of a company to rally to the aid 
of that company in the face of a rival's bid to buy that company. This can 
occur even when the employees are expressing dissatisfaction with their 
organization. 
A wily leader, feeling that his or her power is slipping or that the group is 
disintegrating, may seek out or manufacture a conflict with an outgroup. 
The result of this conflict is that the group will be thrown back into a stage 


A developmental view o
f
 the search f
or identity 
63 
of identification. For example, President George Bush was accused of 
accelerating the confrontation between the United States and Iraq in an 
effort to consolidate his power base and tum Americans' attention away 
from a deteriorating economy and onto the issue of national identity and 
pride. 
A sudden windfall or a dramatic increase in resources may push groups 
into a stage of individuation. Such a windfall requires that the group grapple 
with the problem of how to divide these resources among the members. 
Individual members may then begin to make their claim to these resources 
based on their skills, contributions to the group, personal needs or other 
personal characteristics. It is interesting to notice that the windfall, in this 
case, may hasten the decay of the group, in its present form. 
Finally, a scandal that arises from within the group may lead to disillu­
sionment and decay. Scapegoating and dissatisfaction with leadership can 
result in members feeling that the group cannot meet their needs, and result 
in their defection to other groups. 
While the progression through stages in order is not inevitable, I would 
like to offer two observations based on the reviews of group behaviour. 
First, it will take some unusual or unexpected event to derail the orderly 
progression. Second, when a group is pushed either back to an earlier stage 
or forward into a new stage, further group development will progress from 
that new stage. For example, if a group undergoing individuation is forced 
suddenly into the identification stage, the next step of that group will be into 
group productivity. The hands of the group clock move from the stage at 
hand, they don't 'catch up' or revert back to assume an earlier cycle of 
development. 
Transition and the nature of development stages 
A second major issue concerns the nature of the various stages and what 
forces the group from one stage into the next. These questions captured my 
interest as I read hundreds of accounts of groups. Was one stage necessarily 
longer than another? How did a group 'decide' that one stage was com­
pleted and it was time to move to the next stage? Were there clear 
distinctions between stages? 
I continually attempted to fit the literature into a concrete structure, but 
the fit was always an uncomfortable one. Like the variety found in the 
transitions between seasons, the changes in group stages followed a variety 
of patterns. In some cases, the transition was abrupt and clear, while in 
other cases it was more gradual, with the group dealing with two issues (for 
example, identification and productivity) simultaneously. Slowly it became 
clear that the stages were not necessarily distinct, bounded entities, but 
rather they were periods in which certain issues predominated the group's 
concern. Although central issues could be defined with considerable inter­
rater reliability, the fact that one issue was dominant did not mean that the 


64 
Social identity: international perspectives 
group avoided issues characteristic of other stages. Again, going back to the 
analogy of seasons, winter in most parts of the world is characterized by a 
preponderance of cold days. But even in the harshest winters, sunny warm 
days may appear for brief periods of time. Such was the nature of the group 
stages. The group discussion during a period would be ruled by concerns 
with a specific issue, but other concerns would also be addressed. However, 
introduction of these 'other' issues usually generated briefer discussions that 
often involved comparatively few members. Thus, a particular stage could 
be identified by the topic of discussion and behaviour that dominated the 
group and by the length o
f
 time s
pent on discussion and the number o
f
 group 
members involved with that issue. 
Similarly with the transition from one stage to another. It was often a 
matter of degree, but the change could be empirically identified by charting 
the topics of group discussion. Such, too, was the case with identifying the 
'trigger' for the transitions. There seemed to be no single event that moved 
the group from one stage to another. In some cases, a group member or 
members would declare that it was time the group moved on to another 
issue. This occurred both in small groups and in larger groups including 
social movements. What was interesting was that this 'trigger' propelled 
the group into another stage only when the group was ready to move. 
There were numerous examples in all sizes of groups of when a member, 
often the leader, requested that the group move on, only to be ignored by 
the group, which clung steadfastly to its present issue. In some groups, the 
trigger was a deadline or an imposed demand from the outside. Often 
members would indicate that a certain issue, generally productivity, must 
be addressed in order to meet a deadline; this moved the group out of 
identification concerns. Boredom was a frequent trigger. Several members 
might complain that the group was going nowhere or that it had already 
addressed certain issues and that they were ready for something new. 
Another sign of malaise and desire to move on was a lack of participation 
by members. 
Failure and success were some of the most interesting events in the groups I 
examined. Failure, regardless of the stage, tended to fixate members in that 
stage for a lengthy period of time. Like the classic Zeigarnik effect (Zeigarnik, 
1927), groups stayed with or quickly returned to issues that they failed to 
resolve. However, if these new efforts were not successful, groups tended to 
move into decay. 
Success was often as difficult to cope with as failure. Success spelled the 
end of a particular stage and moved the group, sometimes prematurely, into 
the next stage. For example, when workers at a computer company in 
Texas began organizing to protest decisions by management, they expected 
to meet stiff opposition. But when management quickly recognized and 
encouraged the worker organization, the group moved from concerns with 
identification to expressing their demands and seeking concessions (produc­
tivity). In this case, the identification phase of development was cut short 
and the workers failed to establish a clear organization. The result was that 


A developmental view o
f
 the search f
or identity 
65 
the union was disorganized and it soon disbanded, leaving many important 
worker/management issues unaddressed. Success and the lack of outgroup 
conflict resulted in the failure of the group to develop. 
From a research standpoint, these observations have serious implications. 
Determining the precise stage of development must be based on careful 
observations of group discussion and behaviour; in this sense, the deter­
minations must be post hoc. Groups will diff
er in the amount of time spent 
in each stage of development. Therefore, direct comparisons across groups 
using only time of existence as the independent variable will be difficult and 
must represent approximations. It is tempting to try to force a more con­
crete calendar onto the group development process. However, this would 
not adequately represent the process and such efforts would not help in 
understanding group development. While the fuzziness of the process is a 
limitation from an empirical sense, it is not fatal to research efforts, as we 
shall see. Predictions about the types of behaviour that should dominate a 
group or time period can be based on the model and these predictions can 
be tested. Analysis of group discussions and behaviours can identify stages 
and allow empirical tests of specific hypotheses. And the model can be used 
to explicate a wide range of intragroup and intergroup behaviours that 
cannot be easily explained by existing, more static approaches to group 
behaviour. 
Group identity: some empirical evidence 
Enough of this talk of 'fuzzy stages' and observation! Let's examine some 
empirical data. But before undertaking this adventure, one point should be 
made. The theory of social identity and its close cousin, self-categorization 
theory (Turner, 1 987), deal with the individual. They argue that individuals 
are driven to develop an identity, and, as a result of this desire, their behav­
iour influences group and intergroup actions. Individuals are the driving 
forces and groups are their vehicles. There is a great deal of support for this 
position, and I have no desire to attack it. However, my observations of 
groups suggest that the theories may be incomplete as they stand. For lurking 
in the background of the research on group development is the suggestion 
that groups, like individuals, strive to develop their identity, separate from 
that of other groups and apart from the unique identities of the individual 
members. 
Indeed, I would suggest that groups are engaged in a struggle with other 
groups and with individual members to establish and maintain a group 
identity. The group identity often outlives the members. It is displayed in 
symbols such as flags, anthems, laws, group structure and costumes that are 
enduring and not subject to the whims of individual members in search of 
their own identities. Groups strive for their independence from other 
groups, and they struggle with group members to keep the identity of the 
group equal to, if not more important than, individual identity. The striking 


66 
Social identity: international perspectives 
motto of the Three Musketeers drives home this point: 'All for one, one for 
all [the group].' 
Indeed, the model of group development may be viewed as a model of 
group identity. The stages represent the group's struggle to establish its 
character, which is threatened by demands from individual members and 
from outside groups. In fact, it is this conflict that ultimately gives the 
developmental process its energy, its dynamic. Where social identity theory 
would suggest that groups change and evolve to meet the demands and 
desires of their members, the group developmental approach argues that 
there is a more reciprocal relationship between the group and its member; 
the needs of individuals affect the development of the group, but the group, 
too, is often the puppeteer, guiding the actions of the members to meet the 
needs of the group. 
This approach argues that the behaviour of individual members is 
affected by the stage in which the group happens to be. Most pointedly, 
when developmental patterns or external events challenge the group to 
establish a clear identity, individual members will respond to support this 
goal. Specifically, this position predicts that in the early stages of group 
development, when the focus is on group identity, group members will 
direct both their intragroup and intergroup behaviours towards group 
identity. These forces stand apart from the need of the individuals to 
develop their personal or social identities. In essence, when the group is bent 
on crafting its identity, individual members are forced to focus on the social 
(group) side of their identities. 
During the early stages of group development, behaviours within the 
group and between groups should be directed towards securing the group's 
identity. On the intergroup side, there should be a rise in intergroup 
competition and conflict designed to create clear group boundaries. Within 
the group, individuals will be 'encouraged' to sacrifice their independence 
for the sake of creating a unified, clearly distinguishable group identity. 
Such actions as high conformity, swift rejection of deviants and perceptions 
of ingroup will foster this goal. Once the identity of the group is achieved, 
the group environment can become one that supports the goals of individual 
members to establish their own unique identities (social and/or personal). As 
a result, latter periods of group development should be characterized by 
fewer demands for personal sacrifice. Three studies, focusing on very 
different aspects of group behaviour, can be offered as evidence of this 
position. 
Preferred intergroup interaction 
Relations between the ingroup and outgroups are important and different at 
each stage of development, but these relations are most vital during the 
stage of identification. Nearly every aspect of group identity is defined in 
terms of the outgroups. As individuals, we may have difficulty defining our 


A developmental view o
f
 the search f
or identity 
67 
place in the universe, but at least we have our skin that defines our physical 
being. Groups, however, have no predetermined boundaries. Hence, not 
only are they faced with establishing their physical boundaries, but they 
must, in a sense, create their own skin. Indeed, this is often the first step in 
establishing the identity of the group. 
Establishing group boundaries not only involves determining who is in 
the group, but it also requires describing who is not a group member. One 
convenient means for making a clear designation of group boundaries is 
through intergroup conflict and competition. Such conflict plays multiple 
roles in the identification process: it forces individuals to declare their 
membership in one group or another; it marks the boundaries of the groups; 
it creates a common bond and purpose between members of the ingroup; 
and it forces the group to clarify what it stands for and what it stands 
against. Therefore, it can be predicted that group members will help their 
groups establish boundaries by seeking conflict and competition with out­
groups during the early identification stage of group development. 
On the other hand, the model of development predicts that intergroup 
conflict will be increasingly distasteful as the group continues its develop­
ment. During the group productivity phase, such conflict will distract 
members from task-oriented activities, although it may still serve the 
purpose of enhancing cohesion within the group. During individuation, 
members are concerned with making social comparisons with members of 
their own group and with members of the outgroup. In attempting to 
establish their worth to their ingroup and negotiating rewards, they must 
often demonstrate their potential value to other groups. Therefore, they 
need information from and acceptance by outgroups. Intergroup conflict 
would be disruptive to this purpose. And during decay, members are 
actively seeking alternative groups. Conflict between their present group 
and the alternative groups would clearly work against their individual goals. 
In an effort to examine these predictions from the model, groups of five 
students assembled to work together for three one-hour sessions. They were 
informed that other groups would also be working in similar settings during 
this period. The groups were allowed to choose from a variety of tasks on 
which to work and they could spend as long as they desired on each task. 
Their behaviour was video-taped during all the sessions and they responded 
to periodic questionnaires throughout. 
An initial analysis of the sessions indicated that the groups were quite 
similar in their pattern of development. During the early periods, time was 
spent getting acquainted, exploring similarities in backgrounds and 
interests, establishing norms and roles, discussing the uniqueness of their 
own groups, and agreeing to best the actions of other groups. As expected, 
the next order of business was on tasks: members examined and discussed 
the options, chose tasks and began talking about their past experience with 
similar tasks, their contributions to the group work, and their expectations 
for the next task. In several of the groups, members jokingly threatening to 
'send' productive members to the outgroups, and also in the form of jest, 


68 
Only 
cooperation 
t 
I 
Only 
Social identity: international perspectives 
competition '--
_
.L..-
_
--'--
_
-'-
_
---'--
_
----'-
_
----"-
_
---'
 
_
_
 '--
_
-'---
----l
 
Beginning 
of group 
Ending 
of group 
ā - - - - - - - - - - - - - - - - - - - - - - - - - - -
Time of group existence 
Figure 5.2 
Nature o
f
 desired interaction with outgroup as a f
unction o
f
 
group lif
e ( Worchel, Coutant-Sassic & Wong, 1993, p. 84) 
some members threatened to defect to the outgroups if their present group 
did not treat them well. During the last session, many of the groups engaged 
in a review of the group efforts, members talked about what they were 
going to do after the experiment, and closure was the topic of the period. 
Most interesting from the perspective of intergroup relations were 
responses on the written questionnaires. Several of these questions queried 
subjects about their feelings about their own group and desired relations 
with the other groups. As can be seen in Figure 5.2, subjects expressed a 
desire to engage in intergroup competition during the early periods of the 
group. As the group aged, their desires changed markedly, and by the time 
of group closing the desire was for cooperation with outgroups. Feelings 
about internal conflict showed the opposite pattern: extreme discomfort 
during the early stages and greater comfort by the final period (Figure 5.2). 
This is the pattern that would be predicted by the model on the assump­
tion that intergroup competition helps differentiate the ingroup from the 
outgroup, and such differentiation is most critical during early stages 
devoted to group formation. The topics of group discussion also fit the 
predicted pattern, focusing first on group identification and productivity 
and later on more individual concerns. It should be pointed out that the 
cyclical pattern of the stages was not observed, but this may have been due 
to the fact that these groups had a defined end point. 
These results are not only interesting from a theoretical standpoint, but 
they may also carry a message for applied group dynamics. If groups 
actively seek out competition and conflict with outgroups at certain periods, 
it follows that efforts to reduce intergroup hostility during these periods will 
not be well received. Reducing intergroup hostility may only be possible 
after groups have been able to establish their identity, and do not f
eel that 
interaction with the outgroup will threaten that identity. Sherif et al. (1961) 


A developmental view o
f
 the search f
or identity 
69 
found that intergroup hostility was reduced only after several events 
involving intergroup cooperation towards a superordinate goal. One expla­
nation for these results is that several cooperative encounters are necessary 
before hostility is reduced. Another explanation argues that the timing of 
the cooperative endeavours is the critical factor in ameliorating this 
hostility. Sherif's groups may have been most accepting of those final 
cooperative situations which occurred towards the end of the session. 
Ingroup and outgroup perceptions 
A number of studies have found that groups see greater homogeneity in the 
characteristics of the outgroup than within their own group (Jones, Wood & 
Quattrone, 198 1 ;  Linville & Jones, 1980; Park & Rothbart, 1 982). Explana­
tions for the effect include the fact that people have more and varied 
contact with ingroup members than with outgroup members. Therefore, 
they can readily see the variability of behaviour of ingroup members, but do 
not experience this variability in the limited contact with outgroup mem­
bers. Further, the contact with ingroup members generally occurs in a 
variety of situations, while contact with the outgroup is often in a single 
type of situation (Quattrone, 1 986). 
This interesting effect poses potential problems for the present model of 
group development. The model suggests that during periods of identification 
ingroup members are most interested in finding commonalities between 
their own group so as to create a group identity. However, later in the 
developmental cycle, especially during individuation and decay, the model 
suggests that members will be searching for differences between ingroup 
members, thereby enhancing their own individuality and giving them reason 
to search for other groups. The basic prediction, then, is that perceptions of 
homogeneity within groups will be dependent on the period of group 
development in which these perceptions are measured. Perceptions of 
ingroup homogeneity should be greatest during the identification stage and 
least intense during individuation and decay stages. 
In order to examine this prediction, groups of eight subjects worked on a 
series of tasks (Worchel, Coutant-Sassic & Grossman, 1992). They were 
then divided into two groups of four, and the two groups competed on 
other tasks. Subjects were asked to rate ingroup and outgroup members on 
a number of dimensions at different times during the competition (immedi­
ately at the start of competition, after one competitive encounter and after 
two competitive events). A final condition involved assigning subjects to 
two groups on arrival, never allowing them to work as a single group, and 
then measuring their perceptions after one competitive event. Based on the 
model of group development, it was predicted that members would perceive 
ingroups as more homogeneous than outgroups during the early phases of 
competition. This was the situation where groups should be most concerned 


70 
Social identity: international perspectives 
Table 5. 1 
Perceived dif
f
erences in ingroup and outgroup similarity 
Traits 
Condition 
Intelligence" 
Friendliness 
Appearance 
Combined-questioned 
immediately 
3.78b,c 
3.84 
1 .24 
Combined-questioned 
after one task 
2.97 
4.03 
0.97 
Combined-questioned 
after two tasks 
-3. 1 1  
-1 .66 
-0.67 
Never combined 
-2.69 
-2.07 
-2.00 
Talkativeness 
2.47 
1 .33 
-0.1 2  
-1 .25 
" Subjects were asked: How similar were the members of your group (outgroup) on the 
following dimensions: intelligence, friendliness, personal appearance and talkativeness. I = 
very dissimilar, 10 = very similar. 
b Scores were determined by ingroup-outgroup. Positive score indicates ingroup more similar; 
negative score indicates outgroup perceived as being more similar. 
c N = 6 groups in each condition. 
Source: Worchel, Coutant-Sassic & Grossman, 1992, p. 198. 
with identity, especially if the group had just been formed from a larger 
group, Later in the competition, there would be pressures to perceive the 
differences within the group as the members focused on task and member 
skills and contributions. 
As can be seen in Table 5 , 1 ,  the results supported these predictions and 
they illustrate the need to consider group development when examining 
group perceptions, Ingroups were perceived as more homogeneous than 
outgroups during the early stages of competition, but the ingroups were 
seen as more heterogeneous than outgroups during the latter stages of 
competition or when the groups had existed as separate units from the 
beginning of the experiment. It is also interesting that groups perceived 
greater absolute differences between the ingroup and outgroup during the 
early stages of competition than during later stages. 
Responses to minorities 
There is voluminous literature demonstrating that minorities can exercise 
considerable influence over group decisions (Moscovici, 1 976; Moscovici & 
Mugny, 1 983; Moscovici & Nemeth, 1974; Nemeth & Staw, 1 989), The 
focus of much of the research is on the most effective tactic for minorities to 
follow (Maass & Clark, 1 984) and on explaining the cognitive impact of the 
minority (Nemeth, 1 986). Surprisingly, the research has not examined many 
group factors in demonstrating the impact of the minority, In fact, minority 
influence work has been conducted outside of group settings altogether. The 
typical experiment examines how subjects respond to learning of a statis­
tically rare or unexpected position, While the minority influence paradigm 


A developmental view o
f
 the search f
or identity 
7 1  
had generated exciting data, it is important to bring the research into group 
settings. Indeed, minorities are defined by their relation to groups. 
The present model of group development suggests that the minority plays 
an important role in the group process, but that the impact of the minority 
is influenced by the stage of group development. On one hand, the voice of 
dissent represents a clear threat to group cohesion and identity. Minorities 
differentiate members within the group, and may present an image of a 
vulnerable or disorganized unit. On the other hand, minorities can offer 
valuable information and new perspectives to the group. Their unmolested 
existence signals that dissent, disagreement and dissatisfaction with the 
group are acceptable. Given these functions, it can be predicted that 
minorities will be more influential and accepted in the group during the 
stages of productivity, individuation and decay. On the other hand, 
minorities will be quickly rejected during the identification stage when the 
group is concerned with developing cohesion and uniformity. Furthermore, 
the influence of the minority will be dependent on the group's previous 
experience with minorities. If the experience has been that minorities are not 
disruptive and help move the group towards its goals, minorities will be 
more readily accepted than when the group's previous experience with 
minorities has been negative. 
My students and I have completed several studies examining minority 
influence. In one (y.I 
orchel, Grossman & Coutant-Sassic, 1993) we found 
that minorities were most influential in groups if they had previously 
demonstrated that their positions were correct and the issue of influence 
dealt with productivity, not group identity. Minorities who were perceived 
as a threat to group identity were excluded from the group; indeed, they 
were viewed as forming another outgroup. It was clear in this study that 
minorities could exercise greater influence on some issues than others and 
that the group focus or activity determined which issues were subject to 
influence. Previous work by Crano (1989) also found that minorities were 
more influential on some issues (objective) than others (subjective). This 
research extends these findings by suggesting that when examined in a group 
context the group will determine the issues on which minority opinion will 
be accepted. 
In a second study (Coutant, Worchel & Grossman, in preparation), 
groups of five subjects met together weekly over a period of six weeks. In 
each session, they were given tasks on which to work. At different points 
during the life of the group, various manipulations were executed; the main 
independent variable was the time at which the manipulation took place 
(during the first session, during middle sessions or during the last session). 
In the case of minority influence, the groups were given problems to discuss 
and required to arrive at a group decision about the solution. A confederate 
in the group played the role of the minority, and after ascertaining the 
modal opinion he proposed an extremely different position. The major 
dependent measure was the influence of the minority on the group decision 
at different periods of the group's life. 


72 
Social identity: international perspectives 
Table 5.2 
Minority in
fluences in groups 
over time 
Time 
Early 
Middle 
Late 
4.24 
4.67 
4.85 
Note: Responses represent the group decision 
on how to deal with a case at hand. Minority 
took a position of lO on a scale of 1 - 1 0, 
therefore higher numbers indicate closer 
agreement with the minority position. 
As can be seen from Table 5.2, the minority had little influence in the 
initial meeting of the group; in fact, the group moved away from his stated 
position. The minority was most influential when the influence attempt 
came during the closing group session. Experience with these ongoing 
groups indicated that the initial sessions were devoted to issues of group 
identity, while in the closing session individual members focused more on 
personal issues and the ending of the group. Hence, these results are in line 
with predictions based on the model of development: minorities will have 
little influence during identification stages and greatest influence during 
periods of individuation and decay. In addition to supporting the specific 
predictions taken from the model, this study and one finding similar results 
(Coutant-Sassic, 1 991) illustrate the need to consider the factor of group 
development when discussing a group 'effect' such as minority influence. 
We cannot speak of minority influence as a general effect; rather the influ­
ence is closely tied to the issue and to the development of the group. In fact, 
it might be interesting to speculate that the process of influence differs under 
the different conditions. The present view of minority influence (Nemeth, 
1986) is that it involves a single process: creating divergent thinking in the 
majority. I would suggest that during the stage of identification when 
the minority is viewed as a threat to group identity convergent thinking 
on the majority position results from the presence of a minority point of 
view. 
The task ahead 
It is time to take stock of where we now stand. In a general sense, the 
present model can be viewed as a model of group social identity. Groups, 
like individuals, must establish their own identity. This identity has two 
dimensions: the identity of the group (similar to the personal identity of 
the individual in social identity theory) and the identity of the group as it 
relates to other groups in its universe (similar to the social identity of 
individuals). The search for identity influences both ingroup and intergroup 


A developmental view o
f
 the search f
or identity 
73 
behaviour as one identity or the other predominates. By way of summary, I 
have attempted to establish a number of points concerning group and 
intergroup behaviour. First, individuals are engaged in a constant struggle 
of balancing their desire to belong to groups and the concern of establishing 
their personal independence and uniqueness. Events or situations that 
threaten their social identity will lead individuals to emphasize their rela­
tionship with their groups. Likewise, threats to their personal identity will 
motivate individuals to seek more intimate interactions, making their 
personal characteristics more salient. This struggle is felt at the group level 
by members alternating between being active participants in group activities 
and withdrawing themselves, physically and/or psychologically, from the 
group. 
Second, groups are dynamic units that develop through a predictable 
series of stages. The development process is cyclical in that it can repeat 
itself many times during the life of the group. The repetition occurs as the 
group loses old members and gains new ones. The amount of time the 
group spends in any single stage is dependent on numerous intragroup and 
intergroup factors. The transition between stages can be abrupt and clear or 
slow and gradual. Numerous events can trigger the movement from one 
stage to another, including success, failure, leader behaviour, time pressures 
and deadlines, and intergroup events. 
Third, stages can be identified by the issue that dominates group behav­
iour and discussion and involves the most group members. Stages not only 
influence the focus of group activity, they also impact the interpersonal 
behaviours that will characterize the group. For example, minority influence 
will be most evident during individuation and decay stages, while con­
formity will be most common during the identification stage. Relating 
developmental stages to individual and interpersonal behaviour offers fertile 
grounds for developing specific predictions. 
Fourth, the development of the group is keenly affected by intergroup 
relations, and these intergroup relations are also influenced by the group's 
development. For example, the group will be most concerned with diff
er­
entiating itself from outgroups during the identification period. This 
differentiation process will be seen in efforts by the group to instigate 
conflict and competition with the outgroup. On the other hand, perceived 
threats to the group's identity from the outgroup may throw the group into 
the identification stage, regardless of when these threats occur. 
Finally, the developmental approach to group behaviour offers insights 
that may be used in intervention strategies. Those strategies that are 
congruent with the group's stage of development will be most effective. For 
example, it was suggested that prejudice and discrimination could be most 
effectively reduced by interventions during the individuation or decay stages 
of development. 
These are just a few of the advantages that result from expanding the 
scope of social identity theory. The theory has a solid base, and it is time to 
explore pushing its boundaries. 


74 
Social identity: international perspectives 
Note 
Work on this chapter was supported in part by a research grant from the Texas Coordinating 
Board of Higher Education (T 
ARP). The chapter was completed while I was a Visiting 
Lecturer at the Department of Psychology, University of Waikato, Hamilton, New Zealand. I 
wish to express my appreciation for the support I received from my colleagues at the University 
of Waikato. I also want to acknowledge the help of several students who have worked with me 
on developing the ideas presented in this chapter: Dawna Coutant, William Webb, Frankie 
Wong, Michele Grossman, Judy Ouellete, Greg Pool and Leah Worchel. 


6 
The Role of Prototypicality in Group 
Influence and Cohesion: Contextual 
Variation in the Graded Structure of 
Social Categories 
Penelope Oakes, S. A lexander Haslam and John C. Turner 
The concept of prototypicality 
The concept of proto 
typicality was one product of the 'Roschian revolution' 
(Neisser, 1 987, p. vii) of the 1 970s, which introduced some key new ideas 
to the analysis of category structure and representation. As is now well 
known, Rosch challenged the philosophically based, classical view of 
category representation with her ideas about, and evidence of, extensive 
variation in category structure. She and her colleagues (for example, Rosch, 
1978; Rosch, Mervis, Gray, Johnson & Boyes-Braem, 1 976) found that 
comparisons between categories reveal that they vary in their relative 
inclusiveness, for example 'dalmation', 'dog', 'animal', 'living thing', are 
categories of increasing inclusiveness. Those at a higher level are called 
superordinate, compared to lower level categories known as subordinate. 
Around the middle of the hierarchy of inclusiveness is the basic level (for 
example, 'dog', 'chair', 'tree') at which, it is argued, most perceptual activity 
takes place (Mervis & Rosch, 1 98 1 ;  see also Lassaline, Wisniewski & 
Medin, 1 992). More importantly for our present purposes, Rosch also 
found that within categories, members vary in their typicality, for example 
robins are seen (by American subjects) as more typical of the category 'bird' 
than are ostriches, but both robins and ostriches are recognized as sharing 
membership in the one 'bird' category. In other words, it appears that 
categories have an internally graded structure, rather than members possess­
ing an even level of shared defining attributes, as in the classical view. It was 
this finding that led to the idea of a category prototype (a best example of 
the category), and the argument that category membership requires a 
certain level of similarity to the prototype. Rather than being tightly 
bounded, categories seem to take the form of 'fuzzy sets', much looser 
groupings which depend upon what Rosch called 'family resemblance', that 
is, members do not all share a given set of defining features, but are related 
through their similarity to a prototype. 


76 
Social identity: international perspectives 
In her most widely cited paper Rosch (1978) tried to counter some 
possible misunderstandings of her ideas. In particular, she emphasized that, 
although her work indicated that prototypes served important functions in 
categorization, the concept did not 'constitute any particular model of 
processes, representations, or learning' (p. 40). Indeed, she commented: 
To speak of a prototype at all is simply a convenient grammatical fiction; what is 
really referred to are judgements of degree of prototypicality 
. . . .  
to speak of a 
single entity that is the prototype is either a gross misunderstanding of the 
empirical data or a covert theory of mental representation 
. . . .  
In short, proto­
types only constrain but do not specify representation and process models. 
(pp. 40-4l )  
Prototypicality in self-categorization theory 
Self-categorization theory (Turner, 1985; Turner, Hogg, Oakes, Reicher & 
Wetherell, 1987) is often introduced as a theory of processes underlying 
psychological group formation, and indeed its original mission was to 
explain the psychological basis of the group. In effect, however, it attempts 
more than this, and one of its most important contributions has been the 
development of a new perspective on the functioning of categorization 
processes in social perception and interaction (cf. Brewer, 1988; Fiske & 
Neuberg, 1990). We shall not attempt to provide a comprehensive overview 
of the theory here, but it is worth setting down some of the main points 
concerning categorization in order to place our discussion of prototypicality 
in its proper context (for more detailed discussion of categorization, see 
Oakes, 1996). 
In discussing the role of categorization in social perception, the theory 
applies and develops Rosch's ideas about variation in both relative 
inclusiveness across categories and relative prototypicality within categories. 
The theory begins with the assumption that self-conception reflects self­
categorization, the cognitive grouping of the self as identical to some class of 
stimuli in contrast to some other class of stimuli. Following Rosch (1978), 
self-categories are assumed to exist at different levels of abstraction related 
by class inclusion. That is, a given self-category (for example, 'scientist') is 
seen as more abstract than another (for example, 'biologist') to the extent 
that it can contain the other, but the other cannot contain it: all biologists are 
scientists, but not all scientists are biologists. Going beyond the simple 
distinction between personal identity (self as a unique individual) and social 
identity (self as a member of a social group) presented in earlier work (for 
example, see Turner & Giles, 1981), self-categories both more and less 
abstract than personal and social identity are envisaged, but for purposes of 
theoretical exposition three levels of abstraction of self-categories are 
distinguished: the interpersonal (subordinate level of abstraction, personal 
identity, self as an individual person), intergroup (intermediate level of 


The role o
f
 pro 
to 
typicality 
77 
abstraction, social identity, self as a group member) and interspecies (super­
ordinate level of abstraction, self as a human being). These are defined not by 
specific attributes but by the level at which people are being compared and 
categorized. For instance, 'altruism' could function as a cue to individual 
identity, to a particular social category, or to being human, depending on the 
context. 
In terms of the processes underlying group formation, the central 
explanatory idea in the theory is that group behaviour is made possible by 
and reflects self-concept functioning at the level of ingroup-outgroup (rather 
than interpersonal or interspecies) categorizations. Developing TajfeJ's 
(1978d) idea of an interpersonal-intergroup continuum, self-categorization 
theory contrasts the social and personal (ingroup-outgroup and inter­
personal) levels of self-categorization, and hypothesizes that shifts between 
these levels are a varying outcome of a continual conflict between personal 
and social identity as they vary in relative salience. Shifts towards social 
identity produce depersonalization of self-perception and behaviour, that is, 
self-stereotyping, perception of increased identity between the self and 
ingroup members and increased difference from outgroup members (on 
relevant dimensions). The theory proposes that it is this process of 
depersonalization that makes group behaviour possible and produces its 
emergent, irreducible properties. 
In retaining (from social identity theory) this emphasis on the two levels 
of personal and social identity, the theory appears to be suggesting that 
there are, in Rosch's terms, two 'basic' levels in the cognitive representation 
of the self and others, the social and the personal. In truth, however, this 
distinction is maintained only for purposes of theoretical clarity in 
explaining the hypothesized basis of psychological group membership. One 
of the essential points of the self-categorization analysis is that the appro­
priate level o
f
 categorization varies with the context, and that to speak of any 
level as more 'basic' than another is misconceived. 
To develop this point further, the theory portrays categorization as a 
dynamic, context-dependent process, determined (in part) by comparative 
relations within a given context (for example, see Haslam, Oakes, Turner & 
McGarty, 1 996). This approach is formalized in the principle of meta­
contrast, which is so called because it involves a contrast between contrasts, a 
judgement of difference between differences. The meta-contrast principle 
predicts that a given set of items is more likely to be categorized as a single 
entity to the degree that dif
f
erences within that set o
f
 items are less than 
dif
f
erences between that set and others within the comparative context (cf. 
Campbell, 1 958; Rosch, 1978; Taj fel 
, 1969a; Tversky, 1 977). This principle 
encapsulates but subtly transforms the classic idea that categories form on 
the basis of intraclass similarities and interclass differences. It goes further 
than earlier work in making clear that such similarities and differences are 
not independent and additive, but aspects of the same meta-contrast (cf. 
Medin, 1 989; see below). The meta-contrast principle is explicit that 
categorization is relative to the frame of reference and hence to the contrasts 


78 
Social identity: international pers
pectives 
Figure 6.1 
A categoria! hierarchy o
f
 vegetation 
available within the salient stimulus field. It depends not just on similarities 
between stimuli (as is so often assumed), but on relative similarities, more 
similarity between certain stimuli than between others. Meta-contrast 
contextualizes categorization by tying it to an on-the-spot judgement of these 
relative similarities and differences. 
For example, consider the system of natural categories in Figure 6. 1 .  
Meta-contrast predicts that we would categorize and perceive a given piece 
of vegetation as, say, a 'tree' (rather than as 'vegetation') to the extent that, 
in the current comparative context, the differences between trees (oaks, 
birches, and so on) are perceived to be less than the differences between 
'trees' and 'shrubs' - the distinction between trees and shrubs is more 
marked, and more relevant, than are the features that trees and shrubs 
share as 'vegetation'. Alternatively, the tree might be categorized and per­
ceived simply as 'vegetation' to the extent that differences between types of 
vegetation (trees, shrubs, and so on) are seen to be less than the differences 
between vegetation and, say, animals. The salient categorization will be 
'oak' when perceived differences between individual oak trees are less than 
the differences between oaks and birches (or some other comparison 
species). 
The meta-contrast principle applies similarly in person perception. 
Consider the hierarchy in Figure 6.2. Here, we might categorize an 
individual as 'Spanish' to the extent that, in the current comparative 
context, perceived differences between individual Spaniards (Carmen, 
Isabella, and so on) are less than those between Spaniards and Germans. 
Alternatively, the salient category might be 'European' in a context 
where perceived differences between various European groups (such as 
Spaniards and Germans) are less than those between Europeans and non­
Europeans. 
These ideas can be expressed in terms of a meta-contrast ratio, that is, the 
average perceived intercategory difference divided by the average perceived 
intra-category difference. For example, this might be: 


The role o
f
 pro 
to 
typicality 
Figure 6.2 A categorial hierarchy o
f
 Europeans 
average perceived difference between Spaniards and Germans 
average perceived difference between individual Spaniards 
79 
The higher this ratio, the more likely it is that Spaniards will be perceived in 
terms of their shared national identity. Turner and Oakes (1989, p. 242) 
provide specific examples of how meta-contrast can be calculated, as do 
Haslam and Turner (1 992) and some of the papers exploring prototypicality 
in more detail that we shall discuss below (e.g. McGarty, Turner, Hogg, 
David & Wetherell, 1 992). 
Note, however, that meta-contrast provides only a partial account of 
categorization. It describes the comparative relations between stimuli which 
lead them to be represented by a category, but it is also important to take 
into account the social meaning of differences between people in terms of 
the normative and behavioural content of their actions, and the relative 
accessibility of particular categorizations (Oakes, 1987; Oakes, Turner & 
Haslam, 1 991). In general terms, the theory explains the salience of any 
given category as a function of an interaction between its relative accessi­
bility (the 'readiness' of a perceiver to use a particular categorization) and 
the 'fit' between the category and reality. Meta-contrast is the theory's 
principle of 'comparative fit' (the match between category and the com­
parative properties of stimuli), but 'normative fit' (the match between 
category and the content properties of stimuli) is also always inseparably at 
work (Oakes et aI., 1991). 
This emphasis on categorization as highly variable and context-dependent 
produces a concomitant emphasis on the context-dependence of perceived 
similarity and difference, the major outcome of categorization. People who 
are categorized and perceived as different in one context (for example, 
'biologists' and 'physicists' within a science faculty) can be recategorized 
and perceived as similar in another context (for example, as 'scientists' 
rather than 'social scientists' within a university) without any actual change 


80 
Social identity: international perspectives 
in their own positions (Gaertner, Mann, Murrell & Dovidio, 1989). This is 
the essence of categorization: it is a cognitive grouping process that trans­
fonns differences into similarities and vice versa. We need some psycho­
logically neutral tenn such as perhaps 'distances' to indicate precognized 
stimulus relations: there are distances between people, but are they 
similarities or differences? Are physicists and biologists similar or different? 
Arising from the comparisons specified in the meta-contrast principle, 
categorization subjectively transfonns 'distances' into similarities and differ­
ences, and from perceived similarities and differences flow, amongst other 
things, perceptions of attraction and dislike, agreement and disagreement, 
cooperation and conflict. In sum, categorization provides the fundamental 
basis of our social orientation towards others. Within the science faculty, 
physicists might reject and deride biologists, claiming they aren't 'real 
scientists', but in comparison with social scientists the two groups may 
present as inseparable allies. 
As changes in the comparative context produce changes in the ingroup­
outgroup relationships, they also affect intra-category structure, and this is 
where the concept of prototypicality becomes important. The meta-contrast 
principle can also be used to define the relative prototypicality of members 
within a group (see Turner & Oakes, 1989, pp. 259-265). In general tenns, 
the more a group member differs from outgroup members and the less he or 
she differs from other ingroup members (that is, the more this person 
exemplifies what in 
group members share and what they do not share with 
the outgroup), the more that individual will be perceived as prototypical of 
the group. So, for example, in a communist party, members wishing to be 
perceived as prototypical communists must differ in politics from members 
of more conservative, capitalist parties, but they must not be so ultra-left 
that they begin to differ significantly from other communists - prototypi­
cality depends upon both inter- and intragroup comparisons. One important 
point is that, because relative prototypicality depends on (amongst other 
things) intergroup comparisons, it will vary along with variation in the 
intergroup context in which judgements are made. For example, the proto­
typical communist in a comparative context including fascists will be 
dif
f
erent from the prototypical communist as compared with liberal 
democrats, or Trotskyists. Self-categorization theory thus agrees with Rosch 
(1978) that fixed prototypes are 'fictions', and emphasizes context­
dependent judgements o
f
 proto 
typicality rather than fixed prototypical 
images which represent groups as constants across changing contexts (cf. 
Brewer, 1988; Brewer, Dull & Lui, 198 1 ;  see below). 
There are further aspects of the analysis of categorization presented within 
the theory (for example, the issue of the relationship between the different 
levels of abstraction, the dependence of comparison and categorization upon 
identity at a higher level of abstraction), but these issues are not central to 
our current discussion and are covered elsewhere (for example, Oakes, 
Haslam & Turner, 1994; Turner & Oakes, 1989). The major point to 
emphasize here is the context-dependence and fundamentally comparative 


The role o
f
 prototypicality 
81 
nature of 
the categorization process, and the consequent variability of 
both 
the level and the content (including the prototypical position) of social 
categories. 
Prototypicality and group polarization 
Group polarization refers to the tendency for group interaction (for 
example, discussion or some related activity) to move the average of group 
members' responses closer to the extreme of an already favoured position. 
For example, after discussing feminism a group of moderately pro-feminist 
individuals would tend to become, on average, more pro-f
eminist than they 
had been initially. There is a large and complex literature related to group 
polarization (Turner, 1 991), and our aim here is not to review that 
literature, nor to enter into any debate about the relative merits of different 
explanations of the phenomenon, but simply to show how the concept of 
prototypicality as conceptualized in self-categorization theory has enabled 
resolution of the apparent contradiction between polarization and other 
social influence phenomena (for more detail see Turner, 1991 ; Turner & 
Oakes, 1989; Wetherell, 1987). 
Group polarization has been seen as inconsistent with a central tenet of 
the traditional informational influence model of conformity derived from 
the findings of Allport, Sherif and Asch - the idea that social norms form 
through a process of interpersonal averaging or convergence. Norms are 
thought of as reflecting what the group shares, and this has been concep­
tualized in terms of a simple conglomeration of what each separate indi­
vidual brings to the group, specifically the average of individual members' 
separate views. Given this view of norms, specialized theories have had to 
be developed to explain polarization as something more than or different 
from conformity to group norms (Burnstein, 1982; Sanders & Baron, 1 977; 
Wetherell, 1 987), because in polarization group members do not converge 
on the average of their individual positions. Something occurring within the 
group moves them towards a position more extreme than the average. 
It has long been evident that polarization could be explained as conformity 
if it could be shown how ingroup norms could sometimes be more extreme 
than the mean (for example, Singleton, 1979). To do this, one needs to 
rethink the interpersonal averaging approach to the development of group 
norms, and there are additional reasons for doing this. The interpersonal 
averaging approach is individualistic (Turner & Oakes, 1986), implying that 
social influence is simply a 'change in individuals induced by individuals' 
(Kiesler & Kiesler, 1969, p. 26). In contrast, self-categorization theory argues 
that group behaviour (including social influence) reflects and is made 
possible by a change in the level of abstraction of self-categorization, a 
qualitative shift in both self-conception and the nature of relations with 
others. In this sense, group interaction does not involve individuals whose 
idiosyncratic opinions can be averaged, but group members whose behaviour 


82 
Social identity: international pers
pectives 
will be guided by the distinctive, emergent, irreducible properties of their 
group. The group norm embodies those group properties, and reflects both 
what group members have in common (on dimensions relevant to the group) 
and the way in which the group can be distinguished from relevant 
outgroups. Group norms (and other distinctive group properties such as 
stereotypes) reflect the comparative relations within which the group defines 
itself as well as, and as much as, they reflect attributes of group members. 
More specifically, the position perceived as normative for a group in a 
given context will be the position that is most prototypical in that context, 
as defined by the principle of meta-contrast (see above). It will be the 
position which is, on average, most diff
erent from that of relevant outgroup 
members and least diff
erent from that of other ingroup members (obtained 
by dividing the individual's average difference from outgroup members by 
his or her average difference from ingroup members). Of course, the posi­
tion least different from that of other ingroup members is the mean, but the 
mean may not be the position most different from the outgroup. Thus, 
under certain intergroup conditions (see below) the prototypical position 
will be skewed away from the mean. 
The extent to which the mean does not represent the position most 
different from the outgroup, and hence the degree of discrepancy between 
the most prototypical position and the mean, depends on the comparative 
context within which the ingroup defines itself. In general, the ingroup 
prototype will tend to coincide with the mean where the mean is at the 
midpoint of the comparative context, that is, the ingroup is occupying the 
'middle ground' on some issue relative to relevant outgroups. When ingroup 
responses occupy a more extreme position in the comparative context (as a 
function of either ingroup responses shifting towards an extreme, or a 
change in intergroup comparison making moderate responses appear more 
extreme), it becomes more likely that the most prototypical position will 
tend to be more extreme than the mean. For example, insofar as a group of 
moderately left-wing students is seen to occupy the middle ground in 
student politics within their institution (that is, there are other salient groups 
to both the left and the right of the moderates), the average view in the 
group will be more prototypical. However, if either individual opinions in 
the group shift to the left, or the comparative context comes to be domi­
nated by, say, fascists and other extreme conservatives, a more extremely 
left-wing position within the moderate group will become prototypical. The 
average position would no longer adequately represent the difference of the 
moderates from the right-wingers. Examples specifying this process of 
calculating the prototypical position (and its variation from the mean 
position) through the principle of meta-contrast can be found in McGarty et 
al. (1992), and a number of other sources (for example, Turner et aI., 1987). 
Several studies provide evidence consistent with the self-categorization 
analysis of group polarization (for example, Abrams & Hogg, 1990; Abrams, 
Wetherell, Cochrane, Hogg & Turner, 1 990; Hogg, Turner & Davidson, 
1 990; Mackie & Cooper, 1 984; McGarty et aI., 1992; Spears, Lea & Lee, 


The role o
f
 pro 
to 
typicality 
83 
1990; Turner, Wetherell & Hogg, 1989; van Knippenberg & Wilke, 1 988). 
For example, McGarty et al. (1 992) present correlational evidence (from a 
computer simulation and two semi-naturalistic group discussion studies) that 
ingroup prototypes do become more polarized as the group becomes more 
extreme in the social context. They also show that the magnitude of group 
polarization is a function of the degree to which groups actually shift 
towards their prototypes together with the degree to which those prototypes 
are polarized. 
One novel prediction from the self-categorization analysis is that the 
direction of polarization can be reversed by changing the comparative 
context, for example moderate left-wingers should move to the left in a 
context dominated by the right wing, but to the right when comparing 
themselves with ultra-leftists. Consistent with this prediction, Hogg et al. 
(1 990) report findings from a decision-making study in which subjects 
perceived the ingroup norm as riskier than the pretest mean when con­
fronted by a more cautious outgroup, but as more cautious than the pretest 
mean when confronted by a riskier outgroup. 
In summary, self-categorization theory explains group polarization as 
conformity to an extremitized group norm, and explains the extremitization 
of the group norm through an account of the way in which comparative 
context influences norm formation. The concept of proto 
typicality has been 
important here in specifying the position that will be most normative for a 
given group in a given comparative context: the position with the highest 
meta-contrast (maximal similarity to ingroup positions and maximal 
difference from outgroup positions) is the most prototypical of the group 
and hence the most consensual and normative. 
Prototypicality and social attraction 
Prior to the 'cognitive redefinition' (Turner, 1982) of the social group 
represented by social identity theory and self-categorization theory, the 
sense of group belongingness and attraction within groups was understood 
to rest simply on inter
personal attraction between individual group mem­
bers. This was the essence of the traditional concept of group cohesiveness 
(Lott & Lott, 1 965), the affective and motivational 'glue' that held groups 
together. However, in the same way that the interpersonal averaging 
conception of group norms has been reconsidered in self-categorization 
theory, the interpersonal attraction account of group cohesiveness has also 
been challenged (Hogg, 1 987, 1 992, 1 993). It has long been evident that 
interpersonal attraction is neither necessary nor sufficient for group 
behaviour (for example, Hogg & Turner, 1985a; Turner, Hogg, Oakes & 
Smith, 1 984; Turner, Sachdev & Hogg, 1 983), but that, at the same time, 
there are strong bonds of attraction within groups (for example, Sherif, 
1967), including 'minimal' groups (for example, Brewer & Silver, 1 978; 
Brown & Turner, 1979; Doise, Csepeli, Dann, Gouge, Larsen & Ostell, 


84 
Social identity: international perspectives 
1972). Self-categorization theory has introduced the idea of depersonalized 
social attraction to distinguish between interpersonal liking (which seems to 
be unrelated to group processes) and the intragroup liking that accompanies 
salient group membership and group interaction. 
Hogg (1992) summarizes the self-categorization analysis of intragroup 
attraction thus: 
the core idea [is] that self-categorization depersonalizes the basis of interindividual 
attraction, such that ingroup individuals (and self) are liked in proportion to their 
perceived group prototypicality. (p. 125) 
Again, we have an emphasis on the qualitative distinction between group 
and individual behaviour. Group cohesiveness and intragroup attraction are 
based on group members' liking of each other as group members, not as 
unique individuals, and it is prototypicality that defines the extent to which 
individuals are or are not group members. Recall that prototypicality 
depends upon both similarity to other ingroup members and difference from 
the outgroup. The closer any individual group member comes to this ideal, 
maximally normative ingroup position, the more he or she will be liked by 
other ingroup members, and it is this qualitatively distinctive depersonalized 
social attraction that holds groups together, rather than the individual 
personal attraction discussed in earlier models (Hogg, 1 992, 1 993). 
Hogg and his colleagues have tested a number of predictions from this 
self-categorization model of depersonalized attraction (for example, Hogg, 
Cooper-Shaw & Holzworth, in press; Hogg & Hardie, 1991 ,  1992). Hogg et 
al. (in press, Study I )  tested the basic hypothesis that, where group mem­
berships rather than interpersonal relationships are salient, liking among 
group members is depersonalized in terms of the group prototype. They 
studied six small, mixed-sex interactive work groups within a large tertiary 
educational institution, who were asked to report patterns of intragroup 
liking under conditions that either heightened or reduced the salience of 
their work group. They also rated the cohesiveness of the group, described 
the group prototype, ranked the prototypicality of fellow members, rated 
their own prototypicality, and rated the subjective clarity of the group 
prototype. As predicted, it was found that group liking and interpersonal 
liking were independent, and that group liking was positively associated 
with the proto 
typicality of both the self and others, perceptions of strong 
group cohesiveness and a clear group prototype. Interpersonai liking, on the 
other hand, was either unrelated or negatively related to these variables, but 
it was associated with perceptions of interpersonal similarity. Overall, then, 
the study produced clear evidence of depersonalized, category-based intra­
group attraction as distinct from interpersonal liking based on interpersonal 
similarity. A second study involving 1 5  mixed-sex groups of between 6 and 
10 introductory psychology students produced similar findings. 
Hogg and Hardie (1991) report a field study of depersonalized attraction. 
Members of an Australian football team completed a questionnaire at a 
practice session which assessed subjective perceptions of prototypical 


The role o
f
 proto 
typicality 
85 
features of the team, and a rating of self-prototypicality. Team members 
then ranked each other in terms of prototypicality, social attraction (elicited 
under conditions of high team membership salience) and personal attraction 
(elicited where interpersonal friendships were emphasized). 
Team members' responses indicated significant agreement about the team's 
prototypical f
eatures. The most important were a sense of camaraderie 
(,mateship', and unqualified mutual support), task-oriented mutual encour­
agement ('revving-up') and participation in good-natured 'horsing-around' 
and socializing. As predicted, in the team-salient condition there was a more 
significant relationship between subjects' liking for others (social attraction) 
and perceived prototypicality than between interpersonal liking and 
perceived prototypicality, and this effect was most marked for members 
who identified most strongly with the team and saw themselves as highly 
prototypical. Social popularity (that is, being consensually liked as a team 
member) was also strongly related to consensual prototypicality, but the 
relationship between personal popularity (that is, being consensually liked as 
a personal friend) and prototypicality was weaker. 
In summary, Hogg and his colleagues present a strong case for distinction 
between social attraction and personal attraction, with the former uniquely 
related to group processes and involving the perception of self and others in 
terms of group prototypicality rather than individual characteristics. The 
mechanisms underlying the relationship between perceived prototypicality 
and attraction are being investigated further (see Hogg, 1992, 1993): for 
example, does mutual self-categorization produce attraction simply through 
the accentuation of prototypical similarity, or is group prototypicality in 
some way inherently attractive, and might this be because intergroup pro­
cesses work to produce ethnocentric prototypes which portray the ingroup as 
more attractive than comparison groups (for example, see Dunning, Perie & 
Story, 1 99 1 ,  for evidence of the self-serving nature of prototypes)? Clearly, 
the precise nature of depersonalized attraction is not yet fully understood, but 
the fact that there is a distinction between group-prototypical and inter­
personal liking has been established, and further emphasizes the irreducible 
nature of group processes. 
The contextual variability of prototypicality 
As we noted above, Rosch (1 978) emphasized that her analysis of the graded 
nature of within-category structure focused on judgements o
f
 prototypicality 
rather than any notion of fixed prototypes, and she insisted that prototypes 
were not to be taken as a model of category representation or processing. 
Similarly, self-categorization theory emphasizes the context-dependence of 
prototypicality, seeing it as a variable outcome of the categorization process. 
Indeed, contextual variability is built in to the self-categorization definition of 
prototypicality, in that it depends upon both within-category similarity and 
between-category difference. Thus, at the very least, prototypicality for a 


86 
Social identity: international perspectives 
given category will vary as the nature of relevant comparison groups varies, 
and as the wider context in which those groups are perceived varies. For 
example, we have already cited evidence from Hogg et al. (1990) that a group 
can perceive itself as prototypically risky or prototypically cautious 
depending upon whether it compares itself with a more cautious or more 
risky outgroup respectively. 
More recently, two experiments conducted by Haslam, McGarty, Oakes, 
Turner and Onorato (1995) directly tested self-categorization theory's 
assertion that the relative proto 
typicality of extreme outgroup members 
would increase to the extent that an intergroup definition of the judgemental 
context was made salient. This prediction follows from the principle of 
meta-contrast discussed above: in intragroup contexts individuals embody­
ing the mean group position should be more representative of the group as 
a whole (since they maximize intragroup similarity), but in intergroup 
contexts polarized members of both ingroup and outgroup should be more 
representative (as they also maximize intergroup difference). 
In Experiment I all subjects watched the same video-taped message from 
a woman who discussed the causes of road accident-induced brain damage 
and the means by which it could be reduced. They were then asked to make 
judgements about the degree to which she was representative of the group 
to which she was said to belong. In one of eight independent conditions she 
was described as either an extreme or a moderate member of either an 
ingroup (a group that wanted to improve road safety) or an outgroup (a 
group that wanted to ban the sale and consumption of alcohol in order to, 
amongst other things, reduce the road toll). The salience of subjects' own 
social identity was also manipulated so that half the subjects were simply 
given information about the target's position (that is, low salience) but half 
also had to state whether or not they themselves were in favour of banning 
alcohol/improving road safety (high salience). 
As anticipated, the perceived proto 
typicality of the target varied with her 
alleged position within her own group (moderate or extreme) and with the 
extent to which an intergroup relationship between subjects and target was 
salient. More specifically, when the target was described as an outgroup 
member she was seen as most representative of her group in the condition 
where subjects' own group membership was salient and she was described as 
an extremist. In other words, in the context of a salient intergroup division, 
an extremist was more representative of the outgroup than a moderate. 
A similar pattern also emerged from a second study in which all subjects 
watched a video of a discussion in which a group of four people debated 
issues about crime and punishment and came to a pro-authority conclusion. 
For example, the group disagreed with the statement 'criminals should be 
helped rather than punished' and agreed with the statement 'the best way to 
solve the drug problem is for the present laws to be made much stricter'. In 
one (low salience) condition subjects then simply made judgements about 
how representative of the group as a whole were (a) the group member who 
was most pro-authority and (b) the group member who was least pro-


The role o
f
 pro 
to 
typicality 
87 
authority. In two further conditions the intergroup nature of the judge­
mental context was made more salient. In the medium salience condition 
subjects were asked to stage their own (generally anti-authority) views 
before they saw the video. In the high salience condition they were given a 
lecture about the psychological basis and negative social consequences of 
authoritarianism before stating their own views, and the views of all sub­
jects were then collated on a blackboard. As predicted, to the extent that the 
experimental manipulations served to make intergroup differences between 
the subjects and the stimulus group salient, the relative prototypicality of 
the extremely pro-authority target increased so that this individual was seen 
as more representative of the group as a whole than the individual who was 
only moderately pro-authority. In higher salience conditions subjects were 
also less likely to agree with the extremist and more likely to characterize 
the group as a whole in negative terms. 
Taken together, these studies confirm the point central to self­
categorization theory that the degree to which a person is perceived to be 
representative of a group is not simply a function of properties of that 
person considered in isolation. Instead, the very same individual will be 
perceived as more or less prototypical of a social category depending on the 
social context within which he or she is defined. More specifically, both 
studies show that extreme members of outgroups will be seen as more 
representative of the groups to which they belong to the extent that 
differences between ingroup and outgroup are made salient and perceivers 
therefore conceptualize the judgemental context in intergroup rather than 
interpersonal terms. 
To what extent does other social psychological work on prototypicality 
take into account this context-dependence, and the idea of prototypicality as 
a variable outcome of categorization? In general, prototypes have been 
discussed as the basis for category representation, with a concomitant 
emphasis on identifying the prototype of a given category and the effects of 
similarity to this prototype on information processing, rather than issues of 
prototype variability. For example, Brewer and her colleagues were 
amongst the first to apply Rosch's ideas in social psychology in a series of 
studies examining prototypes of elderly people (Brewer et a!., 198 1 ;  Brewer 
& Lui, 1984). This work offers the following view of prototypes: 
Categories are represented cognitively by prototypes - actual or imaginary 
instances of the category that contain attributes most representative of items 
inside the category and least representative of items outside the category. Once a 
prototype of a category has been formed, membership in the category is assessed 
in terms of 'prototypicality' or perceived similarity to the prototypic instance. 
(Brewer et aI., 198 1 ,  p. 656) 
The assumption is that category representation is in terms of prototypes 
(cf. Rosch, 1978), and that, once formed, prototypes are the fundamental 
basis of categorization decisions, that is, it is similarity to the represented 
prototype that determines whether or not an item is perceived as a member 
of a category. In later work, Brewer (1 988) has developed this analysis of 


88 
Social identity: international pers
pectives 
prototypes further to argue that categories are represented in terms of 
pictoliteral prototypes, picture-like images of the ideal category member. 
Thus, the idea of contextual variation in prototypicality within categories is 
not raised in Brewer's work, and the suggestion seems to be that categories 
are represented by fixed visual images. 
Lord and his colleagues discuss prototypes in their leadership categoriza­
tion theory (for example, Fraser & Lord, 1988; Lord, Foti & De Vader, 
1 984; cf. Hogg, 1996). They describe leadership prototypes as cognitive 
structures (for example, Fraser & Lord, 1988, p. 292), and argue that 
perception of an individual as a leader depends upon the extent to which he 
or she is similar to a leadership prototype. Lord et al. (1984) discuss the idea 
of basic-level leadership prototypes, and suggest that these could be used as 
'a standard for appropriate behaviour' (p. 359). Clearly, then, the emphasis 
here is on the identification of relatively context-independent prototypical 
standards against which behaviour is evaluated. Similarly, in their treatment 
of leadership Rush and Russell (1 988) refer to prototypes as 'easily accessed 
schemata which serve as abstract representations of the most representative 
(stereo 
typic) member of a particular category' (p. 89). Again, the emphasis 
is on prototypes as stable, abstract representations of categories rather than 
variable outcomes of the categorization process. 
In contrast, cognitive researchers have identified a number of problems 
with the view that categories are organized around fixed prototypes. 
Barsalou (1987, 1989) has reported extensive evidence of the contextual 
variability of prototypicality. For example, Barsalou, Sewell and Ballato 
(1986) found a low level of agreement between subjects' typicality judge­
ments given twice in the same context but at two different times separated 
by a few weeks. Barsalou and Sewell (1 984) asked American undergraduate 
subjects to judge typicality from both the American and the Chinese point 
of view. They generated graded structures that were, on average, uncor­
related across categories. For example, from the American perspective the 
most typical birds were robin and eagle, but from the Chinese perspective 
they were swan and peacock. Barsalou and Sewell also correlated the 
average graded structure provided by Emory University undergraduates 
with that provided by Emory University faculty for the same categories and 
found a low overall correlation (r = .2). However, when asked to provide 
typicality judgements from faculty's point of view, the undergraduates 
produced a perfect replication of the faculty's own responses, and graduate 
students were able to reproduce exactly both undergraduate and faculty 
typicality judgements. Barsalou (1 987) concludes from this work: 
The graded structure within categories does not remain stable. Instead a category's 
graded structure can shift substantially with changes in context. This suggests that 
graded structures do not reflect invariant properties of categories but instead are 
highly dependent on constraints inherent in specific situations. (p. 107) 
Similarly, Medin (1989) identifies the treatment of categories as context­
independent as a major problem for theories which assume prototype-based 


The role o
f
 pro 
to 
typ icality 
89 
representation. He cites the work of Roth and Shoben (1 983) which again 
shows the contextual variability of judgements of prototypicality. For 
example, Roth and Shoben found that tea is judged to be more typical of 
the category 'beverage' than milk in the context of secretaries taking a 
break, but the reverse ordering obtained when the context was truck drivers 
taking a break. 
An intriguing series of 
experiments by Medin, Wattenmaker and Hampson 
(1987) demonstrates how difficult it is for subjects to work with 
categorizations based simply on relative similarity to a prototype, or family 
resemblance. For the first four experiments, stimuli were constructed to form 
two categories within which items were linked by resemblance to a prototype. 
This was done in terms of the degree to which individual items had properties 
which matched or mismatched the identified prototype. For example, the 
categories might be two groups of cartoon-like animals, and the varying 
properties might be the shape of the head, length of the tail, body markings 
and number of legs (see Medin et a!., 1987, for details). Overall, these 
experiments provided virtually no support for the idea of family resemblance­
based categorization. Asked to sort the full set of stimuli into two equal-sized 
groups, most subjects simply sorted them on the basis of one feature and 
ignored the identified prototypes completely. 
In light of these negative findings, Medin et a!. re-evaluated their assump­
tion that the simple matching and mismatching of isolated properties was 
sufficient to capture what goes on when people make categorization deci­
sions. In the remaining three experiments reported in the paper, they gave 
subjects information about interproperty relationships which could serve to 
explain why integration across separate components of the stimuli was 
appropriate. For example, in Experiment 7 cartoon-like drawings were again 
used, with variation in body size, foot type (paws or webbed feet), body 
covering (feathers or hairs), ear size and mouth type (beak or mouth). Again, 
two prototypes were identified, associated with a pattern of correlated 
attributes which suggested that one category might be related to birds and 
flying. Subjects, given the task of sorting 1 0  figures into two equal-sized 
groups, were told that half the animals were flyers and half were non-flyers. 
This instruction was designed to give the subjects a basis on which to relate 
the separate attributes to each other. It was found that although the majority 
of the sortings were still unidimensional, a significant minority (9 out of 24 
subjects) now did sort by family resemblance, that is, into categories struc­
tured around the identified prototypes. Further, subjects' sortings were 
supported by explanations which drew on the information about flying, even 
when they did not follow the predicted family resemblance pattern. Medin et 
a!. attribute their increased success in these latter experiments to the fact that 
the further information provided (for example, about flying) both made 
interproperty linkages salient and allowed subjects to associate individual 
items through their relationship to a higher order theme. 
These findings are consistent with Medin's cogent argument (for example, 
Medin, 1989; Medin & Wattenmaker, 1987; Murphy & Medin, 1 985) that 


90 
Social identity: international pers
pectives 
categorization cannot be based on judgements of similarity through match­
ing of isolated attributes. Indeed, he asserts that similarity 'is more like a 
dependent variable than an independent variable' (Medin, 1989, p. 1474) in 
the categorization process (cf. McGarty & Turner, 1992). He has suggested 
that perceivers have wide-ranging 'background theories' about their world 
and how it works, and that these theories play a crucial role in 'category 
cohesiveness', that is, in defining the way in which categories 'hang together' 
as coherent wholes, and therefore in determining why we have the 
categories we do rather than others. For example, Murphy and Medin 
( 1985) point out that on the basis of simple attribute matching we could 
identify several similarities between a plum and a lawnmower (both cannot 
hear, both have a smell, both can be dropped . . .  
) which might suggest they 
should share a category membership, but this is not consistent with our 
general theories about the world, so we understand that the 'similarities' 
identified are not sufficient to determine categorization. On the other hand, 
we can construct meaningful categories which bring together items with 
very little apparent similarity. Barsalou (1983) provides the example of a 
category comprising children, money, photo albums and pets, which 
becomes sensible when one considers it in the context of a knowledge base 
which includes the fact that the category represents 'things to take out of 
one's house in a fire' (see also Rips, 1989, for further discussion of the idea 
that similarity is neither necessary nor sufficient to determine category 
membership). 
These arguments about the insufficiency of similarity have been used to 
challenge the classical view of categorization, but Medin (1989) points out 
that the more recent prototype theories (and exemplar-based theories) are in 
fact just as dependent upon similarity as a determinant of categorization. It 
is similarity to the prototype (or a specific exemplar) that is assumed to 
determine categorization (for example, Brewer et aI., 198 1 ; Lord et aI., 
1984). Medin argues for knowledge-based rather than similarity-based 
categorization, although he sees a role for similarity in contributing to what 
he calls 'psychological essentialism', which functions to constrain the 
'deeper substance' (Medin, 1989, p. 1479) of the knowledge-rich theories 
which drive the categorization process (see Medin, 1989; Medin & Ortony, 
1989, for further discussion of psychological essentialism). 
In summary, cognitive psychologists have argued that theories which 
assume that categories are represented by fixed prototypes, and that 
categorization is a matter of attribute matching to a prototype, cannot deal 
with evidence of contextual variability in category content and structure. 
They also fall foul of evidence that attribute matching (simple similarity 
judgement) is insufficient to account for category cohesiveness and cat­
egorization decisions. Medin (1989) comments: 
Prototype theories . . . fail to reflect the context sensitivity that is evident in 
human categorization. Rather than getting at the character of human conceptual 
representation prototypes appear to be more of a caricature of it. (p. 1472) 


The role o
f
 pro 
to 
typ icality 
91 
In this context, it is important to emphasize that self-categorization 
theory is not a 'prototype theory' in the sense referred to in Medin's work. 
As noted above, the theory sees proto 
typicality as a wholly context­
dependent outcome of the categorization process, not the basis for category 
representation, and agrees with Medin that this categorization is a product 
of perceivers' background knowledge in interaction with current stimuli (see 
Oakes et aI., 1994; Turner, Oakes, Haslam & McGarty, 1994). The simi­
larity and difference judgements involved in the meta-contrast principle are 
made in the context of background theories about the ways in which 
category members tend, in general, to be similar and different - this is 
embodied in the 'normative fit' principle discussed briefly above (Oakes et 
aI., 1991). One of the major contributions of the theory has been its analysis 
of perceived similarity and difference as outcomes (rather than simple 
determinants) of categorization, and we see perceived proto 
typicality as one 
aspect of the way in which categorization contextualizes simple attribute 
judgements, gives them meaning in relation to other stimuli being judged 
(Turner & Oakes, 1989; Turner et aI., 1995). 
Conclusion 
This chapter has considered the concept of prototypicality from the per­
spective of self-categorization theory. We have reviewed ideas and evidence 
from two main areas of self-categorization research in which prototypicality 
has played an important role. In polarization research, it has been shown 
that polarization can be explained as conformity to the prototypical position 
within the group, a position extremitized through the influence of com­
parative context on ingroup definition. Research on intragroup attraction 
has established a distinction between social and personal attraction, the 
former reflecting attraction within groups based on depersonalization and 
prototypicality. In both these research areas, prototypicality is seen as a 
context-dependent outcome of the categorization process, as determined by 
both comparative (meta-contrast) and normative (fit with the content of 
background theories) factors. Because the principle of meta-contrast allows 
accurate specification of the most prototypical attributes within a given 
comparative context, precise predictions about conformity and attraction 
within groups have been made possible. 
Recognition of the internally graded, variable structure of categories was 
one of the revolutionary contributions of Rosch's work. However, many 
subsequent developments of that work have transformed the dynamic idea 
of proto 
typicality judgements into a concept of idealized 'caricatures' 
functioning as relatively fixed representations of categories, and determining 
the categorization process through judgements of item-prototype similarity. 
Self-categorization theory maintains a more Roschian perspective in this 
regard, with its emphasis on prototypicality rather than prototypes, and its 
rejection of the view of categorization as a matter of 
judging similarity to a 


92 
Social identity: international perspectives 
prototype. In predicting categorization through the interaction of com­
parative and normative factors, the theory shares with the work of Medin 
and others an emphasis on the intellectual (Neisser, 1 987) basis of 
categorization. 
Whilst the theory's conceptualization of prototypicality has proved very 
fruitful in the research areas reviewed here, it is nonetheless true that more 
work remains to be done in order to demonstrate the inherently variable 
nature of category representation (cf. Barsalou, 1 987, 1 989). This represents 
one significant direction in which our current work is advancing (Haslam, 
Turner, Oakes, McGarty & Hayes, 1 992; Oakes et aI., 1994; Oakes, Haslam 
& Reynolds, in press). 


7 
Group Beliefs as an Expression of Social 
Identity 
Daniel Bar-Tal 
Social identity 
One of the seminal theories of social psychology which contributes to the 
understanding of intragroup and intergroup processes is social identity 
theory (Tajfel, 1978d, 1981 ,  1982b). The theory posits that individuals not 
only adopt a personal identity as unique persons, but also form a social 
identity which reflects their membership in various groups to which they 
belong. Specifically, social identity is defined as 'that part of an individual's 
self-concept which derives from his knowledge of his membership of a social 
group (or groups) together with the value and emotional significance 
attached to that membership' (Tajfel, 1 978b, p. 63). Social identities, in the 
form of such social categories as nationality, religion, gender, profession, 
ethnicity or political orientation, are internalized and constitute a poten­
tially important part of the individual's self-concept. They provide meaning­
ful and significant self-references through which individuals perceive 
themselves and the world around. 
Of special importance is the assumption stating that individuals have a 
need to achieve positive social identity. Specifically, it is assumed that indi­
viduals strive for positive self-image and that in the course of identifying with 
a group this need translates into a tendency to view one's own group 
favourably. This view is achieved not necessarily through positive 
perception, but rather through comparison on appropriate dimensions. 
The comparison allows one to perceive the ingroup as 'better' than relevant 
outgroups. 
The process of social identification is further elaborated in self­
categorization theory (Turner, 1 987), an extension of social identity 
theory. The theory suggests that individuals categorize themselves as 
members of social categories, and then define, describe and evaluate 
themselves in terms of these categories. In the process of self-categorization, 
a set of cognitive representations of self is formed. Part of the cognitive 
representations reflect the perceived membership of individuals in groups. 
Since individuals belong to various groups, they possess multiple rep­
resentations of self. The categorization of the self results in a perceptual 


94 
Social identity: international perspectives 
accentuation of intra-category similarities and inter-category differences on 
dimensions believed to be associated with the categorization. 
Theories of social identity and social categorization focus on the defini­
tion of 'we-ness' by group members in terms of 'We belong to a group'. 
Group members in this process perceive themselves as group members, 
identify themselves as such, and differentiate between their own group and 
other groups. This primary process of 
cognitive-emotional nature shapes the 
social identity of individuals. Nevertheless, on the basis of social identity 
and social categorization theories, a question should be asked: What are the 
other cognitive representations, in addition to social categorization of the 
self, which underlie a person's social identity? The answer to this question is 
provided in an analysis of group beliefs (Bar-Tal, 1 990). The purpose of the 
present chapter is to present group beliefs as a framework for analysing a 
group member's social identity. Specifically, first the chapter will delineate 
the conception of group beliefs; second, it will outline their characteristics; 
third, it will describe group beliefs' formation; and, finally, implications will 
be presented. 
The conception of group beliefs 
'Group beliefs are defined as convictions that group members (a) are aware 
that they share, and (b) consider as defining their "groupness'" (Bar-Tal, 
1 990, p. 36). The first part of the definition suggests that two beliefs have to 
be shared by group members. One belief pertains to any content that is the 
subject of the group belief and the other pertains to specific knowledge 
saying that the former is shared by group members. The second part of the 
definition suggests that group beliefs are those beliefs that group members 
perceive as defining the essence of their group. It is assumed that they 
provide the cognitive basis that group members view as uniting them as one 
entity. In fact, group beliefs serve as a foundation for group formation and, 
later, as a bond for group existence. 'We are exploited', 'Communism is the 
best system for human beings', 'Jesus is God's son', 'Iraq is our enemy', are 
examples of possible group beliefs. 
In this framework, the belief 'We are a group', expressing social identity 
of group members, is called the f
undamental group belief However, in 
addition to the fundamental group beliefs, group members share group 
beliefs of various contents pertaining to a variety of other subjects. These 
beliefs are part of cognitive representations which together with self-social 
categorization (that is, the fundamental group belief) define social identity 
of group members and thus underlie their we-ness and uniqueness. 
Individuals regard group beliefs as characterizing them as group members 
and as defining the boundary of the group. On their basis individuals may 
categorize themselves and may be categorized by others as group members. 
Acceptance of group beliefs is one of the important indicators of group 
membership. Sharing group beliefs implies, most frequently, that a person is 


Group belie
f
s
 as an expression o
f
 social identity 
95 
a group member. Moreover, being a group member signals to people that a 
person shares group beliefs. A person joining a group must, at least exter­
nally, express acceptance of group beliefs. Thus, on the one hand, group 
beliefs unify group members and, on the other hand, differentiate them 
from outgroups. The unification is reflected in the recognition of similarity 
(Bar-Tal, 1 993). Group members are aware that they share the same unique 
beliefs that define their 'groupness' and thus become aware of the similarity. 
Group members who may differ with regard to many characteristics are 
aware that they share the same unique beliefs that define their 'groupness' 
and thus become aware of the similarity. 
Also, since in many cases groups try to be differentiated, group beliefs 
often provide a criterion for differentiation. Group beliefs draw the line 
between the in 
group and outgroups. As Sherif (1951) pointed out: 'From 
the point view of inter-group relations, the most important consequence of 
group structuring is the delineation of in-group from out-groups. The devel­
opment of in-group and "we-experience" is accompanied by the demarca­
tion and setting of boundaries from out-groups' (p. 395). 
An experiment by Allen and Wilder (1975) demonstrated that group 
formation (including the formation of a group belief) is a more potent 
factor in discrimination against outgroup than the awareness that the 
outgroup members have some similar beliefs. Also, belief similarity about 
the basis for group classification (that is, formation of a group belief) had a 
significant effect on discriminative behaviour. Special ingroup favouritism 
was displayed when the in 
group had similar beliefs. 
Contents of group beliefs 
Group beliefs, as implied by the definition, are held by group members, who 
share these beliefs and consider them as defining their group. Thus, group 
beliefs should be considered as a category of beliefs and should be 
characterized by the same features as other beliefs. First of all, the contents 
of group beliefs will be discussed. 
Beliefs in the present conceptual framework, viewed as propositions to 
which individuals attribute at least some truth, constitute units of knowl­
edge. This is a wide-scope definition which considers any ideas, thoughts or 
opinions about any topic as beliefs. They can thus cover various contents 
which can be classified to such categories as 'values', 'ideologies', 'inten­
tions', 'goals', and so on. Group beliefs may also refer to different topics or 
ideas that a human mind can think. They may concern such subjects as 
religious doctrines, political ideologies, philanthropic goals, human rights, 
identity characteristics, prescriptions of behaviour (that is, norms), and 
others. As examples, four categories of group beliefs will be presented -
those which refer to group norms, values, goals and ideology. 


96 
Social identity: international perspectives 
Group norms 
Group norms, defined as 'an idea in the minds of the members of a group, 
an idea that can be put in the form of a statement specifying what the 
members or other men should do, ought to do, are expected to do, under 
given circumstances' (Homans, 1950, p. 123), regulate group members' 
behaviour and provide criteria for judging it. Norms, as defined, do not 
necessarily have to be group beliefs. Norms may regulate group members' 
behaviour without being considered as characteristic of the specific group. 
In such groups, group beliefs may pertain to contents other than norms. In 
other groups, at least part of the norms may define their characteristics and 
be group beliefs. Nevertheless, when norms function as group beliefs, group 
members believe that their patterns of behaviour are unique to them and 
characterize their membership. 
In an extensive description of the Amish society, 
I Hostetler (1 968) points 
out several norms that can be considered as group beliefs. That is, Amish 
people believe that particular prescriptions of behaviours define their 
'groupness' and contribute to their social identity. For example, Amish 
beliefs prescribe the necessity of separation from the world. Therefore, an 
Amish man does not marry a non-Amish woman, does not do business with 
an outsider, and even does not enter into intimate relationships with a 
person outside the Amish community. In addition to norms of separation, 
the Amish group also perpetuates norms of non-resistance. Group members 
are forbidden to take part in violence or war, and they withdraw from 
conflicts or resistance. They refuse to swear any oath, to bear arms, or to 
hold public offices. 
In another example, Ardener (1983) described the group beliefs of 
Canadian Doukhobors who emigrated from Russia at the beginning of the 
twentieth century. They believe that the Perfect Life is attainable on earth 
through the rejection of materialism. They have not recognized human 
authority structure and therefore are opposed to state legislation or oath 
taking. Their uniqueness has been reflected in deprivation, suffering and 
poverty, which became, according to Ardener, part of their self-definition, 
corresponding to their identification with Christ himself. In this vein, to 
form their own identity, Doukhobors resorted to nudity as a social artefact 
for identifying their group and to the use of arson against institutions and 
then against their own property. Ardener suggested that both practices 
helped to maintain the boundary of the group and to enhance solidarity in 
the face of outgroups. 
Group values 
An analysis similar to the one presented about norms applies to group 
values. Groups may hold values, but they do not necessarily have to 
function as group beliefs. A value is defined as 'an enduring belief that a 
specific mode of conduct or end-state of existence' (Rokeach, 1 973, p. 5). 
The former is called an instrumental value, while the latter is called a 


Group belie
f
s
 as an expression o
f
 social identity 
97 
tenninal value. Values guide the selection of the means and ends of specific 
actions, and serve as criteria by which objects, actions or events are 
evaluated. Examples of values are freedom, truth, individualism or equality. 
They reflect the ideals to which individuals and groups aspire. 
It is possible to characterize groups by the values that their group 
members share (Kluckhohn, 1 95 1 ;  Parsons, 1 968). Sociologists have focused 
on shared values, referring to them as value systems in a given society. 
Nevertheless, a given shared value is not necessarily held by every person in 
the group. When a sufficient number of group members subscribe to the 
value, it may not only characterize the group, but may also become an 
important detenninant of the group members' behaviour. 
In some groups the values, like nonns, may be fonnally fonnulated (that 
is, in writing) and maintained as group beliefs, while in other groups they 
are never fonnally defined, but may be considered as group beliefs and 
carried latently through the social processes of socialization and influence. 
However, in both types of groups, when values are shared by group mem­
bers and viewed by them as characterizing the group, they are group beliefs. 
In this case, the value or values define the uniqueness of the group and 
differentiate it from outgroups. 
To specify values that characterize a nation is a complex task. Never­
theless, social scientists commonly have suggested, on the basis of either 
collected data or their own observations, that self-sufficient individualism 
is one of the salient characteristics of American society (for example, 
Diamond, 1976; Gillin, 1 955; Williams, 1970). This value emphasizes the 
responsibility and the right of each individual to pursue personal goals with 
as much freedom from restraints as possible. Americans are socialized to 
internalize this value and to believe that it characterizes their society. 
Many groups present values as part of their credo. In fact, many of the 
group nonns and goals functioning as group beliefs are based on values. 
This stress on values can be observed in religious groups, as well as in many 
political groups. An example of a political group that was founded on the 
basis of a value is Amnesty International. In its statute, the value of human 
rights underlies more specific goals. The beginning reads: 
I .  Considering that every person has the right freely to hold and to express his 
or her convictions and the obligation to extend a like freedom to others, the 
objects of Amnesty International shall be to secure throughout the world the 
observance of the provisions of the Universal Declaration of Human Rights, 
by: 
a) irrespective of political considerations working towards the release of and 
providing assistance to persons who in violation of the aforesaid provisions 
are imprisoned, detained or otherwise physically restricted by reason of their 
political, religious or other conscientiously held beliefs or by reason of their 
ethnic origin, sex, color, or language, provided that they had not used or 
advocated violence (hereinafter referred to as 'Prisoners of Conscience'); 
b) opposing by all appropriate means the detention of any Prisoners of 
Conscience or any political prisoners without trial within a reasonable time or 


98 
Social identity: international perspectives 
any trial procedures relating to such prisoners that do not conform to 
internationally recognized norms; 
c) Opposing by all appropriate means the imposition and infliction of death 
penalties and torture or other cruel, inhuman or degrading treatment or 
punishment of prisoners or other detained or restricted persons whether or 
not they have used or advocated violence. (Amnesty International pamphlet, 
London, 1978) 
Although the Ku Klux Klan is vastly different from Amnesty International, 
its group beliefs are also based on values. The principal beliefs of the Ku 
Klux Klan are presented in the following way: 
We invite all men who can qualify to become citizens of the Invisible Empire, to 
approach the portal of our beneficent domain, join us in our noble work of 
extending its boundaries, and in disseminating the gospel of Klankraft, thereby 
encouraging, conserving, protecting and making vital the fraternal relationships in 
the practice of an honorable clannishness; to share with us the sacred duty of 
protecting womanhood; to maintain forever the God-given supremacy of the 
White Race; to commemorate the hold and chivalric achievement of our fathers, 
to safeguard the sacred rights, privileges and institutions of our civil government; 
to bless mankind and to keep eternally ablaze the sacred fire of a fervent devotion 
to a pure Americanism. (Cited in Vander Zanden, 1960, p. 459) 
Thus, the values of womanhood, white supremacy, mankind and American­
ism are the values that the group intends to protect. 
Group goals 
Group goals are defined as beliefs of valued or desired future specific states 
for the group (Cartwright & Zander, 1968; Etzioni, 1 975). On the basis of 
this definition, we can assume that with few exceptions almost all groups 
have goals. In most of these cases, goals may serve as group beliefs in 
addition to other beliefs. Individuals form groups to achieve certain goals. 
Subsequently, goals are often considered as a raison d'etre for group 
formation, frequently keep group members together, provide a basis for 
solidarity, and give direction for activity. These functions are particularly 
explicit in certain voluntary groups that are formed to advance specific 
goals (Sills, 1958). Thus it is not surprising that group goals frequently serve 
as basic group beliefs. They increase the identification of the group 
members with their group and define the boundary for group membership. 
In this vein, March and Simon (1958) noted that: 
The greater the extent to which goals are perceived as shared among members of 
a group, the stronger the propensity of the individual to identify with the group 
and vice versa. (p. 66) 
Two examples of group goals may provide illustrations of how they 
function as group beliefs. Trans-Species Unlimited (TSU) is a national 
animal rights group 'dedicated to the total elimination of animal abuse and 
exploitation'. This group belief leads TSU to direct actions and legislative 


Group belie
f
s
 as an expression o
f
 social identity 
99 
campaigns to end such practices as the use of leghold trap and live bird 
shoots and commercial exploitation of rabbits for meat, fur and research 
The Richard III Society was founded in England in 1924 with the 
following aims: 
1 .  To promote in every possible way historical research into the life and times of 
King Richard III. 
2. To secure a reassessment of the historical material relating to this period and 
of the role in English history of this monarch. 
3. To circulate all relevant historical information to members of 
the Society and 
to educational authorities. (copied form the pamphlet of The Richard III 
Society, New Orleans, undated) 
The principal goal of this group is to present the truthful history of King 
Richard III in view of past distortions and biases. 
Groups differ with regard to the goals that they try to achieve. There are a 
variety of goals that can be classified into different categories. One type of 
category has received special attention by political scientists - namely, goals 
to influence governmental decisions. Groups that have these goals are called 
interest groups (Salisbury, 1975), and then the term interest denotes a 'shared 
attitude toward what is needed or wanted in a given situation, observable as 
demands, or claims upon other groups in the society' (Truman, 1 95 1 ,  p. 33). 
These shared demands or claims, which reflect goals, serve as a basis for 
group formation and function as group beliefs. The interests can pertain to 
various contents involving either tangible or intangible claims, and concerns 
of 
either a particular policy objective or wide societal goals (Salisbury, 1 975). 
Interest groups can be based on either a few or many group beliefs. In 
some groups, the group beliefs refer to one broad topic that is relevant to 
several issues. For example, the Consumer Federation of America reflects a 
viewpoint that favours the protection of the rights of individual consumers 
through the regulation of corporate activities, and the US Chamber of 
Commerce reflects a strong belief in the character of the American free 
enterprise system, opposing excessive government regulation of business. 
Other groups' interests are formed on the basis of narrow group beliefs 
reflecting single issues. For example, the National Rifle Association has a 
strong belief in the American citizen's unrestricted right to bear arms; the 
National Right to Life Committee believes in the sanctity of the foetus. 
Group ideology 
One of the most discussed concepts for describing a group's uniqueness is 
ideology. Ideology is often referred to as the mental characteristic of 
a group. Indeed, various definitions of ideology refer to its commonality - a 
state indicating that ideological beliefs are shared by group members. It 
reflects their common experience and serves as a basis for group members' 
cooperation, morale, order and as a rationale for their behaviour. Ideology 
refers to an integrated set of beliefs constituting a programme, a theory of 
causes and effects, and premises on the nature of humanity and societal 


100 
Social identity: international perspectives 
order (Apter, 1964; Lane, 1 962; Shils, 1968). For example, Toch (1965) 
defined ideology as: 
a set of related beliefs held by a group of persons. The ideology of a social 
movement is a statement of what the members of the movement are trying to 
achieve together, and what they wish to affirm jointly . . .  
the ideology of a social 
movement defines the movement, and contrasts it with other movements and 
institutions. (p. 21) 
When group members hold an ideology, it frequently serves as a group 
belief. By definition, it consists of a set of ideas that characterize the way in 
which a group posits, explains and justifies the ends and means of its 
organized social actions. It usually provides an identity to group members, 
defines the group cohesion, and describes its exclusivity; in most cases, an 
identification with a specific ideology indicates a membership in a particular 
group (Lane, 1962). 
The contents of an ideology can be political, social or even religious. An 
ideology usually refers to images of the desired society and the means and 
conditions needed to achieve it. For example, the ideology of the Fasci 
Italiani di Combattimento (the Italian Fascist Party) placed a special 
emphasis on a nation or a state - the fascist state. The individuals and 
groups are secondary to the state, which embodies the culture and spirit of 
the people and is considered to be the highest and noblest value. The state 
provides an orientation for the individuals - it is only there that they can 
fulfil their identity. According to the fascist ideology, life is a continuing 
struggle for national supremacy. In order to achieve the nation's goals, the 
state has to exercise complete authority over all areas of public and private 
life, from the education of children to control of the economy. The state has 
the right to control and restructure the society. Accordingly, the state 
requires total loyalty. There is no place for groups or individuals to object, 
since the state is the source of political, economic and moral action 
(Mussolini, 1 935). These principles served as group beliefs of the Italian 
Fascist Party, and, with its ascendance to power in 1 922, the party tried to 
turn them into group beliefs for the entire nation. 
For seven decades, the principles of communist ideology served as group 
beliefs for the members of the Communist Party of the Soviet Union. This 
ideology was the official creed and the most accessible characteristic in 
every domain of life in Russia. The basic principles of communism were 
inculcated in Soviet citizens from the early years of education and main­
tained through all the state communication channels. The ideology 
advocates a classless social system of equality for all members of society, 
which provides everyone with material and cultural benefits according to 
their growing needs (Triska, 1 962). 
In sum, this limited review of the contents did not exhaust all possible 
categories of group beliefs. As indicated, group beliefs can have any 
contents, and the above categories serve only as relevant examples of the 
ones that are frequently used by social scientists. Group members share 


Group belie
f
s
 as an expression o
f
 social identity 
101 
group beliefs that characterize them and differentiate them from other groups. 
As illustrations of possible contents, this review was not based on group 
members' reports identifying their group beliefs, though it is possible that 
they actually serve as group beliefs. Future research should treat the formal 
statements as hypotheses that need to be validated through empirical 
evidence. 
Characteristics of group beliefs 
Group beliefs can be characterized by the same features as other beliefs. 
Degree of confidence and level of centrality are among the most important 
characteristics of beliefs (Bar-Tal, 1990; Bem, 1 970; Kruglanski, 1 989). 
These characteristics are of special relevance for group beliefs and therefore 
discussed at some length. 
Con
fidence 
Group beliefs are usually held with great confidence because they are con­
sidered to be facts and verities. This occurs because group beliefs define the 
essence of the group and thus the reality of the individuals who view 
themselves as group members. In fact, in most cases, individuals voluntarily 
join groups, such as religious denominations, political organizations or 
professional associations, on the basis of group beliefs. The act of 
joining a 
group indicates that group members accept, at least formally, group beliefs 
and hold them as valid. 
A reduced confidence in group beliefs may shatter group members' 
reality, especially in cases when group membership is very important for 
individuals or when the change is sudden. The questioning of group beliefs, 
especially central ones, may have significant implications for group life. 
Moving group beliefs from the status of verities to the status of hypotheses 
by the majority of group members may change the essence of the group or 
its raison d'etre, or even cause its division or disintegration. It is, therefore, 
not surprising that groups often use various mechanisms to maintain high 
confidence in group beliefs. 
It is recognized, though, that groups differ with regard to their insistence 
on confidence. While orthodox groups insist on high confidence in group 
beliefs (Deconchy, 1 984), other groups may be less strict on this matter. The 
former groups do not allow any doubts or sanction their expression. The 
other groups are more tolerant of changes and therefore their group beliefs 
may be gradually altered. 
Centrality 
The centrality of group beliefs implies that they are often accessible in 
group members' repertoire and that they are relevant for consideration 


1 02 
Social identity: international pers
pectives 
in making various evaluations, judgements or decisions, including behav­
iours (Bar-Tal, 1 986). The centrality characteristics of group beliefs has at 
least two different aspects. First, the characteristic may refer to the 
importance of the set of beliefs for the group members, and, second, it may 
refer to the importance of the specific belief within the set of group beliefs. 
The former aspect is based on an assumption that some groups are more 
important for their members than others. Individuals are usually members 
in a number of groups and not all of them are of equal importance to 
them. The importance of the group for group members is reflected in the 
centrality of group beliefs. That is, when the group is important for group 
members, in most cases the group beliefs are central. For example, it can be 
assumed that group beliefs are very central in groups such as the Amish 
(Hostetler, 1 968) and Lubovitcher Hassidim (Levy, 1 975). Group 
membership in these groups is the most important characteristic for the 
members and the group beliefs are frequently accessible and often taken 
into consideration. In other groups, group beliefs may be unimportant for 
most of the group members. 
The other aspect of characterizing group beliefs as central refers to the 
assumption that not all beliefs are of equal centrality for group members. 
Some group beliefs may be of greater importance than others and thus may 
be more central. It means that when a group has a set of elaborated group 
beliefs, some of them contribute more to the group characterization and its 
definition than others. The more central group belie
f
s
 are considered as 
prototypic in group characterization and there
f
ore are called basic group 
belie
f
s. Their weight in defining the essence of the group is crucial. Beyond 
the simple self-definition 'We are a group', these beliefs provide the group's 
credo. For instance, in certain groups, group beliefs that refer to group 
goals or identity may be more central than group beliefs that refer to values 
or group history. In a specific example, it may be assumed that while the 
Mormons have numerous group beliefs, one of the basic ones is the 
conviction that the Book of Mormon is another Testament of Jesus Christ, 
a religious and secular record of ancient American civilization (O'Dea, 
1 957). 
The centrality characteristic is not a given, but often is maintained by 
external factors. Availability and saliency of group beliefs in the group 
members' environment may influence their centrality. The first element 
refers to the frequency with which the content of group beliefs is exposed 
to group members. The other element expresses the prominence and 
vividness of the group belief when it is presented to group members. 
When group beliefs are repeatedly presented to group members in a 
prominent and vivid manner, they are often accessible in group members' 
repertoires (Higgins & King, 1981). Groups often keep group beliefs 
accessible in the minds of the group members by constantly repeating the 
contents of group beliefs. Moreover, they even sometimes develop cultural 
and educational mechanisms to maintain group beliefs as central in group 
members' minds. 


Group belie
f
s
 as an expression o
f
 social identity 
1 03 
Formation of group beliefs 
Formation o
f
 the f
undamental group belief 
The crucial phase in the f
ormation of group belie
f
s
 is the emergence o
f
 the 
f
undamental group belief indicating group existence (that is, ' We are a 
group'). This phase is preceded by various social perceptual and cognitive 
processes. According to the present view, numerous reasons can cause 
individuals to start believing that they constitute a group. Individuals usually 
have to f
ormulate at least one belief which provides the antecedent f
or the 
f
undamental group belief stating that a group exists (that is, ' We are a 
group'). Such an antecedent belief may later become a group belie
f 
Any belief can serve as an antecedent f
or individuals to begin considering 
themselves members o
f
 a group. The belief 'We are a group' may be instigated 
by any of the following beliefs: 'We act interdependently', 'We have the same 
fate', 'We have the same characteristic', 'We live in the same place', 'We have 
the same goal', 'We believe in the same religious doctrine', 'We have the 
same ancestors', 'We accept the same ideology', 'We are treated in the same 
way', and others. The antecedent belie
f(s) may be f
ormed on the basis o
f
 
actual experience and perception or on the basis o
f
 reliance on inf
ormation 
which comes f
rom sources that propagate these belie
f
s. In the first case, 
individuals may actually become aware of the similarity in what may be 
called natural evolvement, whereas in the other case, an epistemic source 
may indicate to the individuals that they are similar and point out that 
similarity. 
For example, Epstein's (1978) work indicates that ethnic identity is based 
on shared beliefs in common attitudes or property, which differentiate the 
group from the outgroups. The shared beliefs are products of common 
perceptions of a group's past and interaction between forces operating within 
a group and those coming from outgroups. Tajfel (1981) hypothesized that 
minority groups tend to emerge as a result of common attitudes or treatment 
by the outside groups which, on the one hand, facilitates the perception of a 
common fate and, on the other hand, indicates the boundaries operating 
between the ingroup and the outgroups. Several other conceptions stress 
individuals' shared needs (for example, Killian, 1 964; Toch, 1 965), such as 
common feelings of frustration, alienation, deprivation, exploitation and 
injustice, any of which may serve as a basis for group formation. 
An experiment by Zander, Stotland and Wolfe (1960) may be seen as a 
demonstration of how the basic group belief is formed. In this experiment, 
the investigators created a group composed of female college students. To 
foster a group entity, the experimenter provided several antecedent beliefs 
regarding similarity, proximity, common goals and identity. Specifically, the 
experimenter systematically manipulated the perception of 'groupness' by 
varying seating arrangements (creating proximity), by pointing to similari­
ties among group members (creating commonality), and by assigning a goal 
for the group (creating competition against another group). Group members 


1 04 
Social identity: international pers
pectives 
were then asked to choose a name for their group. These antecedent beliefs, 
together with the name of 
the group, produced the fundamental group belief 
'We are a group'. The study also showed that these individuals behaved as 
group members. Thus, the investigators concluded that when a group is 
formed, large proportions of the self 'become involved in the group and are 
affected by identification with the group' (p. 475). 
The effects of group experiences on the formulation of the fundamental 
group belief are underlaid by the process of perception. It has been sug­
gested that principles of perceptual organization contribute greatly to the 
self-perception of individuals as a group (Asch, 1 952; Campbell, 1958). 
Specifically, Campbell (1958) proposed the use of four Gestalt principles 
(proximity, similarity, common fate and pregnance) to explain why indi­
viduals begin to see themselves as one entity. These principles, which lead 
discrete elements to be perceived as parts of a whole, can be viewed as 
characteristics of individuals, and therefore allowing definition of the group 
and differentiation between the ingroup and the outgroups. Indeed, 
Brewer's (1979) review of group formation under limited conditions indi­
cates that factors such as interdependence, intragroup similarity or shared 
fate affect the feeling of 'groupness'. Moreover, these factors not only 
determined group formation, but also influenced group members' behaviour 
in the direction of ingroup bias. 
A somewhat similar approach to group formation was recently suggested 
by cognitive social psychologists. They placed their focus on the knowledge 
that individuals possess (for example, Pryor & Ostrom, 1 987). Thus, infor­
mation about the collective of individuals (that is, their attributes) serves as 
a basis for their self-classification as group members. This information 
denotes the similarity of group members and their uniqueness in com­
parison to outgroup members. 
Moreland (1987) extended the perceptual-cognitive perspective to four 
different bases for self-classification of individuals as group members, and, 
therefore, for the formation of the fundamental group belief 'We are a 
group'. One basis derives from the conditions and resources in the environ­
ment. Factors such as money, time, propinquity or social networks may 
affect group formation. For example, people who live or work in close 
proximity may form a group which can be based on common goals or 
common interests. Also, a social network of relationships may facilitate 
group formation, as happens among friends or members of professional 
organizations. 
Behavioural basis is evident when people become dependent on one 
another for the satisfaction of their needs. Various theoretical approaches, 
such as evolutionary perspective, social exchange theory, social evaluation 
perspective or psychodynamic theories, suggest that individuals form groups 
in order to satisfy their various needs. Needs of survival, defence, social 
support, predictability, world understanding, uncertainty reduction, anxiety 
avoidance or personal adjustment are only a few examples of needs which 
can be satisfied through group membership. 


Group belie
f
s
 as an expression o
f
 social identity 
1 05 
Groups may also be formed on the basis of affection as reflected in 
shared feelings. The positive feelings can be expressed in interpersonal 
attraction of group members, attraction towards group goals or activities, 
and love and respect of the group's leader. 
Finally, the cognitive basis indicates that groups are formed when people 
realize that they share important personal characteristics such as attitudes, 
values, interest or goals. Summarizing this basis, Moreland (1 987) pointed 
out that: 
Research on personal factors suggests that small groups will form when people (a) 
have acknowledged their shared characteristics before, or (b) have done so 
recently, or (c) have found it useful to think of themselves in that way. Research 
on situational factors suggests that small groups will form when (d) people are 
reminded of shared characteristics, or (e) their outcomes seem to depend on those 
characteristics, or (f) the characteristics that they share are unusual in some way. 
(p. 103) 
On the basis of the present analysis, it becomes obvious that social 
categorization is the psychological process which underlies group f
ormation in 
the minds o
f
 the individuals who are group members. As noted earlier, 
Tajfel's theory of social identity and Turner's theory of self-categorization 
elaborate on this process (Tajfel, 1978d, 1981 ;  Turner, 1987). 
In this vein, several studies demonstrate that individuals form a psycho­
logical group solely on the basis of shared social category, in the process of 
comparing themselves to others. In these experiments, performed within the 
framework of Tajfel's social identity theory and Turner's self-categorization 
theory, individuals were categorized as group members and they acted 
accordingly (for example, Billig & Tajfel, 1 973; Hogg & Turner, 1985b; 
Tajfel, Flament, Billig & Bundy, 197 1 ;  Turner, Sachdev & Hogg, 1983). 
In a study by Tajfel et al. (1971), individuals were divided into groups on 
the basis of a fairly trivial task (expressing a preference for abstract 
paintings of Klee and Kandinsky or guessing the number of dots in rapidly 
projected clusters). Later, they were put in individual cubicles and asked to 
divide points worth money between two persons who were identified as 
either belonging to their group or not. This minimal group situation was 
enough to produce ingroup favouritism. Individuals tended to award more 
points to members of their group. These results were replicated, even in the 
experiment by Billig and Tajfel (1973), when the division into groups was 
done on the basis of the toss of a coin. 
The important contribution of these studies to the understanding of group 
formation is the demonstration showing that the perception of the group as 
an entity does not have to be based on actual experiences, but a mere piece 
of information may be a sufficient condition for self-labelling as a group. 
Individuals may be told that they are given the same label (or category), 
and this information may cause them to form the belief 'We are a group' 
(for example, Billig & Tajfel, 1973). 
In this respect it was analysed how groups form the fundamental belief 
which underlies the formation of a group. Common experience and 


106 
Social identity: international pers
pectives 
perceived information may influence individuals to form a new shared 
reality. The appearance of the fundamental belief 'We are a group', which 
was the centre of the work of Tajfel and Turner, reflects the emergence of 
new social identity and new self-categorization. This is a psychological 
beginning of group existence. 
But the present conception suggests that social identity is not based solely 
on the fundamental group belief which reflects self categorizations of 
individuals as group members. As was presented, additional beliefs called 
group beliefs play a determinative role in the formation and maintenance of 
social identity. These group beliefs establish the raison d'etre for group 
existence and provide the rationale for individuals to be group members. 
Formation o
f
 additional group belie
f
s
 
Once the fundamental group belief 'We are a group' is formed, other group 
beliefs are usually added, though the contents of some additional group 
beliefs may be formed as antecedent beliefs prior to the formation of the 
fundamental group belief. As already indicated, the contents of group 
beliefs can be of various categories and of unlimited scope. Groups differ 
with regard to the repertoire of group beliefs that they form. They differ in 
the contents and in the quantity of group beliefs that they have. Each group 
has its own set of group beliefs. Although groups may have similar group 
beliefs, two separate groups must have at least one group belief which 
makes them different from each other. With regard to quantity, groups may 
have few group beliefs or a very long list of group beliefs. The selection of 
group beliefs depends very much on a group's goal(s), common experiences, 
ideas regarding symbols, emerging norms and values, constructed rational­
izations for group formation, perceived similarities, perceived uniting forces, 
and so on. 
One of the few studies of small groups from which the emergence of 
group beliefs can be inferred is the study of norm formation by Sherif 
(1936). He found that a group of strangers facing an ambiguous stimulus 
tended to converge their judgements into a uniform norm. That is, common 
experience in a group situation served as a basis for establishing a common 
social reality. Sherif concluded from these results that individuals who are in 
contact, in a defined place, form common values or norms as a basis for 
group formation. It may be further assumed that some of these norms and 
values may serve as group beliefs, since they may characterize the group 
and underlie its uniqueness. In this vein are findings obtained by Festinger 
and Thibaut (1951), who observed that 'belonging to the same group tends 
to produce changes in opinions and attitudes in the direction of establishing 
uniformity with the group' (p. 92). 
Another line of research by Sherif and his colleagues that contributes to 
the understanding of group belief formation was the Robbers' Cave 
experiments, which investigated intergroup conflict and cooperation (Sherif, 


Group belie
f
s
 as an expression o
f
 social identity 
107 
Harvey, White, Hood & Sherif, 1 961). The experiments also studied the 
formation of a group. In one of them, 22 fifth graders were randomly 
divided into two groups and brought to separate camps so that group 
formation would be the consequence of the bonding process within the 
ingroup without the contact of an outgroup. Within seven to eight days the 
formation of the two groups emerged and stabilized. As a point of interest, 
it can be noted that the two groups formed various group beliefs which 
defined their uniqueness and character. Thus, for example, the first group 
selected the name 'Tom Hale Rattlers', chose a flat, referred to several 
places as their territory, and developed a norm of 'toughness' (reflected in 
cursing and avoidance of reporting injuries). The second group called itself 
'Eagles', selected their own song, and selected places in their territory for a 
campfire and swimming hole. 
As Sherif et al. (1961) pointed out, these beliefs were produced from 
scratch. They were formed as a consequence of the interaction processes 
found in intragroup relations. They served as unique characteristics of the 
group and defined the substance of the 'groupness'. The particular group 
name, the flag, the norm or the territory, in addition to the belief 'We are a 
group', defined the essence of the group. For group members, these group 
beliefs served as a cognitive basis for their feeling of 'groupness'. The belief 
that they are a group, as well as other group beliefs, became part of their 
reality. It can be said that although the mere division into groups created 
the fundamental group belief, the addition of group beliefs not only 
strengthened the group identity, but also provided meaning to the we-ness 
reality. Group beliefs, thus, are an inseparable part of social identity. They 
provide the essence to the mere feeling of belonging to a group. 
It should be noted though that group beliefs do not have to be formed 
after the group was founded. Already it has been indicated that certain 
beliefs may function as antecedents for individuals to form a group and 
later become group beliefs. Also, individuals formulate beliefs which may 
serve as a basis for group formation. 
Numerous political parties, religious groups and volunteer organizations 
are based on group beliefs which were formulated before the group was 
formed. In this process, the belie
f
s
 are first f
ormulated, and then an attempt 
is made (sometimes it is even unintentional) to f
orm a group in which 
these belie
f
s
 serve as group belie
f
s. The following examples illustrate this 
process. 
The Jehovah's Witness group was founded in 1 872 by Charles Taze 
Russell in a suburb of Pittsburgh, Pennsylvania. Russell, who was brought 
up as a Presbyterian, was exposed to ideas of Adventism and Second 
Adventism until he formulated his own beliefs. These beliefs, which were 
later developed and modified by his successor, Joseph Franklin Rutherford, 
originally served as a basis for the formulation of a new denomination 
(Stroup, 1945). 
Cannon (1973) described how the Black Panthers formulated their group 
beliefs: 


1 08 
Social identity: international pers
pectives 
One afternoon in October, 1966, Huey told Bobby Seale, 'We need a program. 
We have to have a program for the people. A program that people can under­
stand. A program that the people can read and see, and which expresses their 
desires and needs at the same time.' That night they sat down in the North 
Oakland Poverty Center, where they were working, and wrote out the l O-point 
program which is still the basis for all Panthers' actions. (p. 339) 
The programme concerned freedom for the black community, full 
employment, the end of exploitation, decent housing, true education, the 
exemption of blacks from military service, the end of police brutality, free­
dom to black prisoners, establishment of black courts and self-determination 
(Cannon, 1973). 
In the present conception, an important question should be asked: Which 
common belie
f
s
 (that is, shared belie
f
s
 by group members) become group 
belie
f
s? In principle, many beliefs can become group beliefs in the group 
members' repertoire. In reality, however, relatively few beliefs achieve this 
status. Three conditions are suggested as an explanation as to which shared 
beliefs become group beliefs. One condition refers to the functionality of 
group beliefs. Belie
f
s
 which are f
unctional f
or group f
ormation and main­
tenance become group belie
f
s, since it is assumed that the need to belong and 
the need to f
orm social identity are important needs which underlie human 
beings' cognitions and behaviours. In this context, of special importance are 
beliefs which differentiate between the ingroup and outgroups, since they 
allow demarcation of the group boundary and provide significant infor­
mation about the group. 
The differentiating beliefs indicate the similarity of group members which 
overrides the differences between them. That is, through the psychological 
process of categorization, group beliefs simultaneously make the group 
members uniquely similar, yet distinctively different from other groups. 
They provide the individuals with a system of orientation towards their 
own group and other groups. This system leads to the perception and 
judgement of individuals of the same category as more similar to one 
another than they really are and enhances the perceptual and judgemental 
contrast between individuals not belonging to this category (Tajfel, 1 978d, 
1 981). An experiment by Allen and Wilder (1 979) demonstrated this 
phenomenon. They created two groups, ostensibly on the basis of artistic 
preferences (which served as group beliefs), and then had group members 
complete an opinion survey in the manner they thought another member of 
their own (or the other) group would respond to the items. They found that 
individuals assumed that another member of their group would express 
beliefs more similar to their own (previously assessed) opinions than would 
an outgroup member. This difference occurred even for belief items 
irrelevant to art. 
In addition, those beliefs which provide information about the group 
become group beliefs. It is no accident that group beliefs consist of contents 
which contain such elements as group history, group goals and common 
characteristics of the group. These contents provide information for group 


Group belie
f
s
 as an expression o
f
 social identity 
1 09 
members about themselves and their commonalities. They provide the 
raison d'etre for group formation and, later, for group maintenance. Also, a 
fulfilment of additional functions by a belief may especially strengthen its 
status as a group belief. Thus, for example, beliefs that help to achieve 
group goals, raise the self esteem of the group members or strengthen 
feelings of security are likely to become group beliefs. 
Another condition pertains to the saliencey o
f
 the belie
f
s
 which become 
group belie
f
s. Belie
f
s' saliency re
f
ers to those belie
f
s
 which draw special 
attention o
f
 group members because o
f
 their prominence and distinctiveness. 
In combination with the first condition, these beliefs are efficiently and 
swiftly absorbed and demarcate the boundary between the ingroup and the 
outgroups. In some groups the salient group beliefs may pertain to physical 
appearance, the most perceivable cues. Special clothing, ornamentation or 
even skin colour may serve as a salient basis for defining one's own group. 
In other groups, salient beliefs may refer to individual's needs (Toch, 1 965), 
group goals or common experiences. 
Finally, in order f
or belie
f
s
 to become group belie
f
s, epistemic authorities 
of group members have to consider them as such, and support them. 
Epistemic authorities are those knowledge sources who exert determinative 
influence on the formation of knowledge (Bar-Tal, Raviv & Brosch, 1 991; 
Kruglanski, 1989). Group members attribute high confidence to beliefs 
coming from epistemic authority, consider these beliefs as truth, and adopt 
them as part of their own repertoires. In our case, the concept applies 
mostly to leaders who perpetuate beliefs to become group beliefs. Political, 
intellectual, religious, social and cultural leaders determine to a 
considerable extent which beliefs become group beliefs and influence 
group members to accept them as such. They frequently select the group's 
goals which may serve as group beliefs, formulate ideologies or religious 
doctrines, decide what events in the past should be remembered as group 
beliefs for the future, select the symbols for the group, and decide on the 
attributes that characterize group members (for example, Y. Bar-Tal, 
1989). 
It is usually in the formative phase of group development that group 
beliefs are established. This is a period of malleability during which group 
beliefs are formulated and reformulated. With time, group beliefs emerge in 
a relatively stable formulation, although always with a possibility of change. 
In any event, as long as the group exists, it will always be faced with the 
talk of disseminating group beliefs and maintaining them among group 
members. 
These processes are necessary, since group beliefs serve as a foundation 
for group existence. The contents of group beliefs provide the raison d'etre 
for a group's existence and delineate its uniqueness vis-a.-vis other groups. 
The fundamental group belief indicates that the group exists. In addition, 
group members may add other group beliefs of a wide scope of contents. 
Group members share group beliefs, are aware of this sharing and believe 
that group beliefs define their 'groupness', that is, social identity. 


1 10 
Social identity: international pers
pectives 
Implications 
The proposed conception implies that group beliefs, as all beliefs, exist in 
the individual's mind. It does not suggest a view of group beliefs as a special 
entity of the 'group mind' category. Beliefs are related to individuals. 
Groups, organizations, societies or any other collectivity do not hold beliefs 
on the collective level - only individual members in the aggregate groups 
do. But the present conception describes a widely recognized phenomenon 
that group members share beliefs and these beliefs may be viewed as 
providing the basis for group members' social identity and as defining the 
essence of that group. Moreover, the presented approach suggests that 
group beliefs are more than a mere sum of group members' personal beliefs 
(see also Bar-Tal, in press). Sharing of a belief by group members and a 
recognition that it characterizes them provide group beliefs with 
distinguished properties. 
First of all, group beliefs serve as a basis for group existence. They 
provide a common basis with which individuals can identify and through 
which they can define their membership in the group. All groups have to 
have group beliefs, and a necessary condition for a formation of any group 
is formulation of group beliefs. This is in order that individuals should feel 
that they have something in common which distinguishes them from 
outgroups. They need at least one group belief - fundamental group belief. 
Tajfel's minimal group paradigm has demonstrated this effect. Using the 
minimal group paradigm, numerous studies have shown that the mere 
categorization of persons as ingroup or outgroup - that is, formation of a 
fundamental group belief - is sufficient to establish group identity and to 
influence group members' behaviour in the direction of triggering ingroup 
favouritism and outgroup bias (for example, Billig & Tajfel, 1973; Tajfel et 
aI., 1971). But in a real life situation group beliefs are needed in order to 
provide individuals with reasons for their group membership. Group beliefs 
supply the basis for cementing social identity. 
Second, group beliefs tend to arouse high confidence in their content. The 
perception of group members that a group belief is shared and characterizes 
them causes them to ascribe it high validity. The recognition that a belief is 
shared by group members reduces the subjective feelings of uncertainty and 
increases the perceived validity of this belief (see Festinger, 1954; Jones & 
Gerard, 1967). Therefore, group beliefs are usually considered by group 
members as facts, truths or verities. Another factor which contributes to this 
view is the perception that group beliefs define the essence of the group. 
Beliefs of this type cannot be held with low confidence, since low confidence 
may shatter the basic foundation of group existence. Therefore, shared 
beliefs which define the essence of the grouping are held with high 
confidence. 
One consequence of holding group beliefs with high confidence is their 
relative freezing in the group members' repertoire. Group beliefs do not 
change easily, since freezing implies closure. That is, group members do 


Group belie
f
s
 as an expression o
f
 social identity 
I II 
not easily entertain alternative hypotheses to group beliefs, but tend to 
collect information which validates them. The change of group beliefs is of 
special difficulty when they are central. The centrality of group beliefs 
indicates their importance for the group's life. Central group beliefs may 
become part of the group's culture and tradition, and then they are espe­
cially enduring. Group members pass them from generation to generation 
and rarely change them. 
Third, since group beliefs define the essence of the group, group members 
tend to organize their personal beliefs in accordance with them. Group 
beliefs provide the frame of reference for other beliefs. This process takes 
place especially in cases when group beliefs are central in group members' 
repertoire. Then, group members draw implications from group beliefs and 
add new personal beliefs consistent with them. In extreme cases of an 
orthodox group, group beliefs serve as anchors around which other beliefs 
are organized. 
Fourth, group beliefs provide the group members with special powers. 
The mere perception that group members share group beliefs indicates 
strength. Converse (1964) suggested in his classic analysis of beliefs that the 
number of people associated with a particular belief system is an important 
factor in political system: 
claims to numbers are of some modest continuing importance in democratic 
systems for the legitimacy they confer upon demands; and much more sporadi­
cally, claims to numbers become important in nondemocratic systems as threats 
of potential coercion. (p. 207) 
In addition, the power of group beliefs is derived from the perception of 
unity and commonality that may characterize group members who are 
aware of sharing beliefs. Knowledge that group members share a given 
belief expresses a unity of the group and indicates common fate. Group 
members derive strength out of this knowledge. They feel united, belonging 
and integrated within the group. One consequence of these feelings of 
strength is that group beliefs may serve as a basis for demands, desires or 
goals of group members. Leaders of the group usually take into con­
sideration the group's beliefs when they make decisions that affect the 
group's life. Being aware that group members share a belief, and are 
influenced by the belief in their group behaviour, leaders pay special 
attention to group beliefs. Group beliefs reflect the direction that group 
members desire to take in their behaviour. Therefore, leaders frequently 
make decisions regarding group behaviour that correspond to group beliefs. 
Fifth, group beliefs may determine the attitudes and behaviours of an 
outgroup towards the group. Group beliefs are important information for 
outgroups about any given group. They characterize the group and may 
imply even the behaviour that the group may take. Group beliefs may 
indicate possible goals, ideology, values, history, norms or characteristics of 
the group. This information enables acquaintanceship with the group and in 
turn influences the type of intergroup relations that may develop. 


1 12 
Social identity: international pers
pectives 
Finally, if one believes that group members act according to their beliefs, 
then group beliefs should be considered as an important source of under­
standing group behaviour. As Krech, Crutchfield and Ballachey (1962) 
suggested: 
Man acts upon his ideas. His irrational act no less than his rational acts are guided 
by what he thinks, what he believes, what he anticipates. However bizarre to 
the behavior of men, tribes or nations may appear to an outsider, to the men, to the 
tribes, to the nations, their behavior makes sense in terms of their own world view. 
(p. 1 7) 
Group beliefs provide the cognitive basis to many group behaviours. They 
may serve as reasons, goals, explanations or justifications for group behav­
iour. In other words, group beliefs may function as guiding forces for a 
group, and, therefore may determine the direction, intensity and persistence 
of group behaviour. This is one of the most important implications of group 
beliefs. 
Conclusion 
Group belief conception provides cognitive and social perspectives to the 
study of a group. The existence of a group is a social reality for group 
members. Group beliefs provide the basis which allows group members to 
view the group as a social reality. The group, then, is a product not only of 
structural characteristics, environment, situational conditions, motivational 
tendencies or social influence, but also of personal cognitive processes. 
These processes determine the essence of the group, since group beliefs 
provide the contents which serves as a basis for group formation and group 
maintenance. 
This conception carries the notion of social identity one step beyond the 
Tajfel and Turner theories. It suggests that social identity is not based 
merely on the categorization process, but also on beliefs which provide the 
rationale for the group existence. The fundamental group belief indicating 
'We are a group' and additional group beliefs of different contents which 
define the nature of the groupness allow individuals to feel, think and act as 
group members. These beliefs underlie the uniqueness of the group and 
diff
erentiate it from outgroups. 
It is important that behavioural scientists study the contents of group 
beliefs and the process through which group members acquire them, become 
aware of sharing them, and are affected by them in their behaviour. This 
line of interest will, on the one hand, add a cognitive aspect to the study of 
a group and, on the other hand, liberate beliefs from their individual closet 
to extent their social meaning. In this respect, the present approach provides 
the bridge between the individual's level of analysis and group level. Group 
members, as individuals, are the ones that acquire beliefs which shape their 
reality. But when group members become aware of the fact that their beliefs 
are shared by other group members, these beliefs acquire special quality. 


Group belie
f
s
 as an expression o
f
 social identity 
1 1 3 
That is, they may have important cognitive, affective and even behavioural 
implications for the group members as individuals and for the group as a 
whole. 
Note 
I .  Most of the examples provided in the following review are drawn from formal statements 
prepared by the groups. Although it is possible that they reflect group beliefs, group members 
were not examined to indicate their responses. Therefore, the presented 'group beliefs' should 
be treated with caution. They only serve as illustrations for presenting the conception. Further 
research among group members may determine the actual status of these beliefs. 


8 
Social Identity and National Identity 
Jose Miguel Salazar 
Of all the variants of social identity, national identity is perhaps the one 
that has had the most dramatic impact on historical events, with both 
positive and negative consequences for humankind. 
The great romantic-nationalist movements of the nineteenth century, as 
well as the movements of national liberation in the first half of the twentieth 
century and the most recent break-up of multinational European states, are 
clear examples of the past and present role played by this type of identity. 
Identity with national or ethnic categories is evidently also a reference point 
used by Tajfel in the development of his theory, as inferred from the 
character of some of his early work (Tajfel, I 
969b, I 
970a). 
Even though the eighteenth and nineteenth centuries have been called the 
era of nationalism (Kohn, 1 962), this designation could also apply to the 
present century, as there is ample evidence of creative manifestations and 
excesses of the national idea in our time. In the aftermath of the Second 
World War, nationalism and independence were bywords of the movements 
to reaffirm identities in countries which were colonies in the previous 
century, particularly in Asia and Africa. Yet not only in the so-called third 
world is evidence found of the ascent of this type of identity; it is increasingly 
noticeable in the awakening of subdued nationalisms in European countries, 
and in the breaking-up of previously functioning multinational states. Also 
in the Americas there are clear manifestations of its increasing importance 
from Canada, to Mexico, to Peru; and even in the Pacific area with the 
awakening of the aboriginal populations of Australia, silent for centuries. 
Whoever suggests that the era of nationalism is finished and that we are 
in a new era of universal humanity must be disconnected from his or her 
surrounding reality. The allegations of 'proletarian internationalism' which 
proposes that the proletariat has no country, as well as of 'cosmopolitan­
ism', which indicates that as a consequence of transnationalization now it is 
the bourgeoisie that has no country, are both questionable. On the contrary, 
the so-called process of globalization and the development of the 'global 
village', based on technological improvements, constitutes a good example 
of the dialectics of the historical-cultural process, since it has had the 
curious effect of reinforcing ethnicity and national identities. 
It has been claimed that nations and nationalism only acquire reality in 
relation to the state and that historically these are intimately connected with 


Social identity and national identity 
1 1 5 
the ascent of the bourgeoisie, and should be considered transient phenomena 
only referred to as a stage of historical development (that is, Kedourie, 1 960; 
Kohn, 1 944; Seton-Watson, 1965). Yet it is possible to differentiate between 
nationalism, or more precisely 'modern nationalism', clearly associated with 
the state, and 'national sentiment', which is more ancient and of a more 
general and less political character (Hertz, 1 944). 
The emphasis on the nation-state is important from the perspective of 
elaborating a political history, but in the study of national phenomena from 
a socio-psychological perspective it is impossible to ignore this other facet. 
Hence in this chapter we shall begin by considering the affective back­
ground of national phenomena. After that, in order to gain a better under­
standing of national identity as such, some developments within the social 
identity approach will be examined in order to answer some questions about 
the problem in hand. 
Considerations about national sentiments 
The definitions of what a 'nation' is have been very varied. The term 
appears in a fifteenth-century Bible translating a Hebrew word that means 
at the same time 'ascendance' and 'posterity'. It is also used to refer to a 
natural quality, a native quality. In the medieval universities a 'nation' 
referred to a group of students from a common region. And the idea of 
belonging to a group based on descendance, differentiable from others, was 
present and important in history long before the appearance of modern 
nationalism (Smith, 1 99 1 ;  Chap. 2). 
Even in the case of empires as ancient as Egypt and China, national 
sentiment was important, at least for members of some strata. Evidence of 
this is found in some hieroglyphs about a Theban prince who expelled the 
hyksos from Egypt around 1 580 Be; they say: 'I will grapple with them, and 
cut open their belly! I will save Egypt and overthrow the Asiatics' (cited by 
Smith, 1 991, p. 46). Even in Greece, being an Athenian or a Spartan had 
emotional and behavioural implications; the feeling of being an alien, 
dramatized in Euripides's Medea, must surely reflect a reality understood by 
the audience of the drama. 
On the other hand, the biological basis of national sentiment, and more 
concretely of ethnocentrism, has been considered by social scientists 
interested in socio-biology. Campbell (1965) in his provocative essay 'Ethno­
centric and Other Altruistic Motives', elucidates the thesis that ethnocentric 
motivation and behaviour is directed to the protection and survival of the 
members of the ingroup, leading to sacrifices and even death. This act by 
definition is an altruistic type of behaviour, going against personally selfish 
interests. More recently Reynolds, Falger and Vine (1987) have published an 
interesting book, The Sociobiology o
f
 Ethnocentrism, that gathers more up­
to-date contributions from this same perspective. 


1 16 
Social identity: international pers
pectives 
There has been great insistence that 'nationalism' is a phenomena of 
modernity, and that it is incorrect to consider it something more trans­
cendental. This is a way of dodging considerations derived from the 
biological condition of humankind and the fragmentary but unquestionable 
evidence of the antiquity of 'national sentiments'. There is no doubt that 
feelings of attachment to the nation-state are a relatively recent phenom­
enon since the objects of attachment ('modem' nations and the 'nation­
state') are also recent, but to deny the attachment to what is perceived as 
one's own (be this a city-state, as in Greece, or a tribe in the Amazonian 
forest) seems exceedingly restrictive. To do this does not take into account 
some basic considerations about the human condition. 
In speaking of a 'nation', reference to the fact of 'birth' is important. 
(Although in some languages this relationship has been lost or never existed, 
it is very clear in Spanish [nacer = to be born; nacion = nation] and in other 
Romance languages.) There is a determination given by belonging to a 
group in which one was born, an ascribed role that confers an important 
anchor to the individual. So although the term 'nationalism' in effect 
appears in common usage around the period of the French Revolution, the 
term 'nation' is much older. The concept of nationalism is endowed with 
clear political references to the nation-state, which is not necessarily the case 
for the concept of nation. 
There are groups that in the accepted sense constitute nation-states, 
but there are other conglomerates, based on the same principle of birth, 
that may generate loyalties and attachments as strong as, or sometimes 
even stronger than, the nation-state. The following three cases may be 
mentioned: 
(a) the positive affect towards a regional group within an existing state 
('regionalism'); 
(b) the attachment to a grouping that ignores existing geographical borders, 
as in the case of the Kurds of the Middle East or the Waiyu in the 
Goajiro Peninsula that Venezuela and Colombia share (,stateless 
nationalism'); 
(c) the positive feeling towards supra-national groupings, as in the case of 
Latin America ('supra-nationalism'). 
We consider that there are certain commonalties in the psychological 
contents of the sentiments expressed in these cases, as well as in the case of 
state nationalism. And it is this broadly defined national referent which is in 
our mind when we examine the question of 'national sentiment' in this 
chapter. 
So what is the basis of this national sentiment? If we take a close look, we 
find that it is based on four elements: (a) territoriality, (b) shared culture, (c) 
historical memory of genealogical communality and (d) the existence of a 
nation-state. 


Social identity and national identity 
1 17 
Territoriality 
Ardrey (1967) extensively develops the point of the importance of territory 
in human behaviour, going back to the biological base of animal terri­
toriality and making use of the concept of territorial imprinting. He defines 
what he calls the 'biological nation' as 'a social group containing at least 
two mature males which holds as an exclusive possession a continuous area 
of space, which isolates itself from others of its kind through outwards 
antagonism, and which through joint defense of its social territory achieves 
leadership, cooperation, and capacity for concerted action' (p. 1 91). His 
argument is centred on an evolutionary position and observations of 
phylogenetic ancestors, which from his point of view give evidence of the 
antiquity of the phenomenon 'nation'. Of course, like many other animal 
analogies, this may be questioned, especially if it is sustained that language 
and culture introduce a qualitative break with evolutionary continuity. Yet 
the observations continue to have interesting implications. 
The excesses of the approaches centred on geographical determinism, in 
vogue in the nineteenth century, and its later rejection, have made us lose 
sight of the importance of the geographical environment in relation to 
behaviour. Even though the geographical referent is more tangible and 
objectifiable than other referents of nation, there are few authors who give it 
the relevance it deserves, one exception being Doob (1964), who bases his 
studies about patriotism in Tyrol on the attachment to the 'Heimat'. Also 
there has been some recent rethinking about the theme amongst environ­
mental psychologists (Altman & Low, 1992; Proshansky, Fabian & 
Kaminoff, 1983). 
There are multiple cases in history of the fight to hold on to territories by 
people who were born in an area, which may not be the best from the point 
of view of economic rationality; and there are many well-known examples 
of the feeling of ostracism of migrants or displaced peoples who must live in 
a territory which is not their own. 
But while it is very concrete, the territorial referent is more evident in the 
regionalism variant of national loyalty. Given geographical variability 
within countries, it is logical that the role of territory and all that is con­
nected to it should be more marked in more restricted loyalties. 
Altogether there is no doubt that territoriality plays an important role in 
the development of national sentiments, 
Culture 
Sharing a culture involves collective ways of solving the problems con­
fronted in interacting with the surrounding world, developed and trans­
mitted from generation to generation. This level of shared experience is also 
important in the development of national sentiments. 
Language is the supreme creation of culture, so much so that for many 
anthropologists it is considered its defining trait. In any case it permits the 
transmission of collective solutions and their preservation. Language is also 


1 1 8 
Social identity: international pers
pectives 
a very early acquisition in human ontogenesis, and so it becomes a basic 
point or matrix for the development of feelings of attachment. 
There are other elements of culture that become obvious referents of 
attachment: music, different elements of the so-called 'material culture', 
customs, shared values. They may all be very powerful in awakening 
positive feelings and loyalties towards the national ingroup. 
Ethnicity 
Given the conceptual weaknesses and dangerous political use of the term 
'race', the concept of ethnicity has become a partial substitute. Smith 
(I 
984a, 1984b) develops the idea of 'myths of descent' as the basis of ethnic 
feeling and ethnic nationalisms. Through these myths, the members of a 
community perceive they have a common origin, usually connected to a 
glorious past. An ethnie is structured around these mythological socio­
cultural elements, the association with a common land, elements of 
common culture, and something of great relevance: a 'collective proper 
name' (Smith, 1 991, p. 21). It must be clear that in talking about myths no 
attempt is made to disprove the importance of the phenomena, since myths 
are not to be ignored. As Mariategui (1925), a very important Peruvian 
thinker, said: 'Myths move mankind through history. Without a myth, the 
existence of mankind has no historical sense' (p. 22). Hence considerations 
about ethnicity cannot just be discarded, and obviously are connected to the 
development and maintenance of national sentiments. 
State 
It is indisputable that the existence of a national state stimulates an 
important type of national sentiment through the systematic use of ideology 
producing institutions partially in the hands of the state like the school 
system and the mass media. This is so pervasive that for some the only 
really important national sentiments are those related to the state. But this 
is too restrictive, as regional, ethnic and more recently supra-national 
identities may have as much reality, from a psychological perspective. 
There is an additional problem in restricting the study of national 
sentiment to what occurs at the level of the nation-state, and it is that the 
idea that a nation equals a state is not a viable equation in our present 
world. It constitutes, at most, an ideal type, but the great majority of states 
are really multinational. Some contemplate it explicitly, as Canada or 
Spain, where there is recognition of the existence of diverse nationalities 
within the state. However, in states such as Venezuela the fiction of a 
culturally homogeneous state is disturbed with news of massacres of 
indigenous tribes, such as the Yanomami. Some states that are explicitly 
multinational apparently function well, although surprises occur when the 
latent centripetal forces gain ascendance over the forces of integration, as in 
the painful case of Yugoslavia. 


Social identity and national identity 
1 19 
In defining a nation, there is implicit the existence of one or several non­
nations and the existence of limits separating them. When a nation is 
defined in terms of a state, it is usually clear what the limits are. But the 
same doesn't happen when you base your definition on any of the other 
elements: the territorial, the cultural or the ethnic. Hence it becomes 
necessary to look for theorizings capable of suggesting ways of solving this 
problem. 
Social identity theory frames the problem of 'borders' in socio­
psychological terms. Consequently this theoretical approach is essential in 
any attempt to understand national phenomena. Hence we shall consider 
very briefly some elements of this approach which may help us to 
understand better the question in hand. 
The social identity approach and its relevance to understanding 
national phenomena. 
The identity theory approach is relevant to the problem of defining group 
borders since it places it at the irrefutable level of human subjectivity. This 
theoretical approach, which in its beginnings Tajfel sometimes referred to as 
CIC theory (categorization-identity-comparison) (Tajfel, 1 98 1 )  has 
stimulated multiple theoretical and empirical developments, examples of 
which can be found in the other chapters of this book. 
The three original basic elements, the existence of a category accepted 
internally and externally, the identification with it, and the process of social 
comparison, are still very fruitful concepts. The first element becomes 
concrete in the form of a label; the second element implies the acceptance of 
the category for oneself with all the affective, cognitive and behavioural 
consequences. Regarding the third element, the process of social com­
parison, it is frequently a positive evaluation of the assumed category, but 
this is not always the case. Tajfel (1981, Chap. 1 3) himself presents evidence 
of studies with children in which there is evidence of devaluation of their 
own ethnic group. Several studies have produced evidence of devaluation of 
the ingroup (Salazar, 1 983a). The postulation of the need to maintain a 
positive group evaluation seems to have been taken from Sumner's (1906) 
ethnocentrism theory, considered by many as constituting a very basic 
means of understanding intergroup relations (Le Vine & Campbell, 1 972). 
Many developments have occurred on the basis of the original elements 
of social identity theory, among these accounts regarding the basis of 
categorizations (Billig & Tajfel, 1 973; Turner, Hogg, Oakes, Reicher & 
Wetherell, 1 987), meta-contrast ratios (Turner et ai., 1 987), the acquisition 
of categories, the hierarchy of identities; the process of depersonalization, 
the salience of categories, and so on, are very relevant to understanding 
national identity. 
Making use of some of these developments, we have attempted answers 
to some questions related to the understanding of national identities. 


120 
Social identity: international pers
pectives 
Questions about national identity 
1 .  What is the motivation that underlies the acquisition o
f
 national identities? 
Kelman (1969) considers that it is based on both perceived instrumentality 
and a need for transcendence; the latter would give a particular force to the 
social identities related to the national. With a bit of poetic licence, it may 
be said that the existence of national communities is based on the desire for 
eternity. So the idea of some to conceive of nationalism as a form of religion 
(Hayes, 1 926) is not far-fetched; nationalism may be conceived as a lay 
religion that offers, like the others, the promise of eternity. 
The nation as a community links people not only with a present inhabited 
by other people they do not and could not possible know, but also with a 
past and a future. This is possible through depersonalization, derived from 
the process of self-categorization. In reality that which transcends, that 
which is eternal, is the group, but when this happens this 'eternity' gets 
transferred to the depersonalized individuals that are part of it. 
This may be seen as an attempt to postulate a basic religious motivation 
in human beings. But whether religiosity is something intrinsic to human­
kind or is a historical phenomenon is not important. What is argued here is 
that national identity is based on an aspect of such religiosity, basically the 
need to transcend. 
In more mundane terms, the 'transcendence' of a nation may be con­
ceived of as attainable by the power it exerts or has been able to exert 
within the international context. Hence national identities have their more 
powerful manifestations related to political states, which are by definition 
conceived of in power terms. 
2. On what basis do the perceptions o
f
 similarities and dif
f
erences arise in 
the case o
f
 national identities? 
It is known that a dialectical relationship exists between categorizations, 
expressed in a label and the perception of similarities. The mere existence of 
a label is not enough, for a base must exist to give justification or backing to 
assuming the category. It is also evident that the dimensions around which 
the comparisons take place vary in terms of the identities considered. 
Humans' capacity for interaction is limited. Dunbar (1992), making a 
projection on the basis of the size of grouping in higher primates, suggests 
that 1 50 is the limit of a viable human group, something like the natural 
size of a human group. Nevertheless, humankind lives in communities much 
larger than that, hence necessarily we live in 'imagined communities' 
(Anderson, 1983). An ethnic community or a nation, is hence one of these 
'imagined communities', but even these must be construed around certain 
elements or dimensions. 
When the question of the basis of national sentiments is analysed, four 
elements come to the fore: territory, culture, ethnicity and state. These 
elements may be considered as dimensions along which meta-contrast ratios 
of similitude-difference may be established. Particularly in the cases of 
culture and ethnicity, and in a lesser degree in the case of territory, it is 


Social identity and national identity 
121 
possible to conceive of continuums in which the meta-contrast ratios 
gradually change. In the case of the state the categories are discrete and 
variations in meta-contrast ratios are less applicable. 
3. Could we conceive o
f
 levels or hierarchies in the case o
f
 nation-related 
identities? 
In order to encompass the similitudes between state nationalism, region­
alism and supra-nationalism, we may speak of 'nation-related identities'. 
These identities tend to occur in concentric circles. The idea of concentricity 
of 'national-related' identities could be considered as a particular case of 
hierarchies within the intermediate level, taking as the higher one that of the 
individual as member of humanity, and the lowest that of the individual 
self. 
We suggest that there exists a common psychological substratum to 
identities centred on a region or a specific regional group, and the one 
centred on a specific nation-state or the emergent supra-national identities. 
We see this as a matryoshka of identities, one inside the other. 
The relative importance of these identities is not permanent and situ­
ational factors play an important role in determining their role or effect on 
behaviour. 
4. How does the process o
f
 social comparison f
unction in the case o
f
 
national identity? 
In the case of national identities the question of social comparisons is 
particularly important. Social comparisons will affect both the salience of 
the category as well as its evaluation, and influence the development of a 
positive social identity. The social comparisons that are established involve 
other nations or ethnic groups. These comparisons may take place along 
different dimensions: size of the countries, political systems, state of the 
economy, 'well-being' or 'happiness' of the inhabitants, and so on. In 
different cases the activated meta-contrasts will vary. 
But in the activation of 'nation-related' concentric identities, a basic 
element to be considered is the degree of transcendence that identity may 
guarantee without losing the essence of the identity. It may be argued that 
the more extensive the reference group, the more positively valued will be 
the identity, since the possibility of transcendence is greater. But it is also 
true that the smaller the unit, the greater the homogeneity, and hence the 
more solid the meta-contrasts. There must exist an equilibrium point 
between both elements, possibly of transcendence and solidity of identity, 
which constitutes an interesting empirical problem. 
There is evidence that when Latin Americans and Venezuelans (or 
Peruvians or Colombians) are evaluated, the social identity is more positive 
in the first case (Salazar, 1 983b), and it may be that this is due to an 
underlying perception of greater possibility of transcendence at that level 
than at the more restricted ones. Something similar seems to be occurring in 
Europe. 
But as has been frequently pointed out, positive social identities may be 
connected to negative manifestations of aggressive ethnocentrism. The 


1 22 
Social identity: international pers
pectives 
development of an identity within a wider context redefines the outgroup, 
and here the European case is illustrative. The strengthening of the sense of 
being European, a psychosocial correlate to the development of the 
European community is evident (Echebarria, Elejabarrieta, Valencia & 
Villareal, 1 992; Fells & Niznik, 1 992). This has led to the channelling 
towards non-Europeans of feelings of rejection that in previous periods had 
been directed to other groups within Europe. The racist manifestations 
against Turks in Germany, against North Africans in France, are cases in 
point. In Spain the development of 
the feeling of being a part of Europe has 
been accompanied by a rejection of Dominicans and other 'South 
Americans' that in another definition of the situation, that is, in terms of a 
Hispanic community, were considered less objectionable. 
The varying processes of social comparison point to the importance of 
exploring the consequences that may be derived from placing the same 
individual in varying referential frames, be these regional, national or supra­
national. 
5. In what conditions do dif
f
erent 'nation-related' identities become salient? 
To begin with, it must be accepted that nation-related identities become 
activated by field factors, situational factors that affect the salience of a 
particular category. In the case of the European community, economic and 
political developments activated a supra-national identity that has been 
latent, and submerged under a history of conflicts. Obviously identities are 
politically manageable. However, for this to occur they must exist in a 
latent form. 
In defining the ingroup, there exists an implicit definition of outgroups. 
Yet these definitions are not static. In emphasizing one level of the nation­
related identities, a group that shares the same identity at another con­
centric level could be defined as an external group. An example makes this 
clear. The identities of a Caraquefio (from Caracas), a Venezuelan and a 
Latin American are nation-related and subsist in concentric circles. When 
my Caraquefio identity comes to the fore, I consider people from other 
regions of my country as external groups, with possible behavioural conse­
quences. This ceases to be so when my Venezuelan identity gets activated, in 
which case a Colombian, or another national group, gets defined as the 
external group. This externality would disappear when Latin American 
identity is invoked, since it encompasses both groups at a higher level of 
abstraction. 
For a Latin American living in Europe, the Latin American identity is 
easily activated; in a Latin American country different from ours the 
Venezuelan identity becomes frequently activated; in a gathering taking 
place in a small provincial town, the regional identity of coming from the 
capital city may come to the fore. It is here that the questions of accessi­
bility and ad
justment come into play. 
But there are qualifications that must be made in relation to the thesis of 
concentricity. It is erroneous to think that there exists perfect balance or 
concordance within the concentric circles. On the contrary, it must be 


Social identity and national identity 
123 
recognized that sometimes these may appear as contradictory and in effect 
enter into conflict. In this sense the Basque case is particularly significant, 
where for many this identity is independent of the Spanish one, and hence 
does not function in concentric circles. According to Linz (1986), a high 
percentage of the inhabitants of the Basque provinces do not consider 
themselves Spaniards. Nevertheless the idea of concentric circles probably 
functions with regard to the European identity (Torregrosa & Ramirez, 
199 1). 
Another problematic situation concerns the conflict between national 
identity and other types of identity, such as those related to partisan politi­
cal loyalties. Here these may take precedence and lead to dramatic situ­
ations as those of a civil war, as has happened so many times in history. 
National identity may also come into conflict with human identity, as 
would be the case with a soldier who refused to comply with the order to 
kill someone on the basis of their national label. 
This brings to the fore the question of 'globalization'. The interrelation 
between the different countries in the world has always existed; what is new 
is the development of technologies that bring us close enough to find out 
what is happening in other places in the world with a speed non-existent in 
previous times. But while it may be true that this facilitates the imitation 
and copying of foreign models, usually from more developed countries, it 
also has created consciousness of similitudes and differences, and does not 
necessarily lead to homogenizations. A more probable outcome, it seems to 
me, would be the development of wider communities, on the basis of 
common elements, in which the sense of national identity is not eliminated, 
but transferred to a higher level. 
In my view, homogenization is not a desirable goal. Rather, I would 
argue that the recognition and acceptance of differences is preferable to the 
universal assimilation of all to a single pattern. 


PART III 
IDENTITY AND GROUP/ 
INTERGROUP PROCESS 
9 
Social Identity and Intragroup 
Differentiation as Subjective Social 
Control 
Jose M. Marques, Daria Paez and Dominic Abrams 
In analysing collective behaviour associated with religious movements III
 
sixteenth-century Europe, the historian 1.1. Delumeau (1993) wrote: 
The Protestant uprising naturally provoked an increasing aversion to heresy, 
which was considered as the ultimate evil . . .  
: the traitor inside is worse than the 
enemy outside. He must be made to leave his hide-out, he must be eliminated as a 
priority and no other punishment is severe enough for him. (pp. 514-51 5) 
This statement illustrates the fact that the deviate members of a large 
community may be more harshly treated than outsiders, because, while both 
depart from the community's values, the former threaten these values more 
severely than the latter. Deviates' behaviour and beliefs endanger the 
collective value standards more to the extent that these deviates cannot 
simply be expelled to the outgroup. 
Punishment of deviates and traitors appears to be an important com­
ponent of a community's social regulation system. Another historical illus­
tration of this phenomenon is witch-hunting in the Middle Ages (for 
example, Hamilton & Rauma, 1995; Yamagishi, 1 995). Witch-hunting, as 
well as more recent instances of collective persecution - like the Inquisition 
in the fourteenth- to sixteenth-century Iberian countries, McCarthyism in 
the United States in the 1 950s, or political purges in the Stalinist period 
in Soviet Union, to mention but a few - often find support from the general 
population, or, at least, from those who identify with the value-system that 
legitimates the persecution, even when the persecuted do not represent an 
immediate or material danger. This support is sometimes tangible but, most 


Social identity and intragroup dif
f
erentiation 
125 
often, emerges in the form of derogatory attitudes and judgements. It likely 
stands for the individuals' commitment to the community's standards and 
self-definition in light of these standards (cf. Doise, Chapter 2, this volume). 
Concomitant with this social self-definition is the effort to ensure and to 
legitimize beliefs on the distinctiveness and relative superiority of the 
community's standards as compared to the outgroup. This process may also 
be viewed as a reaction against social change and as an immediate psycho­
logical response against threatening minorities cf. Perez and Mugny, 
Chapter 10, this volume). It may be a major factor in the maintenance of a 
clear-cut and secure social identity as well as a guarantee of the individual's 
psychological well-being as a group member. As sociologist Georg Simmel 
(1955) put it, in groups structured around their members' attitudes. 
the opposition of a member to an associate is no purely negative social factor, if 
only because such opposition is often the only means for making life with actually 
unbearable people at least possible. . . . It allows us to prove our strength 
consciously and only thus gives vitality and reciprocity to conditions from which, 
without such corrective, we would withdraw at any cost. Opposition achieves this 
aim even where it has no noticeable success, where it does not become manifest 
but remains purely covert. Yet while it has hardly any practical effect, it may yet 
achieve an inner balance . . .  
, may exert a quieting influence, produces a feeling 
of virtual power, and thus save relationships whose continuation often puzzles the 
observer. (p. 19) 
In the present chapter, we argue that identification with ingroup norms 
and values may lead individuals to subjectively engage in processes analog­
ous to social influence upon actual or represented ingroup instances. These 
subjective social influence processes may allow individuals to restore their 
confidence in the group and, hence, to re-establish a positive social identity. 
The social psychological analysis of small groups and large social 
categories 
Social psychological research has focused on different factors involved in 
the explanation of group processes. Traditionally, this research evolved 
along two separate lines which, for simplicity, we may designate as the 
'small-group approach' and the 'social identity framework' (including self­
categorization theory). The former has focused mainly on processes like 
productivity, decision-making, role and status differentiation, cohesiveness 
or socialization (cf. Brown, 1 988; Cartwright & Zander, 1 968; Levine & 
Moreland, 1994). Most of this work has been conducted in the realm of 
face-to-face groups. The latter has emphasized the effects of social cat­
egorization in the psychological dynamics involved with individuals' sense 
of group membership. This work has often been conducted with large social 
categories and their representations (for example, Tajfel, 1978d). Only 
recently have these two lines been subject to integrative efforts (Hogg, 1992; 
Hogg & Abrams, 1 988; Marques & Paez, 1994). 


126 
Social identity: international pers
pectives 
Group affiliation, social influence and intragroup differentiation 
Research in small groups was strongly influenced by the work of Festinger 
and colleagues. Festinger (1950) proposed that affiliation in groups fulfils a 
social reality f
unction, that is, uncertainty reduction through selective 
affiliation with those who share similar beliefs, and a locomotion function, 
that is, facilitation of goal achievement through the association with those 
people whose goals are similar to the individual's. Indeed, people in small 
groups generate consensual opinions (Festinger, Schachter & Back, 1950) 
and reject deviates who resist group pressures to comply with the group's 
position on a relevant matter of opinion (Schachter, 1951), or who inten­
tionally fail to contribute to the group's goal achievement (Jones & 
DeCharms, 1967; cf. Levine, 1980). Most importantly, evidence shows that 
increases in external pressures upon the group, in group uniformity and in 
the discrepancy between majority and deviate positions increase the 
majority's pressures towards uniformity (Cartwright, 1968; Janis, 1968; 
Levine, 1 980; Moreland & Levine, 1982; Shaw, 1976). Evidence also suggests 
that social reality and group locomotion are ensured within the group 
through two forms of social control, traditionally known as inf
ormational, or 
private, and normative, or public, influence (Jones & Gerard, 1967). How­
ever, research in small-group processes has rarely addressed itself - at least 
conceptually - to the role played by intergroup relations in processes 
occurring within groups (Hogg, 1992; Hogg & Abrams, 1988; Marques & 
Paez, 1994). 
Social categorization, social identification and intergroup 
differentiation 
In tum, research on large-scale categories has emphasized intergroup more 
than intragroup processes. This may be due to the influential work of Tajfel 
and colleagues (for example, Tajfel, 1978d; Tajfel & Turner, 1979; Turner, 
1975a). This work started from the assumption that social behaviour may 
be conceived of as occurring along one continuum, whose poles are inter­
group and interpersonal behaviour: 
What is meant by 'purely' interpersonal is any social encounter between two or 
more people in which all the interaction that takes place is determined by the 
personal relationships between the individuals and by their respective individual 
characteristics. The 'intergroup' extreme is that in which all behaviour of two or 
more individuals towards each other is determined by their membership in 
different social groups or categories. (Tajfel, 1978a, p. 41) 
Whereas interpersonal behaviour involves the perception of features 
specific to the individuals involved, intergroup behaviour implies self- and 
other-stereotyping. In categorizing themselves as members in a social 
category, people focus on those features that they share with all category 


Social identity and intragroup dif
f
erentiation 
1 27 
members. Brown and Turner (1981) advanced this idea under the heading 
of the self-stereotyping hypothesis. 
In brief, the social identity framework proposes that group representa­
tions correspond to prototype-like representations, because intergroup 
similarities and intragroup differences are irrelevant in intergroup situations. 
Clearly, this assumption restrains a view of intergroup and intragroup 
processes as simultaneously occurring in social life. Although this is not 
entirely true (Hogg, 1 992; Turner, 1 99 1), it nevertheless remains that the 
social identity framework (including self-categorization theory) has diffi­
culties in dealing with the simultaneous emergence of intergroup and 
intragroup behaviour. 
Group prototypes as denotative norms 
Among other well-known assumptions, the social identity framework states 
that, in social life, people attempt to achieve cognitive clarity by exagger­
ating intergroup differences and intragroup similarities (Hogg, 1 992; Doise, 
Deschamps & Meyer, 1 978; Tajfel, 1 969a, 1 978d; Tajfel & Wilkes, 1 963). 
Basic to this process is comparative fit (Oakes, Haslam & Turner, 1 994, and 
Chapter 6 this volume; Oakes & Turner, 1 990). Comparative fit stands for 
the actual differences and similarities between the stimuli present in the 
judgement setting that allow the establishment of correlations between 
continuous stimulus properties and discrete categories (cf. also Tajfel, 
1969a). 
According to Turner, Hogg, Oakes, Reicher and Wetherell (1 987; cf. also 
Hogg, 1 992; Hogg & McGarty, 1 990), comparative fit generates the cog­
nitive construction of group prototypes. This process has been formalized in 
terms of a meta-contrast princi
ple (Turner et al. 1987; cf. also Hogg, 1 992; 
Hogg & Abrams, 1 988; Hogg & McGarty, 1 990). Given a distribution of 
group member positions on a relevant intergroup dimension (for instance, a 
personality attribute or an attitude dimension), the meta-contrast principle 
allows perceivers to ascertain whether intergroup differences are, on the 
average, higher than intragroup differences. Formally, a ratio can be com­
puted for each group member, so that the group's most prototypical 
member is the one who shows the highest ratio and thus occupies that 
position which simultaneously maximizes intergroup differences and 
minimizes intragroup differences. This position corresponds to the group 
prototype. From this point, in 
group and outgroup members are subjectively 
assessed in terms of their respective prototypes. As a result intergroup 
differences are emphasized to the detriment of interpersonal differences 
among the members of the same group and interpersonal similarities among 
ingroup and outgroup. 
In addition, according to self-categorization theory, individuals expect 
group members to match their respective groups' prototypes. Once social 
categories become salient, perceivers may judge the 'normative fit' (Turner 


1 28 
Social identity: international pers
pectives 
et aI., 1987), that is, the correspondence between the activated categories and 
the content properties of stimuli, in the situation. Lack of normative fit 
would lead perceivers to switch to more adequate categorization criteria 
(Oakes et aI., 1994), that is, to alternative intergroup dimensions which best 
allow the generation of clear-cut intergroup boundaries, and more eff
ectively 
account for the judgemental situation. Different criteria of categorization 
will be used to the extent that, depending on the stimulus properties of the 
judgemental setting, they succeed or fail in cognitively maximizing 
intergroup differences and intragroup similarities (Haslam & Turner, 1992; 
Turner et aI., 1987; cf. also Bruner, 1957). To give an example, in a political 
rally, comparative fit could entail a 'liberal versus conservative' categor­
ization. But if in that rally one noticed that, regardless of their political 
affiliation, female speakers supported women's rights whereas men addressed 
themselves to other issues, one would likely switch from the initial 'liberaV 
conservative' dimension to an alternative 'female/male' categorization. 
In brief, for self-categorization theory, group prototypes fulfil structuring 
functions. The meta-contrast generates cognitive clarity in that it allows 
individuals to perceive the social world in terms of meaningful, differ­
entiated units. Concomitantly, individuals deal with emerging intragroup 
differences by changing the criteria of categorization to alternative dimen­
sions which hinder such differences in order to steadily maximize intergroup 
distinctiveness. They are strongly motivated to maximize intergroup distinc­
tiveness, and this requires that social situations be cognitively construed so 
that intragroup differences are minimized. 
But, according to self-categorization theory, group prototypes also fulfil 
emotional functions (Hogg, 1992). Following the above-mentioned self­
stereotyping hypothesis, self-categorization theory states that individuals 
identify with groups to the extent that they perceive a match between the 
self and the ingroup prototype. When self-stereotyping occurs, the ingroup 
becomes fully representative of the self, and the self becomes perceptually 
interchangeable with other ingroup members (Turner, 1 981). According to 
Hogg (1992; Hogg & Abrams, 1988), a major implication of this fact is that 
social identification must be logically conceived of not as an outcome of 
interpersonal attitudes or similarity among group members, but, rather, 
as the result of a positive orientation towards the group prototype. As he 
put it, 
when a specific social identity is the salient basis for self-conception, self­
perception and conduct become ingroup stereotypical and normative, perceptions 
of relevant outgroup members become outgroup stereotypical, and intergroup 
behaviour acquires - to varying degrees, depending on the history of relations 
between the groups - competitive and discriminatory properties. (Hogg, 1992, pp. 
90-91) 
In short, when the ingroup is cognitively salient, a positive orientation 
towards the self becomes intrinsically equivalent to a positive orientation 
towards the group. Therefore, we may see social identification as a positive 
orientation towards the ingroup prototype, as it materializes in the 


Social identity and intragroup dif
f
erentiation 
1 29 
stereotyped self (Hogg, 1 992; Hogg & Abrams, 1 988). Furthermore, as 
individuals assimilate themselves to an ingroup prototype, their thoughts 
and behaviour conform to this prototype. The prototype is thus internalized 
as a norm defining the appropriate features and behaviour standards of the 
social self (Turner, 1 99 1 ). 
Denotative versus prescriptive group norms in inter- an intragroup 
judgements 
The above idea is important to our present concerns. From it, we may accept 
a definition of 'group norm' as a descriptive principle of intergroup 
differentiation, or cognitive clarity concomitant with social categorization. 
But norms may also be viewed as prescriptive standards, which are closer to 
overt or cognitively exerted social influence (Kenny & La Voie, 1 985). In this 
case, norms may be viewed as standards on which ingroup members anchor 
their judgements (Forsyth, 1 995) about other ingroup members in order to 
ascertain, not intergroup distinctiveness, but rather the legitimacy of the 
ingroup's superiority (Marques & Paez, 1994). In this vein, we shall speak of 
denotative norms to refer to the former, and prescriptive norms to refer to the 
latter kind of process. Notice, however, that prescriptive norms should 
function as a prerequisite for intergroup differentiation. Accepting this point, 
the major question is through what kinds of mechanisms do individuals step 
from a descriptive norm, which generates intergroup differentiation and 
identification with the ingroup, to a prescriptive norm, according to which 
such identification is regulated and reinforced? We now elaborate on this 
question. 
The 'black sheep effect' 
Our main empirical argument for the distinction between denotative and 
prescriptive group norms comes from evidence on the so-called 'black sheep 
effect' (Marques, 1990; Marques, yzerbyt & Leyens, 1 988). This evidence 
(Marques, 1 990; Marques & Paez, 1994, for reviews), shows that people 
differentiate more strongly between likeable and unlikeable ingroup members 
than between likeable and unlikeable outgroup members. For instance, 
Marques et al. (1988) asked their Belgian subjects to judge attractive and 
unattractive Belgian and North African students on a series of traits. The 
results showed that attractive Belgian students were judged more favourably 
than attractive North African students, whereas unattractive Belgian 
students were judged more un 
favourably than unattractive North African 
students. Similar results were found by Marques and Yzerbyt (1 988). These 
results showed that law students evaluated a good performance of another 
law student more favourably than a similar performance by a philosophy 
student, and evaluated a poor performance by a law student more unfavour­
ably than a similar performance by a philosophy student. Also, Marques, 


1 30 
Social identity: international perspectives 
Robalo and Rocha (1992) found that high-school pupils evaluated more 
extremely, both negatively and positively, likeable and unlikeable pupils of 
their own school than similar pupils of a rival high school. 
Taken at face-value, these findings might suggest an interpretation based 
on normative fit. That is, when faced with likeable and unlikeable in 
group 
and outgroup members, subjects would switch from the initial ingroup­
outgroup categorization to one which proved more informative given the 
stimulus-material they were provided with (that is, likeable and unlikeable 
characteristics or good and poor performance). However, other findings 
raise difficulties with this explanation. For instance, Marques et al. (1988, 
Experiment 2) showed that subjects over-rated likeable in 
group members 
and derogated unlikeable ingroup members, relative to their outgroup 
counterparts, only on dimensions relevant to the definition of the ingroup as 
a whole. On other equally socially desirable or undesirable dimensions, but 
which did not contribute to the distinctiveness of ingroup and outgroup, 
subjects were equally extreme in their evaluations of ingroup and outgroup 
members. In another study, Marques (1990) showed that subjects differ­
entiated more strongly between normative and counter-normative ingroup 
than outgroup members, but only when they considered the normative 
dimensions as directly relevant to the ingroup's image. Moreover, contrary 
to an interpretation based on a switch from the original ingroup-outgroup 
dimension to another one (distinguishing, for example, between likeable and 
unlikeable persons), Marques and colleagues found that ingroup bias (that 
is, positive differentiation of ingroup from outgroup as a whole) generally 
co-occurs with higher differentiation between likeable and unlikeable 
ingroup rather than outgroup members. 
In discussing evidence on the black sheep effect, Marques and Paez 
(1994) drew from the general assumptions of the social identity framework. 
However, they raised a problem with this framework as regards the 
explanation of the entire set of results. Research on intergroup relations has 
long shown that ingroup members are expected to be more likeable than 
outgroup members (for example, Brewer, 1979). Therefore, it is true that an 
explanation based on a search for intergroup differentiation and clarity 
of group boundaries may partly account for more negative judgements of 
undesirable (that is, outgroup typical) than of desirable (that is, in 
group 
typical) ingroup members. A simple reading of the social identity frame­
work would lead one to state that socially undesirable (that is, non­
prototypical) ingroup members are disliked because they contribute nega­
tively to the clearness of group boundaries. In tum, socially desirable (that 
is, prototypical) ingroup members would be liked, since they bolster such 
clearness (Hogg, 1992). However, this reasoning would fall short of 
explaining why unlikeable (that is, outgroup typical) outgroup members are 
not preferred to likeable (that is, ingroup typical) outgroup members. 
Indeed, in assuming that the subjects' basic motivation is to enhance inter­
group differentiation, one should expect them to express more unfavourable 
feelings towards undesirable ingroup and desirable outgroup members. 


Social identity and intragroup diff
erentiation 
1 3 1  
These feelings would ensue from a negative reaction to those who threaten 
the intergroup boundaries. By the same token, subjects would express more 
favourable feelings towards desirable in 
group and undesirable outgroup 
members, who, by definition, bolster such boundaries. But, clearly, this was 
not the case. To Marques and Paez (1994), the social identity framework 
could account for the black sheep effect if it encompassed the idea that 
people derogate socially undesirable ingroup members as part of a cognitive­
motivational strategy to purge from the group those ingroup members who 
negatively contribute to positive intergroup distinctiveness. In other words, 
derogation of unattractive, or deviate, ingroup members, and upgrading of 
attractive, or normative, ingroup members relative to outgroup members 
supports, or at least is concomitant with, intergroup differentiation. Based on 
this idea, they proposed an explanation based on what we can designate as 
'subjective group dynamics'. 
The notion of 'subjective group dynamics' 
The notion of 'subjective group dynamics' was inspired by an articulation 
between self-categorization theory and theory and research on small-group 
processes (Marques & Paez, 1994). In brief, following self-categorization 
theory, self-stereotyping generates full interchangeability between ingroup 
members (including the self) as regards ingroup-defining standards. In other 
words, self-stereotyping would generate 'promotional interdependence' 
(Deutsch, 1968; Lott & Lott, 1965), or subjective interdependence between 
the self and other ingroup members. As a result, the value assigned to the 
social self concept depends on the value assigned to the salient ingroup 
members in the judgemental context. In this vein, the self should become 
fully interdependent from other ingroup members. 
The above idea states a cognitive parallel between rejection of deviates in 
small groups and negative evaluations of unlikeable ingroup members in 
social categories which people conceptualize in terms of subjective belong­
ingness (Hogg & Abrams, 1988; Hogg, 1992; Tajfel, 1978d). Moreover, the 
effects occurring in this latter case should be strengthened relative to those 
observed in face-to-face groups. In face-to-face groups, deviates are 
downgraded because they endanger the subjective validity of social reality 
and the likelihood of group goal achievement. In groups characterized by 
subjective membership, social reality exclusively amounts to self-definition. 
The only relevant social reality is the positive value assigned to the social 
self. Any potential jeopardy of this special aspect of social reality would, 
most likely, be experienced as more threatening than contingencies arising 
from other aspects of social reality. 
The subjective full interdependence between ingroup members allied with 
the fact that self-definition engulfs the social reality of the cognitive group 
would have logical consequences at the judgemental level. Ingroup members 
who are perceived in line with expectations supporting the ingroup's relative 


1 32 
Social identity: international pers
pectives 
superiority should generate a state of subjective validity (Turner, 1 991) of 
the positive social self. This would explain praise of likeable in 
group 
members as occurs in the black sheep effect. But other ingroup members 
whose perceived characteristics or behaviour hinders positive intergroup 
distinctiveness would spur strong hostility. This would explain derogation of 
ingroup deviates. 
'Subjective group dynamics' and the cognitive construction of 
deviance 
Recent research on perceptions of group variability indirectly supports the 
idea that derogation of ingroup deviates emerges as a way to subjectively 
secure perceived group cohesiveness and hence to validate the superiority of 
the ingroup relative to the outgroup as a whole. First, there is evidence that 
when social identity is threatened or dominated, individuals tend to perceive 
the ingroup as more homogeneous than the outgroup. Lorenzi-Cioldi and 
colleagues (for example, Lorenzi-Cioldi, 1988, 1993; Lorenzi-Cioldi, Eagly 
& Stewart, 1995), for instance, have shown that members of dominated 
groups tend more to see themselves as similar to an ingroup prototype (they 
see their group as an 'aggregate'), and focus more on the abstract features 
that characterize the ingroup. In turn, members of dominant groups tend to 
see themselves more as individualized instances of the category (they see 
their group as a 'collection') and focus more on interpersonal differences 
between ingroup members (Doise & Lorenzi-Cioldi, 1990). Consistently 
with this phenomenon, evidence collected by Simon and colleagues (Simon 
& Brown, 1987; Simon & Hamilton, 1994) shows that members of minority 
or dominated groups consider themselves to be more similar to other 
ingroup members and judge the ingroup to be more homogeneous than do 
members of majorities or dominant groups. In the same vein, Doosje, 
Ellemers and Spears (1 996) have found that members of low-status groups 
perceive the ingroup to be more homogeneous than the outgroup when their 
social identification is strong. In turn, members of high-status groups, or 
members of low-status groups whose identification is low, perceive the 
ingroup to be more heterogeneous. 
Derogation of ingroup deviates as a function of lack of perceived 
ingroup cohesiveness 
Clearly, the above-mentioned research allows us to parallel findings in 
small-group settings and evidence found using larger social categories. In 
both cases, threats to the group generate a striving on the part of group 
members to increase group cohesiveness, probably as a means to reinforce 
the positive aspects of their social identity (Marques & Paez, 1994). In small 
groups, derogation of group deviates allows undesirable members to be 


Social identity and intragroup diff
erentiation 
1 33 
purged from the group. One may thus logically reason that a similar process 
would occur when the basis for judgements is membership of cognitive 
categories. 
Marques, Brito, Correia and Serodio (1996, Experiment 2) have recently 
explored this idea. They divided their subjects in minimal groups according 
to a bogus test of imagination, and asked them to fill in an attitude ques­
tionnaire. The questionnaire was composed of several statements ordered 
from highly socially desirable to highly socially undesirable, and subjects 
were asked to choose, among those statements, the one with which they 
agreed the most. One week later, subjects were informed of their group 
membership (imagination type), and examined responses given by five 
ingroup or outgroup members to the attitude questionnaire. Figure 9.1 
outlines the information provided to subjects about ingroup or outgroup 
members in the homogeneous and heterogeneous group conditions. As can 
be seen, in the homogeneous ingroup condition, all group members but one 
(the socially undesirable member) gave exactly the same response as the 
subject (item 6). When the ingroup was heterogeneous, two ingroup 
members gave the same response as the subject, two other members gave 
neighbouring responses, and one gave a socially undesirable response (item 
2). This was reversed in the outgroup condition. 
The subjects were then asked to judge one socially desirable member and 
one socially undesirable member (respectively, members X and Y, in Figure 
9. 1). Marques and colleagues (1996) predicted that subjects would derogate 
the unlikeable ingroup member more strongly in the heterogeneous ingroup 
than in the homogeneous ingroup condition, because, in the former condi­
tion, the socially undesirable ingroup member would be more threatening to 
the validation of the group's identity. In turn, no differences would emerge 
from evaluations of outgroup members. The data supported this idea. As 
can be seen in Figure 9.2, subjects derogated the undesirable ingroup 
member more than the undesirable outgroup member in the heterogeneous 
condition, and derogated the former member more strongly in this con­
dition than in the homogeneous ingroup condition. 
These results indicate that rejection of ingroup deviates may be a direct 
function of lack of ingroup cohesiveness around valued ingroup standards. 
When the ingroup is cohesive - that is, majority members uniformly adopt 
a valued position so that the group's norms are secured - group members 
seem to feel the presence of deviates as less threatening to the group than 
when the majority is more heterogeneous, and hence the group's normative 
system is more easily impaired by this presence. 
Derogation of ingroup deviates and upgrading of outgroup deviates 
as a prescriptive means to legitimize ingroup norms. 
In more general theoretical terms, the above-summarized study suggests that 
intragroup differentiation is concomitant with intergroup differentiation 


1 34 
Social identity: international pers
pectives 
Homogeneous group 
Heterogeneous group 
Social desirability continuum of 
Ingroup 
Outgroup 
Ingroup 
Outgroup 
statements 
1 .  'Homosexuals should be exiled to 
appropriate facilities from which they 
A 
should never be released so as not to 
endanger normal people' 
2. 'Homosexuals should be housed in 
institutions and allowed to leave only 
Y 
A B C  
Y 
B 
on certain special occasions, as long as 
X 
X 
they do not endanger other people' 
3. 'Homosexuality is an i llness and 
homosexuals should be forced to 
C 
undergo medical treatment' 
4. 'Homosexuality is shameful and gays 
should at least try to disguise their 
tendency' 
5. 'Although they are normal, gays should 
C 
be helped to find a better way' 
6. 'Gays, as everyone else, have the right 
X 
Y 
X 
Y 
to choose their own sexual life' 
A B C  
B 
7. 'Homosexuality is adapted to modern 
society and, therefore, gays should be 
A 
appointed to decision-making 
positions' 
Notes: 
1 .  Statement '6' = subject's stand on the issue. 
2. Distance of ingroup deviate member and outgroup deviate member (D) from their 
groups' modal positions = 14.001. 
Figure 9. 1 
Mani
pulation o
f
 group unif
ormity and group members' status: 
distribution o
f
 ingroup and outgroup members' responses 
rather than just the outcome of a switch from one categorization level (for 
instance, 'type of imagination') to another (for instance, 'tolerance towards 
sexual minorities'). Obviously, the notion of 'norm' used in this context is 
akin to that proposed by Turner and colleagues to designate group proto­
types, since it would correspond to the definition of the acceptable ingroup 
position on a social comparison dimension. However, we also propose that 
such norms encompass a prescriptive component: individuals would be 
normative not only as regards the 'denotative' prototypical definition of their 
group (a cognitive parallel, for instance, to Jones and Gerard's [1967] notion 
of 'informational influence'), but also in terms of a 'prescriptive' demand for 
compliance with this prototype (a cognitive parallel to Jones and Gerard's 
[1967] notion of 'normative influence'). 


Favourability 
7 
Social identity and intragroup dif
f
erentiation 
Member: 
-- Likeable 
1 35 
6 
7 
6 
5 
4 
3 
2 
- - - Unlikeable 
5 
4 
3 
2 
Ingroup 
Outgroup 
Homogeneous group 
Ingroup 
Outgroup 
Heterogeneous group 
Figure 9.2 
Evaluations o
f
 attractive and unattractive ingroup and outgroup 
members as a f
unction of group unif
ormity (Marques, Brito, Correia & 
SerOdio, 1996, Experiment 2) 
The legitimizing role of outgroup 'converts' 
In another set of studies, the present authors attempted to show that this 
process is indeed the outcome of an interplay between denotative and 
prescriptive norms, as defined above. Specifically, whereas denotative norms 
serve intergroup differentiation purposes, prescriptive norms serve ingroup 
legitimization purposes. As a result of prescriptive ingroup norm salience, 
subjects would not only downgrade ingroup members who fail to attune to 
these norms, but also upgrade outgroup members who attune to them. 
In one study, Marques, Abrams and Paez ( 1996, Experiment 3) asked 
their subjects to examine a case of murder, purportedly as part of a study 
on jury decision-making. Subjects were informed that two opposed patterns 
of decision-making existed (patterns X and Y) and that their membership of 
one pattern would be detected on the basis of their criteria in judging the 
crime. Subjects were also to rank the six characters involved, from most to 
least responsible. One week later, subjects were informed that, according to 
their reasoning about the murder (but not according to their ranking), they 
belonged to one pattern. In one condition (norm condition), subjects were 
told that a norm existed according to which those who belonged to their 
pattern should rank the characters exactly as they had done, and that the 
opposed pattern should show the reverse ranking. In another condition (no 
norm condition), subjects did not receive this information. In both con­
ditions, it was made explicit that the ranking did not indicate to which 
pattern individuals belonged. The only valid criterion was the way the 
rankings were justified by them. Subjects were then presented with the 
responses of five ingroup or five outgroup members. In the ingroup con­
dition, these responses were constructed such that: (a) each subject's 


1 36 
Social identity: international perspectives 
Modal member 
Deviate member 
Rank 
Subject's 
Ingroup Outgroup 
Ingroup Outgroup 
A 
A 
B 
D 
C 
2 
C 
C 
D 
B 
A 
3 
F 
F 
E 
E 
F 
4 
E 
E 
F 
F 
E 
5 
D 
D 
C 
A 
B 
6 
B 
B 
A 
C 
D 
Figure 9.3 Example of ranking o
f
 characters A-F used to build modal and 
deviate members 
response fully matched the ingroup norm; (b) four ingroup members 
displayed exactly the same response; and (c) one ingroup member (deviate 
member) displayed a response similar but not identical to the outgroup. In 
the outgroup condition, the pattern of information was reversed. To clarify, 
Figure 9.3 exemplifies the way information was constructed to manipulate 
ingroup and outgroup members' modal or deviate status relative to the 
group norm. 
Among other results, judgements of ingroup and outgroup as a whole 
were not affected by the norm information. In line with the classical 
ingroup bias, subjects evaluated the ingroup more favourably than the 
outgroup. That is, the categorization according to the X and Y pattern 
entailed intergroup differentiation. However, evaluations of modal and 
deviate group members varied according to whether subjects were or were 
not aware of the ranking norm. In the no norm condition, modal and 
deviate ingroup members were always judged more favourably than out­
group members. But, in the norm condition, modal ingroup members and 
deviate outgroup members were judged more favourably than deviate 
ingroup members and modal outgroup members. In other words, when no 
prescriptive norm existed, subjects behaved according to the same 
intergroup differentiation principle, in judgements of the whole groups, in 
judgements of modal members and in judgements of deviate members. 
However, when they were provided with information about such norms, 
their judgements diverged in the direction of the legitimization of the 
ingroup prescriptive norm (see Figure 9.4). 
These results suggest that denotative and prescriptive norms simulta­
neously operate in judgements of groups and their members. They could not 
be accounted for simply in terms of the denotative character of the meta­
contrast principle. Indeed, this principle would state that subjects would 
prefer normative ingroup members and normative outgroup members to 
counter-normative ingroup or outgroup members, because the former 
reinforce intergroup differentiation whereas the latter hinder such differ­
entiation. If, however, another quality of norms (that is, prescriptive norms) 
served to legitimate the values underlying social identity, then subjects 


Social identity and intragroup dif
f
erentiation 
Favourability 
7 
6 
5 
4 
3 
2 
Ingroup 
Outgroup 
Norm conditions 
Member: 
-- Modal 
- - - Deviate 
Ingroup 
Outgroup 
No norm conditions 
Figure 9.4 
E
f
f
ects o
f
 norm awareness on judgements o
f
 modal and deviate 
ingroup and outgroup members 
1 37 
would prefer outgroup members, who, by subscribing to ingroup norms, 
would hinder intergroup differentiation, to outgroup members, who, by 
subscribing to outgroup norms would reinforce such differentiation. 
Whereas denotative norms serve to differentiate between ingroup and out­
group as a whole, prescriptive norms serve to legitimate such differentiation. 
In support of this interpretation, we found that, in the norm condition, 
identification with the group was associated with the derogation of the 
deviate ingroup member (r = 0.32) and to the upgrading of the deviate 
outgroup member (r = 0. 1 8; meta-analytical average r regardless of 
correlation valence = 0.26). Instead, in the no norm conditions, identi­
fication with the ingroup was associated with a more favourable evaluation 
of the modal outgroup member (r = 0.68, p < .02). This result suggests that, 
in a minimal group situation, subjects' judgements follow the meta-contrast 
principle. They judge group members depending on their membership rather 
than on their contribution to the legitimization of ingroup norms, and value 
ingroup distinctiveness to the detriment of prescriptive aspects of normative 
ingroup conformity. However, when prescriptive norms enter the picture, 
subjects appear to simultaneously value intergroup distinctiveness and 
group members' normativity. 
In another study (Marques, Abrams & Paez, 1996, Experiment 2), instead 
of manipulating subjects' awareness of a prescriptive norm, we manipulated 
the normative components of the judgemental setting: whereas some 
subjects were informed that their judgements of modal and deviate members 
would be scrutinized by typical ingroup members (ingroup accountability 
condition), others were informed that these judgements would be scrutinized 
by typical outgroup members (outgroup accountability condition). We 
reasoned that, if judgements uniquely depended on the meta-contrast 
principle, outgroup accountability would make intergroup comparison more 
salient. As a result, subjects would upgrade ingroup members relative to 


1 38 
Favourability 
7 
Social identity: international pers
pectives 
6 ~ 
5 
4 
3 
/ 
/ 
2 
Member: 
-- Modal 
- - - Deviate 
Ingroup 
target 
Outgroup 
target 
Ingroup 
target 
Outgroup 
target 
Ingroup accountable 
Outgroup accountable 
Figure 9.5 
E
f
f
ects o
f
 in 
group and outgroup accountability on judgements o
f
 
modal and deviate ingroup and outgroup members 
outgroup members, regardless of their normative status, and would show a 
stronger ingroup bias in judgements of the entire ingroup and outgroup. 
However, if judgements were guided by a prescriptive principle, ingroup 
accountability would enhance the prescriptive status of the judgements. As 
a result, subjects would derogate deviate ingroup members and upgrade 
deviate outgroup members more in this condition, while simultaneously 
showing a stronger ingroup bias in judgements of ingroup and outgroup as 
a whole. The results supported this idea. Subjects upgraded the ingroup as 
whole relative to the outgroup more in the ingroup accountability condition 
than in the outgroup accountability condition. At the same time, they 
derogated the deviate ingroup member and upgraded the deviate outgroup 
member more than when they were accountable to the outgroup (see Figure 
9.5). 
Again, correlational data showed that, in the ingroup accountability 
condition, the less subjects identified with the ingroup, the less they 
upgraded the deviate outgroup member. In the outgroup accountability 
condition, the correlation was also negative but non-significant (r = -0. 1 1 ,  
n.s.). Together with the whole set of results, these data indicate that, in a 
strongly normative context (ingroup accountability or high normative 
awareness), subjects appear more motivated to legitimate in 
group norms by 
upgrading outgroup members who endorse, and by downgrading ingroup 
members who oppose, ingroup norms. 
A model of intergroup and intragroup differentiation 
The results of the studies summarized above suggest that, as predicted by 
self-categorization theory, and, specifically, the meta-contrast principle (e.g. 


Social identity and intragroup dif
f
erentiation 
1 39 
Hogg & McGarty, 1990), social categorization motivates individuals to 
engage in intergroup differentiation. But these results also suggest that the 
prescriptive context in which intragroup judgements arise strongly affects 
the subjects' responses. Ingroup members were downgraded when the 
ingroup lacked cohesiveness, as well as when prescriptive norms were made 
salient or the situation explicitly evoked the anticipated evaluation of rele­
vant others (that is, ingroup members). It thus seems reasonable to suppose 
that derogation of ingroup members is a part of a more general process of 
'subjective group dynamics'. 
'Subjective group dynamics' and 'referent informational influence' 
The relationship between the observed patterns of judgement and the 
constructs of social identity as a group's social reality as well as of absolute 
interdependence among ingroup members seems straightforward. Social 
categorization entails a sense of interdependence between the self and the 
representation of other ingroup members as a means to maintain a positive 
social identity. In turn, this social identity is built around cognitively rep­
resented group characteristics and behaviour standards whose maintenance 
is pivotal to positive intergroup differentiation. As a result, group members 
should be motivated to exert subjective control over ingroup members in 
order to sustain a sense of legitimacy of the ingroup's superior position 
relative to other groups. A threat to such legitimacy, and hence to the self, 
would arise with any perceived lack of conformity within the ingroup in 
light of the relevant ingroup standards. Subjectively control would be 
exerted especially over those ingroup members who threaten the overall 
positiveness of the group and its relative superiority over the outgroup. 
In brief, we propose that group judgements may, in circumstances more 
complex than strictly minimal intergroup situations, arise from an 'implicit 
theory' of group processes in the individual's mind. With Figure 9.6, we 
have attempted to sketch the proposed process. 
The basic idea underlying the process depicted in the figure is that 
derogation of some ingroup members occurs when, following social cat­
egorization individuals know the normatively appropriate behaviour to be 
adopted by ingroup members in order to protect a positive intergroup 
differentiation (Abrams, 1990). This corresponds to re
f
erent inf
ormational 
influence (Hogg & Turner, 1987; Turner, 1991; cf. also Hogg & Abrams, 
1988). Ingroup members who are perceived as not observing these behaviour 
requirements will trigger 'subjective group dynamics'. They will enhance 
subjective social control in order to secure the positive ingroup differen­
tiation. In the absence of perceived norm violation, the activated contrasting 
category labels will suffice to generate intergroup differentiation. This 
phenomenon can be explained in terms of a 'denotative' meta-contrast, and 
corresponds to the intergroup level, which represents the classical process of 
intergroup differentiation and social discrimination. 


1 40 
Social identity: international perspectives 
1 .  Social categorization and intergroup salience 
STIMULUS 
SETIING 
Fit 
(Comparative) 
(Normative) 
secure social 
identity 
(ingroup bias) 
inter-category 
salience 
social identity as 
'social reality' and 
superordinate goal 
norm 
violation 
? 
yes 
denotative 
cognitions 
M ETA-CONTRAST 
self-stereotyping and 
perceived 
interchangeability 
between self and ingroup 
2. Intragroup salience and subjective control 
perceived absolute 
interdependence and 
'common fate' 
legitimization of 
positive social 
identity 
Figure 9.6 
A two-step process o
f
 inter- and intragroup judgement 
But when an intragroup distinction becomes salient and threatens inter­
group differentiation, subjects will derogate ingroup members who do not 
comply with the ingroup standards. This corresponds to the intragroup 
level. In this vein, the intragroup level would correspond to a subjective 
effort intended to legitimize the intergroup differentiation, and to warrant 


Social identity and intragroup diff
erentiation 
141 
the ingroup's unifonnity with regard to the prescribed standards that sustain 
ingroup's superiority. 
Conclusion 
We propose that in most circumstances nonnative intragroup differentiation 
may function as a foundation for intergroup differentiation and a positive 
social identity. This process can be viewed as the psychological equivalent 
of public actions of nonn enforcement, like those we mentioned at the 
outset of this chapter. In this sense, nonn violation is functional to group 
life, to the extent that the social control mechanisms it evokes help to define 
relevant criteria of group membership. Because the prescriptive nonn only 
appears as such once it has been violated (Forsyth, 1995), deviance plays a 
basic role in defining the group's objectives and interests by generating 
(subjective or objective) social control (cf. Hamilton & Rauma, 1995). 
Moreover, active participation in the punishment of deviance may reinforce 
the individual's commitment to the nonn, thus reinforcing social identi­
fication. In purely psychological tenns, this process might function more or 
less as predicted by Bern's (1967) self-perception theory: when they lack a 
definite attitudinal position about an object, individuals may infer their 
attitude from their behaviours ('if I have punished a deviate person I infer 
that I have a clear attitude in favour of the nonn'). Moreover, punishment 
rituals towards deviates may be viewed as part of an initiation process, to 
reinforce the positive evaluation of the group (cf. Zimbardo & Leippe, 
1991 
). 
It is obvious that the presently proposed process is still in need of 
theoretical and empirical scrutiny. But we hope that the ideas outlined in this 
chapter may serve as a valuable contribution to the study of the processes 
occurring in small and large groups alike, and a starting point for research­
ing into the cognitive processes underlying the social construction of nonns 
and deviance in daily life. 


1 0  
Categorization and Social Influence 
Juan Antonio Perez and Gabriel Mugny 
A relevant dimension in studies of social influence is whether or not the 
targets of influences see themselves as sharing the same social category as 
the influence source. The most commonly advanced hypothesis is that the 
effects of categorization (whether at the cognitive level of accentuating 
intragroup similarities and intergroup differences, or at a more motivational 
level of ingroup favouritism and/or outgroup discrimination) should facili­
tate the influence of an ingroup source and diminish that of an outgroup 
source. However, it appears to be difficult to draw any general conclusion 
given that in practice experimental effects have not all consistently been in a 
single direction, especially if manifest and latent influences are distinguished 
(Moscovici, 1980). The first objective of this chapter is to review those 
studies which have varied the degree of identity between source and target, 
and the second is to present the principal conceptualizations of the links 
between categorization and influence. 
The effects of categorization on influence 
There is first of all a series of studies which have manipulated the cat­
egorization (in 
group or outgroup) of sources having a majority status 
(Doise, Gachoud & Mugny, 1986; Perez, Mugny & Moscovici, 1986; 
Volpato, Maass, Mucchi-Faina & Vitti, 1990). Ingroup majorities have 
more (direct and immediate) influence than outgroup majorities. In those 
studies which have also considered indirect dimensions of influence, this 
ingroup majority advantage disappears in interaction with other factors. 
Studies in which subjects are simultaneously exposed to minority and 
majority positions do not contradict and indeed confirm these findings. Thus, 
in the study by Perez and Mugny (1987), the in 
group majority achieved 
greater direct influence than the outgroup majority, but at the indirect level 
this simple effect disappeared in interaction with the categorization of 
the minority. In the Clark and Maass (1988b) study, categorization of the 
majority was not varied - it was an in 
group - but it could be seen that it had 
more influence at the public level than either an ingroup or an outgroup 
minority. At the private level an interaction could again be observed with the 
categorization of the minority as either ingroup or outgroup. 


Categorization and social influence 
143 
In studies where the categorization involved is based on arbitrary criteria 
(such as whether one over- or underestimates the number of pupils attend­
ing secondary school, whether one is in the same classroom, and so on) the 
in 
group has more influence at the public level than the outgroup. This 
'simple categorization' effect can even be observed when the source has a 
minority status (cf. two studies by Martin, 1988b, 1988c, and one by 
Papastamou et aI., cited in Mugny & Perez, 1991a, pp. 24-27), as well as 
when the minority or majority character of the source is not made salient 
(Wilder, 1990, study I). 
In two studies on attitudes towards abortion, in which gender was used as 
the basis for categorization, it was found that a male minority group had 
more influence than a female minority group. In the Maass, Clark and 
Haberkorn (1982) study they constituted an ingroup source (although when 
the variance was decomposed the effect did not reach significance), while in 
the Perez and Mugny (l 985b) study they represented an outgroup. This 
gender effect may be specific to the theme of abortion because in a task 
involving estimation of increases in study grants, Martin (l 988b) found that 
a female minority group produced more public influence among other 
women than did a male minority group, while men were more influenced by 
a male source. 
If one turns to consideration of 
another variable, the distance between the 
attitude of the target and that of the source, one finds similarly variable 
results. In the Clark and Maass (1988a) studies on homosexual rights, an 
outgroup (homosexuals) produced more influence than an ingroup source 
(heterosexuals) among those subjects most favourable to homosexual rights; 
the ingroup source, however, produced more influence than the outgroup 
source among subjects who had the least favourable attitudes. Mackie, 
Worth and Asuncion (1 990) observed that the ingroup had a stronger 
influence than the outgroup but only among subjects who were most 
opposed and when the message contained a 'strong' argument. Using the 
theme of xenophobic attitudes, Mugny, Kaiser and Papastamou (1983) 
found the opposite interaction: the outgroup (a group of foreigners) had 
more influence than the ingroup (a group of fellow compatriots) on the 
most xenophobic subjects; an in 
group minority had more influence than an 
outgroup minority on those subjects whose own position was already closest 
to the anti-xenophobic message. 
In four studies in which perception of the source was 'individualized' or 
'personalized', ingroup and outgroup achieved the same degree of influence 
(Doise et aI., 1986; Martin 1988a; Papastamou et aI., cited in Mugny & 
Perez, 1991a; Wilder, 1990). 
Negotiating style (cf. Mugny, 1982) tends to interact with the identity of 
the source. In one study on 'green' attitudes, Mugny and Papastamou 
(1982) found that when influence targets are led to see themselves as similar 
in identity to the source of influence, the latter has more influence when the 
argument is made in a rigid style than when it is made in a more flexible 
style, when subjects are led to see themselves as dissimilar, it is the source 


144 
Social identity: international perspectives 
with the flexible style that is more influential. In a study on attitudes 
towards foreigners, Mugny and Perez (1985) observed the same interaction 
effect. In the same vein, Mugny, Kaiser, Papastamou and Perez (1984) 
discovered that a source sharing the same nationality arguing in favour of 
rights for foreigners had more influence if the argument was based on 
radical socio-political principles, while a source consisting of foreigners had 
more influence when the same case was argued from humanitarian 
principles. 
Another problem concerns the effects of categorization in different 
influence conditions. In studies in which the school to which the source 
belongs is varied (same school as the subjects or a rival school), results when 
they have been significant have shown that the in 
group has more impact at 
the public level than the outgroup (Mackie et aI., 1990; Martin, 1 988a; 
Wilder, 1990). Martin observed that at the private level the outgroup had as 
much influence as the ingroup, both kinds of source producing more change 
than that of a control condition without a persuasive message. 
In some studies measures of influence at both a direct and an indirect 
level have been available. Thus, in a paradigm based on musical prefer­
ences, Aebischer, Hewstone and Henderson (1984) found that an outgroup 
majority (from a different and devalued education establishment) had more 
indirect influence than a valued ingroup majority source, while there was no 
difference in direct influence. In a study of environmentalist attitudes (Perez 
& Mugny, 1985a), it was found that when the persuasive message was seen 
to have derived from a single individual selected at random and not 
representative of the group, direct influence was greater than when the 
source was seen to represent the opinion of the entire group. However, at an 
indirect level the effects were reversed: the message representative of the 
group had more impact than the unrepresentative message. In a study 
involving attitudes on abortion and contraception (Perez & M 
ugny, 1 986a), 
the connotation of a source was manipulated (the source was characterized 
in terms of entirely positive attributes or entirely negative attributes), while 
the source was also an ingroup minority (same age-group as the subjects: 
young people) or an outgroup minority (different category: adults between 
30 and 50 years old). The minority outgroup with positive connotations was 
more influential at the direct level than the outgroup with negative 
connotations, and more influential also than the ingroup with a positive 
image. At the level of indirect influence the most significant effect was that 
of the negative-image in 
group, as compared to the positive-image ingroup. 
In two studies Volpato et al. (\ 990) measured the effects of categorization 
on creativity. They led subjects to believe that a committee either in their 
own town (ingroup) or in another town (outgroup) had produced two 
communiques, one by the majority and the other by a minority, respectively 
opposing or favouring the establishment of a final exam at the completion 
of baccalaureat studies. They were exposed only to the argument in favour 
of this proposal. The results showed that a smaller number of subjects 
indicated they were in favour of the source's proposition when this was 


Categorization and social in
fluence 
145 
presented as originating from an ingroup minority. The proposItion was 
more widely accepted when it was attributed to an ingroup majority or 
indeed an outgroup minority. Although the direct influence of an ingroup 
minority was shown to be less than in these other two conditions, in these 
two studies this same condition produced a higher proportion of subjects 
who invented new solutions to solve the problem. 
Finally, on the matter of differentiated influence, Mugny et al. (1 983) 
reported that under certain conditions an outgroup arguing on humanitarian 
grounds had greater delayed influence than the same outgroup using a socio­
political argument. 
These, in summary, are the effects which have been observed of categor­
ization on social influence. It must be acknowledged that their diversity and 
complexity is such that the most prudent conclusion would be to admit that 
there is no systematic effect associated with categorization as such, nor any 
straightforward relation between categorization and type of influence. The 
hypothesis that there is an ingroup bias in influence has therefore to be 
regarded as inadequate; the variables moderating the effects of categoriza­
tion appear to be as numerous as the paradigms. Perhaps this disparity of 
manipulations reflects a belief that 'simple categorization' (cf. Tajfel, 
Flament, Billig & Bundy, 1971) is sufficient to produce intergroup effects, 
and, by analogy, differences in social influence. This leads us to consider 
what are currently the most elaborate explanations in this field, and to ask 
to what extent they can account for this diversity. 
Independence and heterogeneity of the source 
Wilder invoked independence and heterogeneity of the source (1 977, 1978, 
1 990; cf. also Harkins & Petty, 198 1 ,  1983, 1987) to account f
or conformity 
in restricted groups, and more recently to explain why one should expect an 
ingroup to have more influence than an outgroup. The idea is that the 
individuals who comprise a group have more influence if they are perceived 
as independent entities than if they are perceived as a collective entity. The 
reasons advanced for this are, on the one hand, that information provided by 
a group is treated with less care (for example, in a less personalized and less 
specific manner), and, on the other, that members perceived as a group are 
considered less credible because they are attributed less independence. 
Recently, Wilder (1990) has used this argument to predict that an ingroup 
will have more influence than an outgroup. His reasoning is that the target 
starts from the presupposition of a greater homogeneity among outgroup 
members than among the members of his or her own group. The target 
would recognize finer distinctions within the ingroup which would lead to a 
perception of members of the ingroup as more independent of one another. 
The same author argued that his is not a motivational explanation of 
ingroup superiority (it does not appeal to a motive of ingroup favouritism or 
of discrimination against the outgroup), but rather a cognitive explanation, 


146 
Social identity: international perspectives 
based on perception of the homogeneity or heterogeneity of members of the 
influence source. A series of four experiments illustrates this model (Wilder, 
1990), and to these may be added two studies by Mackie et al. (1990) which 
show that the message attributed to an ingroup is given greater attention 
than one attributed to an 'other' group. 
One difficulty posed by this analysis is the implication that an ingroup is 
nothing but an aggregation of separate individuals. Thus the comparison 
made is between an aggregate of individuals and an outgroup rather than 
between an ingroup and an outgroup. This does not take into account recent 
developments in the theory of levels of self-categorization (Turner, Hogg, 
Oakes, Reicher & Wetherell, 1987), according to which three such levels may 
be distinguished: a superordinate level at which individuals see themselves as 
identical to all other human beings; an intermediate level at which the social 
identity of the individual is defined by the assertion of differences between 
groups or categories of membership and non-membership; and finally a 
subordinate level which defines personal identity in terms of categorization 
of the self as a unique being, distinct from all others including members of 
one's own group. From this perspective, 'personalized' perception is not 
synonymous with ingroup categorization. If this argument is conceded, it 
would appear that Wilder's model, rather than explaining the ingroup­
outgroup difference, is more suited to accounting for the superior influence 
of a personalized source (whether belonging to an ingroup or an outgroup) 
as compared to that of a source perceived in more 'depersonalized' or 
homogeneous terms. 
The results of a study (Perez, Mugny & Navarro, 1991) justify our stress 
on the importance of these distinctions. In this study subjects were required 
to react to a message either in personal terms (subordinate level of se1f­
categorization), or in 'social terms', which is to say at the intermediate level, 
in one of two ways: either to react as a representative of their membership 
group (in this case young people) or as a member of the outgroup (adults). 
The results show that direct influence is greater when the message is treated 
in either personal or ingroup terms. These findings are consistent with the 
predictions of Wilder's model. However, at the indirect level, while influence 
remains strong in the personalized condition, it disappears in the ingroup 
condition, an effect which confirms the importance of not confounding 
the personalized level with the intergroup level. Additionally, indirect influ­
ence was found in reaction to the message in intergroup terms. This set of 
results is interpretable in terms of the notion of ingroup paralysis: under 
the conditions provided subjects conform to positions prototypical of the 
ingroup, which has the effect of paralysing examination of other positions 
from a personal point of view or from any other point of view which is not 
that of 
the ingroup. In this case the ingroup introduces a resistance to change 
and innovation. In a series of studies Sanchez-Mazas, Perez, Navarro, 
Mugny and Jovanovic (1993) have confirmed that when subjects are led to 
the self-perception that in terms of their attitudes they are identical to their 
own group, this paralyses any change in their attitudes. 


Categorization and social in
fluence 
147 
Influence according to self-categorization theory 
Self-categorization theory treats influence as an intergroup phenomenon 
(cf. Turner, 198 1  
a, 1991). The fundamental factors are the level of self­
categorization and the ways in which individuals seem to operate to 
determine the validity of their opinions and behaviour. In the first place 
individuals begin by anticipating that their judgements and actions in a 
given task will coincide with those of other members perceived as similar to 
the self and who are regarded as a comparison group appropriate to the 
situation. Individuals will be confident that their own judgements are valid 
to the degree that similar others are observed to express the same opinions 
or behave in the same way in the situation. 
Uncertainty and loss of confidence in their judgements and behaviours 
increase when individuals find themselves in disagreement with other group 
members whom they perceive, despite the disagreement, as similar to 
themselves, which is to say as an ingroup. Insofar as this disagreement 
cannot be attributed to differences between the self and the group this 
theory predicts a change in judgements or behaviours in the direction of 
those of the group or of those which best represent the ingroup consensus; 
those which are most prototypical. 
No such expectation exists when the others are perceived as different, 
which is to say as an outgroup. Disagreement in this case can be attributed 
to differences in group membership, according to the principle of categorial 
differentiation (Doise, 1976). Consequently, disagreement with the outgroup 
produces little uncertainty, and it is not anticipated that individuals in these 
circumstances would change their judgements or their behaviour. 
In the application of his theory to minority influence, Turner (1991) 
claims in addition that for a minority to exert influence it must present 
itself, or be perceived, as forming part of the ingroup (and not an out­
group), which is to say a subgroup that defends a pro-normative position 
within the ingroup. He predicts a reduction in direct influence if the 
disagreement the minority introduces in the group is polarized to the point 
of turning it into an intergroup, and not simply an intragroup, conflict. 
This theory tends also to account for differences between direct and 
indirect influence in terms of the level of self-categorization which obtains at 
any particular moment. Thus a minority categorized as an outgroup will 
have less influence, whether direct or indirect. In order for minorities to 
produce direct influence they must be perceived as part of the ingroup. On 
the other hand, if a minority produces a conversion effect (indirect influ­
ence), this will be because the comparison is made within a broader context 
in which the minority is perceived as an integral part of the ingroup, and 
because at this superordinate level of categorization, its alternative norm is 
congruent with the norms and values of the ingroup. 
This theory offers one of the most complete explanations for the tendency 
of an outgroup to have less influence than an ingroup. An initial problem, 
however, is that the theory tends to predict both a direct and an indirect 


148 
Social identity: international perspectives 
influence of an ingroup majority and does not take account of the latent 
paralysis effect which appears to be precisely the effect of overt conformity 
(Sanchez-Mazas et aI., 1993). A further difficulty is that it does not allow 
for the fact that an ingroup minority produces less direct influence than an 
outgroup minority or one which is not explicitly categorized as an ingroup. 
Additionally, the explanation it offers for conversion does not seem to be 
entirely convincing. In particular, in several studies, it has been found that 
in the same situation the same source which does not produce a direct 
influence does produce indirect influence. Is one supposed to believe that in 
this situation the influence source is perceived simultaneously at two levels 
of categorization and that, for example, when faced with an item reflecting 
direct influence subjects see the source as an outgroup, but when they 
encounter an indirect item they perceive the same source as part of the 
ingroup? There is a similar difficulty in seeing how it is possible that at a 
public level the influence source could be categorized as an outgroup but at 
a private level as an ingroup, or how a public context could prime an 
intergroup context while in a private context priming a superordinate 
intragroup context. These possibilities certainly cannot be excluded but no 
research has yet supported such dynamics. 
Dissociation theory 
The need to relate different levels of influence to one another is taken into 
consideration by dissociation theory, which we will now present. Briefly, 
this theory makes a general distinction between two levels of influence: first, 
manifest influence, which includes public, immediate and direct influence; 
and, second, latent influence, which covers influence at the private, delayed 
and indirect levels. It provides a basis for relating processes that account 
for manifest influence (and which the subject can control) to the processes 
underlying latent influence (which the subject is less able to control 
consciously). 
Manif
est in
fluence and positive social identity 
Manifest influence is a function of normative and informational dependence 
(Deutsch & Gerard, 1955). The importance of normative dependence relates 
to the individual's motivation to acquire, to present or to preserve a positive 
self-image or self-concept. Given that one function of manifest influence is 
protection of a positive social identity, one is influenced at this level by 
sources able to exercise the strongest normative pressures or in situations in 
which modifications in judgements allow one to avoid costs or gain more 
credit than if one maintained one's independence. Susceptibility to manifest 
influence also flows from the uncertainties felt by the subject, such that the 
more the source is able to engender uncertainty in the subject (by virtue of 
the former's credibility, ingroup identity or majority status), the g-:eater the 
source's influence. 


Categorization and social influence 
149 
Latent influence and social constructivism 
By latent influence we do not have in mind only private influence but all 
influence reflecting constructivist activity. Thus, for example, it includes the 
case in which subjects change on dimensions which can be inferred through 
the operation of a given mental activity. This concerns the attention the 
target pays to what the source says (to its message, its judgement, its 
attitude, and so on), the deployment of an innovative form of thought 
which allows the target to derive or discover new functions and ramifi­
cations on the basis of the information provided, to establish new connec­
tions between the content given and other contents, as well as the 
elaboration of a new representation of the question. Some authors have 
discussed these kinds of activity under the heading of validation processes 
(Butera, Legrenzi, Mugny & Perez, 1 990-1991; Moscovici, 1980), others 
under the heading of divergent thinking (Nemeth, 1986). Huguet, Mugny 
and Perez (1990-I 
991) also use the concept of decentration. Within the set 
of paradigms dealing with attitudes and opinions, these activities have been 
described as inferences about the organizing principle underlying a position 
(Perez & M 
ugny, 1986b). In all these cases, however, the same basic 
hypothesis underlies these various designations: mental activity is required if 
a latent mode of influence is to arise out of social interaction, and it will 
occur only in the context of certain kinds of social relationship and not 
others. 
Postulates o
f dissociation theory 
To the degree that manifest influence is directly observable by others or is, 
ultimately, something of which subjects themselves are conscious, then 
subjects' identity is at stake whenever they are influenced by any source, and 
dissociation theory recognizes that identification is more probable in the 
face of an ingroup source. This hypothesis indicates that a message 
emanating from an ingroup is more likely to be accepted, even if it is not 
read or processed, than a message originating with an outgroup source. 
This more positive attitude towards whatever originates in the ingroup 
can, especially when the subject is highly involved or motivated, give rise to 
a higher degree of attention to a message attributed to an ingroup and thus 
a more elaborated processing of the message as compared to one from an 
outgroup. Faced with an ingroup, because identification is more intense the 
message can then be accepted with a less defensive attitude and the subject 
can feel more motivated to pay attention to it. One may therefore predict 
that in certain conditions there will be direct but also indirect influence. This 
will only happen, however, if the situation does not induce a paralysis and if 
identification with the ingroup is positively valued and poses no conflict of 
identification. 
In order for the ingroup to have indirect influence, one basic condition 
must be satisfied: subjects must perceive a divergence between their own 
position and that of their group. One risk for an ingroup is that if 


1 50 
Social identity: international perspectives 
individuals perceive the group to be in agreement with their own position, 
and this accords with an anticipated homogeneity of attitudes, they will 
develop a socio-cognitive paralysis, the consequence of which is to inhibit 
the activity necessary for the emergence of 
indirect influence. The attitudinal 
proximity which individuals often establish through their manifest 
conformity to the ingroup can therefore by this same process limit the 
latter's impact to a pattern of influences similar to compliance (positive 
manifest influence combined with zero latent influence), as has been 
observed in a number of studies (cf. Perez et aI., 1991; Sanchez-Mazas et aI., 
1 993). One way of motivating subjects to process the ingroup message is 
therefore to lead them to perceive a normative conflict between their own 
position and that of their group, which flows not from a simple self­
categorization but from an intragroup conflict. 
Another difficulty for the ingroup arises in the case of identification 
conflict (cf. Mugny et aI., 1983; Mugny & Perez, 1991b), in which indi­
viduals are aware of belonging to a group which has little positive value and 
may even be negatively valued, or in which they risk being identified with 
such a group. In these cases, identity with or similarity to this ingroup 
threatens the maintenance or construction of a positive social and personal 
identity. Would individuals let themselves be publicly and manifestly 
influenced by a socially discredited group posing such a conflict of identi­
fication? In fact it is not enough to respond that in all probability they 
would not let themselves be influenced; it is also important to recognize 
what would happen at a latent level in such a situation. A basic postulate of 
dissociation theory is that the more that socio-cognitive activities and 
attention are occupied in creating a disidentification or social differentiation 
from a source of influence, the less these are concerned with the influence 
message. In the event that individuals come to give any attention to the 
message and to analyse it, this is undertaken on the basis of a motivation to 
construct a differentiation so that a negative processing of the message tends 
to predominate. 
Let us imagine a situation in which the target of influence does not wish 
to be identified with the source of influence and that in one case this 
involves an ingroup source and in another an outgroup source. 
In the situation where the influence source is perceived to be an ingroup 
with which the influence target does not wish to be seen to be identified (for 
example, the case of a discredited minority), the available data suggest that 
in 
group favouritism is indeed not the dominant strategy. Studies by 
Marques (1990) on what he has described as the black sheep phenomenon 
provide a direct illustration of the way in which a negatively valued ingroup 
can be the object of more discrimination than an equally negatively valued 
outgroup. The reason for this is that the more intense the identification is 
(for example, by virtue of a strong identity or similarity between individual 
and group), the more effort and activity will be required, should the 
occasion arise, to try to disidentify. This is the reason why ingroup influence 
may be zero at the latent as well as at the manifest level. 


Categorization and social influence 
1 5 1  
It can thus prove to be more positive for social identity to discriminate 
against an ingroup rather than an outgroup, or to favour the outgroup to 
the detriment of the ingroup. Indeed, an outgroup source can paradoxically 
have a net advantage, particularly at the latent level, over an ingroup 
source. Given that the ties of identification are less close with an outgroup, 
it seems reasonable to predict that disidentification with such a group 
requires less activity and effort, because in this case a ready-made differ­
entiation is provided; it is founded on differences of group membership. The 
mental activity generating latent effects will no longer be confounded with 
the activity of differentiation, the two being immediately dissociated. 
A major hypothesis of dissociation theory is that even in situations 
involving intense conflict a source categorized as an outgroup will be more 
likely to produce dissociation, to the degree that social differentiation is 
assured by the existence of a pre-established categorization, or that the 
differences are so great that to shift in the direction of the source's responses 
carries no risk of a 'categorial confusion' (Lemaine, 1974). This dynamic 
is also involved when the space for comparison of subject with source is 
multidimensional, such that there is no negative interdependence between 
source and target of influence (Perez & Mugny, 1990). But this does not 
imply that an outgroup source will have direct influence. In reality, at the 
direct level it is the categorization processes that operate, as it were quite 
naturally. However, to the degree that this social differentiation is dissoci­
ated from message reception, conversion effects towards outgroup positions 
can appear. If for one reason or another the subject does not come to 
dissociate reception of the message from construction of this differentiation, 
the outgroup will no longer have any significant indirect effect (Butera, 
Huguet, Mugny & Perez, 1994). 
Conversion dynamics and the nature of the task 
Let us acknowledge that at the psychological level the influence situation 
produces a dissociation of the relation with the source from the processing 
of the message. One question remains then unanswered, namely: what can 
motivate the subject, particularly when faced with an outgroup, to process 
the content of the message? 
A first possibility is that epistemic expectations intervene which are 
specific to the nature of task to which the outgroup's divergent judgement 
relates. For example, in objective, non-ambiguous tasks (of the type used by 
Asch, 1956) for which only one correct response, whatever its provenance, is 
assumed to be possible, the influence mechanism of an outgroup generally 
rests upon the motivation of the subject to re-establish uniformity of 
judgements. In such a situation and given the impossibility of changing the 
source, it is the subject who ends up changing, always at the latent level, in 
order to reconcile social differentiation and epistemological beliefs con­
cerning the necessity of consensus (Butera et aI., 1994). 


1 52 
Social identity: international perspectives 
When tests of ability are involved (for example, estimating distances) and 
subjects presuppose that there do exist objectively correct answers but do 
not know what they are and so feel uncertain of their ability, outgroups can 
achieve an influence by the triangulation effect already described: 
outgroups' judgements can be perceived as valid by virtue of their socially 
independent origins (cf. Goethals, 1972). 
When the tasks involve opinions, differentiation from the outgroup at the 
level of attitudes itself constitutes an epistemological requirement (Perez, 
Mugny, L1avata & Fierres, 1993) which accounts for manifest discrimina­
tion. The indirect effects of an outgroup (in particular when this is a 
minority) arise from the fact that dissociation allows a centration on the 
contents and initiates a debate about ideas (Billig, 1985). Examination of a 
contrary argument, even if this is in order to resist it, implies in particular 
that one increases one's exposure to information from the other. Here then, 
as in the case of denial (Falomir, Mugny & Perez, 1996), a process of social 
cryptoamnesia may be engaged through which, paradoxically, ideas may be 
internalized whose social origins may be forgotten as a result of dissociation 
(cf. M 
ugny & Perez, 1991 
b). 
Certain social minorities (for example, black people) may also constitute 
outgroups protected by the Zeitgeist in such a way that discrimination 
towards them is socially censured, at least for a significant proportion of 
society. In this case, a double effect can be produced. At a manifest level 
there may be on the one hand, a greater influence (reflecting social 
desirability) which does not, however, generalize to the latent level, and, on 
the other, an effect close to socio-cognitive paralysis. Paradoxically it is 
when subjects discriminate against minorities that they experience a conflict 
the resolution of which can provoke a change in latent attitudes (cf. Perez, 
Mugny, L1avata & Fierres, 1993). The occurrence of racist, xenophobic or 
sexist acts can therefore be in conflict with norms and values of justice and 
social equality accepted more generally by subjects. The self-reproach 
arising from this discrimination can thus act as a mechanism for change in 
latent attitudes. 
Conclusion 
It is clear that intergroup dynamics, at least with respect to social influence, 
are not confined to the habitually recognized effects of social categorization 
(accentuation of intragroup similarities and intergroup differences). They 
also involve strategies for constructing or maintaining a positive social 
identity, and can additionally disfavour the in 
group or favour the outgroup, 
in particular as a function of whether or not the social context threatens the 
targets' personal and social identities in their relations with salient entities in 
the categorial field. 
Of the two factors that are fundamental to understanding social influence 
in an intergroup context, one certainly is the dynamic of identification or 


Categorization and social in
fluence 
1 53 
differentiation, but the other is the conflict created by the divergent position 
of a source of influence. Simply categorizing a source as ingroup to produce 
identification is insufficient to guarantee influence, whether at the manifest 
or the latent level. It is the divergence and conflict which determine matters 
in one direction or the other and these alone are able to account for latent 
processes of constructivism. 
In all cases, divergence creates a conflict by reason of epistemological 
presuppositions about the task, in particular expectations about consensus 
or dissensus. Categorization therefore is not alone in playing an active role, 
processes of influence have their own autonomy by virtue of the relational 
but also socio-cognitive conflicts which they introduce. From this point of 
view, dissociation theory seems to provide a tool which allows us, to a 
greater degree than models offering a single-process view of intergroup 
effects, to organize the multiple dynamics which can arise from the inter­
action between categorization and social influences. 
Note 
This text was prepared with the financial support of the Swiss Fond National de la Recherche 
Scientifique, the Spanish Comision Interministerial de Ciencia y Tecnologia (CICYT, SEC95-
0628) and the French Centre National de la Recherche Scientifique. 


1 1  
Social Identity and Interpersonal 
Relationships 
Miguel Maya 
Attraction and interpersonal relationships 
The emotional-sexual relationships established between men and women 
resulting from attraction (from here on referred to as 'relationships between 
couples') can, at first, be considered as being situated close to the inter­
personal end of the interpersonal-intergroup continuum, already mentioned 
in previous chapters of this book. This can be deduced both from the 
classical research carried out in social psychology and the formulations of 
some authors classified as social identity theorists. 
The traditional approach towards studying attraction and interpersonal 
relationships has always placed special emphasis on considering them from 
the perspective of the individual, without bearing in mind the social or 
intergroup dimension. This is reflected not only in the most commonly used 
phrase to define this study area - 'interpersonal attraction' - but also in the 
approaches and theories used to explain these phenomena. 
The main theoretical focal points used to explain attraction are the 
theories of cognitive consistence, exchange and interdependence (Kelley & 
Thibaut, 1978) and those based on the principles of association and 
reinforcement (Byrne, 1971). All these, as well as the factors used to 
measure attraction (proximity, similarity, reciprocity, among others), refer 
fundamentally to peculiar characteristics of individuals involved in the 
interaction. 
Some authors working within the social identity theory framework have 
also argued that this particular type of male-female relationships can be 
placed close to the interpersonal end of the continuum (Hogg & Abrams, 
1988; Huici, 1984). There is a tendency to consider male-female relation­
ships as intergroup when the treatment shown towards people belonging to 
another sex group is undifferentiated, which implies minimal or almost non­
existent knowledge of them. In this case it is likely that stereotyped percep­
tion of the outgroup will appear, in addition to intergroup conflict and 
ingroup favouritism. On the other hand, interpersonal male-female rela­
tionships tend to be characterized by a one-to-one relationship, a married or 
heterosexual couple. 


Social identity and interpersonal relationshi
ps 
1 55 
Relationships between couples as intergroup relationships 
This heading could also be termed: 'Interpersonal relationships as inter­
group relationships', which would be no mere play on words. Our intention 
here is to show that despite the fact that relationships between couples have 
been traditionally referred to as 'interpersonal relationships', and have thus 
had to be placed close to the interpersonal end of the continuum, there in 
fact exist a number of powerful intergroup aspects within them. Indeed, 
these have been acknowledged by the above-mentioned authors. 
For example, Huici (1984) argues that the existence of power and status 
differences between women and men in society, the consensual agreement 
on differential perceptions of each sex and the virtual universality of the 
discrimination based on sex tend to locate interactions between members of 
gender categories close to the intergroup end of the continuum. Hogg and 
Abrams (1988) or Williams and Giles (1978) come to similar conclusions. 
Here we would like to emphasize that the consideration of these power 
differences between sexes or the existence of outgroup discrimination 
(features closely related to the intergroup extreme) does not necessarily need 
to appear associated with anonymous or collective interactions or those 
which occur in the professional or public spheres. An interaction between 
two people, although charged with positive affection, with completely idio­
syncratic characteristics and situated within the sphere of intimacy, may still 
reveal powerful intergroup features. 
The definitions of intergroup behaviour given by both Sherif (1 967) and 
Tajfel (1978a), clearly show that what distinguishes interpersonal from 
group or intergroup behaviour is not the number of people implied. Take, 
for example, the hypothetical case of the businessman X who turns down 
the woman Y's application to work in his office because he considers that 
women are less intelligent and capable than men. If this case involved only 
a man and a woman, it would appear to be clearly based on an intergroup 
relationship, given that the businessman's behaviour (male) with regards to 
the candidate (female) was exclusively concerned with her belonging to the 
above group and not with her personal qualities. As Brown (1 988) shows, in 
order to establish at what point of the interpersonal-intergroup continuum 
is situated an interaction between two people from different social 
categories (such as male and female), we need to study in close detail the 
nature of that very interaction. If the behaviour and attitude of the man 
towards the woman (or vice versa) is that stereotypically associated with 
each gender, we can then begin to talk of intergroup behaviour. If this turns 
out to be impossible, the interaction's idiosyncratic features will emerge in a 
more interpersonal relationship. 
In most societies (although our data refer only to the Western world), 
boyfriends and girlfriends, wives and husbands and fathers and mothers 
behave differently, although coinciding to a great extent with the stereo­
types and roles that society associates with men and women (Peplau, 1983). 
A good example of this is that commonly heard saying: 'I can't do that, 


1 56 
Social identity: international perspectives 
because I'm a married man (woman) .' In this way, we can safely say that 
couples, even those characterized by high intimacy levels, also reveal 
important intergroup aspects. We will now go on to discuss our approach in 
more detail, referring to two areas in which uniformities can be seen in both 
male and female behaviour in a relationship between couples. 
(1) Conception o
f
 a couple. The majority of studies, carried out with middle­
class American college students, reveal great similarities in the expectations 
and conceptions that boys and girls have of couples and heterosexual 
relationships in general, although some differences have been found. While 
women from all cultural and economic backgrounds generally grant a great 
value to affection and companionship in their relationships, only male 
college students - but not those from a working-class background -
coincided with these feelings (Peplau & Gordon, 1985). Likewise, women 
grant a great importance to intimate communication in their relationships 
with men (Parelman, 1983). In general, men tend to have more traditional 
beliefs with regard to their relationships with women, in keeping with male 
roles and stereotypes (Peplau & Gordon, 1985). This result is as expected 
given that the traditional concept of a couple would appear to benefit men 
as opposed to women. 
(2) Pro
f
essional development and economic res
ponsibility. In accordance 
with traditional gender stereotypes, economic responsibility for both the 
couple's and family's welfare falls on the shoulders of the male. Greater 
priority is therefore given to his professional career. This is clearly reflected, 
along with other signs, in the proportion of men and women who work 
outside the home. 
Kotkin (1 983) found that 70% of married women at university had, in his 
study, sacrificed something (for example, economic support, change of 
address, renouncement of own career) to aid their partner's university 
studies or professional life, while only 10% of men had made the same 
sacrifice for their female partners. It will, therefore, come as no surprise that 
women intending to study at university or work hard at a professional career 
are often less committed to their couple (Rubin, Peplau & Hill, 1981). 
However, aside from the priority given to men's professional aspirations 
over women's, we are also concerned with the fact that simply getting 
married and having a family can act as an obstacle for women's professional 
success. For example, when Harrison, Moore and Rucker (1985) studied the 
biographies of more than two thousand eminent men and women who had 
appeared in various editions of Who's Who, they found that it was less likely 
for successful women to marry and have children, and more likely that they 
would get divorced. In the same way, Etaugh and Riley (1983) have shown 
that single women have higher income levels, from both employment and 
education, than their married counterparts, while the reverse model is true in 
the case of men. 


Social identity and inter
personal relationships 
1 57 
Another research area which would appear to reinforce the notion that 
women retreat from professional life within a couple deals with the image 
men have of successful women (in keeping with the traditional gender 
stereotypes), especially those working in typically male fields. In general, 
these studies portray a negative image, meaning that, to male eyes, these 
women would appear to be less attractive as partners (Pfost & Fiore, 1990). 
Thus, it would seem that although couples can be categorized as inter­
personal (as they involve two people, possess strong levels of intimacy and 
affection and can vary greatly in nature), they also possess strong intergroup 
elements: they emphasize the similarities of male attitudes and behaviours, 
on the one hand, and the similarities of female attitudes and behaviours, on 
the other, and in doing so they follow the traditional gender stereotypes. The 
fact that stereotyped concepts of gender are used in couples to evaluate and 
establish their behaviour (more so than other concepts) is hardly surprising, 
given that a great many of these stereotypes refer to male-female rela­
tionships. Thus, it would seem perfectly acceptable for a traditionally 
minded male to ask his wife to 'take the initiative' with regard to domestic 
problems and their children's education, or even in the realm of her work 
outside the home, but not in her direct dealings with him. 
Relationships between interpersonal and intergroup aspects 
There exist three fundamental ways of connecting the interpersonal and the 
intergroup: (1) as orthogonally related; (2) as qualitatively different aspects 
of reality; and (3) as parts of a continuous flow. We shall examine each in 
turn. 
1. Orthogonality o
f
 both dimensions 
Stephenson (1984) was one of the first researchers who, from within social 
identity theory, established the possibility that the interpersonal and inter­
group need not be placed at opposite ends of a single dimension but rather 
act as two independent dimensions. In this way, a relationship could be 
placed at a given point in a bi-dimensional space, resulting from the 
intersection of both dimensions, as shown in Figure 1 1 . 1 .  
It is possible that there exist certain emotional-sexual relationships 
between men and women in which neither interpersonal nor intergroup 
aspects are of great importance. This would be the case in an incipient 
relationship in which contact between the two members of the couple is 
rather low - the fact that both interpersonal and intergroup requirements 
are low does not imply that they are of equal importance. In this case, it is 
likely that the intergroup aspect would predominate over the interpersonal. 
Other relationships can reflect a high intergroup requirement and a low 
interpersonal one: marriages of convenience provide a good example, or 


1 58 
Social identity: international pers
pectives 
INTERGROUP DEMANDS 
High 
'Compromise' 
'Conflict' 
High 
INTERPERSONAL 
DEMANDS 
'Collusion' 
'Apathy' 
Low 
INTERGROUP DEMANDS 
Low 
INTERPERSONAL 
DEMANDS 
Figure 1 1 . 1  Inter
personal and intergroup demands according to Stephenson (1984) 
those marriages typically found in the Victorian age, in which intimacy was 
virtually non-existent and almost the entire relationship and the behaviour 
of the couple had been already established by social criteria. Relationships 
with a high interpersonal and a low intergroup component would be those 
in which the couple challenges established social criteria concerning behav­
iour and attitudes expected from men and women involved in a love affair 
and create a complete different and unique relationship. Finally, it should 
also be possible to find relationships in which both the interpersonal and 
intergroup aspects are present. A graphic example of this could be the 
relationship portrayed in Liliana Cavanni's film The Night Porter. This 
story concerns an ex-Nazi official and a former female prisoner who main­
tain a passionate relationship that reflects (to a great extent) intergroup 
requirements since a great part of their behaviour is stereotypical (domi­
nance, aggression, and so on, in the male and submission, passivity, and so 
on, in the female). However, at the same time, interpersonal aspects are 
equally important: the relationship is deeply idiosyncratic and strange, 
considering they belong to entirely separate groups, their romance thus 
being distinguished from the millions of male-female relationships estab­
lished during the Second World War. 
An example of the simultaneity of interpersonal and intergroup aspects 
appears in our study (Moya, 1987) on social episodes. Following Forgas's 


Social identity and interpersonal relationshi
ps 
1 59 
approach (1 979) a number of subjects (60 female and 36 male psychology 
students, 9 male and 13 female student nurses and 23 housewives) were 
asked to indicate the perceived similarity among a set of situations rep­
resenting usual interactions with members of the opposite sex and rate these 
same situations on 1 5  bipolar scales. Both the situations and the scales had 
been obtained in a previous session and were all representative of the 
heterosexual interactions of each group. Examples of these interactions 
include 'talking on the telephone with a boyfriend/girlfriend', 'having sexual 
contact', or 'studying with colleagues'. Here, our interest lies in those 
episodes in which the member of the opposite sex is a partner. For example, 
in the case of the housewives, these episodes included: 'speaking with my 
husband', 'having a drink with my husband', 'making love with my 
husband', and in the case of the student nurses, 'eating with my partner', 
'having arguments with my partner', 'making love with my partner' and 
'waking up next to my partner'. 
The results show that the subjects perceive these episodes as intergroup 
interactions, in the sense that they consider their behaviour here as being 
very similar to that of members of the same sex (remember that uniformity 
is one of the fundamental characteristics of intergroup behaviour). Thus, on 
the scale of 0 to 9 points whose extremes were 'I behave differently to how a 
member of my sex would' and 'I behave in a similar way . .  
.', the scores 
obtained in those episodes involving a couple were the following: 7.3 in 
male psychology students and 7.7 in female; 4. 1 5  in male student nurses and 
6.3 in female; and 6.92 in housewives. At the same time, however, these 
interactions were perceived as being highly interpersonal, since the subjects 
saw themselves as expressing great intimacy, emotional involvement and 
dignity. They also believed that they expressed themselves spontaneously in 
the relationship and felt treated as persons. This is confirmed by the results 
of a multidimensional scale analysis (INDSCAL by Carroll & Chang, 1970) 
of the ratings of the interaction episodes on the 1 5  bipolar scales. See, for 
example, the episodic space obtained by the housewives, where two 
dimensions proved sufficient to represent the space. 
The first dimension was defined by the scales 'passive-active' (-.86), 
'intimate-non-intimate' (-.79), 'boring-interesting' (-.71), 'intense-super­
ficial' (-.58) and 'frustrating-rewarding' (-.51). To our way of thinking, this 
dimension clearly refers to the emotional aspect of the episodes - which 
could be termed intimacy - and is related with the person's degree of 
emotional involvement in the situation. The episodes 'chatting with our 
child', 'making love with my husband', 'playing with our child' and 
'interactions with friends' were situated at one extreme, while 'speaking with 
the children's teacher' and 'going to the doctor' were placed at a midway 
point on the dimension. Finally, at the other extreme were positioned the 
shortest interactions with virtual strangers (shopkeeper, baker, gas man). The 
second dimension was defined by aspects of dignity, equality and control 
perceived by the person in the episode, the scales defining this dimension 
being: 'relaxed-tense' (-.89), 'pleasant-unpleasant' (-.8), 'cooperative-


1 60 
Social identity: international perspectives 
competitive' (-.8), 'free-coerced' (-.79), 'I feel treated as a person-I don't 
feel treated as a person' (-.77), 'I know how to behave-I don't know how to 
behave' (-.75), 'I express myself as I am-I don't express myself as I am' 
(-.74) and 'reciprocated-not reciprocated' (-.71). At one extreme of this 
dimension were located the following episodes: 'a man is pursuing her', 
'quarreling with our child', 'going to the doctor', 'speaking to the children's 
teacher' and 'speaking to an office worker'. At the midway point more 
superficial episodes were placed ('going to the hairdressers', 'chatting with a 
neighbour') and at the other extreme were situated 'making love with my 
husband' and the relationships with children and neighbours. 
As can be seen, there is no interpersonal-intergroup dimension along 
which the interactions of these housewives with men can be placed sub­
jectively so that interpersonal episodes could be placed at one extreme -
relationship with partner or with child - and the intergroup ones at the 
other - a man is pursuing her, going to the hairdresser. An interpersonal 
relationship (for example, with their child) is perceived in an entirely 
different way depending on whether the housewife is playing or arguing 
with himlher. 
2. The interpersonal and intergroup as qualitatively diff
erent aspects 
o
f identity 
Kay Deaux (1992, 1993) has recently established that social and personal 
identities, rather than being clearly distinguishable, are in fact deeply inter­
related. According to Deaux, social identities can be defined as those roles 
or categories within which a person feels represented. Personal identity 
would refer to those features or forms of behaviour which a person feels 
adequately describe himlherself; these characteristics are often intimately 
associated with one or more social identities. In this way, a male X can 
regard himself as being very much a typical man, similar to the majority of 
men in his social surroundings (social identity), and at the same time 
consider himself as 'affectionate' (personal identity). This latter character­
istic is probably more associated with his belonging to the category of 
father, grandfather or spouse than to that of a military officer (imagine that 
this male belongs to all these categories or plays all these social roles). In 
this way, personal identity is defined, at least in part, by our membership of 
a group, at the same time as these group memberships will, in themselves, 
be charged with personal meaning. 
It should also be borne in mind that identity can be perceived as a 
hierarchical structure in which groups of identities are related to categories 
of characteristics or attributes. For example, the identity of a female Y 
would be exemplified as shown in Figure 1 1 .2. The upper part of the figure 
shows social identities and the lower - as if it were a mirror - the com­
ponents of a personal identity. 
One of Deaux's primary concerns surrounds this hierarchical structure of 
social identity. The higher a determined identity, the more behavioural 


I 
Woman I 
1 Maintains interpersonal 1 
relationships 
1 
Professional & 1 
independent 
1 
Close 1 
1 Superficial 1 
1 Professional 1 
1 
With a progressive 1 
ideology 
I 
1 Involved in 1 
Housewife 1 1 Popular 1 
1 University 1 
1 
Boss 1 
1 
Feminist 1 1 
Lesbian 1 
a couple 
with men 
student 
- Worried about 
- Devoted to others 1- Attractive 1 
1- Ambitious 1 1- Ambitious 1 
- With social 
- Independent 
her appearance 
- Kind 
- Sexy 
- Emotional 
- Aware of her 
- Attractive 
feelings 
1- Kind 
1 
- Committed 
1- Nice 
1 
- Good-looking 
1- Helpful to 
others 
1 
- Works hard 
- Competitive 
worries 
- With many 
I 
I 
- Ambitious 
girlfriends 
I 
I 
1- Works hard 
1 
- Follows her own interests 
1- Fights against injustice 1 
- Active 
1- Follows her 
own interests 1 
- Sensitive 
- Practical 
- Affectionate 
- Responsible 
Figure 1 1 .2 
Example o
f
 a hypothetical identity taken f
rom Deaux's f
ormulation and Maya's data (Maya, 1 989; Maya & Peplau, in 
preparation) 
R 
'" 
is' 
--
) 
 
S. 
I:l 
;::
 
:::.
 
S· 
ő 
~ 
(1) 
' 
;::
 
I:l 
--
€ 
::r 
T. 
;::
 
} 
.;:;. 
'" 
0\ 


1 62 
Social identity: international perspectives 
Table 1 1 . 1  Meta-contrast ratios in self-perception 
and the f
our stimuli used by Moya and Pep/au 
UCLA 
Granada 
Person involved in a couple 
1 .09 
1 .25 
University student 
1 .22 
1 . 1 3  
Professional 
1.03 
0.88 
Committed to their family 
0.95 
1.04 
repercussions this will have for an individual. Although it is likely that 
certain exceptional identities occupy high positions in almost all individuals 
in a society (such as being male or female), it is also possible that there exist 
great individual or subcultural differences. Maintaining her position in a 
couple may, for a certain woman be thought of as belonging to an entirely 
secondary category included in, or subordinate to, her membership of 
another group, such as that of a high social status. On the other hand, for 
another woman, being part of a couple could be considered as the most 
general category, into which all other memberships or social roles fall. It is 
worth remembering that any characteristics associated with one social 
identity can be associated with others and that this association between 
social identities and personal characteristics can vary from one individual to 
the next or from one social group to another. 
Some of these points were verified in recent research (Moya & Peplau, in 
preparation). A total of 204 subjects (95 men and 1 09 women) took part 
in the first test, all students of an introductory course in psychology at the 
University of California at Los Angeles (UCLA). In the second test, 45 men 
and 1 25 women participated, all third-year psychology students at the 
University of Granada. All subjects were asked to rate themselves on 22 
bipolar scales (9 represented instrumental attributes, 9 expressive and 4 
neutral). They were then asked to rate four different types of people of the 
same sex on the same 22 scales (university student, person involved in a 
lasting relationship, a committed family man/woman and a professional). 
With these scores, 4 meta-contrast ratios were calculated, one for each 
stimulus, following Turner, Hogg, Oakes, Reicher and Wetherell's proposal 
(1987). If the meta-contrast ratio is lower than I, the stimulus in question is 
perceived as being a member of the outgroup. The higher above I, the 
greater the stimulus is perceived as a member of the ingroup. The meta­
contrast ratios obtained by the American and Spanish students appear in 
Table 1 1. 1 .  
UCLA students only considered persons committed to their family as 
outgroup members (RMC lower than 1), while for those from Granada, 
only the professional was considered as such. Likewise, both sets of students 
perceived themselves as being closer to the person involved in a relationship. 
An analysis of variance was performed using as the dependent variable 
the meta-contrast ratio. The independent variable was subjects' current 
involvement in a relationship and had two values: involvement versus non-


Social identity and interpersonal relationshi
ps 
1 63 
involvement. For UCLA students involved in a relationship, the meta­
contrast ratio with the person committed to the family was .89. For those 
not involved, the corresponding figure was 1 .01 F(I ,  1 55) = 7.26 was 
significant at p < .01 .  No similar differences appeared in the Granada 
sample. 
The importance of taking into account the specific contents that 
individuals associate with each social identity has also been shown in other 
results of this research. According to our hypothesis, those involved in an 
emotional relationship with a member of the opposite sex rated themselves 
as higher in the expressive items than those not involved. On the other 
hand, and contrary to expectations, this was also the case with instrumental 
items (although here the differences did not reach statistical significance). 
These tendencies were found in both the American and the Spanish sample. 
In our judgement, these results reveal the existence of different social 
identity configurations in two distinct social mediums (the American and 
the Spanish) since being a university student or a committed family member 
seems to occupy different positions in identity structure, while the relation­
ship between these identities in each sample also presents certain differences 
(as in the opposition present in the American sample between those 
involved in a relationship or not). 
3. The continuous flow o
f
 identities 
Turner and coIleagues (1987) conceive of the existence of different levels of 
abstraction in the categorization of '1', separated from each other and 
discretely formed, in such a way that the individual passes from one to 
another according to how the situation varies. It is therefore the dynamic 
character of identity which is emphasized here, an aspect that has been 
present in social identity theory but which is rarely reflected in practical 
research carried out under its inspiration (Condor, 1989). Tajfel (1982a) 
emphasized how social identifications could be subject to temporal fluctu­
ations: 'the psychological existence of a group for its members consists of a 
complex sequence of appearances and disappearances, of factors emerging 
and blurring into thin air' (p. 485). Turner considered his model of identity 
as neither static, fixed, global or corrected, but totaIly the reverse: 'the 
categories of "Self" are not absolute, but rather "Self" is dynamic, relational, 
comparative, fluid, context-specific and variable' (Turner, 1988, p. 1 14). 
Thus, it is possible that the interactions maintained by members of a 
couple change continuously from some category memberships to others (as 
conceived by self-categorization theory) or fluctuate incessantly along the 
interpersonal-intergroup continuum. For example, deciding what to do one 
Sunday afternoon, it might occur to a boy to invite to the cinema the girl he 
has just started seeing, to then pay for her ticket and attempt some physical 
or sensual contact (foIl 
owing certain stereotypical criteria). He suddenly 
finds out that the girl cannot stand going to the cinema and would much 


1 64 
Social identity: international perspectives 
rather have an afternoon chat in a quiet cafe, he thus rapidly changes his 
plans. In this way, a form of behaviour which commenced with certain 
intergroup components has quickly found itself being affected by the 
particular characteristics of the people involved. 
Final comments 
This chapter's purpose has been two-fold. On the one hand, it has 
attempted to show that the study of intergroup relations and the processes 
and structures related to identity can be particularly enhanced if the 
interpersonal relationships, in which the individuals are involved, are taken 
into account at the same time (for example, relationships between couples). 
On the other, it has indicated how the study of interpersonal relationships 
can be enhanced if we simultaneously bear in mind identity and intergroup 
aspects. We believe that an articulation of intergroup and interpersonal 
processes could be more useful in the understanding of certain phenomena 
than considering these two aspects as antagonistic or contrary. 
However, this articulation between interpersonal and intergroup aspects 
seems relatively complex. We have made reference to some approximations 
that contain interesting suggestions regarding how these two components 
should be articulated. Yet, rather than coming up with definitive answers, 
these approximations only suggest paths down which one can advance, thus 
granting an important role to empirical research. 
Focusing now on the case of couples, the arguments we have put forward 
show that a greater understanding of them could be attained if, in addition 
to considering the specific dynamics of the relationship, other factors were 
also taken into account. These could include the social identity structure of 
those involved in a couple, as well as the content of their different social 
identities. Nevertheless, the study of these identities demands approxima­
tions which consider their relational character (Rosenberg & Gara, 1985). 
So, although identities of a general character exist (such as 'being a man' or 
'being a woman') and guide the behaviour of individuals without involving 
other people (for example, they serve the individual when deciding how to 
dress or what hobbies to take up), there are others (such as 'being in a 
couple' or 'being a father') which cannot be understood without taking into 
account a relationship with other people (for example, his/her partner or 
the specific composition of the family as in the mentioned examples). 
Consider, for example, the different conception of self felt by a 'woman 
in a relationship' before and after her partner had been paralysed in an 
accident. 
In conclusion, this approach merely connects up with an old tradition of 
social psychology (Morales, 1989) which has included: the battle over 
structural and process aspects of identity (for example, Manford Kuhn and 
Ralph Turner versus Blumer and Goffman within symbolic interactionism); 
the extraordinarily discerning analysis of social and personal identity carried 


Social identity and inter
personal relationships 
1 65 
out by, for example, Goffman (1 968); and establishing the impossibility of 
understanding 'self' or the identity, disregarding the specific social inter­
actions maintained by individuals. 
Note 
I .  The content of Figure 1 1 .2 has an illustrative purpose. The structure of the identity 
within has been taken from Deaux (1993) and the content of the different identities has been 
elaborated from the results of research carried out by Moya (1989) and Moya and Peplau (in 
preparation). 


1 2  
Social Identity and Aspects of Social 
Creativity: Shifting to New Dimensions of 
Intergroup Comparison 
Steve Hinkle, Laurie A. Taylor, Lee Fox-Cardamone and 
Pamela G. Ely 
Most accounts of social identity theory (SIT) outline three response patterns 
which groups and group members can make to unfavourable social 
identities: individual mobility, social competition and social creativity (see 
Hogg & Abrams, 1988; Tajfel & Turner, 1986). Perhaps the earliest dis­
cussion of social creativity can be found in two of Tajfel's 1978 publications 
concerning group differentiation (l 978d) and the social psychology of 
minorities (l978e). In these works, Tajfel discussed low status or minority 
groups' response to their unsatisfactory social identities. When social cir­
cumstances supported leaving the ingroup, social mobility or moving on to 
another, presumably high-status, ingroup was proposed. When mobility was 
not sociologically or psychologically likely, Tajfel proposed three 
possi 
bili 
ties. 
(a) to become, through action and reinterpretation of group character­
istics, more like the superior group; 
(b) to reinterpret the existing inferior characteristics of the group, so that 
they do not appear as inferior but acquire a positively valued distinc­
tiveness from the superior group; 
(c) to create, through social action and or diffusion of new 'ideologies' new 
group characteristics which have a positively valued distinctiveness 
from the superior group (Tajfel, 1978c, pp. 93-94). 
Later, this list was elaborated to include a fourth possibility: 
(d) group members may also seek positive distinctiveness by 'changing the 
out-group (or selecting the out-group) with which the in-group is 
compared - in particular, ceasing or avoiding to use the high-status 
out-group as a comparative frame of reference' (Tajfel & Turner, 1979, 
p. 43). 
In more recent discussions of SIT, the first of these is typically referred to as 
'social competition', while the latter three are, generally, considered as 


Social identity and aspects o
f
 social creativity 
1 67 
aspects of 'social creativity' (see Hogg & Abrams, 1988; Tajfel & Turner, 
1986).' 
The purpose of this chapter is to take a detailed look at one of these 
aspects of social creativity: the establishment of new group characteristics or 
dimensions of intergroup comparison as a means to institute positive 
distinctiveness for the ingroup. First, theoretical issues pertinent to this 
aspect of social creativity will be summarized. The chapter will then review 
past research pertinent to establishing positive ingroup distinctiveness 
through the use of new dimensions for intergroup comparison. New, as yet 
unpublished, research we have conducted with several other colleagues will 
then be presented. The chapter will conclude with some evaluative com­
ments on this aspect of SIT and SIT research along with discussion of 
applications and issues for future research. 
Some theoretical issues 
Legitimacy and stability 
At the core of many SIT predictions are the notions of status legitimacy and 
stability. The relationships between groups can be perceived as legitimate or 
illegitimate, stable or instable (Tajfel 1978a). In general, Tajfel saw legiti­
macy and stability as covarying. That is, according to SIT, legitimate and 
stable intergroup relations tend to co-occur as do illegitimate and instable 
intergroup relations (Tajfel, 1978a; for contrasting perspective, see Turner & 
Brown, 1978, who treat legitimacy and stability as independent aspects of 
the intergroup situation). SIT goes on to associate social competition with 
illegitimate/instable intergroup relations. Hence, the social creativity 
strategies, including establishing positive ingroup distinctiveness on new 
dimensions of intergroup comparison, are postulated by SIT as being most 
relevant to legitimate and stable intergroup contexts (Tajfel, 1978b). Here, 
where groups with threatened or negative social identities see the status quo 
as justified and/or not likely to change, direct social competition to alter the 
status quo on prevailing comparison dimensions is not apt to be viewed as 
likely to succeed or justifiable. Hence, the creation of new comparison 
dimensions as a path to positive social identity is a strategy of obvious 
motivational significance. 
Group status 
Another theoretical issue posited as relevant to the use of new intergroup 
comparison dimensions is group status. While SIT is not as clear as it could 
be, it appears as though Tajfel believed that the use of new comparison 
dimensions was most likely for low-status groups in legitimate/stable status 
circumstances, but most likely for high status groups when they perceive 
their advantaged position as illegitimate and/or instable (Tajfel, 1978c). 
However, not all of the early SIT theorizing is consistent on this. For 


1 68 
Social identity: international perspectives 
example, Caddick (1982) sees illegitimacy and instability as instigating the 
search for new dimensions of distinctiveness for both high- and low-status 
groups. It also appears that, irrespective of illegitimacy and instability, early 
SIT research viewed the use of new comparison dimensions as being more 
characteristic of low- than of high-status groups (Brown & Ross, 1982). 
Selection o
f
 new comparison dimensions 
Perhaps one of the most useful and detailed early theoretical discussions in 
SIT of the use of new dimensions for intergroup comparison is that of van 
Knippenberg (1978). In particular, he offers some valuable ideas regarding 
the selection of new comparison dimensions. Assuming legitimacy and 
stability, existing status differences will continue and be difficult to deny. As 
a result, van Knippenberg proposed that new comparison dimensions must 
be capable of enhancing a group's social identity, but without affecting 
already existing status dimensions. That is, in selecting new comparison 
dimensions to effect a more favourable social identity, the new dimensions 
will afford the ingroup positive distinctiveness while being uncorrelated with 
the existing status dimension where the ingroup does not compare 
favourably. 
Lemaine, Kastersztein and Personnaz's (1978) position is somewhat 
different from those of Tajfel (1978c) and van Knippenberg (1978). The 
latter two authors take no position regarding the relative importance of new 
versus old dimensions of comparison. In contrast, Lemaine et al. clearly 
hold that groups will strive to make new bases for intergroup comparisons 
dominant. For Lemaine et a!., groups will not be content to simply 
acknowledge their own inferiority on existing status dimensions while 
claiming superiority on a new status dimension. Rather, the ingroup will go 
further and strive to make the new basis for comparison of greater 
importance and, thus, the dominant basis for status comparisons. In a later 
paper, van Knippenberg (1984) modifies his earlier stance in a manner 
consistent with the views of Lemaine et a!., at least with respect to the 
classic conflictual intergroup situations emphasized by SIT. He also pro­
poses socially cooperative intergroup relations where groups coexist in a 
climate of mutual appreciation. Here, van Knippenberg (1 984) proposes 
that groups can and will acknowledge both their own superiority and 
inferiority relative to other groups on various comparison dimensions and 
that all such dimensions can be valued by all groups. He does, however, 
note that ingroups and outgroups may weight the relative importance of the 
sundry dimensions differently. 
New comparison dimensions versus other responses to un
f
avourable 
identities 
Another important theoretical issue concerns the prepotency of turning 
to new bases for intergroup comparison relative to other responses to 
unfavourable identities such as mobility, social competition and other social 


Social identity and aspects o
f
 social creativity 
1 69 
creativity strategies, including shifting the locus of comparison to new 
outgroups and re-evaluating the valence of existing, unfavourable com­
parisons. Tajfel (1978c) clearly sees mobility as the strategy of choice in 
circumstances where mobility is viewed as possible and desirable (that is, 
with a social mobility ideology; Hogg & Abrams, 1988; Tajfel & Turner, 
1979). However, there is little in his writings to clearly order social com­
petition and social creativity into a hierarchy of strategic choice. Certainly, 
he says virtually nothing concerning the relative strength of the various 
forms of social creativity. In contrast, Ball, Giles and Hewstone (1984) 
propose that mobility attempts will precede social creativity strategies, 
which, in tum, will take place prior to social competition. However, Ball et 
al. say nothing about the ordering of the various forms of social creativity. 
In general, the lack of theory regarding use of the various social creativity 
strategies is an unfortunate omission in SIT. 
Process issues 
Finally, what does SIT say relevant to the process of creating new bases for 
positive intergroup distinctiveness? First, group members must somehow 
develop a shared perspective on the attributes serving as new bases for 
intergroup comparisons and the ingroup's standings on these new attributes 
must, in fact, be positive (Tajfel, 1978c). The next step involves persuading 
the outgroup to accept the new bases for intergroup comparison and the 
ingroup's positive standing on these attributes. That is, consensual accept­
ance by both the in- and outgroup is necessary for the new dimensions of 
intergroup comparison to fully validate the ingroup's positive social identity 
(Tajfel, 1978c). Lemaine et al. (1978) add to SIT theorizing about such 
processes. They hold that the initial intragroup aspect of this process will, 
ideally, take place without the knowledge of the outgroup. Such seclusion 
shields the ingroup seeking new bases for intergroup comparisons from 
criticism and helps reduce uncertainty regarding the new comparisons. 
Past empirical research: a selected review of the literature2 
In spite of the complex SIT theorizing concerning the use of new com­
parison dimensions as a response to negative social identities, most of the 
relevant empirical literature has yet to address the theoretical postulates 
outlined above. As we shall see, the dominant characterization of this 
literature is simply the demonstration that groups with threatened social 
identities will, at times, tum to new bases for making intergroup com­
parisons. 
In all likelihood the most frequently cited empirical work is that of 
Lemaine and his colleagues (Lemaine et aI., 1978). In one study, children at 
summer camps participated in a building competition where some groups 
were relatively disadvantaged by being given inadequate materials. While 


1 70 
Social identity: international perspectives 
unable to build as large or elaborate huts as other groups, disadvantaged 
groups began developing gardens, thereby establishing a new basis for 
intergroup comparisons more favourable to their ingroups. 
In a review, Hinkle and Brown (1990) cited 12 papers, using both 
laboratory and field contexts with subject populations from nurses to 
university lecturers; all demonstrated that the same group when making 
intergroup comparisons on multiple dimensions will evince in 
group 
favouritism on some of the dimensions, outgroup favouritism on others, 
and no favouritism at all on still other dimensions. The simultaneous 
occurrence of ingroup and outgroup favouritism on differing dimensions is 
consistent with the notion of using new bases for intergroup comparisons to 
establish positive social identities when comparisons on other dimensions do 
not favour the ingroup. In addition to the studies reviewed by Hinkle and 
Brown, other similar results are reported in Capozza, Bonaldo and 
DiMaggio (1982). The diversity of subject populations and paradigms 
across these various researches points to the robustness of this form of 
social creativity. 
Perhaps one of the most interesting early SIT studies bearing on the use 
of new dimensions for making intergroup comparisons is Brown and Ross 
(1982). The study's primary emphases are on both social competition and 
social creativity. In particular, Brown and Ross's consideration of creativity 
emphasizes the re-evalution of the importance of various comparison 
dimensions rather than the topic of the present chapter, the use of new 
comparison dimensions. Nonetheless, while the latter issue is not focal to 
Brown and Ross, they do present some important relevant data. Before 
discussing their findings, some background regarding the paper's hypotheses 
and design is necessary. 
Brown and Ross discuss illegitimacy and instability as inducing a 'threat 
to identity' or insecure social identity which, in response, produces either 
social competition or social creativity ('the strategies . . .  
were not thought 
to be mutually exclusive', p. 1 62). Their study's primary manipulation was 
presented in terms of high, moderate or low threats to identity, though 
careful examination of their procedures indicates that the study's primary 
independent variable can just as easily be termed 'status illegitimacy'. A 
second important independent variable in this research was a manipulation 
of ingroup status as being either superior or inferior. One of the 
experiment's primary predictions was 'that increases in differentiation along 
. . .  
alternative dimensions would . . .  
vary in proportion to the threat to 
identity experienced, and that these changes would be most prominent in 
the lower status groups' (p. 1 62). 
When ingroup bias on alternative dimensions of intergroup comparison is 
compared before and after implementation of the status illegitimacy 
manipulation, the data clearly show that inferior status groups in the high 
and moderate illegitimacy (that is, threat to identity) conditions manifested 
substantially more ingroup favouritism on the alternative comparison 
dimensions. This shift towards greater ingroup bias on the alternative 


Social identity and aspects o
f
 social creativity 
1 7 1  
dimensions did not occur for inferior status groups in the low illegitimacy 
condition, nor did it occur for superior status groups in any of the three 
conditions of status legitimacy.3 
These results are partially supportive of early SIT theorizing in that the 
use of new comparison dimensions to attain positive group distinctiveness is 
more characteristic of low- rather than high-status groups. On the other 
hand, Brown and Ross's results contradict Tajfel's (1978c) view that this 
and other forms of social creativity would be most prominent for low-status 
groups when there are legitimate/stable status relations. Similarly, their 
findings are also inconsistent with Caddick's (1982) prediction that the 
various social creativity strategies would occur for both high- and low-status 
groups when the intergroup status structure was both illegitimate and 
instable. It appears from Brown and Ross's data that the most empirically 
tenable theoretical position for SIT is that social creativity, at least in the 
form of using new dimensions for making ingroup-favouring intergroup 
comparisons, is most likely to be observed in low-status groups under 
conditions of status illegitimacy and/or instability (though the specific 
effects of illegitimacy and instability as distinct variables remain to be 
tested). 
Other interesting data from Brown and Ross bear on the issue of the 
valuing of new dimensions of intergroup comparison. Recall Lemaine et 
al.'s (1987) and van Knippenberg's (1984) hypotheses that groups shifting to 
new dimensions of intergroup comparison to attain positive distinctiveness 
would also see these new dimensions as of greater importance than previous 
bases for intergroup comparison. In general, Brown and Ross found that 
while inferior status groups were, with status illegitimacy, quick to use 
alternative comparison dimensions to evince ingroup favouritism, there was 
little evidence to support the notions that these new dimensions' importance 
increased or that they were viewed as being of greater importance than the 
original intergroup comparison dimensions. This result is consistent with 
similar findings reviewed by Hinkle and Brown (1990). Indeed, the existing 
evidence seems, if anything, more consistent with the view that low-status 
groups are willing to salvage damaged or unfavourable social identities by 
shifting to or emphasizing new in 
group-favouring dimensions of com­
parison even while acknowledging that the latter are of lesser importance. 
An example of this might be the losing sports team that emphasizes its own 
superiority in terms of better sportsmanship (see Lalonde, 1992) or more 
enjoyment of the game, even while willingly owning that it is more 
important to win. 
Caddick (1982) presents similar findings with respect to illegitimate status 
differences and the occurrence of ingroup favouritism on new dimensions of 
intergroup comparison. An illegitimate status difference led to increased 
expression of ingroup favouritism on several comparison dimensions other 
than the original basis for intergroup comparison. Interestingly, and in 
contrast to Brown and Ross, ingroup favouritism on the new comparison 
dimensions occurred for members of both low- and high-status groups. 


1 72 
Social identity: international perspectives 
Caddick develops a thought-provoking analysis of this result based on the 
fact that in his experimental procedure the high-status group is clearly 
responsible for the illegitimacy of the initial comparison between the 
groups. Caddick suggests that this may have compromised the high-status 
group's social identity through guilt feelings induced by violation of values 
such as fairness or justice. To compensate for their compromised favourable 
social identities, the high-status groups adopted new bases for making 
ingroup-favouring intergroup comparisons. 
Some new empirical findings 
Our own investigation of social creativity issues began rather unexpectedly 
through an effort to replicate findings from a study by Condor, Brown and 
Williams (1987). One component of Condor et al.'s experiment involved 
groups of unacquainted subjects working on a laboratory task to generate 
ideas for recruiting students to their university under an instructional set 
emphasizing task issues such as efficiency and productivity. Following this 
exercise, subjects watched a short video of another group engaged in the 
same endeavour and then made in- and outgroup ratings on task-oriented 
dimensions such as the group's competence and energy and socio-emotional 
dimensions such as the groups' friendliness and warmth. Condor et al. 
found much greater ingroup favouritism on the task-oriented dimensions of 
intergroup comparison than on the socio-emotional dimensions. This was 
expected since the task-oriented dimensions are more closely associated 
with the experimental instructional set than the socio-emotional dimen­
sions. 
Hinkle and Crook (1 987) attempted to replicate Condor et al. as closely 
as possible and found a complete reversal of their findings. That is, signifi­
cantly greater levels of ingroup favouritism occurred on the socioemotional 
dimensions than on the task-oriented dimensions. Why? There were only 
two important differences between the procedures of Condor et al. and 
Hinkle and Crook. First, the subject populations differed, with the latter 
study using American university students while the former's subjects were 
students from a British university. While it is virtually impossible to rule out 
differences in the populations as an account for the reversal of findings, one 
of the results from Hinkle and Crook points to the comparability of the two 
subjects populations. Their subjects saw the task-oriented dimensions as of 
greater importance than the socio-emotional dimensions. Thus, Hinkle and 
Crook's findings appear to be a social creativity effect in that their subjects 
are manifesting the most ingroup favouritism on a secondary, less important 
dimension of intergroup comparison. 
There was a second difference in the procedures of Condor et al. and 
Hinkle and Crook. Rather than presenting the outgroup via a video, the 
latter simply asked subjects to think of an average group of university 
students working on the same task and rate this 'imagined' outgroup. 


E 
(/) 
'." 
'§ 
0 
> 
† 
a. 
:::l
 
e 
CJ) 
.E: 
Social identity and aspects o
f
 social creativity 
35 
30 
25 
20 
1 5  
1 0  
5 
0 
-5 
Below 
average 
Comparison dimension 
• Task-oriented 
o Socio-emotional 
Average 
Outgroup status 
Well above 
average 
Figure 12. 1 
/ngroup f
avouritism on dif
f
erent comparison dimensions as a 
f
unction of outgroup status (Hinkle, Fox-Cardamone, Taylor & Crook, 
/988) 
1 73 
Assuming a rather high level of achievement motivation amongst these 
university student subjects, it may be that their vision of an 'average' group 
is of one whose performance does indeed threaten that of their own group. 
This interpretation corroborates the view that Hinkle and Crook's subjects 
are engaging in a social creativity process. If the outgroup's task perform­
ance is imagined to be quite good relative to ingroup performance, subjects' 
shift to the less important socio-emotional dimensions to manifest ingroup 
favouritism may be seen as an effort to offset the unfavourable intergroup 
comparison on the task-oriented dimensions. 
Experimental demonstrations o
f
 social creativity e
f
f
ects 
To follow up on the interpretation outlined above, Hinkle, Fox-Cardamone, 
Taylor and Crook (1988) replicated Hinkle and Crook's experiment and 
extended it by adding two additional experimental conditions. While some 
groups of subjects were asked to think of an average group of university 
students as the outgroup stimulus, others were asked to think of below 
average or well-above average groups. It was predicted that below average 
outgroups would not threaten the distinctiveness of the ingroup on the 
important, task-oriented dimensions. Hence, subjects in this condition were 
expected to manifest greater ingroup favouritism on the task-oriented 
dimensions than the socio-emotional dimensions. In contrast, the prediction 
for the average and well-above average outgroup conditions was that such 
outgroups would threaten the ingroup's distinctiveness on the important task­
oriented dimensions. As a result, social creativity would lead to greater 
ingroup favouritism on the secondary dimensions of intergroup comparison, 
that is, the socio-emotional dimensions. As can be seen in Figure 12. 1 ,  this is 
exactly what was found. 


1 74 
Social identity: international pers
pectives 
Interestingly, Hinkle et al. collected data on the perceived importance of 
the task and socio-emotional comparison dimension a
f
ter subjects had made 
their ingroup and outgroup ratings. There was no evidence for any change in 
the importance of the dimensions. The task-oriented dimensions were still 
seen as being significantly more important than the socio-emotional 
dimensions in all three experimental conditions. Thus, the social creativity 
seen in the average and well-above average outgroup conditions does not 
appear to actually compensate for unfavourable comparisons on the task­
oriented dimensions. In these conditions subjects are not saying, 'Your group 
is better at the task, but it is more important to get along well and we are 
better at that.' Rather, they appear to be saying, 'While your group is better 
than ours at the task, at least we get along better.' The social creativity seen 
in this study does not appear to compensate for an unfavourable social 
identity, but simply to take a step towards salvaging the ingroup's social 
identity. As with earlier research reviewed above, this result is inconsistent 
with Lemain et al.'s (1978) and van Knippenberg's (1984) views that the shift 
to new, ingroup-favouring dimensions of intergroup comparison would 
necessarily be accompanied by perceiving the new dimensions as of greater 
value or importance than those upon which the ingroup did not compare 
favourably. 
The Hinkle et al. research can be criticized for the use of an imagined 
manipulation of outgroup status. Recently, two of us (Laurie Taylor and 
Steve Hinkle) conducted an experiment similar to those reported above, but 
incorporating a more concrete manipulation of outgroup status. Again, 
groups worked under a task-emphasis instructional set to generate ideas for 
the recruitment of students to their university, and the ingroup and out­
group were rated on both task-oriented and socio-emotional dimensions of 
intergroup comparison. However, rather than asking students to imagine 
outgroups of varying status levels, outgroup status was manipulated by 
providing subjects with a list of some of the recruitment ideas generated by 
the other group. These lists were prepared in light of pilot test data such 
that they contained only the lowest (below average outgroup status) or 
highest (above average outgroup status) rated ideas from the pilot study. In 
the below average outgroup condition, ingroup favouritism was greater on 
the task-oriented dimensions than on the socio-emotional dimensions. In the 
above average outgroup condition, a social creativity effect was again 
observed: greater ingroup favouritism on the socio-emotional comparison 
dimensions than on the task-oriented dimensions (see Figure 12.2). 
Conclusions, applications and directions for future theory and 
research 
General conclusions 
The strongest and most consistent finding across numerous studies using 
various populations and paradigms is that social creativity in the form of 


E 
If) 
ok 
0;::
 
::J 
0 
> 
† 
Q. 
::J 
2 
CJ) 
.E: 
Social identity and aspects o
f
 social creativity 
30 
25 
20 
1 5  
1 0  
5 
0 
_
_
_
_ 
--j Comparison dimension 
Below average 
• Task-oriented 
D Socio-emotional 
Above average 
Outgroup status 
Figure 12.2 Ingroup f
avouritism on dif
f
erent comparison dimensions as a 
f
unction of outgroup status (unpublished data f
rom Taylor & Hinkle) 
1 75 
responding to a negative social identity with ingroup favouritism on new 
comparison dimensions is a reliable and robust phenomenon. Furthermore, 
the experimental investigations of Hinkle et al. and Taylor and Hinkle 
demonstrate that low ingroup status causes greater ingroup favouritism on 
new comparison dimensions than on the original dimensions upon which 
the ingroup does not fare well. 
Status, legitimacy and stability 
Past this broad conclusion, the existing research findings bearing on the 
theoretical issues raised at the outset of this chapter are limited. The issue of 
group status in conjunction with its legitimacy and stability has received the 
most attention and the early SIT theoretical formulations have received 
mixed support. The evidence from studies such as Brown and Ross (1982) 
and Caddick (1982) supports the notion that low-status groups are more apt 
to manifest group favouritism on alternative dimensions of intergroup 
comparison than high-status groups. How legitimacy and stability enter the 
theoretical equation has been far less clear. Caddick's results indicate that 
illegitimacy increases ingroup-favouring comparisons on new dimensions 
for both high- and low-status groups. Brown and Ross found this to be the 
case for low-status groups only. Clearly, other variables must be moderating 
the relationship between illegitimacy, group status and the occurrence of 
ingroup favouritism on new comparison dimensions. Relevant theory and, 
more importantly, research are needed. Caddick's speculation concerning 
the responsibility of a high-status group for the existence of an illegitimate 
intergroup status structure appears to be a potential moderator worthy of 
further empirical attention. 
At present, there appears to be no support for Tajfel's ( I978c) early 
proposal that low-status groups would be particularly likely to use new 


1 76 
Social identity: international perspectives 
dimensions of intergroup comparison when status relations are legitimate 
and stable. Perhaps with a legitimate/stable status structure it is very 
unlikely for low-status group members to perceive any change in the status 
quo relative to superior groups, including the use of new comparison 
dimensions. It seems plausible that another social creativity strategy, a shift 
to different comparison outgroups, may be the more likely response to an 
unfavourable social identity under a legitimate/stable status structure. A 
superordinate issue to any SIT discussion regarding legitimacy and stability 
is the need for more thoroughly elaborated theorizing and research on the 
relationships between legitimacy and stability. 
Importance o
f
 comparison dimensions 
The other theoretical issue which has received some empirical attention has 
to do with the importance or value placed on new intergroup comparison 
dimensions. Recall that both Lemaine et al. (1978) and van Knippenberg 
(1984) hypothesized that a shift to ingroup favouritism on new intergroup 
comparison dimensions would be accompanied by the new bases of com­
parison acquiring greater value or importance than the original comparison 
dimensions. The intuitive strength of this postulate makes it quite difficult to 
abandon. However, there is substantial evidence to the contrary. Groups 
appear to frequently evince ingroup favouritism on new, less valued com­
parison dimensions while acknowledging that they do not measure up to the 
outgroup in other, apparently more important, comparisons. This is an 
intriguing result calling out for more research to further illuminate this 
finding and perhaps identify moderator variables specifying when new 
comparison dimensions will or will not be seen as being of greater import­
ance than the initial comparison dimensions. 
Van Knippenberg's (1984) discussion of socially cooperative intergroup 
relations in contrast to SIT's traditionally competitive intergroup relations 
seems to offer a good deal of theoretical promise with respect to this as it 
seems plausible that the psychological importance of placing greater value 
on new intergroup comparison dimensions may only be relevant to com­
petitively oriented intergroup relations. 
Similarly, Hinkle and Brown's (1990) distinction between autonomous 
and relational processes of social identity construction may also be theor­
etically relevant (see also Brown, Hinkle, Ely, Fox-Cardamone, Maras & 
Taylor, 1992; Hinkle, Brown & Ely, 1992). Groups with autonomous 
identity construction orientations may establish entirely satisfactory social 
identities through seeing themselves favourably on relevant comparison 
dimensions independent of whether other groups rank below or above them. 
As a result, the relative value or importance of various comparison dimen­
sions may not be psychologically relevant with the autonomous orientation. 
Another potentially useful approach to developing more research and 
theory on the importance of different comparison dimensions takes a 
temporal perspective. It seems plausible that the immediate response to an 


Social identity and aspects o
f
 social creativity 
1 77 
unfavourable intergroup comparison is, to use a term introduced earlier, 
to salvage a positive social identity. If our research group fares disas­
trously in the latest government league tables of research quantity and 
quality, the first step is to establish a toehold towards re-establishing a 
favourable social identity. One way to do this is by claiming high status 
for the ingroup on any dimension, even an inferior one. 'Yes, the quality 
and quantity of our group's research is not on par with others, but we are 
doing some limited, innovative work and our teaching is very good. And, 
yes, we recognize that research innovation and teaching quality really are 
not the sine qua non for research groups.' This salvages the group's esteem 
in the short term and provides the foundation for subsequent develop­
ments such as actually improving the group's research productivity or 
convincing relevant entities in the social structure that research innovation 
and teaching quality ought to be the more important bases for intergroup 
comparIson. 
The processes o
f
 social creativity 
There is, to our knowledge, no research on Tajfel's ( l978c) and Lemaine et 
al.'s (1978) proposals that selection of new comparison dimensions proceeds 
first as an intragroup process and, according to Lemaine et aI., surrep­
titiously and, then, as an intergroup process where the ingroup attempts to 
persuade the outgroup regarding the importance of the new comparison 
dimension and the ingroup's superior status. The comments in the previous 
section on temporal aspects of social identity processes again seems 
relevant. It may be that groups with unfavourable identities must first take 
steps to salvage positive distinctiveness, then work within the group to 
develop genuinely favourable intergroup comparisons, and, finally, shift to 
the issue of persuading other groups as to the validity and importance of the 
new bases of comparison. Research paradigms without a longitudinal 
component are unlikely to detect or shed light on such processes. Unfor­
tunately, longitudinally oriented studies are virtually non-existent in the SIT 
tradition. 
Hinkle and Brown's (1990) notion of autonomous orientations may also 
be pertinent. For autonomous groups, since they are interested in the 
ingroup in comparison to abstract standards rather than in comparison to 
other groups, the need to persuade external audiences of the validity of new 
intergroup comparisons is apt to be irrelevant. 
An interesting and closely related issue is raised by van Knippenberg 
(\ 984), who argues that responses to unfavourable social identities are not 
always directed towards immediate enhancement of the group's social iden­
tity. Specifically, van Knippenberg discusses the role of self-presentational 
processes in the expression of ingroup or outgroup favouritism. Self­
presentational issues seem equally relevant to social creativity processes 
including the selection of new bases for making intergroup comparisons. 
Abrams and Hogg (1988; also Hogg & Abrams, 1990) have outlined similar 


1 78 
Social identity: international pers
pectives 
ideas which go well beyond social identity and self-presentation as motiva­
tional issues relevant to the expression of ingroup or outgroup favouritism. 
Strategy choice in response to unf
avourable social identities 
While there are a few studies on mobility versus either social competition or 
social creativity (or both, for example, Ellemers, Wilke & van Knippenberg, 
1993; Jackson, 1993; Wright, Taylor & Moghaddam, 1990), there is no 
research and virtually no relevant theory on selection of the various social 
creativity strategies. These are important theoretical issues. SIT's inability 
to predict the specific form of response to an unfavourable social identity is 
a significant omission and diminution of its utility. While there may be good 
historical reasons for why these issues were not well addressed in the early 
years of SIT's development, the contemporary status of SIT would be 
enhanced by progress on this front. 
Potential applications 
We are unaware of any genuinely applied literature on the use of new 
intergroup comparison dimensions. However, existing studies and theor­
etical discussions are rife with possibility. Perhaps the richest potential for 
application resides in issues of group morale, cohesion and motivation. 
How might poorly performing sports teams, business organizations and 
other similar entities deal with attendance problems of low group morale 
and motivation? The generation of new, favourable dimensions of inter­
group comparison offers one strategy. While the existing research indicates 
that a losing team's development of the view that they enjoy what they are 
doing more than their competitors does not fully psychologically 
compensate for its poor won-lost record, which such a team is still apt 
to see of premier importance, it still may serve as a starting point for 
rebuilding team spirit and motivation. While the theoretical basis for this 
has not yet been tested, the hypothesis that social creativity results in 
enhanced cohesiveness and, subsequently, better group performance is an 
intriguing one. Were notions such as this begun to be examined in applied 
contexts, other interesting issues would certainly come quickly to the fore. 
For example, in all previous research of this area, the creation or ingroup­
favouring use of new comparison dimensions has always arisen spon­
taneously from the studies' inherent intragroup dynamics. In contemplating 
applications, it is quite likely that the instigation to the use of new 
dimensions of intergroup comparison would result from the intervention of 
a group's leadership such as managers, administrators, coaches, and so on, 
or outside sources such as consultants. There is little if any helpful existing 
research or theory. Links to work in organizational psychology and the 
leadership literature would likely be beneficial in determining whether or 
not leadership efforts to stimulate social creativity would be effective or how 
such endeavours could be best implemented. 


Social identity and aspects o
f
 social creativity 
1 79 
Final comments 
In conclusion, research on social creatIVIty in the fonn of selecting new 
comparison dimensions for making ingroup-favouring evaluations has 
played a small, but important, role in SIT research, a role which appears to 
be increasing in its significance and potential applications. The robustness of 
this fonn of social creativity is well documented and some important com­
plexities have begun to be addressed. The existing empirical findings provide 
support for certain SIT precepts, but not others. Our own view on this is 
that in instances where the literature has not supported SIT it has served the 
valuable function of precipitating modifications of the theory that have 
enhanced its power and viability. However, as should be clear fonn earlier 
discussion in this chapter, there are many interesting issues still in need of 
both conceptual and empirical attention. SIT can only be enriched as these 
issues stimulate further work in the study of intergroup processes. 
Notes 
1 .  Note that other authors label or group responses to unfavourable social identities 
somewhat differently (for example, Taylor & Moghaddam, 1987). In preparing the present 
chapter, the authors have endeavoured to follow as closely as possible the conceptual 
organization offered in Tajf
el's early statements of SIT. 
2. Turner and Brown (1978) is a frequently cited paper on legitimacy and stability and their 
relationship to social competition and social creativity. However, in the present context, their 
measures are of limited relevance. The creativity indices concern the number of new possible 
comparison dimensions generated by the SUbjects. However, subjects did not actually make any 
ingroup or outgroup ratings on these new dimensions. Hence, the study's results do not bear 
directly on the use of new bases of comparison to establish favourable social identities. 
3. Note that Brown and Ross themselves do not emphasize this pattern in their results. In 
fact, they do not discuss it at all. However, the described pattern is very clear from examination 
of the data in their Tables 2 (p. 168) and 5 (p. 1 7 1 ). 


1 3  
Mere Categorization and Competition: A 
Closer Look at Social Identity Theory and 
the Discontinuity Effect 
Stephen M. Drigotas, Chester Insko and John Schopfer 
The word 'group' can have many meanings. It can be used to represent 
anything from face-to-face interaction to aggregates of unacquainted 
people. Its imprecision is further enhanced by the fact that one may be 
assigned to many groups (based on sex, race, nationality, family, and so on) 
without any feelings of belonging or particular affinity for any one of them. 
Because of this apparent amorphousness, researchers through the years have 
attempted to delineate when an aggregate of individuals becomes a group. 
Horwitz and Rabbie (1989) credit Lewin (1948) for early theoretical interest 
in perceived collectivity as the defining nature of groupness. Lewin main­
tained that interdependence of fate is the determinant of who belongs to 
which group, not simply degree of similarity-dissimilarity of features. 
Campbell (1958) called the shift from perceived aggregate to perceived 
collectivity 'entitativity' and suggested various Gestalt mechanisms for the 
perception of many as one. At this point we feel that it would be premature 
to settle on what precisely is meant by the term 'group'. 
The behaviour of individuals within groups can be as diverse as the 
considerations of what constitutes a group. Historically some social scien­
tists have focused on the supposedly animalistic nature of crowd behaviour 
as providing possible evidence for a group mind unfettered by moral 
constraints (cf. Allport, 1924; Durkheim, 1 898; LeBon, 1 896; McDougall, 
1920). On the other hand, after extensive observation McPhail (1991) argues 
that the notion of 'madding' crowd is a 'myth'. Still within contemporary 
social psychology there is evidence that under some circumstances 
individuals within groups behave in a manner that is less than praiseworthy. 
For example, individuals within groups are more likely socially to loaf 
(Latane, Williams & Harkins, 1979), give more shocks to a 'learner' 
(Zimbardo, 1969), and not help in an emergency situation (Latane & 
Darley, 1970). 
Our approach to the general problem of group behaviour has been 
specifically to compare interindividual behaviour with intergroup behaviour. 
Consistent with Roger Brown's (1954) suggestion of a discontinuity between 
individual and group behaviour, we have adopted the term 'discontinuity' 


Mere categorization and competition 
181 
for the tendency of intergroup behaviour to be more competitive, or less 
cooperative, than interindividual behaviour. Within the context of certain 
matrix games there is now an abundance of evidence for a descriptively 
large discontinuity effect (for a review see Schopler & Insko, 1992). What is 
less clear is the reason for the discontinuity effect. 
I While different theorists 
have suggested divergent causes for the competitiveness of groups, the 
purpose of the present chapter is to compare and contrast two of the major 
approaches to understanding the phenomenon: social identity theory 
(Tajfel, 1970b, 1978a, 1978b, 1982c; Tajfel & Turner, 1979, 1986; Turner, 
198 1  
b) and the research tradition which has specifically investigated the 
discontinuity effect (Insko & Schopler, 1987; Insko, Pinkley, Harring, et aI., 
1987; Insko, Pinkley, Hoyle, et aI., 1987; Insko, Schopler, Hoyle, Dardis & 
Graetz, 1990; Insko et ai., 1988; Insko et aI., 1992, 1993; McCallum et ai., 
1985; Schopler & Insko, 1992; Schopler, Insko, Graetz, Drigotas & Smith, 
199 1 ;  Schopler et aI., 1993). 
Social identity theory 
An overview 
Tajfel's social identity theory (SIT) is based upon the assumption that we all 
have a need for favourable self-identity, and that the status of the groups to 
which we belong helps us achieve such positive self-identity. Moreover, 
Tajfel and his associates claim that such positive evaluations are made on a 
comparative basis - we nourish our positive self-identities when our group 
does better than another group (Tajfel, 1978b, 198 1 ,  1982c; Tajfel & Turner, 
1979, 1986; Turner, 1981b). Thus, intergroup competitiveness is directly 
linked to the maintenance of self-identity. The SIT theorists also claim that 
mere categorization of subjects into 'groups' results in the orientation 
towards own-category favouritism. An abundance of research (Billig & 
Tajfel, 1973; Tajfel & Billig, 1974; see Brewer, 1979 and Diehl, 1 990, for 
reviews) has demonstrated that when subjects are categorized into one of two 
mutually exclusive groups and respond to a non-correspondent allocation 
task, they typically make choices favouring an own-group member over an 
outgroup member - ostensibly in the pursuit of bolstering self-identity 
through the relative advantage for own group. 
A discontinuity perspective on social identity theory 
A serious problem inherent in social identity theory is its inability 
theoretically to account for the diff
erences between intergroup interaction 
and interindividual interaction. Research involving the discontinuity effect 
has long demonstrated that group interactions can in many circumstances 
be more competitive than individual interactions. However, as noted 
elsewhere (Insko & Schopler, 1 987; Schopler & Insko, 1992), individuals 
should also receive positive bolstering of self-esteem from doing better than 


182 
Social identity: international pers
pectives 
another individual. But if this is true, why from a social identity theory 
perspective should individuals not be just as competitive as groups? It is true 
that social identity theory was developed to address intergroup relations. 
The point is, however, that there is nothing in social identity theory itself 
that indicates why the processes inherent in two-on-two relations should not 
also be present in one-on-one relations. If besting another group should 
bolster one's self-esteem and one's self-definition, the same should also be 
true of besting another individual. Why then, from a social identity per­
spective, are not interindividual relations just as competitive as intergroup 
relations? Furthermore, if it is true that individuals receive a self-definitional 
boost from besting another individual, by implication social identity theory 
does not adequately account for between-group competitiveness. Social 
identity theory's self-definitional premise is plausible, but does not by itself 
account for between-group competitiveness. The theory is not so much 
implausible as incomplete. 
The minimal group paradigm 
Investigations of social identity theory's prediction that mere categorization 
is sufficient to evoke relativistic social comparisons have used a minimal 
group paradigm in which subjects are categorized and then asked to allocate 
outcomes to own- and other-category members. The most common tech­
nique for categorization involves dividing subjects into sets, or groups, on the 
basis of their fabricated preferences for the paintings of Klee versus those of 
Kandinksy. Subjects are asked to rate pairs of Klee-Kandinsky slides and are 
then assigned to different rooms based upon their supposed preferences. 
Once in the separate rooms subjects are asked individually to select one 
column of payoff points from a matrix, and thus assign a payoff amount to 
an own-group member and an other-group member. While Tajfel and his 
associates use a somewhat complicated 'pull' procedure for scoring (see 
Turner, 1978a, for a complete description of the assessment procedure), only 
the four resultant orientations are discussed here. They include (a) maximum 
ingroup profit (MIP), (b) maximum difference in favour of the ingroup 
(MD), (c) fairness (F) and (d) maximum joint profit (MJP). It is the 
maximum difference choice that represents the greatest relative advantage 
and best exemplifies the need for ingroup favouritism. Their research has 
generally concluded that MD is the most powerful predictor of choice, 
although evidence for the other four choices does exist. In fact, Turner, 
Brown and Tajfel (1979) summarized their results as indicating that the MD 
choice is only non-significantly more influential than MIP and MJP. 
Nevertheless, the SIT theorists hold strong to their contention that between­
group competitiveness is the result of the need for positive social identity 
through the relative advantage of ingroup performance. 
As indicated below, there is a question whether or not mere categoriza­
tion is sufficient to produce relativistic social comparison. Beyond that there 
is a definitional matter as to whether aggregates of individuals should be 


Mere categorization and competition 
1 83 
defined as groups to the extent that they engage in relativistic social 
comparisons. One could interpret Tajfel and his associates as believing, first, 
that entitativity is created (on the independent variable side) by categor­
ization, and, second, that entitativity is indexed (on the dependent variable 
side) by relativistic social comparisons. While Tajfel and his associates do 
not use the term 'entitativity', they do believe that mere categorization does 
induce 'genuine awareness of membership in separate and distinct groups' 
(Tajfel, 1 978a, p. 35). 
The alternative allocation task experiments 
A series of experiments using the Multiple Alternative Matrices (MAMs) 
were conducted during the early eighties to assess directly the maximizing 
difference orientation exhibited within the minimal group paradigm. 
Bornstein et al. (1 983a) developed the procedure which enables the meas­
urement of seven different social orientations. The labels for the orientations, 
taken from the analyses of Kelley and Thibaut (1978), are differentiated 
based upon whether the orientations favour own group, other group or 
neither group. The orientations favouring own group include maximizing 
own gain (max. own), maximizing relative own gain (max. reI. own), and 
maximizing joint gain favouring own group (max. joint own). The orien­
tations favouring the other group include maximizing other's gain (max. 
other), maximizing other's relative gain (max. reI. other), and maximizing 
joint gain favouring other (max. joint other). Finally, the orientation 
favouring neither group minimizes the difference between group outcomes 
(min. diff.). 
The obvious advantage of this approach is its delineation of interpersonal 
motives. Each orientation represents a mutually exclusive motivation for 
interaction. Thus, the overlapping orientations problem inherent in the 
minimal group paradigm is easily resolved through use of the MAMs and a 
better test of SIT is possible. The max. reI. own choice best represents SIT's 
maximum difference (MD), max. own is similar to maximum ingroup profit 
(MIP), min. diff. is similar to fairness (F), and max. joint best represents 
maximum joint profit (MJP). 
Point values for own versus other group can be assigned to represent the 
seven orientations without motivational overlap. In other words, the rep­
resentations of the seven orientations are mutually exclusive as to the 
motivation for choosing them. Therefore, the specific predictions of SIT 
concerning categorization and self-esteem gained through maximizing 
relative advantage could be tested in an unconfounded manner through the 
use of the Multiple Alternative Matrices. 
A series of studies using the MAMs was conducted to test the viability of 
SIT in predicting actual motivational behaviour in the minimal group 
paradigm. Because the details of these studies have been presented in 
previous articles and a chapter (Insko & Schopler, 1 987), only the highlights 
are presented here. 


1 84 
Social identity: international pers
pectives 
Bornstein et al. (1983a) conducted a series of three experiments using the 
MAMs that represented a substantial challenge to social identity theory. In 
the first two studies, college students were given a cover story concerning 
their participation in an experiment as part of a larger investigation of 
artistic preferences. The rest of the experiment followed the Tajfelian pro­
cedure exactly except that subjects made monetary choices for one own­
group and outgroup member using the multiple alternative matrices. The 
results indicated that the subjects were not choosing max. reI. own more 
than would be expected by chance (across the seven orientations and 10 
trials), and they actually chose it significantly less than would be expected 
by chance. Recall that max. reI. own is analogous to MD, the main strategy 
believed to be evoked by mere categorization. Instead, subjects significantly 
tended to choose both max. joint own and min. diff. (with significantly 
more min. diff. for females than males). 
In the second experiment subjects were also asked to go back through the 
matrices and indicate a second preference on each matrix, and then a third 
preference. Subjects did not select max. reI. above chance for either the 
second or third preferences. Furthermore, the second experiment also 
included a condition in which subjects responded using the Tajfel matrices 
instead of the MAMs. Results using these matrices indicated support for the 
MD choice, as in previous studies using the less differentiated matrices. 
The third experiment was conducted using junior high students (9th and 
10th grade) from a middle-class US suburb. No cover story was used. 
Results again indicated no significant deviation from chance for max. reI. 
own. As in the first two experiments, min. diff. was selected significantly 
more than chance. However, unlike in the first two experiments, max. own 
was also selected significantly more than chance. 
Given the slightly different results for junior high students, Wittenbraker 
(1983) conducted an experiment using 5th, 9th and 12th grade students 
from the same school system used in the previously mentioned third study. 
He used the same mere categorization procedure except that coloured bar 
graphs were used to represent outcomes for Klee versus Kandinsky 
members. Once again, results indicated a significant effect for min. diff. 
There was also a replication of the max. own effect for 9th grade males. 
Finally, there was a significant effect for max. joint own for both 5th and 
12th graders. 
2 
Social identity theory's response 
In response to the Bornstein et al. series of experiments, Turner (1983a, 
1 983b) raised three objections. The first two were procedural in nature while 
the third constituted a major theoretical challenge. The two procedural 
objections were that the cover story distorted the results and that in the 
MAMs format subjects allocated money to just one own group and one 
outgroup. Two different studies (Bornstein et aI., 1983b) were conducted in 
response to these objections. The first compared the old cover story with a 


Mere categorization and competition 
1 85 
new cover story and a condition with no cover story at all. The results 
indicated no effect for the three-level manipulation of cover story. The 
second experiment contrasted a procedure using allocation to one own 
group and one other group to one using allocation to different own group 
and outgroup members. The results indicated non-significant effects for 
constancy of recipient. Furthermore, both studies again found significant 
effects for both min. diff. and max. joint own and a non-significant effect 
for max. reI. 
The major theoretical objection to the use of the MAMs concerned what 
Turner called the maximizing assumption. Turner maintained that the 
MAMs differentiate between social orientations only to the degree that 
individuals wish to maximize a value reflected by a particular orientation. 
Instead, subjects may make a specific choice not because of its social 
orientation as reflected by the maximizing assumption, but because of a 
different orientation altogether. For example, a subject may choose max. 
own not out of a desire to maximize own profits but instead out of a desire 
to win to a lesser extent than is possible through max. reI. own. Thus, the 
MAMs would reflect social values only to the degree that subjects adhere to 
the maximizing assumption. 
If one is able to cast doubt on whether subjects actually adhere to the 
maximizing assumption, reinterpretation of previous results using the 
MAMs allows for possible support for social identity theory. Recall that 
research with adult subjects found an abundance of evidence for max. joint 
own but no evidence for max. reI. own or max. own. Obviously, max. reI. 
own and max. own both involve advantaging of own group. But what 
about max. joint own? If the choice is interpreted to reflect a concern with 
both own-group and outgroup outcomes, the evidence of desire for 
superiority is ambiguous at best. However, Turner argues that subjects may 
choose max. joint own not out of concern for maximizing both own and 
other outcomes but out of a desire to win to a lesser extent than is possible 
in max. reI. own. Thus, if the maximizing assumption is rejected, both the 
max. reI. own and max. joint own choices support the social comparison 
prediction of social identity theory. 
Insko, Pinkley, Harring, et al. (1987) conducted a study designed to 
determine whether subjects actually invoked the maximizing assumption 
when choosing either max. own or max. joint own. An identical format was 
used to test the subjects except that the MAMs were presented one at a 
time. If and when a max. own or max. joint own was selected the subject 
was given a questionnaire containing an open-ended item and a closed­
ended item. The open-ended item simply asked subjects to state their reason 
for their immediately prior choice. The closed-ended item forced subjects to 
choose between the maximizing assumption and the Turner rationale. For 
max. own these items were 'to gain the most amount of money possible' 
and 'to win by less than the biggest margin'. For max. joint own these items 
were 'to gain the highest combined amount of money for everyone, with my 
group receiving slightly more', and 'to win by less than the biggest margin'. 


1 86 
Social identity: international perspectives 
The results for both the open-ended and closed-ended items indicated 
overwhelming support for the maximizing assumption. For example, 1 8  of 
the 21 subjects who chose max. joint own had open-ended responses 
consistent with the maximizing assumption (p < .0006 by binomial test). 
The MAMs versus the Ta
j
f
elian pull procedure 
If subjects do accept the maximizing assumption of the MAMs, why do the 
MAMs yield different results from the pull procedure? Currently, we do not 
have a definitive answer to this question. We can, however, point to two 
obvious differences in the assessment procedures. First, because the array of 
alternatives on the MAMs are unconfounded, it should be relatively clear to 
the subject that he or she is selecting among distinctive choices. 
Second, unlike the pull procedure, the MAMs simultaneously present 
subjects with the full array of possible preferences to be assessed. Thus a 
subject is able to make a choice with complete knowledge of all of the 
possibilities that are available. We find it plausible that in such a situation 
the subject's ability to indicate his or her clear preference will be optimized, 
that is, that the choice will be an informed choice. 
Consider, for example, a subject who selects max. reI. own on the 
MAMs. It should be apparent to the subject that the choice is clearly 
different from max. own, max. joint own, max. other, and so on. The 
subject should clearly understand that a max. reI. own choice is not a choice 
to maximize own category outcomes, not a choice to maximize joint 
outcomes, not a choice to maximize other category outcomes, but rather a 
choice to maximize the relative difference between own- and other-category 
outcomes. If it is advantageous to use an assessment procedure which 
maximizes the subject's understanding of what is being chosen, why should 
investigators continue to rely on the pull procedure? 
SIT summary 
Taken together, this series of experiments using the MAMs provides a very 
dissatisfying picture of the contention that mere categorization is enough to 
instil relativistic competition through social comparison mechanisms. First, 
five different studies using a variety of populations, including adults, found 
that mere categorization does not lead to maximizing relative choices when 
differentiated social orientation matrices are used instead of the traditional, 
motivationally confounded, Tajfelian matrices. Instead, subjects tend to 
choose max. joint own. Furthermore, research has shown that subjects do 
appear to use the maximizing assumption inherent in the interpretation of 
the MAMs. In other words, with closer inspection of social motives, mere 
categorization does not produce the relative advantage competition that 
Tajfel believed to characterize real groups. Given this revelation, the 
question then becomes: what exactly is needed beyond mere categorization 
to produce intergroup competition? 


Mere categorization and competition 
The discontinuity effect 
The possibility o
f
 schema-related belie
f
s
 concerning the 
appropriateness o
f
 competitive behaviour 
1 87 
As mentioned earlier, the discontinuity effect refers to the tendency for 
intergroup interactions to be more competitive than interindividual inter­
actions. Within the domain of investigated matrices, the effect appears to be 
robust and reliable. Moreover, research concerning the discontinuity effect 
has helped identify what is needed beyond mere categorization to produce 
intergroup competition. We specifically hypothesize that all of the evidence 
triangulates on the suggestion that competition flows from circumstances 
implying the believed appropriateness of competition. We, furthermore, 
speculate that a set of such beliefs in the appropriateness of competition 
constitute a schema. By a schema we mean a set of beliefs serving to organize 
and guide memory for past events and expectations regarding future events. 
In postulating such a schema we are clearly going beyond the evidence. More 
definitive evidence for a schema would be provided by data indicating that 
perceived effects are magnified or altered in memory. We have no such data; 
rather, we have data which suggest that experienced competitiveness can 
produce further competitiveness - even when the further competitiveness 
involves a different opponent. 
Competitive experience with one opponent and subsequent 
competitive relations with another opponent 
Insko, Pinkley, Harring, et ai. (1987) conducted an experiment comparing 
intergroup interactions versus interindividual interactions in a Prisoner's 
Dilemma Game (PDG; see Figure 13.1) followed by the usual mere cat­
egorization procedure. Previous research using the PDG had indicated that 
groups behaved competitively while individuals behaved cooperatively in 
the identical setting (McCallum et aI., 1985). Thus, the behaviour of 
subjects who experienced competition (in group-on-group interaction) or 
did not experience competition (in one-on-one interaction) on a previous 
task could be compared in the mere categorization task. It was hypothesized 
that the prior experience would carry over to the categorization task, 
resulting in increased levels of competitiveness (max. reI. own or max. own). 
The results indicated that there was more competition between groups than 
between individuals with the PDG and that this behaviour carried over to 
the mere categorization procedure in the form of increased max. own (but 
not max. reI.) on the MAMs. 
Reciprocation of competitive responses, or escalation of conflict with the 
same opponent, has been documented by others (for example, Kelley & 
Stahelski, 1970). It is important to realize, however, that the Insko, Pinkley, 
Harring, et ai. result occurred despite the fact that the subjects were led to 
believe that the subject composition of the Klee and Kandinsky categories 
in the mere categorization situation was not necessarily the same as the 


188 
Social identity: international pers
pectives 
Cooperation 
Competition 
Cooperation 
Competition 
$0.57 
$0.66 
$0.39 
$0.66 
Figure 13.1 
A Prisoner's Dilemma Game matrix (PDG) 
subject composition of the groups in the PDG situation, and thus that the 
obtained results were not a simple result of reciprocation. Why, then, did 
competitive behaviour carry over from one situation to the other? One 
possibility is that there was a perseveration of competitive behaviour. We 
believe, however, that such an explanation is at best an oversimplification. 
Rather we interpret the results as suggesting that subjects shifted their 
beliefs regarding what was appropriate or normative. As a result of prior 
competitive group-on-group interaction, subjects believed that when sub­
sequently categorized in a different group it was appropriate to behave 
competitively towards another category or group. 
To some readers it may appear that in postulating that competitive 
experience results in a change in beliefs regarding the appropriateness of 
competition we are going beyond the evidence, and indeed we are. Why do 
we need to postulate anything other than simple perseveration of com­
petitive behaviour? There are two answers to this question. First, the reader 
should note that there is a wealth of published research indicating that 
behaviour change can result in attitude and belief change (for a review of 
this literature see Chapter 1 1  in Eagly & Chaiken, 1993). In addition the 
evidence for this linkage is particularly strong in those situations in which 
the choice to engage in the behaviour is 'free' or apparently unconstrained, 
and that appears to be the case in the above-described research. Second, the 
readers should note that competitive experience also involves competitive­
ness in the opponent, and a series of studies by Insko et aI. (1 992) relating 
to max. reI. indicates that it is an opponent's competitiveness which is 
particularly important in producing subsequent competitive behaviour. 
The design of the Insko et aI. (1992) studies used a three-choice matrix 
(labelled a PDR matrix; see Figure 1 3.2) which reflected three different 
social orientations: max. joint, max. own and max. reI. The max. own 
choice on this matrix is suggested by Campbell's (1965) realistic conflict 
theory, which implies that the dominant social motive in intergroup conflict 
is maximizing 'real' outcomes. The max. reI. choice obviously reflects the 


Max. joint 
Max. own 
Max. reI. 
Mere categorization and competition 
Max. joint 
Max. own 
$0.99 
$1 44 
$1 . 1 4  
Confounded 
max. rei. 
$1 .44 
$1 . 1 4  
$0.84 
$0.84 
$1 44 
Max. rei. 
$0.84 
$0.69 
$0.99 
$1 . 1 4  
Confounded 
max. reI. 
$1 44 
$0.84 
$1 . 1 4  
$0.99 
$0.99 
$0.99 
1 89 
Note: The third choice in the PDR matrix is max. rei. The third choice in the PDC 
matrix is confounded max. reI., which is max. rei. that is confounded with max. own. 
Figure 13.2 A Prisoner's Dilemma max. reI. matrix (PDR) and a 
Prisoner's Dilemma con
f
ounded matrix (PDC) 
Tajfelian notion of maximizing relative advantage to bolster social identity. 
The results of the first study supported both Campbell and Tajfel's theory 
concerning group behaviour: relative to individuals, groups made both 
more max. own and max. reI. choices. There was some hint, however, that 
the max. reI. behaviour only occurred after the completion of the practice 
trials, which were played for points instead of money against the other 
group (or individual). 
One possible reason for this initial reluctance to choose max. reI. might 
relate to tactical considerations. Some groups may try tactically to 'set up' 
the other group by not choosing max. reI. until the later trials. Therefore, in 
experiment 2 we eliminated this confound by having subjects play their 
practice trials against an experimenter, and invoked a procedure whereby 
each group (or individual) played one trial with each of three other groups 
(or individuals). This procedure was designed to eliminate possible tactical 
considerations because subjects knew they would only interact once with 
each other group (or individual). The procedure had the added benefit of 


190 
Social identity: international pers
pectives 
Table 13.1 
Max. rei. as a f
unction o
f
 matrix type (PDR versus PDC), 
trials and groups versus individuals, experiment 2 
Trial I 
Trial 2 
Trial 3 
Groups 
Individuals 
Groups 
Individuals 
Groups 
Individuals 
PDR 
PDC 
0.08 
0. 12 
0.08 
0.63 
0.17 
1 .38 
0.18 
0.28 
0.50 
1 .06 
0.00 
0.27 
Note: Because four groups (or individuals) interacted on any given trial, max. reI. scores have a 
potential range from 0 to 4. 
eliminating a possible interpretation of the results as being due to recipro­
cation of competition, or the escalation of conflict with an opponent. 
It is likely that the max. reI. that occurred after the practice trials reflects 
the fact that a number of interactions are needed to instil the belief that 
competition is appropriate within the context of the PDR matrix. A further 
complication is that subjects may have been initially unwilling to choose 
max. reI. out of a reluctance to reduce own outcomes. The max. own choice 
produced higher own outcomes than the max. reI. choice (see the PDR 
matrix in Figure 1 3.2). Experiment 2 addressed this problem by using, in 
addition to the PDR matrix, a new matrix, the PDC matrix. On the PDC 
matrix own outcomes are identical (or confounded) for the max. own and 
max. reI. choices (see Figure 13.2). Thus, for this matrix the selection of 
max. reI. imposed no penalty on own outcomes. 
The results, once again, supported both Campbell and Tajfel: groups 
selected max. own and max. reI. more often than individuals. However, 
there were significant max. reI. sequence effects for both the PDR and PDC 
matrices (see Table 1 3.1). On the PDC matrix, the tendency for groups to 
choose max. reI. more than individuals was non-existent on the first trial, 
increased on the second trial, and then levelled off on the third trial. On the 
PDR matrix, the max. reI. effect was non-existent on the first and second 
trials, but was there on the third trial. There were no sequence effects for 
max. own on either matrix. There was a cooperation effect inverse of that 
for max. reI. for both matrices, but this effect can be regarded as a 
consequence of the shifts in max. reI. 
The findings for both studies are consistent. Max. reI. selections did not 
commence until some experience with the matrices had occurred. We should 
note here that, at least in study 2, this could not have been the result of 
tactical considerations that arise when repeatedly facing another group. 
These effects occurred even when subjects knew they would face each other 
group only once. 
It is also important to recognize that the increased competitiveness cannot 
be the result of simple reciprocation of conflict. Again, subjects knew they 
would face each other group only once. Therefore, there was no chance for 
retaliation against a group that competed. Reciprocation and escalation of 
conflict seem unlikely explanations of the competition sequence effects. 


Mere categorization and competition 
191 
A further analysis of the tria1-by-trial patterns in the data indicated that 
the tendency to chose max. reI. (confounded with max. own) on the PDC 
matrix on the second trial occurred most obviously for those groups whose 
opponents on the first trial had chosen max. own. Having been hit with a 
competitive choice from the first opponent, groups, but not individuals, hit 
their second opponent with a more competitive (but not own outcome 
reducing) max. reI. choice. On the other hand, the tendency to chose pure 
max. reI. on the PDR matrix on the third trial occurred for those groups 
who had engaged in mutual max. own responses with their second trial 
opponents. Having exchanged mutual max. own choices with their second 
opponent, groups, not individuals, hit their third opponent with pure max. 
reI. (max. reI. that did reduce own outcomes and was thus unconfounded 
with max. own). Such findings are consistent with the general supposition 
that prior competitive experience can activate a schema regarding the 
believed appropriateness of a highly competitive, max. reI. response. The 
evidence suggests that after groups engage in mutual max. own responses 
they 'reach for a bigger gun' (pure max, reI.) to use against their next 
opponent. At that point groups are willing to reduce their own outcomes 
just to increase the likelihood of doing relatively better than their new 
opponent. 
Recall the above-described Insko, Pinkley, Harring, et aI. (1987) experi­
ment in which participants interacted first in the context of the PDG and 
second in the mere categorization situation. In that experiment there was 
the possibility that the max. own choices in the mere categorization 
situation were a simple result of response perseveration. Note, however, that 
in the present experiments (Insko et aI., 1992) the response-perseveration 
possibility can be ruled out for the simple reason that the max. reI. choice 
did not follow prior own max. reI. choice. For the PDC matrix the max. reI. 
response occurred following the max. own choice by the opponent., and for 
the PDR matrix the max. reI. response occurred following mutual max. own 
responses. The key thus appears to be the experience of competitiveness by 
the opponent. 
The discontinuity research paradigm 
Studies investigating the discontinuity effect are typically laboratory experi­
ments that observe interaction using a Prisoner's Dilemma outcome matrix 
(or some variation thereof). The research is usually conducted in a 
laboratory suite in which same-sex individuals or groups are located in 
different rooms connected to a central room. After extensive training 
concerning the interdependent nature of the PDG, a trial or series of trials is 
conducted. Subjects are given time to examine a version of the PDG, then 
have some time to discuss possible action with the other group (often face to 
face with representatives but sometimes over a communication system), and 
then return to own-group discussion and actual decision-making. In group 
conditions the decision is typically made by the group as a whole or based 


1 92 
Social identity: international perspectives 
upon a majority vote. The decisions are then collected and the subjects paid 
accordingly. The discontinuity effect has been shown across a large number 
of studies regardless of: number of trials; size of payoff; gender; intergroup 
contact using representatives or all group members; whether the low values 
in the matrices are positive or negative; the type of matrices used (PDG, 
PDG-Alt., Mutual Fate Control); and whether there is equal or non-equal 
power. 
Groups are not always more competitive than individuals 
A further issue that merits comment relates to whether groups are 'always' 
more competitive than individuals. Clearly in many instances groups are 
cooperative and individuals are competitive and it would be foolish to 
maintain otherwise. Even though the above point has been repeatedly made 
(for example, Insko et aI., 1993; Schopler & Insko, 1992), it is sometimes 
asserted that we believe otherwise. For example, Rabbie and Lodewijkx 
(1994) claim that 'Schopler and Insko (1992) have concluded, based on their 
research with the PDG, that groups are invariably more competitive and 
less cooperative than individuals' (p. 1 58). Actually, Schopler and Insko 
(1992) and Insko et al. (1993) argued that Kelley and Thibaut's (1978) index 
of correspondence provides a basis for predicting when groups will be more 
competitive than individuals. Kelley and Thibaut's index of correspondence, 
in a situation in which the range of outcomes for the two players are 
the same, is the correlation between the outcomes across the cells. With the 
PDG and with other matrices that we have investigated, this correlation 
is moderately negative. Clearly, one would not expect discontinuity with 
matrices involving totally correspondent outcomes (+ 1 correlation). With 
such matrices both groups and individuals should coordinate their responses 
and there thus would be no discontinuity effect. Likewise in a zero-sum 
situation with totally non-correspondent outcomes (-1 correlation) both 
individuals and groups have no alternative to competing, and there would 
again be no basis for a discontinuity effect. Given this theoretical argument, 
the empirical challenge is to determine whether the circumstances in which 
groups are more competitive than individuals are, indeed, situations with 
somewhat, but not totally, non-correspondent outcomes. This is an issue 
that is currently being investigated. 
A further consideration that is relevant to the issue as to whether groups 
are always more competitive than individuals is the fact that our investi­
gations of the discontinuity effect have almost always been in situations in 
which individuals and groups communicate with each other prior to making 
simultaneous and irrevocable choices. Typically the communication has 
been face-to-face communication, but sometimes the communication has 
been via an intercom. Furthermore, Insko et al. (1993) obtained evidence 
indicating that in one-trial situations the absence of communication via an 
intercom produced a larger decrease in cooperation between individuals 
than between groups. Further evidence indicated that the presence of 


Mere categorization and competition 
193 
communication produced a greater decrease in expected competItIOn 
between individuals than between groups. Such evidence is consistent with 
the possibility, as discussed below, that there is an outgroup schema leading 
groups not to trust each other and thus for communication between groups 
to be of relatively low credibility. All of which is to suggest that without 
some form of communication the discontinuity effect should be markedly 
smaller and may, under some circumstances, be non-existent. Typically, of 
course, relations between individuals and relations between groups do occur 
in the context of some form of communication. 
The theoretical basis f
or discontinuity: ingroup (greed) and outgroup 
(f
ear) schemata 
As mentioned earlier, research investigating the discontinuity effect allows 
for careful consideration of the roots and nature of competition between 
groups. Previously mentioned research has provided evidence for the link 
between the belief in the appropriateness of competition (group schema) 
and competitive behaviour. Actually there is reason to believe that the 
schema concerning the believed appropriateness of competition can be 
subdivided into two schemata, the ingroup, or greed, schema and the 
outgroup, or fear, schema. At this point we would like to ruminate more 
extensively on these two theoretical bases for the discontinuity effect, 
The ingroup (greed) schema 
If a set of beliefs regarding the appropriate­
ness of competition exist, it is reasonable to suppose that some of those 
beliefs relate to what is appropriate and expected within the ingroup. But 
what might those beliefs be? One possibility relates to the beliefs that are 
postulated by the altruistic rationalization hypothesis. According to this 
hypothesis, group members assume that they are able to rationalize their 
competitiveness in intergroup interactions by believing that such behaviour 
is for the benefit of in 
group members. Thus, such competition is altruistic in 
regards to other members of the ingroup. The hypothesis bears some 
resemblance to Snyder's concept of attributional ambiguity (Snyder, Kleck, 
Strenta & Mentzer, 1979; Snyder & Wicklund, 198 1). Snyder maintains that 
self-interested behaviour is more likely to occur when there is attributional 
ambiguity as to whether the behaviour is self-interested in nature. Unfor­
tunately, there is evidence that when subjects are explicitly given the 
opportunity to share outcomes, and thus make altruistic rationalizations, 
they do not become more competitive. Insko, Pinkley, Hoyle, et al. (1987) 
conducted two studies employing a condition in which individuals were 
bound together by outcome interdependence. Within this condition, there 
were three simultaneous dyad interactions, with all subjects having a separ­
ate room, three on each side of the centre hall. However, subjects were told 
that, while they would interact alone with an individual on the other side 
of the hall, their money would be divided equally with the other two people 
on the same side of the hall. According to the altruistic rationalization 


1 94 
Social identity: international pers
pectives 
hypothesis, this outcome interdependence condition should have created 
more competition than the individual interaction condition. However, the 
means for the two conditions were virtually identical, thus making the 
altruistic rationalization hypothesis implausible. 
A more compelling argument for the creation of a competition-allowing 
ingroup schema revolves around the social support f
or shared self-interest 
that is available for groups but not for individuals. A belief in the avail­
ability of within-group support is quite likely based on a wealth of within­
group experience. We thus postulate that a history of mutual support for 
short-term self-interest has produced an ingroup schema that supports 
greedy behaviour. Individuals, by definition, are not afforded this support 
and therefore, apart from fantasised social support, are less likely to develop 
this schema. In characterizing social support for shared self-interest as a 
schema we are explicitly postulating that individuals in groups expect that 
they will be supported when they suggest a course of action consistent with 
the group's self-interest. Beyond this, however, we also postulate that greed­
based competitiveness is most likely in a situation in which the other group 
is expected to be cooperative and is therefore vulnerable. (Combativeness 
that occurs in a context in which the other group is expected to be com­
petitive and is therefore dangerous is more reasonably attributed to fear 
than to greed.) Schopler et al.'s (1993) study 2 was designed to demonstrate 
the resultant competitive behaviour when there is social support for self­
interest, and an expectation that the other group will be cooperative. 
Confederates were placed in each group and told to give a single suggestion 
as to what the group should choose in a standard PDG. The confederates 
were told to suggest to the other two subjects either that they should choose 
X (the cooperative choice) or Z (the competitive choice). The use of 
confederates also allowed for manipulation of the feedback concerning what 
the other group actually chose. By selecting the confederates as represen­
tatives it was possible to give the subjects false feedback that the other 
group was either choosing all X, all Z, or a 50-50 mixture of the two. The 
feedback manipulation was conceived as a manipulation of the other 
group's expected behaviour. The prediction was of an interaction between 
the feedback and suggestion factors such that only with cooperative 
feedback would the cooperative suggestion succeed in reducing intergroup 
competitiveness. Within the context of the repeated cooperative feedback 
where the other group is perceived as vulnerable, the cooperative suggestion 
theoretically should have reduced the within-group social support for 
combativeness, and thus reduced competitiveness in this condition relative 
to all the other conditions. The predicted pattern of results occurred. 
The outgroup (f
ear) schema While social support may foster greed through 
the bolstering of a competition-allowing ingroup schema, belief that the 
other group will be competitive is also expected to create competition within 
the PDG paradigm. Competition as the result of a competitive outgroup 
schema has been termed the 'fear' hypothesis. 


Mere categorization and competition 
Cooperation 
Withdrawal 
Competition 
Cooperation 
Withdrawal 
Competition 
Figure 1 3.3 
A Prisoner's Dilemma Game alternative matrix (PDG-Alt. ) 
195 
There is some evidence that individuals do maintain beliefs describing 
other groups as competitive, aggressive and dishonest in nature. Doise 
(1969) reported that before any interaction subjects attribute fewer coopera­
tive motives to the opposing group than to themselves or their own group 
members. Hoyle, Pinkley and Insko (1989) found that the source of this 
negative expectancy was not whether one was alone or in a group, but 
whether the opponent was a group or an individual. The expectations for 
the outgroup were significantly more competitive and aggressive regardless 
of whether one was alone or in a group. We do not have definitive evidence 
that such expectations, or beliefs, constitute a schema, but we are explicitly 
postulating that they do. 
There is evidence relating the postulated outgroup schema to the discon­
tinuity effect itself. This evidence comes from research using a new three­
choice matrix referred to as a 
Dilemma Game (PDG-Alt.) matrix (see 
Figure 1 3.3). The PDG-Alt. matrix includes a third choice in addition to the 
cooperative and competitive choices of the PDG matrix. This new choice 
guarantees intermediate outcomes regardless of the opponent's choice. This 
new choice, labelled 'withdrawal', is thus a safe choice that should be 
rationally preferred when there is fear of the opponent's competitive intent. 
Insko et al.'s (1990) second study used the PDG-Alt. matrix. The results 
indicated a discontinuity effect for both competition and withdrawal. Not 
only did groups compete and withdraw more than individuals, they com­
peted and withdrew to approximately the same extent. Thus, the discon­
tinuity effect cannot be explained by greed alone. The negativity of the 
outgroup schema (that is, the expectation that another group IS more 
dangerous than another individual) also affects group behaviour. 


196 
Social identity: international perspectives 
What is not clear, however, is the basis for this negative outgroup schema. 
One could suggest a number of hypotheses regarding its development. The 
first could simply be experience. We may develop general negative outgroup 
schemata based upon negative experiences with specific outgroups. Such 
experiences may cloud interactions with any group, just as the competitive 
PDG experience made competition more likely in the mere categorization 
paradigm, and just as max. reI. did not emerge until after a few trials using 
the PDR and PDC matrices. Consistent with this speculation, Pemberton, 
Insko and Schopler (1997) obtained evidence indicating that intergroup 
relations are recalled as more competitive than interindividual relations, and 
also that intergroup relations are experienced as more competitive than 
interindividual relations. The experiential data were obtained from 'diaries' 
that subjects carried with them for seven days (Reis & Wheeler, 1991). 
A second hypothesis concerns the role of false consensus. This negative 
view of outgroups in general may stem from our knowledge of our own 
competitive tendencies within groups. Thus, we may only need to look at 
ourselves and by projection develop a prediction of how the other group 
may behave. Further research should expand our understanding of these 
matters. 
Regardless of how the negative outgroup schema develops, its effects on 
behaviour are predictable. When the only choice is between cooperation 
and competition, as in the standard PDG, groups lash out in fear. It may be 
that the root of many group behaviours that are intensely competitive in 
nature are simply the result of fear of the other group and not a reflection of 
ingroup greed. This is certainly apparent when there is the chance to 
withdraw from the situation with certainty of outcomes, as in the PDG-Alt. 
In such situations groups tend to withdraw as often as they attack. 
Reducing discontinuity through recategorization: preliminary results 
There is an 'older' literature and a 'newer' literature on the overcoming of 
ingroup ethnocentricism. The 'older' literature tended to specify remedial 
interventions in terms of interactive behaviour between two groups. There 
were such global recommendations as joint interactions to attain super­
ordinate goals (Sherif, Harvey, White, Hood & Sherif, 1961), and there 
were the specific behavioural rules contained in Osgood's (1962) Graduated 
and Reciprocated Initiatives in Tension-reduction (GRIT). The 'newer' 
literature focuses on the cognitive mechanisms presumed to be antecedents 
to intergroup bias, and seems to vary with respect to the directness with 
which the intervention affects relevant cognitions. The most direct methods 
are those that degrade the categorized representation by such methods as 
recategorization (Gaertner, Dovidio, Anastasio, Bachman & Rust, 1993; 
Gaertner, Mann, Murrell & Dovidio, 1989) or criss-crossing category 
membership (Brewer, Ho, Lee & Miller, 1987). With the exception of 
modelling GRIT (Lindskold, 1983), both traditions tested for a reduction in 
ethnocentrism with measures of individual perceptions of ingroup and 


Mere categorization and competition 
197 
outgroup members. To our knowledge no one has attempted to assess 
whether the intervention was successful in increasing such behaviours as 
intergroup cooperation. We believe the ultimate effectiveness of any tech­
nique resides in its ability to create intergroup cooperation between intact 
groups. 
We have pilot tested some recategorization procedures that, in terms of 
reduced intergroup conflict, were spectacular failures. The pilot studies 
began with the formation of four three-person groups, who were sub­
sequently combined into two rivalrous six-person groups or, in a different 
version, two, independent, six-person groups each working cooperatively on 
a task. The recategorization worked well, in the sense of increasing liking 
for former outgroup members and creating satisfactory working relations 
with them. But, when we regrouped participants into their initial three­
person groups and asked them to interact with previous group members on 
a PDG, their rate of competition was just as large as the rate for groups 
whose members were never recategorized. All of which makes us suspect 
that if at the end of camp Sherif and his associates (Sherif et aI., 1961) had 
put the Eagles and Rattlers into a mixed-motive situation, they would have 
been just as competitive as they were in the tournament. Quite possibly 
Sherif and others have underestimated the power of the social setting. 
Conclusions 
That individuals gain self-definition and bolster self-esteem by belonging to 
groups is certainly a reasonable working hypothesis. Furthermore, we find it 
plausible that the need to support high self-esteem through the demon­
strated superiority of one's own group may be at least a partial explanation 
for between-group competitiveness. This main tenet of social identity theory 
has not been questioned. However, the belief that mere categorization into 
groups produces such intergroup competitiveness, a second tenet of the 
theory, has been questioned. An abundance of research has demonstrated 
that when more differentiated social orientations are used, mere 
categorization does not produce the relative competitiveness predicted by 
social identity theory. 
Instead, research detailing the discontinuity effect has suggested that what 
is need to produce the intergroup competitiveness beyond mere categoriza­
tion is a belief that competition is appropriate. There is evidence that 
competitive behaviour increases following competitive experience. This is 
true even when the opponent is a group other than the one with whom the 
prior competitive experience occurred. Such evidence is consistent with the 
possibility that beliefs in the appropriateness of competition can produce 
competition between categories. We have speculated that such beliefs flow 
both from an ingroup schema relating to expected social support for self­
interested, or greedy, behaviour, and from an outgroup schema relating to 
the expected competitiveness, and dangerousness, of an outgroup. 


198 
Social identity: international pers
pectives 
Notes 
Work on this chapter was supported by a National Institute of Mental Health Grant 
(5ROI MH53258) to the last two authors. 
I .  Given the existence of difference in the interaction of two groups and the interaction of 
two individuals, we are inclined to define aggregates of people as groups whenever behaviour 
between these aggregates is different from behaviour between individuals. However, in order to 
avoid circularity it is necessary to use the term 'group' in two different senses. Groups in the 
sense of aggregates, or sets of individuals, is assumed in the definition of discontinuity. The 
existence of discontinuity, however, provides evidence for the existence of groups in the more 
psychological meaning of the term. 
2. In seeming conflict with the above results, Platow, McClintock and Liebrand (1990) 
found that categorized subjects would select max. reI. on the Bornstein et al. (1983a, 1983b) 
matrices. Two differences between this experiment and all, or much, of the above research are 
that Platow et al. used points and not money for outcomes, and did not have a cover story. The 
combination of points and the lack of a non-game cover story may have led some subjects to 
interpret the situation as game-like, and thus to attribute relative value to the points. We 
believe that researchers should be careful not to create a game-like atmosphere and should use 
outcomes that have value outside of the experimental context. 


PART IV 
THE IMPACT OF 
CULTURE ON IDENTITY 
AND CATEGORIZATION 
1 4  
Discrimination and Beliefs on 
Discrimination in Individualists and 
Collectivists 
J. Francisco Morales, Mercedes L6pez-Saez and Laura Vega 
The concept of individualism 
Individualism refers not to the individual as an empirical subject or 
'individual specimen of the human species', but rather to the individual as a 
'moral, independent, autonomous being, . . .  
bearer of our supreme values'. 
In this second sense, individualism is the prevailing contemporaneous 
ideology of person and society (see Dumont, 1987, pp. 37ff.). 
As opposed to individualism, which ascribes to the individual the main 
value in society, for holism the supreme value lies in society as a whole. 
Historically, individualism appears in holistic societies as a source of 
opposition to them and as a sort of supplement. Here, the Indian 'renouncer' 
is an important landmark. In order to achieve the utmost truth, the 
renouncer relinquishes social life and its limitations and devotes himself to 
his own destiny. Similarly, in the Hellenistic period, against the self-sufficient 
polis of 
Plato and Aristotle, the idea was postulated that in order to become a 
wise person it was necessary to renounce the world. At that time, non­
learned people were considered as victims of the worldly life. 
From the beginning of the modem era, society is usually conceived as an 
'association', which suggests that the individuals who belong to it have 
associated with each other 'voluntarily', and thus they are prior to the groups 
and/or the relations they establish among themselves. This conception entails 


200 
Social identity: international perspectives 
two characteristic features: equality and freedom. There is a belief that one's 
own selfish interest is the best guarantee of social order and of general 
satisfaction for all members of society. 
Nevertheless, quite often, the way many people have used freedom has 
prevented others from achieving their own. Here is the central 'paradox' of 
individualism, according to Camps. Ethics must necessarily be individua­
listic, because it has to preserve the individual. But it must be achieved in 
such a way that it protects everyone at the same time, not some at the 
expense of others. When this is not achieved, individualism is equivalent to 
'anti-ideology', insolidarity, insensitivity to unequalities, lack of interest in 
public affairs. And that is what happens to most developed societies, 'whose 
only expectation is the perpetuation of group' (see Camps, 1 993, pp. 14f.). 
Cultural aspects of individualism 
Hofstede (1980) reports the results of the HERMES project, which involved 
the study of employees from 40 different countries, all of them belonging to 
the same multinational business corporation. From the responses given by 
this sample of subjects to 14 work goals, this author identified a factor 
representing a dimension of individualism-collectivism. 
The individualism-collectivism factor loaded positively on the following 
goals, from highest to lowest loading: (a) enough free time for oneself and 
one's family; (b) freedom to follow one's own way at work; (c) having a job 
which provides a personal sense of performance; and (d) living in a place 
desirable to oneself and one's family. The goals with negative loadings, also 
in a decreasing order, were: (a) opportunities to learn new skills; (b) good 
physical working conditions for work; (c) use of skills and abilities at work; 
and (d) good marginal profits. According to Hofstede, goals with positive 
loadings have in common that they emphasize personal independence vis-a­
vis the organization. On the other hand, goals with negative loadings refer 
to things the organization can do for the individual. 
The factor analysis conducted by Hofstede has been called 'ecological', 
because the loading of each goal is the correlation coefficient between each 
country's factor score and its mean score in that goal. This correlation is 
calculated across all 40 countries. From factor scores an Individualism 
Index is calculated for each country, ranging from 0 to 100 due to a 
mathematical transformation. 
Psychological aspects of individualism 
Hui (1988) postulates that individualism is also an individual tendency or 
predisposition, since there is, within any culture, subcultural variation 
among individuals regarding this dimension. This author's work has two 
main characteristics. First, the development of a scale to measure the 


Discrimination and belie
f
s
 on discrimination 
201 
individualism-collectivism construct, named INDCOL, consisting of eight 
subscales. Thus, the construct is considered as multifaceted and is measured 
in relation to eight specific targets: spouse or partner, relatives, parents, 
neighbours, close friends, co-workers, acquaintances and siblings. 
Second, INDCOL measures beliefs, attitudes, behavioural intentions and 
behaviours. As Hui himself notes, this has both theoretical and methodo­
logical consequences. From a methodological standpoint, generality of 
measurement increases, although it is at the expense of a part of the scale's 
internal consistency. From a theoretical point of view, individualism­
collectivism is not considered as a unitary disposition, but rather as a 
syndrome. In fact, by adding up the scores of the eight subscales, the 
General Index of Collectivism is obtained. Hui defines this index as the 
degree to which the person believes the group, and not the individual, is the 
basic unit of survival. 
Based on INDCOL and other scales developed by himself, Triandis, 
Bontempo, Villareal, Asai and Lucca (1988) produced the Three Factor 
Questionnaire. They applied this questionnaire to samples from the United 
States and found, as did Hui, that individualism is a multifaceted construct, 
since it appears characterized by the following features: (a) priority given to 
one's own goals over those of the ingroup; (b) less attention paid to the 
views of the ingroup; (c) self-reliance with competition; (d) distance from 
ingroup; (e) making decisions on one's own before asking others' point of 
view; (f) less concern for ingroup. Therefore, in the event of a conflict 
between ingroup and individual, individualism will lead the person to do 
what he or she, and not the ingroup, considers appropriate. 
Correlates of individualism: self, relationship with in 
group and 
outgroup, emotional experience 
Besides the above-mentioned work by Hui (1988) and Triandis et al. (1988), 
many others have studied the differences in social behaviour between 
individualists and collectivists. Among the best known the following can be 
cited: Hui and Triandis (1986), Markus and Kitayama (J991a), Triandis 
(1 993), Triandis, Leung, Villareal and Clack (1985), Triandis et al. (1 986) 
and Triandis, McCusker and Hui (1990). Table 14.1 summarizes the main 
differences regarding self and relationship with ingroups and outgroups. 
In order to better understand the content of this table, two points must be 
noted. First, not all ingroups are equally important for collectivists. On the 
contrary, there is usually a preference or priority ranking among them. 
According to Triandis (1993), in many societies the order is (a) family, (b) 
nation, (c) religious group and (d) work group. However, in Japan the 
religious group does not exist, and the work group ranks higher. Together 
with the priority order of ingroups, we must take into account the existence 
of vertical collectivism and horizontal collectivism. The former implies 


202 
Social identity: international pers
pectives 
Table 14. 1 
Summary of some dif
f
erences between individualists and 
collectivists 
Diff
erence target 
Individualists 
Collectivists 
Content of self 
Individual diff
erences 
Social categories 
Way of achieving self-
'I can do whatever I want' 
'I am not a burden to my 
actualization 
group' 
Basic unit of survival (belief) 
Individual 
Group 
Regulation of behaviour 
Personal attitudes and cost-
Ingroup norms 
benefit analysis 
Personal goals vs ingroup 
Personal > ingroup 
Ingroup > personal 
goals 
Diff
erence between ingroup 
Weak 
Strong 
and outgroup 
Ingroup and outgroup 
Outgroup more 
Ingroup more homogeneous 
homogeneity 
homogeneous 
Kind of relations 
Horizontal 
Vertical 
submission to ingroup authority. The latter means a fusion with in 
group 
identity. Both kinds of collectivism may coexist, but the prevalence of one 
or the other depends on the type of society concerned. 
Markus and Kitayama (199Ia) designate individualist cultures as 'inde­
pendent self' cultures, and collectivist ones as 'interdependent self' cultures. 
According to these authors, there are important differences between both 
types of culture as far as emotional experience is concerned, since the 
fundamental emotions, their sustaining basis and the functions they fulfil 
do not coincide. Thus, in 'independent self' cultures, the most important 
emotions are ego-focused, specifically anger, frustration, pride, guilt and 
haughtiness, their basis consists on the subject's internal attributes, and their 
function is the expression of inner feelings. On the other hand, emotions in 
'interdependent self' cultures are focused on others. Among them shame 
and sympathy stand out, the basis of these being a special sensitivity to 
others, an ability to take the perspective of the other and an active search 
for interdependence. Their function is to express the outcome of 
interpersonal relationships. 
Individualism and social identity: Hinkle and Brown's proposal 
As the various chapters in this book have shown, social identity theory 
(SIT) deals with processes of group comparison and identification, rela­
tionships between groups, the formation of psychological group, and the 
relative weight of personal and social identity, among many other processes. 
Results obtained by those who have studied the repercussions of the 
individualism-collectivism orientation, summarized in Table 14.2, lead us 
to expect a convergence between studies on individualism and those carried 
out within the framework of SIT. 


Discrimination and belie
f
s
 on discrimination 
Table 14.2 Relation between identification with the 
group and intergroup dif
f
erentiation as a f
unction of 
individualism/collectivism and autonomous/relational 
orientation 
Orientation 
Orientation 
Autonomous 
Relational 
Individualist 
.03 
.33* 
n = 17 
n = 24 
Collectivist 
.47** 
.79*** 
n = 25 
n = 1 8  
* p < . 10; ** p < .05; * * *  
p < .01 
203 
Hinkle and Brown (1990) were the first to pay attention to this con­
vergence. From a theoretical point of view, they state that processes of 
social identity studied by SIT will take place mainly in 'collectivist kind of 
groups' (p. 66), since it seems reasonable to expect only collectivist groups 
or individuals to worry about the outcome of comparisons between ingroup 
and outgroups. The general underlying idea is that a collectivist orientation 
seems important for the manifestation of social identity processes (Hinkle & 
Brown, p. 67). 
From an empirical point of view, Hinkle, Brown and Ely (1992) found 
that the relation between identification with the group and intergroup 
differentiation is significantly greater in collectivists than in individualists. 
This is a proof, they argue, that only collectivists are really concerned with 
the outcome of comparisons between ingroup and outgroups (See Table 
14.2). 
Individualism and identity theory: other perspectives 
SIT is also a theory of the formation of a psychological group. Since the 
earlier experiments using the minimal group paradigm, both Tajfel and 
Turner interpreted the discrimination taking place in that situation as 
evidence of group formation. Insko and Schopler (1987, p. 2 1 5) note that 
Tajfel considers the effect of mere categorization as a proof of entitativity, 
meaning the feeling of shared group membership. As early as 1 98 1 ,  Brown 
and Turner proposed using an interpersonal-group continuum instead of 
the famous interpersonal-intergroup continuum. The reason was, according 
to these authors, that 'the same processes may tend to distinguish both 
intragroup and intergroup processes of interpersonal relationships' (Brown 
& Turner, 198 1 ,  p. 40). In other words, intergroup differentiation implies 
group formation and vice versa. 
Nevertheless, research by Triandis and his co-workers suggests that one 
of the main differences between individualists and collectivists is the 


204 
Social identity: international perspectives 
relationship they maintain with their own groups. In principle, the idea 
must be rejected that individualists find it difficult to join a group. Even 
accepting the schematic picture of Table 14. 1, the very fact of having an 
ingroup and an outgroup implies group identification, as minimal as this 
may be. Even more, it would not be incorrect to predict a greater ease for 
individualists to form a group, having in mind that their differentiation vis­
a-vis outgroup people is weaker. For the same reason, it seems difficult for 
collectivists to be able to form a group with people not belonging formerly 
to any of their few ingroups. 
At this point we must mention an interesting result reported by Hinkle, 
Brown and Ely (1992, p. 107-108). For subjects showing an autonomous 
as well as individualist orientation, the correlation between identification 
and differentiation was .03, as can be seen in Table 14.2. However, these 
subjects did show positive discrimination in favour of their group. There­
fore, we tried to find out: (a) whether individualists show discrimination in 
the minimal group paradigm situation; (b) whether there are differences 
between them and collectivists in their showing intergroup discrimination. 
First experiment 
Subjects were 52 students from adult education courses, aged between 20 
and 60, and being of a low socio-economic level. All of them had previously 
completed the Spanish version of Triandis et al.'s (1988) Three Factor 
Questionnaire, validated in this country. Two groups of 26 subjects were 
formed, each consisting of 1 3  subjects above and 1 3  subjects below the 
median of the questionnaire. Thus, both groups were homogeneous in this 
variable. One of the groups was put to the standard categorization pro­
cedure. This was the experimental or categorized group. The other one, not 
categorized, was the control group. Therefore, there were two variables -
categorization and individualism - each having two values. Subjects 
answered only once to the six Tajfel matrices and to the 10 Insko matrices. 
The MANOV 
A carried out on Tajfel's strategies showed that the 
experimental group (categorized) was more discriminatory than the control 
group (not categorized). Thus the experimental group was significantly 
higher in maximum ingroup profit (MIP) over maximum joint profit (M1P): 
6.17 vs I, F(I, 39) = 1 2.93, p < .01 .  It was also higher in maximum 
(ingroup) difference (MD): 5.29 vs -.58, F(I, 39) = 1 8.61, P < .01 and in 
MIP over fairness (F): 5.96 vs -.26, F(I, 39) = 16.96, P < .01 .  On the other 
hand, the control group was higher in fairness or equity: 7.84 vs 2.38, F(I, 
39) = 7.78, P < .01 .  
The MANOV 
A carried out on Insko's strategies yielded similar results. 
The experimental group was higher in max. reI. own: .86 vs - 1 .0 
I, F(I, 39) 
= 10.83, P = .06. The control group was higher in max. joint other: .25 vs 
-.93, F(I, 39) = 9. 1 5, p < .06, and also in min. diff.: 1 .83 vs .61, F(I, 39) 
-3.76, p < .06. 


Discrimination and belie
f
s
 on discrimination 
205 
No effects of individualism were found, nor interaction between cat­
egorization and individualism. This means that the classic effect of ingroup 
favouritism, first found by Tajfel and afterwards by many other researchers, 
appears again in our study, both for individualists and for collectivists in the 
two types of matrices. 
In order to find out whether there were any differences in the kind of 
discrimination showed by individualists and collectivists, within-cell analysis 
of the preferences in the use of different strategies were carried out, both in 
experimental and in control group. No significant differences appeared 
either. 
Second experiment 
Within-cell comparisons in the first experiment were based on a small 
number of subjects (13). We carried out a second experiment to increase 
that number. We decided not to include the control group, since the effect 
of categorization seemed beyond doubt, and focused only on the individual­
ism variable. Ninety-two subjects participated, of whom 47 were indi­
vidualists and 45 were collectivists. 
The procedure was similar: subjects completed the six Tajfel matrices and 
the 10 Insko ones, which provided the first measure of the dependent 
variable, named first choice. However, a change was introduced. After the 
booklet where they had marked their preferences had been taken away, 
subjects were given a second one where they were asked: (a) to complete the 
16  matrices again, recording their previous choice in each, but explaining 
why they had made it - since there were 1 6  different matrices, subjects had 
to provide 1 6  reasons; this second measure was called justified choice; (b) to 
record the choice they thought people of their group had made, ingroup 
choice; and (c) to record the choice they thought people of the other group 
had made, outgroup choice. 
Next, we will present the results obtained for each of these measures. But 
before we do so, we must stress that the effect of categorization, in the 
absence of a control group, is evident for every measure in the prevalence of 
discriminating over non-discriminating strategies. Thus, in the first choice 
only the fairness/equity strategy in Tajfel's matrices is higher than discrimi­
nating ones: 6.87. But both MIP over MJP (3.89), and MD (3.47), as well as 
MIP over fairness (3.01) are higher than MJP (.30) and MJP + MIP (.47). 
Similar results are found in the other three measures of the dependent 
variable. Insko's matrices show a similar pattern, with the exception of, 
only, the strategy max. joint own, which is as high as the discriminating 
strategies max. own and max. reI. own. 
Results regarding first choice 
The MANOV 
A carried out on Tajfel's matrices yielded significant values. 
The univariate Fs show that individualists are lower than collectivists in 


206 
Social identity: international pers
pectives 
MIP over fairness, 1 .60 vs 4.44, F(I ,  89) = 6.98, p < .01 ,  which is evidence 
of their lesser degree of discrimination. At the same time, they are higher in 
fairness, 7.69 vs 6.04, F (1, 89) 
= 2.00, p < . 1 6. MANOVA was not 
significant for Insko's matrices. 
However, in these latter matrices differences do appear when relative 
preferences of individualists and collectivists are compared by means of the 
Wilcoxon test. Thus, we find that individualists are different from collec­
tivists in that they prefer the strategy of fairness (min. diff.) over max. reI. 
own and max. joint own; also in their preference for max. joint own over 
max. own. Collectivists, on the other hand, choose max. reI. own more 
often than max. joint own. All this converges in a description of indi­
vidualists as more prone than collectivists to fairness and to joint and non­
discriminating strategies, bearing in mind that this differential tendency is 
found within a discrimination framework shared by both kinds of subjects, 
which explains the lack of differences in the MANOV 
A. 
Results regarding justified choice 
The MANOV 
A did not yield any significant differences, either in Tajfel's 
matrices or in Insko's ones. 
As we have mentioned, after subjects had completed the matrices, they 
were asked a number of questions. The first one was about the reason why 
they had chosen a particular strategy in allocating points between two 
persons. More specifically, the question was: 'Why did you choose that cell?' 
Each subject had to record the reason corresponding to each cell. Since 
these amounted to 1 6, each subject provided 1 6 reasons to explain his/her 
choice-allocating strategies. 
After analysing the reasons given by all subjects, we found they could be 
grouped into 10 main categories: 
1 
Fairness or equality 
34% 
2 
Equality to balance 
4.2% 
3 
Equality when it is the only choice not detrimental to one's 
group 
0.58% 
4 
Maximum ingroup profit 
24.8% 
5 
Maximum ingroup difference 
8.53% 
6 
Maximum ingroup joint profit or maximum joint own 
1 5.31% 
7 
Maximum joint other 
3.2% 
8 
Maximum outgroup difference 
0.77% 
9 
Others 
0.84% 
10 Chance 
7% 
Six of these reasons correspond exactly to the strategies of matrices. These 
are f
airness or equality (F) (min. diff.), maximum ingroup pro
fit (MIP) (max. 
own), maximum ingroup dif
f
erence (MD) (max. reI. own), maximum ingroup 
joint pro
fit (MJP) (max. joint own), maximum joint other (max. joint other) 
and maximum outgroup diff
erence (max. reI. other). 


Discrimination and belie
f
s
 on discrimination 
207 
These six strategies tend to be the most often used by subjects. In fact, a 
high percentage of them repeat the reason given for the first cell in the 
remaining 1 5, that is, does not show 'alternation' in its reasons. In these 
cases, which we have called 'Use of Single Strategy', or USS, subjects tend 
to use, almost always, some of these six reasons. 
On the other hand, reasons 2, 3 and 9, respectively categorized as 
'equality to balance', 'equality when it is the only choice not detrimental to 
one's group' and 'others', are subsidiary to the former ones. This means that 
they are used only in cases of 'alternation' of strategies, or SA, where 
subjects do not want to limit themselves to only one reason and resort to 
other complementary ones. 
Let us see an example. Subject 34 uses a perfect alternation of reasons. 
On eight occasions he employs reason 4 (MIP) (max. own), which is dis­
criminatory. In the remaining eight, he uses reason 2 (equality to balance), 
which is egalitarian. This subject's words are worth being cited. When 
he chooses reason 4, he justifies his choice as follows: 'To give points to the 
people having the same code as myself'. When he chooses reason 2, he 
states: 'In order not to increase too much the difference in favour of my 
group.' 
In the case of this subject, as well as in other similar ones, reason 2 is 
clearly subsidiary to reason 4. The subject wants to benefit his/her group 
more, but does not want to do so in an extreme or radical way. Therefore, 
this justice or equality pursued arises from a wish to compensate a previous 
discrimination. 
Reason 3 is even more subsidiary than reason 2. It is usually employed by 
subjects with a discriminatory orientation when they meet a Tajfel matrix in 
which they cannot benefit their group more than the opposite group. This 
is the case of subject 39, who shows a discriminatory pattern of reasons, 
choosing 1 5  times reason 4 (MIP) (max. own), and only once reason 3. In 
justifying his discriminating strategy, this subject states: 'To give more 
points to my group.' However, when choosing the equality strategy in 
Tajfe\'s last matrix, he says: 'To give as few points as possible to the other 
group.' Effectively, in this latter matrix it is not possible to discriminate in 
favour of one's own group. 
It is clear, then, that reason 3 justifies equality in a subsidiary way, as a 
sort of lesser evil to prevent one's own group from getting less points than 
the other. 
The last subsidiary reason is 9, which we have called 'others'. Sometimes, 
subjects look for justice allocating one group more points in one cell and 
less in the next one, since they think that a strict equality strategy will be 
detrimental to both groups in rendering their whole absolute profit lower. 
Thus, subject 1 8  states: 'I have tried to balance the two groups. Since 
compensation is sometimes difficult, it is rather a random criterion.' 
Reason 1 0, 'chance', embodies those reasons not related to the strategy 
chosen. Five subjects used only this reason and were removed from the 
analyses, as was another who chose it most of the times. 


208 
Social identity: international pers
pectives 
Nevertheless, the great majority of reasons (82.64%) were distributed 
among the four main categories, which closely coincided with the main 
allocating strategies. These four reasons are fairness (34%), maximum 
ingroup profit (24.8%), maximum difference (8.53%) and maximum joint 
profit (1 5.3%). 
Although each subject had to justify 1 6 different choices, a thorough 
analysis of the pattern of justifications showed that most subjects used only 
one reason for all 1 6  occasions or alternated between two. There were no 
differences between individualists and collectivists in the regularity of use of 
justification patterns: 74.4% of the former and 73.3% of the latter used one 
or two reasons to justify their respective choices. 
When only the justifications of these 'regular' subjects are considered, the 
four above-mentioned categories explained 96.5% of the reasons used 
(instead of 82.64% which appeared when all subjects were computed). Dis­
criminatory strategies, on the one hand (maximum ingroup profit and 
maximum difference), and non-discriminatory ones, on the other (fairness 
and maximum joint profit), were added up. This calculation led to the 
discovery that in individualists the use of discriminatory reasons (28.1%) is 
less than the use of non-discriminatory reasons (69.2%), a difference which 
is significant. However, in collectivists, though the use of discriminatory 
reasons is also less (42% and 54%, respectively), the difference is far from 
being significant. 
Ingroup choice 
By means of the Wilcoxon test we tried to find out whether there were 
differences between individualists and collectivists in their preferences for 
any strategy in the various choices. In other words, when individualists 
proceed from first choice and justified choice to ingroup choice, do they 
change their preferences as compared with collectivists? If this is true, the 
conclusion would be that individualists see ingroup members in a different 
way from how collectivists see members of their own ingroup. 
Accepting p < .05 as a criterion of statistical significance, we found that, 
in effect, there were strong differences between collectivists and individual­
ists. They certainly coincide in thinking that members of their own group use 
the fairness/equity strategy of Tajfel's matrices less often than themselves, 
which means they consider themselves as fairer than their ingroup's mem­
bers. But individualists, as opposed to collectivists, think that ingroup mates 
are higher than themselves in the three Tajfel discriminatory strategies, 
namely MIP over MJP, MD, and MIP over fairness. 
In Insko's matrices something similar happens. Collectivists consider 
themselves as more just than in 
group members. And so do individualists. 
Moreover, collectivists see themselves as less discriminatory than ingroup 
members in max. reI. own. And so do individualists. But these think they 
are less discriminatory than ingroup members in max. own, and here they 
are opposed to collectivists. 


Discrimination and belie
f
s
 on discrimination 
209 
Outgroup choice 
As in the previous choice, the Wilcoxon test was employed to look for 
differences between individualists and collectivists in the use of strategies in 
this choice as compared to the two fonner ones. 
In Tajfel's matrices, results are exactly the same as in the previous choice. 
Collectivists consider themselves as more prone to fairness (F) than 
outgroup members. But collectivists do not think outgroup members to be 
more discriminatory than themselves. Individualists, however, do think so 
in the following strategies: MIP over MJP, MD, and MIP over F. 
Results are also very similar in Insko's matrices. Both collectivists and 
individualists consider themselves as more fair than outgroup members and 
less discriminatory in max. reI. own. But individualists, as opposed to 
collectivists, think they are less discriminatory than outgroup members in 
max. own. 
First choice-justified choice and ingroup choice-outgroup choice 
comparisons 
Considering the six Tajfel strategies and the seven Insko ones, as well as the 
two groups, individualists and collectivists, a total of 26 first choice-justified 
choice comparisons and 26 ingroup choice-outgroup choice comparisons 
were carried out. None of them yielded significant differences, which, in 
view of the previous results, suggests that subjects are very homogeneous 
when answering for themselves (whether having to justify their answer or 
not) and when they answer in the place of others (be these from in 
group or 
from outgroup). 
Conclusion 
Two conclusions can be drawn from the results of our two experiments. 
First, both individualists and collectivists discriminate in favour of their 
own group in a minimal group situation. As we have seen, there is a slight 
tendency for individualists to be less discriminatory. However, on the 
whole, both kinds of subjects provide evidence of group fonnation, along 
the lines suggested by Insko and Schopler (1987). 
Second, as far as beliefs on discrimination are concerned, it seems clear 
that individualists consider their point allocation to be less discriminatory 
than that of collectivists. The reasons they provide to justify their second 
choice seem to imply that individualists interpret their allocating behaviour 
as a search for justice and a wish to share with the other group. On a much 
lower number of occasions they admit to being motivated by a wish to 
dominate the opposite group. 
Moreover, the divergence between individualists and collectivists appears 
again in the clear-cut distinction the fonner establish between themselves 


210 
Social identity: international perspectives 
and people of ingroup and outgroup. Collectivists tend to consider them­
selves as more just in their comparisons to ingroup and outgroup, but this 
result can be easily explained as due to social desirability. Individualists also 
see themselves as more fair and, at the same time, as less discriminatory. 
This explicit and public acceptance of their non-following ingroup norm 
seems to speak for a claim of independence vis-a.-vis the group, of a certain 
critical positioning against it, despite the fact that actual allocating 
behaviour is discriminatory and, as such, is a proof that the person feels as 
a member of the group to all purposes. 
All the aforementioned leads us to the conclusion already outlined in 
previous pages. It is not a matter of individualists experiencing insurmount­
able difficulties to form groups or identify themselves with them. Rather, 
what happens is that they do so while trying to preserve their individual 
initiative, in this particular case restricted to beliefs. Of course, other inter­
pretations of our data are possible, this allowing and, at the same time, 
encouraging future research on a subject which is of interest to understand 
processes mediating behaviour of people in groups and in society as a 
whole. 
Note 
This chapter was translated from the Spanish by Elena Gaviria Stewart. 


1 5  
Constructing Social Identity: The Role of 
Status, Collective Values, Collective Self­
Esteem, Perception and Social Behaviour 
Daria Paez, Cristina Martinez-Taboada, Juan Jose Arr6spide, 
Patricia Insua and Sabino Ayestaran 
Social identity theory (SIT) and Turner's self-categorization theory (SeT), a 
more cognitive version of the former, have stimulated research on group 
processes and are currently some of the most heuristic theories in this 
domain. Nevertheless, some psychological aspects of SIT have not received 
systematic empirical support. The most important ones are the following: 
(a) the belief in the relationship between salience of the categorization and 
intergroup discrimination; (b) the relationship between strong identification 
and perceptual and behavioural biases which favour the ingroup; (c) the 
relationship between discrimination and an increase in self-esteem (Messick 
& Mackie, 1 989). These results suggest that the relationship put forward by 
SIT, and also partly by SeT, between salience and favourable behaviour 
towards the ingroup, coupled with an increase in self-esteem, is not so 
general as was once thought. However, we must state that these criticisms 
are valid only for a more individualistic approach to SIT and SeT. 
Following Turner, Hogg, Oakes, Reicher and Wetherell (1987), psycho­
logical processes of identification, self-esteem, discrimination, and so on, are 
only one aspect of intergroup behaviour. Socio-structural position and 
socially shared belie
f
s
 on the legitimization of power/status differences and 
individual/collective strategies to cope with these social differences are two 
main explanatory processes of intergroup behaviour. From this perspective 
it is a mistake to focus only on the intra 
psychological relationships of 
identification, self-esteem and discrimination. 
This chapter will analyse certain themes of debate which have arisen with 
respect to the concept of social identity: When does it regulate a person's 
behaviour and perception? When does it produce a favourable perception of 
the in 
group and discriminatory behaviour towards the outgroup? Hinkle, 
Brown and Ely's (1992) revision of 
the literature has shown that there is not 
always an association between group identification and ingroup favouritism, 
and that there is no consistent association between discrimination in favour 
of one's group and increasing one's self-esteem. Hinkle et al. confirmed that 


212 
Social identity: international perspectives 
subjects who share collectivistic values are those who show more association 
between identification and ingroup favouritism. Crocker and Luhtanen 
(1990) have found that the variable linked to intergroup discrimination is 
collective and not personal self-esteem: it is those subjects with a high 
collective self-esteem who show a discriminatory bias in situations which 
threaten one's social identity. Abrams (1990, 1 994) has shown the moder­
ating role of private self-consciousness and strong identification: those 
subjects with a strong group identification and high private self­
consciousness are those who show a stronger behaviour regulation on the 
basis of one's social identity. We will present Spanish data which confirm 
and specify the role of these moderating variables in the relationship 
between social identity, perception and social behaviour. Finally low-status 
groups' socio-cognitive functioning does not always respond to the logic 
derived from SIT's statement of the need to defend one's identity: instead of 
showing ethnocentrism subjects present centrism or outgroup favouritism. 
Turner's perspective emphasizes the role of socially shared belie
f
s
 and socio­
structural position in order to explain these results. In our case we will focus 
on the role played by individualistic versus collectivistic values and in the 
role of collective self-esteem. This is why we will analyse social status and 
the predominance of the collectivistic cultural values as factors which will 
moderate the regulation of one's behaviour by social identity. 
We believe it is specially important to contrast the macro 
social and 
transcultural validity of the above-mentioned theories. This is one of the 
reasons why first of all we will devote special attention to the studies 
conducted in Spanish and Latin American contexts on the issues of 
collective identity, stereotypes and discrimination (Morales & Paez, 1 996). 
Individualism-collectivism, status and the importance of personal and 
social identity 
SIT states that there is a personal-intergroup continuum. On the intergroup 
pole, subjects see themselves as interchangeable members of the category or 
group. On the personal pole, subjects will be defined by their singularity and 
individual specificity. It is thought that the coping strategies which the 
members of low- or high-status groups rely on will be located as opposites 
in this dimension: on one extreme we will find individual mobility, on the 
other side social competition. Nevertheless, research carried out in Latin 
America has shown that Chilean workers simultaneously agreed with social 
conflict strategies and using strategies aimed towards individual improve­
ment of their siblings' position by means of education (Quevedo, 1 972). 
This study did not analyse the correlation between the agreement with both 
strategies. Even though this correlation could be negative, and so confirm 
the hypothesis which states the opposite nature of both these strategies, the 
fact that the average agreement with both strategies was high confirms that 
they may coexist on a group level. The opposition between the salience of 


Constructing social identity 
213 
identification levels which have different inclusiveness (that is, regional and 
supra-national) has not been confinned as a general phenomenon. We 
must state that some subjects show strong regional identification and a low 
identification on a higher level. But other subjects simultaneously show a 
strong identification with different levels of inclusiveness (as was the case 
of a simultaneously strong identification with the Polish and European 
national identities, Mlicki & Ellemers, 1996). Finally it is possible that in 
collectivistic cultures the difference between social and personal identity is 
much less than in individualistic cultures. SIT posits that an important 
dimension of identity is the social dimension. Nevertheless, the identity or 
individual self-concept may be more or less shaped by social identity. 
There is both a trait and situational variability with regard to the 
integration of social identity in personal identity or the fonner's relevance. 
The importance of social identity and its integration within personal 
identity will be mediated by the subjects and the historical context which 
contextualizes them. The importance of social identity is stronger in 
collectivistic and sociocentric cultures in comparison to individualistic and 
egocentric ones. 
Individualism is a set of beliefs, values and cultural practices in which 
people manifest an emotional independence from groups and organizations. 
On the other hand, collectivism as a cultural syndrome is associated with an 
emotional dependence towards a group. A collectivistic culture is that in 
which there is a stable relationship between the individual and certain 
infonnal groups to which one belongs, a subordination of individual objec­
tives to those of the collective, and an adjustment between the person's and 
the group's objectives. There is a tendency towards not leaving the group 
even though the requirements may be high. An individualistic culture is that 
in which a person has negotiable relationships with various groups. S/he 
may abandon these groups when they are very demanding (or potentially 
may be so), and the individual's goals are more important than the group's 
goals (Morales, Lopez & Vega, 1992; Triandis, 1992, 1994; Triandis, 
Bontempo, Villareal, Asai & Lucca, 1988). Cooperation with the group, 
sharing resources and spaces with members of certain stable ingroups, col­
lective achievements and a strong affective link all characterize collectivistic 
subjects. Interpersonal competition, having important private spaces and 
resources, individual achievements and distancing oneself from the group 
are all characteristics of an individualistic subject (Hinkle et aI., 1992). 
From the point of view of the importance of social or collective identity 
for one's self-image or for the social representation of the individual (see 
Doise, Chapter 2 in this volume), individualistic cultures stress a concept of 
the individual as a separate, autonomous, self-contained and independent 
being. Collectivistic cultures emphasize a representation of the person as an 
entity centred in his or her relationships with others, connected to, and 
dependent on, the social context (Markus & Kitayama, 1991 
b). 
The group's social status also affects the degree to which individuals 
define themselves in tenns of their social or personal identity. High-status 


214 
Social identity: international perspectives 
groups perceive themselves as a set of individuals. Low-status groups define 
themselves as an aggregate of subjects who share a series of similar 
characteristics (as Lorenzi-Cioldi et al.'s studies have shown). Members of 
dominant groups will define their identity mainly on the basis of personal 
factors, while members of dominated groups will define their self-concept on 
the basis of mainly social factors (Lorenzi-Cioldi & Doise, 1 994). As a 
summary we may state that dominated groups and collectivistic cultures will 
reinforce the relevance of social identity over personal identity. On the other 
hand, individualistic cultures and dominant groups will pay more attention 
to, and have a better evaluation of, an individual who possesses resources 
while also stressing the relevance of personal identity over social identity. 
Low status, salience of the intergroup comparison, outgroup 
favouritism and group perception 
Culture and status have an influence not only on the importance of the 
social categories in the representation of one's self, but also in assessing 
the representation of group perception. Collectivistic cultures, which posi­
tively value the group, reinforce the perception of intragroup homogeneity, 
while individualistic cultures, which positively value an autonomous subject, 
reinforce the perception of ingroup variability (Devos, Comby & 
Deschamps, 1 995). 
From the point of view of intra- and intergroup social perception, we 
must remember that there is a tendency towards perceiving the ingroup as 
better in certain attributes. For instance Rodriguez (1992) found that 
subjects from Venezuela evaluated themselves better than subjects from 
Colombia in four out of six positive attributes, whilst the Colombians did 
the same in five out of six attributes. 
In general, subjects differ positively in those dimensions which define the 
ingroup, although they may accept outgroup superiority in another set of 
dimensions. This social perception which favours the in 
group is modelled 
by social status: high-status groups tend towards a stronger perceptive 
differentiation in favour of the ingroup than those subjects who hold a 
lower status, as Mullen, Brown and Smith's (1992) meta-analytic review 
showed. 
In relation to the variability of stereotypical attributes, the ingroup is seen 
as more complex, variable and heterogeneous than the outgroup. There is a 
tendency towards perceiving the outgroup as more homogeneous and the 
ingroup as more heterogeneous. As Mullen and Hu's (1989) meta-analytic 
review has confirmed, this effect is further stressed in the case of natural 
groups. Status also has an influence in the perception of variability: minority 
or threatened groups will perceive their group as less variable. This is a way 
of reinforcing its distinctiveness and cohesion. Results have also shown that 
groups perceive themselves as more homogeneous on those attributes or 
dimensions which define them (Leyens, yzerbyt & Schadron, 1 994). 


Constructing social identity 
215 
'Cold' socio-cognitive theory (Leyens et ai., 1994) suggests that having 
had a direct experience with the group will lead to a more variable, 
heterogeneous or individualized representation of the ingroup. 
From an SIT perspective, cognitive categorization as a member of a 
group, or the identification with this group, will stress intergroup differences 
and intragroup similarities. Individualization or aggregation would be a 
function of the defence of one's social identity, on the basis of the existing 
power and status relationships. Social comparison will influence not only 
the perception of differences but also homogeneity. It will not always be the 
outgroup who will be perceived as less variable, simple or homogeneous. 
The ingroup will be perceived as more homogeneous in those attributes 
which define their social identity and in situations in which this identity is 
under threat (Leyens et ai., 1994; Marques & Paez, 1 994). 
In a study conducted by us we confirmed the predominance of this 
normative logic of identity defence, instead of a cognitive logic of experience 
(see Marques, Paez and Abrams, Chapter 9, this volume). We asked 
psychology students (N = 109) to evaluate the ingroup's (psychologists) and 
outgroup's (engineers) variability in stereotypical attributes (Table 15. 1). 
These attributes had been generated in previous studies and suggested that 
being friendly and sociable are typical traits awarded to psychologists, whilst 
being competitive, intellectual and having a strong leadership capacity are 
typical of engineers. On the basis of these answers we calculated the Pd 
(perceived difference) and Var (variability). The estimated variabilities are 
significantly different, although not regularly higher in the ingroup. 
Variability was higher in the outgroup on those attributes stereotypical of 
the ingroup and vice versa. In other words, psychologists perceived them­
selves as more homogeneously friendly and sociable in comparison to 
engineers, while this last group was seen as more homogeneously competitive 
and intellectual, although in this last attribute differences were smaller and 
other studies have found that it is associated with the psychologists' 
stereotype. In those variables in which we did not find significant differences, 
means were in the hypothesized direction. Variability was smaller for 
engineers in leadership and in being normative. 
The tendency towards accentuating ingroup favouritism on important or 
consensual dimensions seems an increase in competition or salience of the 
comparison conditions, as Doise and Sinclair's study (reanalysed by van 
Knippenberg & Wilke, mentioned in Morales & Huici, 1994, p. 737) has 
shown. Nevertheless in a study conducted in Brazil by Nascimento-Schulze 
(1993) on bank clerks, a social comparison situation led to a worse image of 
the group than a situation in which there was no social comparison, and the 
same happened when the context made group membership more salient. As 
we can see, this is contrary to the fact that group salience should activate a 
process of favouritism or defence of the group's image. The author 
interprets these results as showing that these groups of bank clerks in social 
comparison and salient conditions are aware of their problematic work 
conditions and of their unstable status, which is reflected in a diminution of 


216 
Social identity: international pers
pectives 
Table 15.1 Perceived ingroup and outgroup heterogeneity 
Variability means of 
Engineers 
Psychologists 
Attributes 
Mean 
SD 
Mean 
SD 
Friendly 
0.88 
0.56 
0.68 
0.54 
Competitive 
0.60 
0.61 
0.82 
0.63 
Sociable 
0.8 1  
0.72 
0.56 
0.52 
Intellectual 
0.64 
0.56 
0.76 
0.53 
** p < .01 ;  *** p < .001 
t (1 1 3) 
3.86*** 
-3.17** 
4.03** 
-2.72** 
their self-image. This socio-cognitive functioning of groups with unstable 
and relatively insecure status does not seem to follow an ethnocentric logic. 
In a conceptual replication of this research, we have confirmed that in 
conditions of salience of the comparison subjects increase the superiority of 
the outgroup in consensual or stereotypical attributes. In order to manipu­
late the salience of the comparison, we asked subjects enrolled in a group 
dynamics course (N = 1 1 8) to perform a cognitive creativity task: to list the 
attributes which characterize a good leadership. Subjects were randomly 
assigned to a condition of activation of the comparison with the engineer 
outgroup, or of ingroup comparison. All the subjects had to read a first set 
of attributes. Depending on the condition, subjects were told that this list 
had been written by psychologists (ingroup) or engineers (outgroup). Apart 
from performing the task, subjects also had to estimate ingroup and out­
group variability on a series of six attributes: leadership, friendly, competi­
tive, normative or followers of the norm, sociable and intellectual. On a 
five-point scale subjects had to decide, for instance, what percentage of 
engineers and psychologists had no, some, regular, a high or an extremely 
high intellectual capacity. This cognitive task allowed us to obtain vari­
ability indices such as the Pd and the Var, whilst also allowing us to infer 
the group's mean on that attribute. All the subjects evaluated both psy­
chologists and engineers. The F of the between-group interaction with 
regard to the group that was being evaluated (ingroup or outgroup) and the 
salience manipulation was significant (F( l ,  1 1 3) = 4.89; p < .05). Specifically 
psychologists were attributed a much lower leadership average when the 
social comparison was made salient. This reflects the realistic perception of 
the inferior nature, in terms of status, of the psychologists compared to 
engineers (see Table 1 5.2). 
In order to confirm that previous results (psychologists perceived them­
selves as more homogeneously friendly and sociable in comparison to 
engineers; the engineers group was seen as more homogeneously com­
petitive in comparison to psychologists) followed a normative or social 
desirability logic, we asked a sample of psychology students (N = 9), 
psychologists and mental health workers (N = 9), and qualified workers, 
including physicians and engineers, and non-qualified workers (N = 1 2) to 


Constructing social identity 
217 
Table 1 5.2 
Perception of ingroup and outgroup by salience versus non­
salience o
f
 social comparison 
Attributes 
Leaders 
Salience of social comparison 
('between' variable) 
Engineers 
Psychologists 
('within' variable) 
Inferred mean 
Inferred mean 
3.31 
2.71 
Non-salience 
('between' variable) 
Engineers 
Psychologists 
('within' variable) 
Inferred mean 
Inferred mean 
3.47 
3.34 
'evaluate to which degree the following characteristics are desirable or 
positive in each jobs' performance'. As we can see in Table 15.3 both the 
engineers' higher status and the normative nature for psychologists on those 
attributes in which there was a smaller variability (friendly, sociable) was 
confirmed. Exactly the opposite took place for the attribute 'competitive' 
(less variability and it was attributed more to engineers). The only exception 
was found in the attribute 'intellectual', in which psychologists have a 
higher score and that in the first study showed more variability, although 
the difference in variability means was the smallest of all. In order to 
confirm that the attribution of the traits followed both a professional social 
desirability logic as well as a group identification logic, we correlated the 
subjects' level of identification with psychologists ('we ask you to indicate 
your identification with the professional category termed "psychologists"') 
with a composite rating of attributes (friendly, sociable, intellectual) of the 
group. This composite score showed a satisfactory reliability (alpha = .67) 
and a significant correlation of 0.35; p < .05 with one's identification with 
the group of psychologists. This identification was individually associated 
with believing that it was desirable that the psychologists should be friendly 
(r = .36; p < .05), with sociable (r = .27; P < .10) and, although of marginal 
significance, with intellectual (r = . 1 5; P < . 1 0). 
On the other hand, identification with psychologists was not associated 
with the agreement that psychologists were competitive and normative (r = 
< .09). However, those features were typical of the engineers. Identification 
with psychologists was associated with the leadership attribute (r = .42; P < 
.05). Bearing in mind one's professional experience, in comparison to 
students, psychologists and mental health experts stated that it was more 
socially desirable for a psychologist to have leadership capacity (mean = 
5.1 1), and both these groups believed in this fact more than did qualified 
and non-qualified workers (mean = 3.75; F(2, 27) = 4.07; P < .03). Group 
identification was linked with a stronger attribution to this group of the 
positive attributes, especially when these define their professional identity. 
Moreover, even when the task was a comparison with a group of engineers, 
those psychologists and mental health experts more professionally involved 
attributed positive traits (e.g. leadership) to the ingroup, though it has been 
shown that these traits tend to be associated with high status groups. In 


218 
Social identity: international pers
pectives 
Table 1 5.3 Perceived social desirability o
f
 attributes f
or engineers and 
psychologists 
Means of social desirability for 
Engineers 
Psychologists 
Attributes 
Mean 
SD 
Mean 
SD 
Friendly 
3.70 
1 .61 
5.90 
1 .26 
Competitive 
5.70 
0.95 
4.17 
1 .58 
Sociable 
4.67 
1 .40 
5.97 
1 .22 
Intellectual 
5.00 
1 .19 
5.87 
0.90 
Leadership 
5.80 
1 .3 1  
4.30 
1 .31 
Normative 
5.50 
1 .07 
4.70 
1 .58 
Means of group status for 
Engineers 
Psychologists 
Status 
6.27 
0.69 
4.00 
1 .08 
I 
= not at all; 7 = very much 
* p < .05; ** p < .01 ;  *** = p < .001 
t (29) 
-7.05*** 
4.63*** 
-4.33*** 
-3.98** 
3.84** 
2.68** 
1 1 .84*** 
summary we may state that the lower variability traits associated with 
the psychologists are in general those which are seen as socially desirable 
for the job of being a psychologist. These traits are also associated with the 
identification with this group and involvement in one's professional activity 
(see Table 1 5.3). 
We may conclude that a certain relative ingroup bias is present in social 
perception and that it follows a normative logic - of defence of a positive 
social identity by means of stressing the positive aspects of the ingroup in 
those attributes used to define the group. Moreover, in subordinated groups 
this tendency is not shown in social comparison situations. But this does not 
imply that subordinated groups do not show in general outgroup favour­
itism. Bourhis, Gagnon and MOise (1994) have shown how members of 
dominated groups choose and evaluate more positively dominant outgroup 
symbols. In studies conducted in Latin America, different authors have 
found that the Venezuelan ingroup was evaluated worse than the European 
(Italian) and North American outgroups (Montero, 1987, 1990). Even 
though the subjects identified themselves with the ingroup (Venezuela), 
subjects had a better evaluation of the European groups (England and 
Spain) and of certain American nations such as Argentina, although they 
had a bad image of other American countries such as Colombia. Due to the 
results obtained from similar psychosocial, anthropological and historical 
studies, which show that in Latin America, Asia and Africa there is a 
devaluation of the national ingroup and favouritism towards First World 
outgroups, the concept of Altercentrism has been coined. Altercentrism is 
an important theoretical concept as it allows us to pay attention to a basic 
problematic found in SIT: how do we explain the socio-cognitive func-


Constructing social identity 
219 
tioning of 
the dominated groups who do not mobilize themselves, or cannot 
be individually included in higher status groups? 
Various explanatory processes put forward by SIT suggest how domi­
nated groups manage their ambivalent image, showing at the same time 
favouritism towards dominant groups. 
Intragroup social comparison with more disadvantaged subsectors, or, 
alternatively, an advantaged social comparison with a group of similar and 
pertinent attributes, would be another mechanism which would allow the 
showing of favouritism towards the dominant outgroups while at the same 
time retaining a partly positive image of the ingroup. Members of sub­
ordinated ingroups (that is, Venezuelans) accept the superiority of dominant 
outgroups (that is, North Americans), although at the same time they stress 
their superiority over other groups who hold a similar position to theirs and 
whom they undervalue (that is, Colombians) (Montero, 1 987; Salazar, 
1989). 
Another phenomenon that characterizes dominated groups is a type of 
symbolic individual mobility - of which Bovarism and the cosmopolitanism 
of the dominant classes of Latin America is a good example. Bovarism 
(permanently believing one is living in Paris, London or New York, 
although one is really living in Puerto Hundido) is a form of symbolic and 
ideal identification with the dominant outgroup. Latin Americans, 
especially, although not only, those from dominant classes, view themselves 
as second-class or 'transplanted' Europeans or North Americans. The use of 
surnames (for instance, the large amount of people called Johnny, William 
or Peter Gonzalez), adopting customs and identifying with the values of the 
dominant cultures constitute a coping strategy of the accepted negative 
national identity. Studies carried out in developed countries have shown 
that this symbolic individual mobility strategy takes place even though there 
are no real possibilities of integrating oneself into the dominant group 
(Wright, Taylor & Moghaddam, 1990, mentioned in Bourhis et a!., 1 994). 
Finally, low-status or dominated groups use cognitive creativity, redefin­
ing as important a series of new dimensions in which the ingroup is superior 
(see Smith & Bond, 1 993). Research conducted by Salazar et a!. on national 
stereotypes in Latin America has confirmed the existence of a status/ 
competence dimension, a sociability dimension and a third, less important, 
education/instruction factor. There is evidence to suggest that the instru­
mental or status/power dimensions and the expressiveness or sociability 
ones are transculturally valid dimensions of social perception (Zebrowitz, 
1990). Various studies have shown that subordinated groups accept the 
superiority of dominant groups on the competence/ability dimensions, or on 
status, power and resources, but they view themselves as superior in the 
expressive or sociability dimensions. For instance, nurses with a high status 
viewed themselves as superior in technical knowledge, while lower status 
nurses perceived themselves as being more friendly (Smith & Bond, 1993, p. 
62). Studies conducted in Latin America confirm that the self-stereotypes 
reaffirm a positive evaluation of sociability attributes: attributes stereo-


220 
Social identity: international pers
pectives 
Table 1 5.4 Attributes f
or engineers and psychologists 
Engineers 
Psychologists 
Attributes 
Inferred mean 
SD 
Inferred mean 
SD 
Leadership 
3.40 
0.83 
3.07 
0.80 
3.13* 
Friendly 
2.98 
0.66 
3.83 
0.77 
-9.08** 
Competitive 
4. 1 1  
0.70 
2.78 
0.80 
14.40** 
Normative 
3.56 
0.90 
2.69 
0.84 
7.57** 
Sociable 
3.00 
0.66 
3.96 
0.56 
-1 1 .75** 
Intellectual 
3.42 
0.83 
3.27 
0.67 
6.49** 
* = p < .01; ** = p < .001 
typically favourable to the Latin American national groups were of the 
sociable type (that is, happy), while the stereotypical positive attributes 
towards European and North American groups were of a more instru­
mental nature (that is, practical, workers and so on) (Rodriguez, 1992). 
Our own research conducted in Spain on psychologists and engineers 
confirmed that the perceptive superiority of a lower status ingroup (psycho­
logists) was found on the expressive dimension (they viewed themselves as 
more friendly and sociable), whilst they also accepted the superiority of the 
higher status outgroup on more instrumental attributes. They perceived the 
engineers as more competitive, intellectual, and so on (see Table 1 5.4). 
The above-mentioned results may be interpreted as the effect of a 
normative process, in the sense that subjects try to defend their group by 
perceiving it as an entity which adjusts itself to the norms and values which 
relevantly define its subculture more than the outgroup does (Marques & 
Paez, 1994; also see Devos et al., 1995). Nevertheless, although low-status 
group members may use symbolic individual mobility, comparison with 
disadvantaged groups and cognitive creativeness in order to explain one's 
ambivalent social image, according to the psychological version of SIT, low 
status must have an influence on one's self-esteem. 
Status, self-esteem, collectivistic values and social identity 
Status is an important variable for SIT due to the fact that it is associated 
with the prestige and self-esteem one may infer from one's group 
membership. According to SIT, belonging to a social category or group is a 
source of self-esteem. Belonging to low-status groups, which possess nega­
tive attributes, would motivate a search for individual mobility, cognitive 
creativity and social competition. What is implicit in this approach is the 
idea that subjects from disadvantaged classes and groups should suffer a 
deficit in self-esteem. This idea is included within the hypothesis stating that 
a deficit or threat to self-esteem would be an antecedent or cause of the 
ingroup's favourable social perception, of ethnocentric attitudinal prejudice 
and behavioural discrimination. Personal self-esteem as an evaluative 


Constructing social identity 
221 
component of personal identity may be differentiated from self-esteem 
referring to group belonging or being a part of a certain social category, in 
other words collective self-esteem. 
One answer to the fact that the salience of the categorization does not 
directly lead to differentiation, or that it leads to an increase in self-esteem, 
is methodological. The relationship between discrimination or intergroup 
differentiation and self-esteem should take place and be proven not on the 
basis of personal self-esteem, but on the basis of collective self-esteem, in 
other words on the basis of the favourable evaluation of the category or 
social group to which one belongs. Nascimento-Schulze's (1993) results with 
Brazilian workers confirm that there is no relationship between intergroup 
discrimination and an increase in general personal self-esteem. Intergroup 
discrimination produced an increase in the group's evaluation or collective 
self-esteem. In this sense, the development of collective self-esteem scales is 
a step forward in order to better understand the relationship between social 
identity and social behaviour. Crocker and Luhtanen, 1990; Luhtanen & 
Crocker, 1992) have developed a scale which measures stable individual 
differences in the level of a person's positive social identity (high collective 
self-esteem) and also a positive personal identity. These authors developed a 
scale with a private collective self-esteem dimension, a public collective self­
esteem dimension, the feeling of being a member of the group, and the 
importance of this identity. These scales had structural validity and alpha 
coefficients which scored higher than . 
73 (in fact the public self-esteem scale 
reached a score of .80). The private collective self-esteem scale and the one 
referring to the importance of being a member of the group, which meas­
ures the most individual aspects of social identity, also correlated signifi­
cantly with personal self-esteem scales. The collective self-esteem scores 
were positively and coherently associated with the agreement with collective 
values - measured using Hui's individualism-collectivism scale (Luhtanen 
& Crocker, 1992). 
Our results with the Spanish translation of this scale also show satis­
factory results in relation to internal consistency. Alpha coefficients were 
satisfactory for the private and public collective self-esteem scales, identi­
fication with the ingroup and belonging to this same group (Cronbach 
alphas of .83, .78, .66 and .71, respectively). Moreover, they also showed 
convergent validity with Hui's collectivistic values scale (see below). 
Experimental research found that, using the above-mentioned scale, 
subjects high in collective self-esteem, contrary to those with low collective 
self-esteem, showed more positive biases in relation to their group in 
situations in which there was a threat to their group identity. In situations of 
failure, subjects with a high collective identity changed their scores in such a 
way that they mitigated the threat to the group's worth. Personal self-esteem 
did not play a regulating role in the answers (Crocker & Luhtanen, 1990). 
These results show that we have to differentiate between the need for 
individual and collective esteem. Moreover, they also suggest that the SIT 
statement referring to the tendency for discriminatory biases which favour 


222 
Social identity: international pers
pectives 
the ingroup with the aim of obtaining a positive social identity is valid only 
for those subjects with a stable tendency towards positively evaluating social 
identity or socially sharing a positive definition of collective self. 
Studies conducted on relative deprivation and political participation 
reinforce the idea that it is the threat to collective self-esteem, and not the 
lack of personal self-esteem, which is associated with the regulation of 
perception and social conduct by social identity. Let us remember that 
deprivation or personal dissatisfaction is not associated with socio-political 
participation behaviours, which we may view as examples of social conflict, 
but that in fact it is associated with anomie and individual affective 
alterations. It is sociocentric or fraternal relative deprivation (in relation to 
the group) which is linked to political participation (Guimond & Tougas, 
1 994). On the other hand, research on political participation shows that it is 
those subjects who are socially integrated, with a perception of political self­
efficiency, but who are also dissidents or critical of the social system which 
they view as the cause of the group's disadvantaged position, who mobilize 
themselves - and not those subjects who are most isolated, deprived or 
disadvantaged (Kinder & Sears, 1985). This is why Crocker, Luhtanen, 
Blaine and Broadnax (1994) have found that it is those subjects with high 
collective (socially shared beliefs or collective affective evaluation with 
regard to the group and not the person) self-esteem (those who are integrated 
and efficient) who show more ingroup bias in threatening conditions (in 
other words, of relative deprivation in which the group does not receive what 
it expected). As we can see, those subjects who regulate their perception and 
behaviour on the basis of their social identity and those who show ingroup 
favouritism and outgroup discrimination are the ones who have high col­
lective self-esteem in threatening situations. 
The differentiation between social and personal identity as well as the 
relative independence of public, private and personal collective self-esteem 
are two mechanisms which may explain why members of dominated and 
stigmatized classes do not show a deficit in personal self-esteem (Phinney, 
1 990). 
An explanatory process for the lack of deficit in personal and collective 
self-esteem in a dominated group included in an individualistic culture is the 
frequently found phenomenon of the acceptance of the disadvantaged 
nature of the group, while at the same time rejecting its individual dis­
advantage (Crosby, Cordova & Jaskar, 1994). For instance, Puerto Rican 
subjects accepted that Puerto Ricans are dependent, but they themselves are 
not like that. Personal self-valuation was mainly positive, while collective 
self-valuation was mainly negative (Rivera, 1991). As Rivera (1991) stated: 
The results of this first study were the pronounced and consistent divergence 
between the collective and personal valuation; the latter was invariably higher than 
the collective valuation. These results suggest that participants may have inter­
nalized, to some degree, the Puerto Ricans' negative stereotype, depersonalizing 
these categories in order to show that the rest of 
the Puerto Ricans are that way, but 
not themselves, who are attributed positive personal valuations. (pp. 10 1-102) 


Constructing social identity 
223 
Another explanatory process for the lack of low personal self-esteem in 
members of dominated groups is that subjects will not link other people's 
opinions about their group with their own personal self-esteem. Crocker et 
al. (1994) have found that the relationship between public (how the subject 
believes that other people evaluate hislher group) and private collective self­
esteem (how the subject evaluates hislher own group) is medium in 
individualistic cultures (Anglo-Americans), positive and strong in collecti­
vistic cultures (Asian-Americans) and null in dominated groups (African­
Americans). Let us add that private collective self-esteem is strongly linked 
in general with personal self-esteem. Nevertheless, in the sample of black 
subjects there was no relationship between public and private collective self­
esteem. These black subjects strongly differentiate how they believe other 
people value their group and how they themselves do so - and probably 
how they value themselves as individuals. In this way it is possible to cope 
with discrimination and prejudice separating collective negative judgements 
from one's image of oneself. On the other hand, for Asian subjects the 
vision the other person has is strongly associated with one's own view of the 
ingroup. This suggests that the relationship between personal self-esteem 
and group self-esteem is more underlined in collectivistic cultures. 
Self-consciousness, individualistic-collectivistic cultural values, social 
identity and behaviour regulation 
As Abrams (1990) has stated, sometimes behaviour is not regulated by 
social identity. Categorization in a group and salience of group membership 
will not be sufficient conditions for ingroup discrimination and its function 
of augmenting self-esteem. Moreover, a person must strongly identify with 
the group. Abrams (1990, 1 994) has shown in different studies that group 
identification and self-consciousness or focalizing attention on the private 
and public aspects of one's self converge in order to regulate behaviour. 
Subjects with strong identification and high private self-consciousness 
regulated behaviour by means of social identity. 
There is correlational evidence which confirms that high identification with 
a group is associated with a good evaluation of this same group. Garza and 
Herringer (1986) developed an instrument which measured important 
dimensions in group identity, the evaluation of the group, its emotional 
valuation and the stability of group identity. In a replication of this study, we 
applied the Spanish version of this scale (Paez, Arrospide, Martinez-Taboada 
& Ayestaran, 1 992). Our sample was composed of 120 subjects, 46% were 
men and 54% women with a mean age of 24.7 years. Forty-seven subjects 
were students who were starting their participation in a group dynamics 
course, and the rest of the sample were relatives of fifth-year psychology 
students enrolled in the Psychology Faculty of the Basque Country 
University. Students answered the questionnaires in class while the relatives 
were interviewed at home by the students as part of their class curricula. 


224 
Social identity: international perspectives 
Table 1 5.5 Factorial analysis o
f
 the 'most important social identity scale' 
(Garza & Herringer, 1986) 
Factor 
2 
3 
EV Good-bad 
.88 
EV Likes-dislikes 
.86 
EV Positive-negative 
.87 
EM Pleasant-unpleasant 
.83 
EM Comfortable-uncomfortable 
.80 
1M Important-unimportant 
.67 
1M Valid-invalid 
.80 
ST 
Permanent-impermanent 
.69 
ST 
Stable-unstable 
.75 
ST 
Secure-insecure 
.81 
ST 
Cooperative-competitive 
.64 
ST 
Dominant-dominated 
.81 
ST 
With power-without power 
.84 
EM Amusing-boring 
.70 
ST Abstract 
-concrete 
.77 
ST 
Private-public 
.67 
EV = Evaluative dimension; EM = Emotional dimension; 1M = Importance; ST = Stability 
We found three clearly differentiated factors in this scale (see Table 1 5.5). 
The first factor explaining 47% of the variance was saturated by the four 
dimensions mentioned by Garza and Herringer (1986): emotion, evaluation, 
importance and stability, and also by item number 1 3, which refers to 
competition-cooperation. The second factor was composed of items related 
to power and dominance, and they explained 9.4% of the variance. The 
third factor included those items referring to the abstract versus the concrete 
and the private versus the public. This final factor explained 8% of the 
variance. 
These results reaffirm that the evaluation of the group, the emotional 
response to the group, the importance of the group for identity and the 
stability of the identification with a group are congruently associated. With 
regard to the relationship between identification and stereotypes, we 
have found that in Latin America those subjects who strongly identified with 
their national groups (samples from Venezuela, Chile and Colombia) 
rejected sloth, while stressing happiness as a description of a Latin American 
person when answering closed items. With regard to open answers, these 
samples underlined characteristics associated with the Latin American's 
positive sociability: happy, hospitable. Moreover, they mentioned history 
and socio-political processes such as oppression and dependency as identity 
elements. Those subjects who shared a low identification mentioned more 
frequently deficitary instrumental attributes, such as sloth and irresponsi­
bility, and they also made more mention of underdevelopment, poverty and 
language as identity factors (Salazar, 1 989). In other words, those people 
with a strong identification show perceptive favouritism stressing the 
stereotype's positive attributes, relativizing the negative ones, while those 


Constructing social identity 
225 
subjects who did not identify share the stereotypical negative attributes of 
their social category. 
As a summary, we may state that group identification and focalizing 
attention will interact in order to regulate behaviour. It is easy to believe 
that it will be those individuals with a high private self-consciousness and 
collective identity who will be more stable in their beliefs and behaviours 
linked to obtaining group goals. They will also be more emotionally 
reactive to stimuli such as group threats. 
Of course not only situational factors or internal tendencies of group 
identity salience and attention directed towards aspects of collective identity 
influence behaviour regulation. Macrosocial phenomena such as collecti­
vistic orientation cultures reinforce the importance of collective or social 
identity, while individualistic cultures reinforce personal or individual 
identity, as we have already seen (Gudykunst, 1988; Smith & Bond, 1 993). 
Collectivistic cultures emphasize an intensive relationship with ingroup 
members, while in individualistic cultures people behave in a more extensive 
than intensive fashion - even with the ingroup members. Consequently, the 
difference between behaviour towards the outgroup and behaviour towards 
the ingroup is not so important (Triandis, 1994). There is partial evidence 
which confirms that individuals from collectivistic cultures are more sensible 
to outgroup and ingroup membership in the regulation of their behaviour 
(Triandis, 1994). People from collectivistic countries show a stronger ten­
dency towards cooperation, and people from individualistic-type countries 
tend more towards competition. Nevertheless, when they must interact with 
outgroup members, collectivistic persons become more competitive. For 
example, Argentinean and Indian children who had to share game rewards 
with subjects who did not belong to their group were more competitive than 
children from the United States (who in general are competitive) (Smith & 
Bond, 1 993). 
In the following experiment we empirically confirmed the importance of 
collectivistic values in behaviour and perception regulation. We asked 
students (N = 48) enrolled in a group dynamics class to perform a cognitive 
creativeness task: to list the attributes which characterize good leadership. 
All the subjects had to read a first set of attributes supposedly previously 
generated by a group. Depending on the condition, subjects were told that 
this list had been generated by psychologists (ingroup) or engineers 
(outgroup). This first list and the one they created after 10 minutes of 
individual reflection and another 10 minutes of group discussion were 
evaluated as: good-bad, complete-incomplete, unreal-realistic, and 
typical-not at all typical. The response scale ranged from I = the first 
attribute of the pair to 7 = the second attribute of the set. For instance I = 
good, 7 = bad. 
Differences in scores between their own list and the previous one were 
taken as an index of favouritism. The difference was established between the 
first score (the score given to the supposedly other group) and the second 
score (given to one's own group). The subtraction was performed in such a 


226 
Social identity: international perspectives 
way that a positive difference meant a better evaluation of one's own group. 
We subtracted from the scale with highest negative scores (bad 
= 7; 
incomplete = 7; not at all typical = 7) the score given to the first set (other 
group) from the score given to the second list (one's own group). For 
example, if the first list ( 
other group) is considered bad = 7 and the second 
list (own group) is good = 1 ,  the positive difference of +6 was a sign of this 
favouritism. When the scale had a higher score and also a positive 
evaluation (realistic = 7), the second score was subtracted (own group list) 
from the first list. If the ingroup list was awarded a 7 (very realistic) and the 
first list (other group) was awarded a 1 (not at all realistic), the subtraction 
(7 - 1) would yield a positive score which would reflect ingroup favouritism. 
In other words, a positive score always meant ingroup favouritism. 
The scores in Crocker and Luhtanen's collective self-esteem scale, Hui's 
collectivistic values and Carver, Antoni and Scheier's ( 1 985) self­
consciousness scale were used as covariables and their influence was 
controlled. The variance and covariance multivariate analysis showed a 
significant effect if the first comparison group was that of engineers or 
psychologists. On a univariate level the comparison with an ingroup or 
outgroup had a significant effect for the positive evaluation of the list (F(5, 
46) = 1 3. 1 ; p < .001) and near significance for the complete list (F(5, 46) = 
2.74; p < . 10). The other group's list was evaluated as better and more 
complete when it was attributed to psychologists (ingroup), and was 
evaluated as worse and less complete when it was attributed to the engineers 
(outgroup). 
The private and public self-consciousness variables, as well as the group 
identification variables, did not interact with the experimental manipula­
tion. Nevertheless the individualistic/collectivistic values did. Subjects with a 
high score in collectivism behaved in a different way depending on whether 
the first list was attributed to the outgroup or to the ingroup. The multi­
variate interaction effect was significant (F( l l ,  38) = 2.33; P < .026). The 
univariate contrasts indicate that those subjects who share more collec­
tivistic values view the first list as more believable (F(1, 48) = 4.88; P < .03), 
realistic (F(1, 48) = 6.66; p < .02), complete (F( I ,  48) = 5.75; p < .02), and 
less typical (F( l ,  48) = 1 1 .04; p < .002) when this list was attributed to 
psychologists. In other words, when the stimulus was associated with the 
general ingroup and subjects were collectivistic (with regard to their values) 
it was evaluated more positively and as more original. Exactly the opposite 
happened with subjects low on collectivism; these subjects viewed the list 
that was attributed to the outgroup (engineers) as less typical, more 
believable, realistic and complete (see Table 1 5.6). 
The lack of self-consciousness and identity effects may be explained in the 
first case because the scale measures the tendency to focus attention on the 
public and private aspects of the self - and not necessarily on the collective 
dimension of the self. In the second case, the scale measures the general 
tendency to identify with groups in general - and not with the group which 
was used in the manipulation (psychologists). Finally the low number of 


Constructing social identity 
Table 1 5.6 Dif
f
erences between engineers and psychologists as a 
f
unction o
f
 their degree of collectivism 
First list 
Engineers 
Psychologists 
Differences in 
Low 
High 
Low 
High 
evaluation 
collectivism 
collectivism 
collectivism 
collectivism 
Believable 
-1 .30 
-0. 12 
-0. 17 
-\ .00 
Incomplete 
\ .50 
0.25 
-0.80 
0.73 
Realistic 
-1 .50 
-0.63 
-0. 18 
-\ .80 
Not typical 
- 1 .30 
0.60 
0.08 
-1.\0 
227 
subjects does not allow us to perform a contrast simultaneously choosing 
subjects with high group identification and high private self-consciousness. 
The above-mentioned research suggests that those subjects who show 
more ingroup bias in conditions in which the intergroup comparison is 
salient are those who share more the collectivistic values. Moreover, an index 
of collectivistic values is associated with social identification, collective self­
esteem and public and private collective self-consciousness, confirming the 
relationship between these processes. 
Applying the private and public self-consciousness items to the attention 
on social identity's external and internal aspects, we have developed a 
collective self-consciousness scale which we validated by comparing it to 
Garza and Herringer's scale (Paez et aI., 1992). This scale was developed on 
the basis of the items extracted from the original private and public self­
consciousness scale - referring to the most important group a person 
belongs to. The items and instructions were as follows: Think of the most 
relevant group to which you belong (family, friends, and so on). (I) I always 
try to analyse myself in relation to my group's objectives; (2) I am interested 
in how I do things when I do them in my group's name; (3) I reflect a lot 
upon myself in relation to my group's values and beliefs; (4) In relation to 
my group's objectives I am very careful with the image I give to other 
people; (5) I am well aware of the image I give in relation to my group; (6) I 
constantly think of the reasons for my behaviour in relation to my group's 
objectives; (7) I normally think of myself as a representative of my group; 
(8) I am worried about the image my group gives; (9) When my group gets 
what it wants I am happy; (10) I dream about myself many times in relation 
to my group; (I I) I do not feel responsible or guilty if a member of my 
closest group (family) fails; (12) I do not feel responsible or guilty if a friend 
fails; (13) When my group does not achieve what it wants I get very 
annoyed; (14) When a member of my group achieves what s/he wants I feel 
well. Questions number 1 1  and 12 are to be inversely scored. The scale was 
passed with a score of I = totally disagree to 7 = total agreement. In our 
sample the reliability of our collective self-consciousness scale was .73. This 
collective self-consciousness scale had a satisfactory convergent validity - a 
correlation of .39, p < .01 with Hui's individualism-collectivism scale; .22, 


228 
Social identity: international pers
pectives 
p < .05 with Garza and Herringer's scale of group identification; and .21 , 
p < .05 with a scale of importance of group membership for one's self (Paez 
et ai., 1992). 
These results suggest that the chronic accessibility of the collective 
dimension of the self, or, in other words, the stable focalization of internal 
and external attention on the collective or group dimensions of identity, is 
coherently associated with a positive evaluation of collectivism, with a 
positive identification with the groups and with a structure of one's self 
based on one's group membership. 
Conclusions 
This chapter has examined the conditions and processes by which social 
identity regulates social perception and behaviour. We have emphasized the 
role of socio-structural variables and of socially shared beliefs in order to 
explain this process. This is why we underlined the importance of social 
status and the predominance of collectivistic cultural values as factors which 
will moderate the regulation of one's behaviour by social identity. 
We stressed the stronger influence which social identity has in collectivistic 
and sociocentric cultures in comparison to individualistic or egocentric 
cultures. We also showed the same pattern of results for dominated or low­
status groups. On the other hand, dominant classes and individualistic 
cultures reflect more the attention and valuation of a subject who possesses 
resources, augmenting the salience of personal over social identity. 
High-status groups will stress the personal aspects of identity, the per­
ception of ingroup individuation and dif
f
erentiation. Dominated groups will 
have a self-concept which will place more emphasis on group membership, 
stressing the perception of intragroup homogeneity - the same as subjects 
from collectivistic cultures. 
Groups in general view themselves as superior in comparison to other 
groups on those attributes which define them or which are typical of them, 
and tend to perceive outgroups as less variable or more homogeneous. 
Nevertheless, in threatening situations and in relation to defining or typical 
attributes of one's collective identity, exactly the opposite takes place: they 
are perceived as more homogeneous. 
Social status was another variable which moderated the relationship 
between social identity and ingroup favouritism. Subjects perceive their 
ingroup as superior to the outgroup on those attributes which define it, 
although they do not necessarily show general favouritism. This favouritism 
will be most clearly found among high-status groups. Although in general 
subjects represent their ingroup as more heterogeneous than the outgroup, 
at the same time they state that it is more homogeneous, or less variable, on 
those attributes which define it. Low-status or -prestige groups show out­
group f
avouritism or altercentrism, while at the same time they view them­
selves as superior on specific dimensions, especially sociability and 


Constructing social identity 
229 
expressiveness. Our own study on psychologists and engineers in Spain 
confirmed that the perceptive superiority o
f
 a lower status ingroup (psy­
chologists) was found in an expressive dimension (they viewed themselves as 
more friendly and sociable), accepting the fact that the high-status outgroup 
(engineers) was superior to them in more instrumental attributes. 
SIT states that group membership is a source of self-esteem, and that 
belonging to a disadvantaged group will motivate favourable social com­
parison and social competition. In a sense, dominated or low-status groups 
do not show a deficit in self-esteem. The differentiation between personal 
and social identity, as well as the relative independence of private, public 
and personal collective self-esteem, are two mechanisms which explain why 
the members of dominated and stigmatized classes do not show a deficit in 
personal self-esteem. 
Also, as we have already stated, the relationship between discrimination 
and the increase in self-esteem is not consistent. The difference between 
individual and collective self-esteem partly explains the lack of association 
between identification, discrimination and self-esteem. Studies conducted on 
collective self-esteem and on fratemalistic relative deprivation and political 
participation suggest that a threatened high collective self-esteem is one of 
the antecedents of a biased social perception and of discriminatory 
behaviour. 
Tendencies towards ingroup favouritism are clearer among those indi­
viduals who share a strong group identification, high collective self-esteem 
and collectivistic values. Those individuals who value belongingness to 
collectivities also think that their group membership is important and they 
positively value this membership for their identity. These are the people 
who have stronger reactions when their social identity is under threat; they 
increase intergroup differentiation, positively compare their ingroup to 
other groups, and increase the identification and attraction of their group. 
Our correlational studies confirmed that individuals with a strong group 
identification show high private and public group self-consciousness, more 
stability, a better valuation, a positive emotional reaction, and attach more 
importance to belonging to a group for their identity. 
Note 
We wish to thank John C. Turner for his comments on an early version of this chapter. 


References 
Abrams, D. (1990). How do group members regulate their behaviour? An integration of social 
identity and self-awareness theories. In D. Abrams & M.A. Hogg (Eds.), Social identity 
theory: Constructive and critical advances. Hemel Hempstead: Harvester Wheatsheaf. 
Abrams, D. (1994). Social self-regulation. Personality and Social Psychology Bulletin, 20, 473-
483. 
Abrams, D., & Hogg, M.A. (1988). Comments on the motivational status of self-esteem in 
social identity and intergroup discrimination. European Journal o
f
 Social Psychology, IS, 
3 17-334. 
Abrams, D., & Hogg, M.A. (1990). Social identification, self-categorization and social 
influence. In W. Stroebe & M. Hewstone (Eds.), European review o
f
 social psychology, Vol. 
1. Chichester: Wiley. 
Abrams, D., Wetherell, M.S., Cochrane, S., Hogg, M.A., & Turner, J.e. (1990). Knowing 
what to think by knowing who you are: Self-categorization and the nature of norm 
formation, conformity and group polarization. British Journal o
f
 Social Psychology, 29, 97-
1 19. 
Abric, J.e. (1994). Pratiques sociales et representations. Paris: Presses Universitaires de France. 
Adamopoulos, J. (1982). Analysis of interpersonal structures in literary works of three 
historical periods. Journal o
f
 Cross-Cultural Psychology, 13, 1 57-168. 
Adamopoulos, J., & Bontempo, R.N. (1986). Diachronic universals in interpersonal structures. 
Journal o
f
 Cross-Cultural Psychology, 1 7, 169-189. 
Aebischer, V., Hewstone, M., & Henderson, M. (1984). Minority influence and musical 
preference: Innovation by conversion not coercion. European Journal o
f
 Social Psychology, 
14, 23-33. 
Allen, V.L., & Wilder, D.A. (1975). Categorization, belief similarity, and intergroup 
discrimination. Journal o
f
 Personality and Social Psychology, 32, 971-977. 
Allen, V.L., & Wilder, D.A. (1979). Group categorization and attribution of belief similarity. 
Small Group Behaviour, 10, 73-80. 
Allport, F.H. (1924). Social psychology. Boston: Houghton Mifflin. 
Allport, G.W. (1961). The ego in contemporary psychology. Psychological Review, 50, 471-
478. 
Allport, G.W., & Odbert, H.S. (1936). Trait-names: A psycho-lexical study. Psychological 
Monographs, 47. 
Altman, I., & Low, S.M. (Eds.). (1992). Place attachment. New York: Plenum. 
Amancio, L. (1989). Social differentiation between 'dominant' and 'dominated' groups: Toward 
an integration of social stereotypes and social identity. European Journal o
f
 Social 
Psychology, 19, 1-10. 
Anderson, B. (1983). Imagined communities: Re
flections on the origin and spread o
f
 nationalism. 
London: Verso. 
Apter, D.E. (1964). Ideology and discontent. In D.E. Apter (Ed.), Ideology and discontent. 
London Collier Macmillan. 
Ardener, S. (1983). Arson, nudity and bombs among the Canadian Doukhobors: A question of 
identity. In G.M. Breakwell (Ed.), Threatened identities. Chichester: Wiley. 
Ardrey, R. (1967). T
he territorial imperative. London: Collins. 
Argyle, M. (1992). The social psychology o
f
 everyday lif
e. London: Routledge. 
Asch, S.E. (1952). Social psychology. Englewood Cliffs, NJ: Prentice Hall. 


Re
f
erences 
231 
Asch, S.E. (1956). Studies on independence and conformity: A minority of one against a 
unanimous majority. Psychological Monographs, 70, 456. 
Asch, S.E., & Zuckier, H. (1984). Thinking about persons. Journal o
f
 Personality and Social 
Psychology, 46, 1230- 1240. 
Ayesteran, S. (Ed.). (1985). Psicologia y enf
ermedad mental: ldeologia y representacion social de 
la enf
ermedad mental. Bilbao: Universidad del Pais Vasco. 
Bakan, D. (1966). The duality o
f
 human existence. Boston: Beacon. 
Ball, P., Giles, H., & Hewstone, M. (1984). Secondary language acquisition: The intergroup 
theory with catastrophic dimensions. In H. Tajfel (Ed.), The social dimension: European 
developments in social psychology, Vol. 2. Cambridge: Cambridge University Press. 
Barsalou, L.W. (1983). Ad hoc categories. Memory and Cognition, 11, 2 1 1 -227. 
Barsalou, L. 
W. (1987). The instability of graded structure: Implications for the nature of 
concepts. In U. Neisser (Ed.), Concepts and conceptual development: Ecological and 
intellectual f
actors in categorization. Cambridge: Cambridge University Press. 
Barsalou, L.W. (1989). Intraconcept similarity and its implications for interconcept similarity. 
In S. Vosniadou & A. Ortony (Eds.), Similarity and analogical reasoning. Cambridge: 
Cambridge University Press. 
Barsalou, L.W., & Sewell, D.R. (1984). Constructing representations o
f
 categories f
rom diff
erent 
points o
f
 view (Emory Cognition Projects Report No. 2). Atlanta: Emory University. 
Barsalou, L.W., Sewell, D.R., & Ballato, S.M. (1986). Assessing the stability o
f
 category 
representations with graded structure. Unpublished manuscript. 
Bar-Tal, D. (1986). The Masada Syndrome: A case of central belief. In N. Milgram (Ed.), 
Stress and coping in time o
f
 war. New York: Brunner/Maze!' 
Bar-Tal, D. (1990). Group belie
f
s: A conce
ption f
or analyzing group structure, processes, and 
behavior. New York: Springer Verlag. 
Bar-Tal, D. (1993). American convictions about conflictive USA-USSR relations: A case of 
group beliefs. In S. Worchel & 1. Simpson (Eds.), Conflict between people and peoples. 
Chicago: Nelson-Hall. 
Bar-Tal, D. (in press). Societal belie
f
s
 o
f
 ethos: A social psychological analysis o
f
 society. 
London: Sage. 
Bar-Tal, D., Raviv, A., & Brosh, M. (1991). Perception of epistemic authority and attribution 
for the choice as a function of knowledge area and age. European Journal o
f
 Social 
Psychology, 21, 477-492. 
Bar-Tal, Y. (1989). Can leaders change followers' stereotypes? In D. Bar-Tal, C. Grauman, 
A.W. Kruglanski & W. Stroebe (Eds.), Stereotyping and pre
judice: Changing conceptions. 
New York: Springer Verlag. 
Bellelli, G. (Ed.). (1 987). La representation sociale de la maladie mentale. Naples: Liguori. 
Bellelli, G. (Ed.). (1994). L'altra malattia: Come la societa pensa la malattia mentale. Naples: 
Liguori. 
Bern, D.J. (1967) Self perception: An alternative interpretation of cognitive dissonance 
phenomena. Psychological Review, 74, 1 83-200. 
Bern, D.l. (1970). Belie
f
s, attitudes, and human a
f
f
airs. Belmont, CA: Brooks/Cole. 
Bern, S.L. (1981). Gender schema theory: A cognitive account of sex typing. Psychological 
Review, 88, 354-364. 
Bern, S.L. (1985). Androgyny and gender schema theory: A conceptual and empirical 
integration. In T.B. Soderberegger (Ed.), Psychology and gender: Nebraska Symposium on 
Motivation, Vol. 32. Lincoln: University of Nebraska Press. 
Billig, M. (1985). Prejudice, categorization and particularization: From a perceptual to a 
rhetorical approach. European Journal o
f
 Social Psychology, 15, 79-103. 
Billig, M., & Tajfel, H. (1973). Social categorization and similarity in intergroup behaviour. 
European Journal o
f
 Social Psychology, 3, 27-52. 
Boltanksi, 1.L. (1982). Les cadres: La f
ormation d'un groupe social. Paris: Minuit. 
Bomstein, G., Crum, L., Wittenbraker, 1., Harring, K., Insko, C.A., & Thibaut, 1. (l983a). On 
the measurement of social orientation in the minimal group paradigm. European Journal o
f
 
Social Psychology, 13, 321 -350. 


232 
Social identity: international perspectives 
Bomstein, G., Crum, L., Wittenbraker, 1., Harring, K., Insko, C.A., & Thibaut, 1. (1983b). 
Reply to Turner's comments. European Journal o
f
 Social Psychology, 13, 369-381 .  
Bourhis, R.Y., Gagnon, A., & MOise, L.C. (1 994). Discrimination et relation intergroupe. In 
R.Y. Bourhis & l.-P. Leyens (Eds.), Stereotypes, discrimination et relations intergroupes. 
Brussels: Mardaga. 
Brewer, M.B. (1979). Ingroup bias in the minimal intergroup situation: A cognitive­
motivational analysis. Psychological Bulletin, 86, 307-324. 
Brewer, M.B. (1988). A dual process model of impression formation. In T.K. Srull & R.S. 
Wyer (Eds.), Advances in social cognition, Vol. 1. Hillsdale, Nl: Erlbaum. 
Brewer, M.B. (1991). The social self: On being the same and different at the same time. 
Personality and Social Psychology Bulletin, 1 7, 475-482. 
Brewer, M.B., Dull, Y., & Lui, L. (1981). Perceptions of the elderly: Stereotypes as prototypes. 
Journal o
f
 Personality and Social Psychology, 41, 656-670. 
Brewer, M.B., Ho, H., Lee, 1., & Miller, N. (1987). Social identity and social distance among 
Hong Kong school children. Personality and Social Psychology Bulletin, 13, 1 56-165. 
Brewer, M.B., & Lui, L. (1984). Categorization of the elderly by the elderly: Effects of 
perceiver's category membership. Personality and Social Psychology Bulletin, 10, 585-595. 
Brewer, M.B., & Silver, M. (1978). Ingroup bias as a function of task characteristics. European 
Journal o
f
 Social Psychology, 8, 393-400. 
Brown, R. (1954). Mass phenomena. In G. Lindzey (Ed.), Handbook o
f
 social psychology, Vol. 
2. Cambridge, MA: Addison-Wesley. 
Brown, R.l. (1988). Group processes: Dynamics within and between groups. Oxford: Blackwell. 
Brown, R.l., Hinkle, S., Ely, P.G., Fox-Cardamone, L., Maras, P., & Taylor, L.A. (1992). 
Recognising group diversity: Individualist-collectivist and autonomous-relational social 
orientations and their implications for intergroup processes. British Journal o
f
 Social 
Psychology, 31, 327-342. 
Brown, R.1., & Ross, G.F. (1982). The battle for acceptance: An investigation into the 
dynamics of intergroup behaviour. In H. Tajfel (Ed.), Social identity and intergrou
p 
relations. Cambridge: Cambridge University Press. 
Brown, R.l., & Turner, 1.C. (1979). The criss-cross categorization effect in intergroup 
discrimination. British Journal o
f
 Social and Clinical Psychology, 18, 371 -383. 
Brown, R.l., & Turner, 1.C. (1981). Interpersonal and intergroup behaviour. In 1.C. Turner & 
H. Giles (Eds.), I
ntergroup behaviour. Oxford: Blackwell. 
Bruner, 1.S. (1957). On perceptual readiness. Psychological Review, 64, 123-151. 
Burnstein, E. (1982). Persuasion as argument processing. In H. Brandstatter, 1. Davis & G. 
Stocker-Kreichgauer (Eds.), Group decision-making. London: Academic Press. 
Butera, F., Huguet, P., Mugny, G., & Perez, 1.A. (1994). Socio-epistemic conflict and 
constructivism. Swiss Journal o
f
 Psychology, 53, 229-239. 
Butera, F., Legrenzi, P., Mugny, G., & Perez, 1.A. (1990-1991). Influence sociale et 
raisonnement. Bulletin de Psycho 
logie, 45, 144-154. 
Byrne, D. (1969). Attitudes and attraction. In L. Berkowitz (Ed.), Advances in ex
perimental 
social psychology, Vol. 4. New York: Academic Press. 
Byrne, D. (1971). The attraction problem. New York: Academic Press. 
Caddick, B. (1982). Perceived illegitimacy and intergroup relations. In H. Tajfel (Ed.), Social 
identity and intergroup relations. Cambridge: Cambridge University Press. 
Campbell, D.T. (1958). Common fate, similarity, and other indices of the status of aggregates 
of persons as social entities. Behavioral Science, 3, 14-25. 
Campbell, D.T. (1965). Ethnocentric and other altruistic motives. In D. Levine (Ed.), Nebraska 
Symposium on Motivation, Vol. 13. Lincoln: University of Nebraska Press. 
Camps, Y. (1993). Parado
jas del individualismo. Barcelona: Critica. 
Cannon, T. (1973). Black Panther party program: What we want, and what we believe. In R.R. 
Evans (Ed.), Social Movements. Chicago: Rand McNally. 
Cantor, N., & Mischel, W. (1979). Prototypes in person perception. In L. Berkowitz (Ed.), 
Advances in experimental social psychology, Vol 12. New York: Academic Press. 
Capozza, D., Bonaldo, E., & Di Maggio, A. (1982). Problems of identity and social conflict: 


Re
f
erences 
233 
Research on ethnic groups in Italy. In H. Tajfel (Ed.), Social identity and intergroup 
relations. Cambridge: Cambridge University Press. 
Carroll, 0.1., & Chang, J.J. (1970). Analysis of individual differences in multidimensional 
scaling via N-way generalisation of 'Eckart-Young' decomposition. Psychometrika, 35, 
283-319. 
Cartwright, D. (1968). The nature of group cohesiveness. In D. Cartwright & A. Zander (Eds.), 
Group dynamics (3rd ed.). London: Tavistock; New York: Harper & Row. 
Cartwright, D., & Zander, A. (Eds.). (1968). Group dynamics (3rd ed.). London: Tavistock; 
New York: Harper & Row. 
Carver, c., Antoni, M., & Scheier, M. (1985). Self-consciousness and self-assessment. Journal 
o
f
 Personality and Social Psychology, 48, 1 1 7-124. 
Clark, R.D., III, & Maass, A. (1988a). Social categorization in minority influence: The case of 
homosexuality. European Journal o
f
 Social Psychology, 18, 347-364. 
Clark, R.D., HI, & Maass, A. (I 
988b). The role of social categorization and perceived source 
credibility in minority influence. European Journal o
f
 Social Psychology, 18, 381 -394. 
Codol, J.-P. (1975). On the so-called 'superior conformity of the self' behaviour: Twenty 
experimental investigations. European Journal o
f
 Social Psychology, 5, 457-501 .  
Codol, J.-P. (1979). Semblables et difJerents: Recherches sur la quete de la similitude et de la 
difJ
erenciation sociale. Thesis, Universite de Provence, Aix-en-Provence. 
Codol, J.-P. (1982). Diff
erentiating and non-differentiating behavior: A cognitive approach to 
the sense of identity. In J.-P. Codol & J.-P. Leyens (Eds.), Cognitive analysis o
f
 social 
behavior. The Hague: Martinus Nijhoff. 
Codol, J.-P. (1984a). On the system of representations in an artificial social situation. In R.M. 
Farr & S. Moscovici (Eds.), Social representations. Cambridge: Cambridge University Press. 
Codol, J.-P. (1 984b). La perception de la similitude interpersonnelle: Influence de 
I'appartenance categorielle et du point de ref
erence de la comparaison. L'Annee 
Psychologique, 84, 43-56. 
Codol, J.-P. (1987). Comparability and incomparability between oneself and others: Means of 
differentiation and comparison ref
erence point. CaMers de Psychologie Cognitive, 7, 1- 18. 
Codol, J.-P. (1990). Studies on self-centered assimilation processes. In J.P. Caverni, J.M. Fabre 
& M. Gonzalez (Eds.), Cognitive biases. Amsterdam: North Holland. 
Condor, S. (1989). 'Biting into the future': Social change and the social identity of women. In S. 
Skevington & D. Baker (Eds.), The social identity o
f
 women. London: Sage. 
Condor, S. (1 990). Social stereotypes and social identity. In D. Abrams & M.A. Hogg (Eds.), 
Social identity theory: Constructive and critical advances. Hemel Hempstead: Harvester 
Wheatsheaf. 
Condor, S., Brown, R.1., & Williams, J. (1987). Social identification and intergroup behaviour. 
Quarterly Journal o
f
 Social Science Af
f
airs, 3, 299-317. 
Converse, P.E. (1964). The nature of the belief systems in mass pUblics. In D.E. Apter (Ed.), 
Ideology and discontent. New York: Free Press. 
Coutant, D., Worchel, S., & Grossman, M. (in preparation). Minority influence over time. 
Coutant-Sassic, D. (1991). DifJ
erential cognitive processing o
f
 a ma
jority vs. a minority in
fluence 
on the same or difJ
erent task. Unpublished master's thesis, Texas A&M University. 
Cowan, G., & Hoffman, C.D. (1986). Gender stereotyping in young children: Evidence to 
support a concept-learning approach. Sex Roles, 14, 2 1 1 -224. 
Crano, W.O. (1989). Judgmental sub
jectivity, objectivity, and minority influence. Paper 
presented to the third workshop on minority influence, Perugia, Italy, June. 
Crocker, J., & Luhtanen, R. (1990). Collective self-esteem and ingroup bias. Journal o
f
 
Personality and Social Psychology, 58, 323-338. 
Crocker, J., Luhtanen, R., Blaine, B., & Broadnax, S. (1 994). Collective self-esteem and 
psychological well-being among white, Black and Asian college students. Personality and 
Social Psychology Bulletin, 20, 503-51 3. 
Crosby, F., Cordova, D., & Jaskar, K. (1994). On the failure to see oneself as disadvantaged: 
Cognitive and emotional components. In M.A. Hogg & D. Abrams (Eds.), Motivation: 
Social psychological perspectives. Hemel Hempstead: Harvester Wheatsheaf. 


234 
Social identity: international pers
pectives 
Cylwik, A. (1992). Przychylnosc wobec innych narodowosci i mniejszosci narodowych a schematy 
la/My. Unpublished master's thesis, Wydzial Psychologii, Uniwersytet Warzawski, Warsaw. 
Damon, W., & Hart, D. (1982). The development of self-understanding from infancy through 
adolescence. Child Development, 53, 841-864. 
Deaux, K. (1992). Personalizing identity and socializing self. In G. Breakwell (Ed.), Social 
psychology o
f
 identity and the self
-conce
pt. London: Academic Press. 
Deaux, K. (1993). Reconstructing social identity. Personality and Social Psychology Bulletin, 
19, 4-12. 
Deconchy, J.P. (1984). Rationality and social control in orthodox systems. In H. Tajfel (Ed.), 
The social dimension: European developments in social psychology, Vol. 1. Cambridge: 
Cambridge University Press. 
Delumeau, J.J. (1993). La peur en Occident. Paris: Hachette: Pluriel. 
Deschamps, J.-C. (1972-1973). Imputation de la responsabilite de I'echec (ou de la reussite) et 
categorisation sociale. Bulletin de Psychologie, 26, 794-806. 
Deschamps, J.-C. (1977). L'attribution et la categorisation sociale. Berne: Peter Lang. 
Deschamps, J.-C. (1979). Diff
erenciation categorielle et differenciation de soi par rapport a 
autrui. Recherches de Psychologie Sociale, 1, 29-38. 
Deschamps, J.-c. (1980). L'identite sociale et les rapports de domination. Revue Suisse de 
Sociologie, 1, 1 1 1-122. 
Deschamps, J.-c. ( l982a). Differenciation entre soi et autrui et entre groupes: Recherches sur la 
'covariation' entre les differenciations inter-individuelles et inter-groupes. In J.-P. Codol & 
J.-P. Leyens (Eds.), Cognitive analysis o
f
 social behavior. The Hague: Martinus Nijhoff. 
Deschamps, J.-C. (1982b). Social identity and relations of power between groups. In H. Tajfe\ 
(Ed.), Social identity and intergroup relations. Cambridge: Cambridge University Press. 
Deschamps, J .-c. (1983-1984). Categorisation sociale et differenciations interindividuelles. 
Bulletin de Psycho 
logie, 37, 489-500. 
Deschamps, J.-c. (1984). Identite socia 
Ie et differenciation categorielle. Cahiers de Psychologie 
Cognitive, 4, 449-474. 
Deschamps, J.-C. (1991). Identites, appartenances sociales et differenciations individuelles. 
Cahiers Internationaux de Psychologie Sociale, 9-10, 49-6\. 
Deschamps, J.C., Clemence, A., & Roux, P. (1983). Differ
enciations inter-groupes et inter­
individuelles. Paper presented at Congreso della Divisione di Psicologia della SIPS, Padua. 
Deschamps, J.-C., & Doise, W. (1978). Crossed category memberships in intergroup relations. 
In H. Tajfel (Ed.), Diff
erentiation between social groups: Studies in the social psychology o
f
 
intergroup relations. London: Academic Press. 
Deschamps, J.-c., & Lorenzi-Cioldi, F. (1981). 'Egocentrisme' et 'sociocentrisme' dans les 
relations entre groupes: Recherche experimentale. Revue Suisse de Psychologie Pure et 
Applique, 40, 108-131. 
Deschamps, J.-c., & Serino, C. (1996). Mondo al singolare: Pros
pettive socio-cognitive 
sull'identita. Naples: Liguori. 
Deschamps, J.-c., & Volpato, C. (1984). Identita individuale nelle relazioni tra gruppi. Giornale 
Italiano di Psicologia, 11, 275-301 .  
Deutsch, M. (1968). The eff
ects of cooperation and competition upon group process. I n  D. 
Cartwright & A. Zander (Eds.), Group d
ynamics (3rd ed.). London: Tavistock; New York: 
Harper & Row. 
Deutsch, M., & Gerard, H.B. (1955). A study of normative and informational social influence 
upon individual judgment. lournal o
f
 Abnormal and Social Psychology, 51, 629-636. 
Devos, T., Comby, L., & Deschamps, J.-c. (\995). Appartenances sociales et jugement de 
variabilite intra-groupe. Cahiers Internationaux de Psychologie Sociale, 27, 9-43. 
Diamond, M. (1976). The American idea of man: The view from the founding. In T. Kristol & 
P. Weaver (Eds.), The Americans: 1976, Vol. 2. Lexington, MA: Lexington Books. 
Diehl, M. (\990). The minimal group paradigm: Theoretical explanations and empirical 
findings. In W. Stroebe & M. Hewstone (Eds.), European review o
f
 social 
psychology, Vol. 1. 
Chichester: Wiley. 


Re
f
erences 
235 
Doise, W. (1969). Strategies de 
jeu a I'interieur et entre des groupes de nationalites diff
erentes. 
Bulletin du Centre d'Etudes et Recherches Psycho 
techniques, 18, 1 3-26. 
Doise, W. (1976). L'articulation psychosociologique et les relations entre groupes. Brussels: De 
Boeck. 
Doise, W. (1986). Les representations sociales: Definition d'un concept. In W. Doise & A. 
Palmonari (Eds.), L'erude des representations sociales. Paris: Delachaux et NiestIe. 
Doise, W. (1988). Individual and social identities in intergroup relations. European Journal o
f
 
Social Psychology, /8, 99-1 1 \ .  
Doise, W., Clemence, A., & Lorenzi-Cioldi, F .  (1993). The quantitative anal
ysis o
f
 social 
representations. Hemel Hempstead: Harvester Wheatsheaf. 
Doise, W., Csepeli, G., Dann, H.D., Gouge, C., Larsen, K., & Ostell, A. (1972). An experi­
mental investigation into the formation of intergroup representations. European Journal o
f
 
Social Psychology, 2, 202-204. 
Doise, W., Deschamps, J.-c., & Meyer, G. (1 978). The accentuation of intracategory 
similarities. In H. Tajfel (Ed.), Diff
erentiation between social groups: Studies in the social 
psychology o
f
 intergroup relations. London: Academic Press. 
Doise, W., Gachoud, J.P., & Mugny, G. (1986). Influence directe et indirecte entre groupes 
dans des choix esthetiques. Cahiers de Psychologie Cognitive, 6, 283-30\ .  
Doise, W., & Lorenzi-Cioldi, F .  (1990). Levels of analysis and social identity. In D .  Abrams & 
M.A. Hogg (Eds.), Social identity theory: Constructive and critical advances. Hemel 
Hempstead: Harvester Wheatsheaf. 
Doise, W., & Lorenzi-Cioldi, F. (1991). L'identite comme representation sociale. In V. 
Aebischer, J.P. Deconchy & E.P. Lipiansky (Eds.), Ideologie et representations sociales. 
Cousset: Delval. 
Doise, W., & Palmonari, A. (1986). L'etude des representations sociales. Paris: Delachaux et 
Niestle. 
Doob, L.W. (1964). Patriotism and nationalism. New Haven, CT: Yale University Press. 
Doosje, B., Ellemers, N. & Spears, R. (1996). Perceived intragroup variability as a function of 
group status and identification. Journal o
f
 Experimental Social Psychology, 3/, 410-436. 
Dumont, L. (1987). Ensayos sobre el individualismo: Una pers
pectiva antropologica sobre la 
ideologia mode 
rna. Madrid: Alianza. 
Dunbar, R. (1992). Why gossip is good for you. New Scientist, 21 November, 28-3\ .  
Dunning, D., Perie, M., & Story, A.L. (1991). Self-serving prototypes of social categories. 
Journal o
f
 Personality and Social Psychology, 6/, 957-968. 
Durand Delvigne, A. (\ 992). Identite et modides sexues des personnes. These de Doctorat d'Etat 
de Lettres et Sciences H 
umaines, Paris. 
Durkheim, E. (1898). Representations individuelles et representations collectives. Revue de 
Metaphysique, 6, 274-302. 
Duveen, G., & Lloyd, B. (1990). Social representations and the development o
f
 knowledge. 
Cambridge: Cambridge University Press. 
Eagly, A.H., & Chaiken, S. (1993). The psychology o
f
 attitudes. New York: Harcourt Brace 
Jovanovich. 
Echebarria, A., Elajabarrieta, F., Valencia, J., & Villareal, M. (1992). Representations sociales 
de l'Europe et identites sociales. Bulletin de Psychologie, 45, 280-288. 
Eiser, J.R. (1983). Psicologia sociale cognitivista. Bologne: II Mulino. (Original English 
publication: Maidenhead: McGraw-Hill, 1980.) 
Ellemers, N., Wilke, H., & van Knippenberg, A. (1993). Effects of the legitimacy of low group 
or individual status on individual and collective status-enhancement strategies. Journal o
f
 
Personality and Social Psychology, 64, 766-778. 
Emler, N., & Hopkins, N. (1990). Reputation, social identity and the self. In D. Abrams & 
M.A. Hogg (Eds.), Social identity theory: Constructive and critical advances. Hemel 
Hempstead: Harvester Wheatsheaf. 
Emler, N., & Reicher, S. (1995). Adolescence and delinquency. Oxford: Blackwell. 
Epstein, A.L. (1978). Ethos and identity. London: Tavistock. 
Epstein, S. (\ 980). The self-concept: A review and the proposal of an integrated theory of 


236 
Social identity: international pers
pectives 
personality. In E. Staub (Ed.), Personality: Basic issues and current research. Englewood 
Cliffs, NJ: Prentice Hall. 
Erikson, E. (1956). The problem of ego identity. Journal o
f
 the American Psychiatric 
Association, 1, 58-121. 
Erikson, E. (1977). Childhood and society. London: Paladin. 
Etaugh, C, & Riley, S. (1983). Evaluating competence of women and men: Effects of marital 
and parental status and occupational sex-typing. Sex Roles, 9, 943-952. 
Etzioni, A. (1975). Complex organizations (rev. ed.). New York: Free Press. 
Falomir, J.M., Mugny, G., & Perez, I.A. (1996). Le paradoxe du denio In J.-C Deschamps & 
J.-L. Beauvois (Eds.), Des attitudes aux attributions. Grenoble: Presses Universitaires de 
Grenoble. 
Farr, R.M., & Moscovici, S. (1984). Social representations. Cambridge: Cambridge University 
Press. 
Fells, J., & Niznik, J. (1992). Conclusion: What is Europe? International Journal o
f
 Sociology, 
26, 201 -206. 
Festinger, L. (1950). Informal social communication. Psychological Review, 57, 271-282. 
Festinger, L. (1954). A theory of social comparison processes. Human Relations, 7, 1 17-140. 
Festinger, L., Riecken, H.W., & Schachter, S. (1956). W
hen prophecy f
ails. Minneapolis: 
University of Minnesota Press. 
Festinger, L., & Thibaut, J. (1951). Interpersonal communication in small groups. Journal o
f
 
Abnormal and Social Psychology, 46, 92-99. 
Festinger, L., Schachter, S., & Back, K. (1950). Social pressures in inf
ormal groups: A stud
y o
f
 
human f
actors i
n
 housing. New York: Harper. 
Fiske, S.T., & Neuberg, S.L. (1990). A continuum of impression formation, from category­
based to individuating processes: Influences of information and motivation on attention and 
interpretation. In M.P. Zanna (Ed.), Advances in ex
perimental social psychology, Vol. 23. 
New York: Academic Press. 
Fiske, S.T., & Pavelchak, M.A. (1986). Category-based versus piecemeal-based aff
ective 
responses: Developments in schema-triggered affect. In R.M. Sorrentino & E.T. Higgins 
(Eds.), Handbook o
f
 motivation and cognition: Foundations o
f
 social behavior, Vol. 1. New 
York: Guilford. 
Forgas, J.P. (1979). Social episodes: T
he stud
y o
f
 interaction routines. New York: Academic 
Press. 
Forsyth, D.R. (\995). Our social world. Pacific Grove, CA: Brooks/Cole. 
Fraser, S.L., & Lord, R.G. (l988). Stimulus prototypicality and general leadership impressions: 
Their role in leadership and behavioural ratings. Journal o
f
 Psychology, 122, 291-303. 
Gaertner, S.L., Dovidio, J.F., Anastasio, P.A., Bachman, B.A., & Rust, M.C (1993). The 
common ingroup identity model: Recategorization and the reduction of intergroup bias. In 
W. Stroebe & M. Hewstone (Eds.), European review o
f
 social psychology, Vol. 4. Chichester: 
Wiley. 
Gaertner, S.L., Mann, J., Murrell, A., & Dovidio, J.F. (\989). Reducing intergroup bias: The 
benefits of recategorization. Journal o
f
 Personality and Social Psychology, 57, 239-249. 
Garza, R.T., & Herringer, L.G. (\986). Social identity: A multidimensional approach. Journal 
o
f
 Social Psychology, 127, 299-308. 
Gergen, K.J. (1965). Interaction goals and personalistic feedback as factors affecting the 
presentation of the self. Journal o
f
 Personality and Social Psychology, 1, 413-424. 
Gergen, KJ. (1982). From self to science: What is there to know? In I.M. Suls (Ed.), 
Psychological pers
pectives on the self. Hillsdale, Nl: Erlbaum. 
Gillin, J. (\ 955). National and regional cultural values in the United States. Social Forces, 34, 
107- 1 1 3. 
Goethals, G.R. (1972). Consensus and modality in the attribution process: The role of 
similarity and information. Journal o
f
 Personality and Social Psychology, 21, 84-92. 
Goffman, E. (I 956}. The presentation o
f
 self in everyday lif
e. Edinburgh: Social Science 
Research Centre - University of Edinburgh. 
Goffman, E. (\ 968). Stigma. Englewood, Cliffs: Prentice Hall. 


Re
f
erences 
237 
Gol̗b, A., & Reykowski, 1. (1985). Studia nad rozwojem standardow ewaluatywnych. Wroclaw: 
Ossolineum. 
Goldberg, L.R. (1990). An alternative 'Description of Personality': The big-five factor 
structure. Journal o
f
 Personality and Social Psychology, 59, 1216-1229. 
Gottman, 1.M. (1981). Time series analysis: A comprehensive introduction f
or social scientists. 
Cambridge: Cambridge University Press. 
Greenwald, A.G, & Banaji, M. (1995). Implicit social cognition: Attitudes, self-esteem and 
stereotypes. Psychological Review, 12, 1-19. 
Greenwald, A.G., & Pratkanis, A.R. (1984). The self. In R.S. Wyer & T.K. Srull (Eds.), 
Handbook o
f
 social cognition, Vol. 1. Hillsdale, NJ: Erlbaum. 
Gudkyunst, W. (1988). Culture and intergroup processes. In M. Bond (Ed.), T
he cross-cultural 
challenge to social psychology. Newbury Park, CA: Sage. 
Guimond, S., & Tougas, F. (1994). Sentiment d'injustice et action collective: La privation 
relative. In R.Y. Bourhis & J.-P. Leyens (Eds.), Stert!otypes, discrimination et relations 
intergroupes. Brussels: Mardaga. 
Hamilton, V.L., & Rauma, D. (1995). Social psychology of deviance and law. In K.S. Cook, 
G.A. Fine & 1.S. House (Eds.), Sociological pers
pectives on social psychology. Boston: Allyn 
& Bacon. 
Hardoin, M., & Codol, J.-P. (1984). Description de soi et d'autrui: Influence de l'ordre de 
description sur les categories de reponse utilisees. Cahiers de Psychologie Cognitive, 4, 295-
302. 
Harkins, S., & Petty, R.E. (1981). The multiple source effect in persuasion: The effects of 
distraction. Personality and Social Psychology Bulletin, 21, 84-92. 
Harkins, S., & Petty, R.E. (1983). Social context effects in persuasion: The effects of multiple 
source and multiple targets. In P.B. Paulus (Ed.), Basic group process. New York: Springer 
Verlag. 
Harkins, S., & Petty, R.E. (1987). Inf
ormational utility and the multiple-source effect. Journal 
o
f
 Personality and Social Psychology, 52, 260-268. 
Harrison, A.A., Moore, M., & Rucker, M.H. (1985). Further evidence on career and family 
compatibility among eminent women and men. Archivio di Psicologia, Neurologia e 
Psichiatria, 46, 140-1 55. 
Harter, S. (1985). The self-perception pro
file f
or children: Revision o
f
 the Perceived Competence 
Scale f
or children. Manual, University of Denver. 
Haslam, S.A., McGarty, C, Oakes, P.J., Turner, J.C, & Onorato, R. (1995). Contextual shifts 
in the prototypicality of extreme and moderate outgroup members. European Journal o
f
 
Social Psychology, 25, 509-530. 
Haslam, S.A., Oakes, P.J., Turner, J.C, & McGarty, C (1996). Social identity, self­
categorization and the perceived homogeneity of in 
groups and outgroups: The interaction 
between social motivation and cognition. In R.M. Sorrentino & E.T. Higgins (Eds.), 
Handbook o
f
 motivation and cognition, Vol. 3. New York: Guilf
ord. 
Haslam, S.A., & Turner, 1.C (1992). Context-dependent variation in social stereotyping 2: The 
relationship between frame of ref
erence, self-categorization and accentuation. European 
Journal of Social Psychology, 22, 251-277. 
Haslam, S.A., Turner, J.C, Oakes, PJ., McGarty, C., & Hayes, B.K. (1992). Context­
dependent variation in stereotyping I :  The effects of intergroup relations as mediated by 
social change and frame of reference. European Journal of Social Psychology, 22, 3-20. 
Hayes, CH. (1926). Nationalism as a religion. In CH. Hayes, Essays on nationalism. New 
York: Macmillan, 1962. 
Hertz, F. (1944). Nationalism in history and politics. London: Routledge & Kegan Paul. 
Hewstone, M., Hantzi, A., & lohnston, L. (1991). Social categorization and person memory: 
The pervasiveness of race as an organizing principle. European Journal o
f
 Social Psychology, 
21, 517-528. 
Higgins, E.T., & King, G. (1981). Accessibility of social constructs: Information-processing 
consequences of individual and contextual variability. In N. Cantor & J.F. Kihlstrom (Eds.), 
Personality, cognition, and social interaction. Hillsdale, NJ: Erlbaum. 


238 
Social identity: international pers
pectives 
Hinkle, S., & Brown, R.J. (1990). Intergroup comparisons and social identity: Some links and 
lacunae. In D. Abrams & M.A. Hogg (Eds.), Social identity theory: Constructive and critical 
advances. Hemel Hempstead: Harvester Wheatsheaf. 
Hinkle, S., Brown, R.J., & Ely, P.G. (1992). Social identity theory processes: Some limitations 
and limiting conditions. Revista de Psicologia Social, Monogra
jico, 99- 1 1 1 .  
Hinkle, S., & Crook, K.F. (1987). Identification, importance o
f
 comparison dimensions and 
intergroup diff
erentiation. Paper presented at the International Conference on Social Identity, 
Exeter, UK, July. 
Hinkle, S., Fox-Cardamone, D.L., Taylor, L.A., & Crook, K.F. (1988). Identification, outgroup 
threat and dimensions o
f
 intergrou
p comparison. Paper presented at the Invitational 
Conference on Aggression, Conflict and Negotiation, Kill Devil Hills, NC. June. 
Hoffman, M.L. (1989). Empathy and pro 
social activities. In N. Eisenberg, J. Rekowski & E. 
Staub (Eds.), Social and moral values. Hillsdale, NJ: Erlbaum. 
Hofstede, G. (1980). Culture's consequences: International dif
f
erences in work-related values. 
Beverly Hills, CA: Sage. 
Hogg, M.A. (1987). Social identity and group cohesiveness. In l.e. Turner, M.A. Hogg, P.l. 
Oakes, S.D. Reicher & M.S. Wetherell, Rediscovering the social group: A self-categorization 
theory. Oxford: Blackwell. 
Hogg, M.A. (1992). The social psychology o
f
 group cohesiveness. Hemel Hempstead: Harvester 
Wheatsheaf. 
Hogg, M.A. (1993). Group cohesiveness: A critical review and some new directions. In W. 
Stroebe & M. Hewstone (Eds.), European review o
f
 social psychology, Vol. 4. Chichester: 
Wiley. 
Hogg, M.A. (1996). Intragroup processes: group structure, and social identity. In P. Robinson 
(Ed.), Social groups and identity: Developing the legacy o
f
 Henri T
a
j
f
el. Oxford: Butterworth 
Heinemann. 
Hogg, M.A., & Abrams, D. (1988). Social identifications: A social psychology o
f
 intergroup 
relations. London and New York: Routledge & Kegan Paul. 
Hogg, M.A., & Abrams, D. (1990). Social motivation, self-esteem and social identity. In D. 
Abrams & M. Hogg (Eds.), Social identity theory: Constructive and critical advances. Hemel 
Hempstead: Harvester Wheatsheaf. 
Hogg, M.A., Cooper-Shaw, L., & Holzworth, D.W. (in press). Studies of group prototypicality 
and depersonalized attraction in small interactive groups. Personality and Social Psychology 
Bulletin. 
Hogg, M.A., & Hardie, E.A. (1991). Social attraction, personal attraction, and self­
categorization: A field study. Personality and Social Psychology Bulletin, 1 7, 175-180. 
Hogg, M.A., & Hardie, E.A. (1992). Prototypicality, conformity and depersonalized attraction: 
A self-categorization analysis of group cohesiveness. British Journal o
f
 Social Psychology, 
31, 45-56. 
Hogg, M.A., & McGarty, C. (1990). Self-categorization and social identity. In D. Abrams & 
M.A. Hogg (Eds.), Social identity theory: Constructive and critical advances. Hemel 
Hempstead: Harvester Wheatsheaf. 
Hogg, M.A., & Turner, l.e. (l985a). When liking begets solidarity: An experiment on the role 
of interpersonal attraction in psychological group formation. British Journal o
f
 Social 
Psychology, 24, 267-281 .  
Hogg, M.A., & Turner, l.C. (1985b). Interpersonal attraction, social identification, and 
psychological group formation. European Journal o
f
 Social Psychology, 15, 51-66. 
Hogg, M.A., & Turner, J.e. (1987). Social identity and conformity: A theory of referent 
information influence. In W. Doise & S. Moscovici (Eds.), Current issues in European social 
psychology, Vol. 2. Cambridge: Cambridge University Press; Paris: Maison des Sciences de 
I'Homme. 
Hogg, M.A., Turner, J.C., & Davidson, B. (1990). Polarized norms and social frames of 
reference: A test of the self-categorization theory of group polarization. Basic and Applied 
Social Psychology, 11, 77-100. 


Re
f
erences 
239 
Holyoak, K.J., & Gordon, P.e. (1983). Social reference points. Journal o
f
 Personality and 
Social Psychology, 44, 881-887. 
Holyoak, K.J., & Gordon, P.e. (1984). Information processing and social cognition. In R.S. 
Wyer & T.K. Srull (Eds.), Handbook o
f
 social cognition, Vol. 1. Hillsdale, NJ: Erlbaum. 
Homans, G. (1950). The human group. New York: Harcourt Brace. 
Horwitz, M., & Rabbie, J.M. (1989). Stereotypes of groups, group members. and individuals in 
categories: A differential analysis. In D. Bar-Tal, e.F. Grumann, A. Kruglanski & W. 
Stroebe (Eds.), Stereotyping and 
pre
judice. New York: Springer Verlag. 
Hostetler, J.A. (1968). Amish society (rev. ed.). Baltimore: Johns Hopkins University Press. 
Hoyle, R.H., Pinkley, R.L., & Insko, e.A. (1989). Perceptions of behavior: Evidence of 
differing expectations for interpersonal and intergroup interactions. Personality and Social 
Psychology Bulletin, 15, 365-376. 
Huguet, P., Mugny, G., & Perez, J.A. (1990-1991). Influence sociale et processus de 
decentration. Bulletin de Psychologie, 45, 155-163. 
Hui, e.H. (1988). Measurement of individualism-collectivism. Journal o
f
 Research in 
Personality, 28, 17-36. 
Hui, e.H., & Triandis, H.e. (1986). Individualism-collectivism: A study of cross-cultural 
researchers. Journal o
f
 Cross-Cultural Psychology, 1 7, 225-248. 
Huici, e. (1 984). The individual and social functions of sex role stereotypes. In H. Tajfel (Ed.), 
The social dimension: European developments in social psychology, Vol. 2. Cambridge: 
Cambridge University Press. 
Huici, C. (1985). Relaciones intergrupales II: EI paradigma del Grupo Minimo. In e. Huici 
(Ed.), Estructura y procesos de Gru
po, Vol. 2. Madrid: UNED. 
Hurtig, M.e., & Pichevin, M.F. (1986). La dif
ference des sexes. Paris: Tierce. 
Hurtig, M.e., Pichevin, M.F., & Piolat, M. (1992). Sex and age as factors of asymmetry in the 
perceived similarity between two persons. In L. Arcuri & e. Serino (Eds.), Asymmetry 
phenomena in interpersonal comparison: Cognitive and social issues. Naples: Liguori. 
Inoff, G., Halverson, e.F., & Pizzigati, K.A.L. (1983). The influence of sex-role standards on 
children's self
- and peer-attributions. Sex Roles, 9, 1205-1222. 
Insko, e.A., Hoyle, R.H., Pinkley, R.L., Hong, G., Slim, R., Dalton, G., Lin, Y., Ruffin, P.P., 
Dardis, G.J., Bernthal, P.R., & Schopler, J. (1988). Individual-group discontinuity: The role 
of a consensus rule. Journal o
f
 Ex
perimental Social Psychology, 24, 505-5\9. 
Insko, e.A., Pinkley, R.L., Harring, K., Holton, B., Hong, G., Krams, D.S., Hoyle, R.H., & 
Thibaut, J. (1987). Beyond categorization to competition: Expectations of appropriate 
behavior. Representative Research in Social Psychology, 1 7, 5-36. 
Insko, C.A., Pinkley, R.L., Hoyle, R.H., Dalton, G., Hong, G., Slim, R., Landry, P., Holton, 
B., Ruffin, P.P., & Thibaut, J. (1987). Individual-group discontinuity: The role of 
intergroup contact. Journal o
f
 Ex
perimental Social Psychology, 23, 250-267. 
Insko, e.A., & Schopler, J. (1987). Categorization, competition, and collectivity. In e. 
Hendrick (Ed.), Review o
f
 personality and 
social psychology: Group processes, Vol. 8. Beverly 
Hills, CA: Sage. 
Insko, e.A., Schopler, J., Drigotas, S.M., Graetz, K.A., Kennedy, J., Cox, e., & Bomstein, G. 
(1993). The role of communication in interindividual-intergroup discontinuity. Journal o
f
 
Conflict Resolution, 37, 108-1 38. 
Insko, e.A., Schopler, J., Hoyle, R.H., Dardis, G.J., & Graetz, K.A. (1990). Individual-group 
discontinuity as a function of fear and greed. Journal o
f
 Personality and Social Psychology, 
58, 68-79. 
Insko, e.A., Schopler, J., Kennedy, J.F., Dahl, K.R., Gratez, K.A., & Drigotas, S.M. (1992). 
Individual-group discontinuity from the differing perspectives of Campbell's realistic group 
conflict theory and Tajfel and Turner's social identity theory. Social Psychological Quarterly, 
55, 272-291 . 
Jackson, L.A. (1993). Achieving positive social identity. Paper presented at the Invitational 
Conference on Self and Social Identity. Highland Park, FL, June. 
James, W. (\890). The princi
ples o
f
 psychology. New York: Holt. 
James, W. (1892). Psychology: A brie
f
er course. Greenwich, CT: Fawcett. 


240 
Social identity: international perspectives 
Janis, I.L. (1968). Group identification under conditions of external danger. In D. Cartwright & 
A. Zander (Eds.), Group d
ynamics (3rd ed.). London: Tavistock; New York: Harper & Row. 
Jarymowicz, M. (1982). Perception of self-others dissimilarity and signs of attractiveness or 
aversiveness of diversity from others. Polish Psychological Bulletin, 13, 39-44. 
Jarymowicz, M. (1987). Perceiving one's own individuality: Estimation and attractiveness o
f
 
one's own distinctiveness f
rom others. Warsaw: Wydawnictwa Uniwersytetu Warszawskiego. 
Jarymowicz, M. (l99Ia). The self-distinctive traits as personal identity attributes. CaMers de 
Psychologie Cognitive, 11, 679-695. 
Jarymowicz, M. (199Ib). La distinctive Soi-Nous-Autre comme base d'identite individuelle et 
socia 
Ie. Cahiers Internationaux de Psychologie Sociale, 9-10, 103-1 16. 
Jarymowicz, M. (I 
992a). Self, We and Other(s): Schemata distinctiveness and altruism. In P. 
Oliner, S. Oliner, L. Baron, D. Krebs & Z. Smoleska (Eds.), Embracing the other: 
Philosophical, psychological, and historical perspectives on altruism. New York: University of 
New York Press. 
Jarymowicz, M. ( l992b). Introduction to the Social Self 
Symposium: Searching 
f
or the social self 
concept. Paper presented to the EAESP/SESP Joint Meeting, LeuveniLouvain-la-Neuve, 
July. 
Jarymowicz, M. (l993a). Categorial vs cross-categorial social identity and outgroup involvement. 
Paper presented to the EAESP General Meeting, Lisbon, September. 
Jarymowicz, M. (I 993b). An attempt to operationalize 'social' versus 'personal identity' 
constructs. In M. Jarymowicz (Ed.), To know self - to understand others. Delf
t: Eburon. 
Jarymowicz, M. (1994a). W stron̘ indywidualnej podmiotowoSci i zblizen z innymi: 
Podmiotowe podstawy spolecznych identyfikacji. In M. Jarymowicz (Ed.), Poza 
egocentrycznt;l perspektywt;l widzenia siebie i swiata. Warsaw: Wydawnictwo Instytutu 
Psychologii PAN. 
Jarymowicz, M. (1994b). 0 formach umyslowego ujmowania MY i ich zwilizku ze 
spostrzeganiem Innych. In M. Jarymowicz (Ed.), Poza egocentrycznt;l perpsktywt;l widzenia 
siebie i swiata. Warsaw: Wydawnictwo Instytutu Psychologii PAN. 
Jarymowicz, M. (1994c). Poznawcza indiwiduacja a spoleczne identyfikacje. In M. Jarymowicz 
(Ed.), Poza egocentrycznt;l perspektywt;l widzenia siebie i swiata. Warsaw: Wydawnictwo 
Instytutu Psychologii PAN. 
Jarymowicz, M., & Greenwald, A.G. (1993). The anchoring function of the self in social 
judgments: Introduction. In M.C. Hurtig, M.F. Pichevin & M. Piolat (Eds.), Studies on the 
self and social cognition. London: World Scientific Publishing. 
Jodelet, D. (Ed.) (1989). Les representations sociales. Paris: Presses Universitaires de France. 
Jones, E.E., & DeCharms, R. (1967). Changes in social perception as a function of the personal 
relevance of behavior. Sociometry, 20, 175-185. 
Jones, E.E., & Gerard, H.B. (1967). Foundations o
f
 social psychology. New York: Wiley. 
Jones, E. E., Wood, G., & Quattrone, G. (1981 ). Perceived variability of personal characteristics 
in in-groups and out-groups: The role of knowledge and evaluation. Personality and Social 
Psychology Bulletin, 7, 523-528. 
Jung, C.G. (1943). The psychology of the unconscious. In Collected Works. Princeton, NJ: 
Princeton University Press. 
Karylowski, J. (1982). On the two types of altruistic behavior: doing good to feel good versus to 
make others feel good. In V. Derlega and J. Grzelak (Eds.), Cooperation and Helping 
Behavior. New York: Academic Press. 
Kedourie, E. (1960). Nationalism. London: Hutchinson. 
Kelley, H.H., & Stahelski, A. (1970). The social interaction basis of cooperators' and 
competitors' beliefs about others. Journal o
f
 Personality and Social Psychology, 16, 66-9\ .  
Kelley, H.H., & Thibaut, J.W. (1978). Interpersonal relations: A theory o
f
 interdependence. New 
York: Wiley. 
Kelly, J. (1988). Entertainment in individual and group behavior. In J.E. McGrath (Ed.), The 
social psychology o
f
 time. Newbury Park, CA: Sage. 
Kelman, H.C. (1969). Patterns of personal involvement in the national system: A social 


Re
f
erences 
241 
psychological analysis of political legitimacy. In 1.N. Rosenau (Ed.), International politics 
and f
oreign policy: A reader in research and theory. Glencoe, IL: Free Press. 
Kenny, D.A., & La Voie, L. (1985). Separating individual and group effects. Journal o
f
 
Personality and Social Psychology, 48, 339-348. 
Kiesler, CA., & Kiesler, S.B. (1969). Con
f
ormity. Reading, MA: Addison-Wesley. 
Killian, L.M. (1964). Social movements. In R.E.L. Faris (Ed.), Handbook o
f
 modern sociology. 
Chicago: Rand McNally. 
Kinder, D.R., & Sears, D.O. (1995). Public opinion and political action. In G. Lindzey & E. 
Aronson (Eds.), Handbook o
f
 social psychology. Vol. 2. New York: Random House. 
Klein, M. (1948). Contributions to psychoanalysis 1921-1945. London: Hogarth. 
Kluckhohn, C (1951). Values and value-orientations in the theory of action: An exploration in 
definition and classification. In T. Parsons & E. Shils (Ed.), Toward a general theory o
f
 
action. Cambridge, MA: Harvard University Press. 
Kobuszewska, E. (1989). Mechanizmy altruizmu w warunkach deindywiduacji. Unpublished 
master's thesis, Wydzial Psychologii, Uniwersytet Warszawski, Warsaw. 
Kohlberg, L. (1976). Moral stages and moralization: The cognitive-developmental approach. In 
T. Lickona (Ed.), Moral development and behavior. New York: Holt, Rinehart & Winston. 
Kohn, H. (1944). The idea of 
nationalism: A stud
y o
f
 its origins and background. New York: 
Macmillan. 
Kohn, H. (1962). The age o
f
 nationalism: The first era o
f
 global history. New York: Harper. 
Kotkin, M. (1983). Sex roles among married and unmarried couples. Sex Roles, 9, 975-985. 
Krech, D., Crutchfield, R.S., & Ballachey, E.L. (1962). Individual in society. New York: 
McGraw-Hill. 
Kruglanski, A.W. (1989). Lay epistemics and human knowledge: Cognitive and motivational 
bases. New York: Plenum. 
Kuhn, M.H., & McPartland, T.S. (1954). An empirical investigation of self-attitudes. American 
Sociological Review, 19, 68-76. 
Lalonde, R.N. (1992). The dynamics of group differentiation in the face of defeat. Personality 
and Social Psychology Bulletin, 18, 336-342. 
Lane, R.E. (1962). Political ideology. New York: Free Press. 
Lassaline, M.E., Wisniewski, E.l., & Medin, D.L. (1992). Basic levels in natural categories: Are 
all basic levels created equal? In B. Bums (Ed.), Percepts, concepts and categories. New 
York: Elsevier Science Publishers. 
Latani:, 8., & Darley, 1.M. (1970). T
he unres
ponsive bystander: Why doesn't he hel
p? New 
York: Appleton-Century-Crofts. 
Latane, B., Williams, K., & Harkins, S. (1979). Many hands make light the work: The causes 
and consequences of social loafing. Journal o
f
 Personality and Social Psychology, 37, 822-
832. 
LeBon, G. (1 896). The crowd. London: Unwin. 
L'Ecuyer, R. (1981). The development of the self-concept through the life span. In M.D. 
Lynch, A.A. Norem-Hebeisen & K. Gergen (Eds.), Self concept: Advances in research. 
Cambridge, MA: Ballinger. 
Lemaine, G. (1974). Social differentiation and social originality. European Journal o
f
 Social 
Psychology, 4, 17-52. 
Lemaine, G., Kastersztein, 1., & Personnaz, B. (1978). Social differentiation. In H. Tajfel (Ed.), 
Dif
f
erentiation between social groups: Studies in the social psychology o
f
 intergroup relations. 
London: Academic Press. 
Levine, I.M. (1980). Reactions to opinion deviance in small groups. In P.B. Paulus (Ed.), 
Psychology o
f
 group in
fluence. Hillsdale, Nl: Erlbaum. 
Levine, 1.M., & Moreland, R.L. (1994). Group socialization: Theory and research. In W. 
Stroebe and M. Hewstone (Eds.), European review o
f
 social psychology, Vol. 5. Chichester: 
Wiley. pp. 305-336. 
Le Vine, R.A., & Campbell, D.T. (1972). Ethnocentrism: T
heories o
f
 con
flict, ethnic attitudes 
and group con
flict. New York: Wiley. 


242 
Social identity: international pers
pectives 
Levy, S.B. (1975). Shifting patterns of ethnic identification among the Hassidim. In J.W. 
Bennet (Ed.), The new ethnicity: Pers
pectives 
f
rom ethnology. St Paul, MN: West Publishing. 
Lewin, K. (1948). Resolving social con
flicts. New York: Harper & Row. 
Lewin, K., Lippitt, R., & White, R. (1939). Patterns of aggressive behavior in experimentally 
created social climates. Journal o
f
 Psychology, 10, 271-299. 
Leyens, J.-P. (1983). Sommes nous tous des psychologues? Brussels: Mardaga. 
Leyens, J.-P., Yzerbyt, V.Y., & Schadron, G. (1994). Stereotypes and social cognition. London: 
Sage. 
Lindskold, S. (1983). Conflict and conciliation with groups and individuals. In A.P. Hare, V. 
Kent & M. Davies (Eds.), Small groups and social interaction, Vol. 2. New York: Wiley. 
Linville, P.W., & Jones, E.E. (1980). Polarized appraisals of out-group members. Journal o
f
 
Personality and Social Psychology, 38, 689-703. 
Linz, J.J. (1986). El con
flicto en Euskadi. Madrid: Espasa-Calpe. 
Lord, R.G., Foti, R.J., & De Vader, c.L. (1984). A test of leadership categorization theory: 
Internal structure, information processing, and leadership perceptions. Organizational 
Behaviour and Human Per
f
ormance, 34, 343-378. 
Lorenzi-Cioldi, F. (1988). Individus dominants et groupes domines: Images masculines et 
jeminines. Grenoble: Presses Universitaires de Grenoble. 
Lorenzo-Cioldi, F. (1993). They all look alike, but so do we . . .  
sometimes. Journal o
f
 Social 
Psychology, 32, 1 1 1-24. 
Lorenzi-Cioldi, F. (1994). Les androgynes. Paris: Presses Universitaires de France. 
Lorenzi-Cioldi, F., & Doise, W. (1990). Levels of analysis and social identity. In D. Abrams 
and M.A. Hogg (Eds.), Social identity theory, constructive and critical advances. Hemel 
Hempstead: Harvester Wheatsheaf. 
Lorenzi-Cioldi, F., & Doise, W. (1994). Identite personelle et identitee sociale. In R.Y. Bourhis 
& J.-P. Leyens (Eds.), Stereotypes, discrimination et relations intergroupes. Brussels: 
Mardaga. 
Lorenzi-Cioldi, F., Eagley, A.M., & Stewart, T. (1995). Homogeneity of gender groups in 
memory. Unpublished manuscript, University of Geneva. 
Lott, A.J., & Lott, B.E. (1965). Group cohesiveness as interpersonal attraction: A review of 
relationships with antecedent and consequent variables. Psychological Bulletin, 64, 259-309. 
Luthanen, R., & Crocker, J. (1 992). A collective self-esteem scale: Self-evaluation of one's 
social identity. Personality and Social Psychology Bulletin, 3, 302-318. 
Luria, A.R. (1 976). La storia sociale dei processi cognitivi. Florence: Giunti-Barbera. 
Maass, A., & Clark, R.D., III. (1984). Hidden impact of minorities: Fifteen years of minority 
influence research. Psychological Bulletin, 95, 428-450. 
Maass, A., Clark, R.D., III, & Haberkorn, G. (1982). The effects of differential ascribed 
category membership and norms on minority influence. European Journal o
f
 Social 
Psychology, 12, 89-104. 
Mackie, D.M., & Cooper, J. (1984). Attitude polarization: The effects of group membership. 
Journal o
f
 Personality and Social Psychology, 46, 575-585. 
Mackie, D.M., Worth, L.T., & Asuncion, A.G. (1990). Processing of persuasive in-group 
messages. Journal o
f
 Personality and Social Psychology, 58, 812-822. 
Makris-Botsaris, E., & Robinson, P.W. (1991). Harter's self-perception profile for children: A 
cross-cultural validation in Greece. Evaluation and Research in Evaluation, 5, 135-143. 
March, J.G., & Simon, H.A. (1958). Organizations. New York: Wiley. 
Marcia, J.E. (1966). Development and validation of ego identity status. Journal o
f
 Personality 
and Social Psychology, 3, 551 -558. 
Mariategui, J.C. (1925). EI hombre y el mito. In J.C. Mariategui, El alma matinal. Lima: 
Amatuam, 1 950. 
Markus, H. (1977). Self-schemas and processing information about the self. Journal o
f
 
Personality and Social Psychology, 35, 63-78. 
Markus, H. (1984). Possible selves. Paper presented at the Boston University Symposium for 
the Interdisciplinary Study of Personality. 


Re
f
erences 
243 
Markus, H., & Kitayama, S. (l99Ia). Culture and the self: Implications for cognition, emotion 
and motivation. Psychological Review, 98, 224-253. 
Markus, H., & Kitayama. S. (l99Ib). Cultural variations in the self-concept. In J. Strauss & 
G.R. Goethals (Eds.), The sel
f: An interdisc i
plinary approach. New York: Springer Verlag. 
Markus, H., & Smith, J. (1981). The influence of self-schemata on the perception of others. In 
N. Cantor & J. Kihlstrom (Eds.), Personality, cognition and social interaction. Hillsdale, NJ: 
Erlbaum. 
Marques, J.M. (1990). The black sheep effect: Outgroup homogeneity in social comparison 
settings. In D. Abrams & M.A. Hogg (Eds.), Social identity theory: Constructive and critical 
advances. Hemel Hempstead: Harvester Wheatsheaf. 
Marques, J.M., Abrams, D., & Paez, D. (1996). Unpublished manuscript. 
Marques, J.M., Brito, P.N., Correia, M.G., & Serodio, R.G. (1996). Evaluation of attractive 
and unattractive ingroup and outgroup members as a function of group uniformity. 
Unpublished manuscript. 
Marques, J.M., & Paez, D. (1994). The black sheep eff
ect: Social categorization, rejection of 
ingroup deviates, and perception of group variability. In M. Hewstone & W. Stroebe (Eds.), 
European review o
f
 social psychology, Vol. 5. Chichester: Wiley. 
Marques, I.M., Robalo, E.M., & Rocha, S.A. (1992). Ingroup bias and the black sheep effect: 
Assessing the impact of social identification and perceived variability on group judgments. 
European Journal o
f
 Social Psychology, 22, 331-352. 
Marques, J.M., & Yzerbyt, V.Y. (1988). The black sheep effect: Judgmental extremity towards 
ingroup members in inter- and intragroup situations. European Journal o
f
 Social Psychology, 
18, 287-292. 
Marques, J.M., Yzerbyt, V.Y., & Leyens, J.-P. (1988). The black sheep effect: Judgmental 
extremity towards ingroup members as a function of group identification. European Journal 
o
f
 Social Psychology, 18, 1-16. 
Martin, R. (I 988a). Ingroup and outgroup minorities: Diff
erential impact upon public and 
private response. European Journal o
f
 Social Psychology, 18, 39-52. 
Martin, R. (1988b). Minority influence and social categorization: A replication. European 
Journal of Social Psychology, 18, 369-373. 
Martin, R. (1988c). Minority influence and 'trivial' social categorization. European Journal o
f
 
Social Psychology, 18, 465-470. 
Maslow, A.H. (1954). Motivation and personality. New York: Harper & Row. 
McAdams, D.P. (1980). A thematic coding system for the intimacy motive. Journal o
f
 Research 
in Personality, 14, 412-432. 
McAdams, D.P. (1985). The 'Imago': A key narrative component of identity. In P. Shaver 
(Ed.), Self, situations and social behavior. Beverly Hills, CA: Sage. 
McCallum, D.M., Harring, K., Gilmore, R., Drenan, S., Chase, J., Insko, CA., & Thibaut, 1. 
(1985). Competition between groups and between individuals. Journal o
f
 Ex
perimental 
Social Psychology, 21, 301-320. 
McDougall, W. (1920). The group mind. London: Putnam. 
McGarty, C, & Turner, J.C (1992). The effects of categorization on social judgement. British 
Journal o
f
 Social Psychology, 31, 253-268. 
McGarty, C, Turner, J.C., Hogg, M.A., David, 8., & Wetherell, M.S. (1992) Group 
polarization as conformity to the prototypical group member. British Journal o
f
 Social 
Psychology, 31, 1-20. 
McGrath, J.E. (1988). Time and social psychology. In J.E. McGrath (Ed.), The social 
psychology o
f
 time. Newbury Park, CA: Sage. 
McGuire, W.J., & McGuire, C.V. (1988). Content and process in the experience of the self. In 
L. Berkowitz (Ed.), Advances in experimental social psychology, Vol. 21. New York: 
Academic Press. 
McPhail, C. (1991). The myth o
f
 the madding crowd. New York: Aldine de Gruyter. 
Mead, G.H. (1934). Mind, self and society. Chicago: University of Chicago Press. 
Medin, D.L. (1989). Concepts and conceptual structure. American Psychologist, 44, 1469-1481. 


244 
Social identity: international pers
pectives 
Medin, D.L., & Ortony, A. (1989). Psychological essentialism. In S. Vosniadou & A. Ortony 
(Eds.), Similarity and analogical reasoning. Cambridge: Cambridge University Press. 
Medin, D.L., & Wattenmaker, W.D. (1987). Category cohesiveness, theories, and cognitive 
archaeology. In U. Neisser (Ed.), Concepts and conceptual development: Ecological and 
intellectual f
actors in categorization. Cambridge: Cambridge University Press. 
Medin, D.L., Wattenmaker, W.D., & Hampson, S.E. (1987). Family resemblance, conceptual 
cohesiveness and category construction. Cognitive Psychology, 19, 242-279. 
Mervis, CB., & Rosch, E. (1981). Categorization of natural objects. In M.R. Rozenzweig & 
L.W. Porter (Eds.), Annual review o
f
 psychology, Vol. 32. Palo Alto, CA: Annual Reviews. 
Messick, D., & Mackie, D. (1989). Intergroup relations. Annual Review o
f
 Psychology, 40, 45-
81. 
Mlicki, P.P., & Ellemers, N. (1996). Being different or being better? National stereotypes and 
identifications of Polish and Dutch students. European Journal o
f
 Social Psychology, 26, 97-
1 14. 
Monteil, J.M. (1992). Intergroup differentiation and individuation: The effect of social 
deprivation. European Bulletin o
f
 Cognitive Psychology, 12, 189-203. 
Montero, M. (1987). A traves del espejo: Una aproximacion teorica al estudio de la conciencia 
social en America Latina. In M. Montero (Ed.), Psicologia Politica Latinoamerica. Caracas: 
Panapo. 
Montero, M. ( 1990). Autoimagen de los venezolanos: Lo positivo y 10 negativo. In H. 
Riquelme (Ed.), Buscando America Latina. Caracas: Nueva Sociedad. 
Morales, J 
.F. (1989). Identidad social y personal. In J. Mayor & J.L. Pinillos (Eds.), Tratado de 
psicologia general. 7: Creencias, actitudes y valores. Madrid: Alhambra. 
Morales, J.F., & Huici, C. (1994). Procesos grupales. In J.F. Morales, M. Moya, J.M. 
Fernandez-Dols, C Huici, J. Marques, D. Paez & J.A. Perez (Eds.), Psicologia social. 
Madrid: McGraw-Hill. 
Morales, J.F., Lopez, M., & Vega, L. (1992). Individualismo, collectivismo e identidad social. 
Revista de Psicologia Social, Monogra
fico, 49-72. 
Morales, J.F. & Paez, D. (1996). Estereotipos, discriminacion y relaciones intergrupos en 
Espana y Latinoamerica. In R.Y. Bourhis & J.-P. Leyens (Eds.), Estereotipos discriminacion 
y relaciones entre grupos. Madrid: McGraw-Hill. 
Moreland, R.L. (1987). The formation of small groups. In C. Hendrick (Ed.), Review o
f
 
personality and social psychology: Group processes, Vol. 8. Beverly Hills, CA: Sage. 
Moreland, R.L., & Levine, J.M. (1982). Socialization in small groups: Temporal changes in 
individual-group relations. In L. Berkowitz (Ed.), Advances in experimental social 
psychology, Vol. 15. New York: Academic Press. 
Moreland, R.L., & Levine, J.M. (1988). Group dynamics over time: Development and 
socialization in small groups. In J.E. McGrath (Ed.), The social psychology o
f
 time. 
Newbury Park, CA: Sage. 
Morris, W.N., Worchel, S., Bios, J., Pearson, J., Roundtree, C, Samaha, G., Wachtler, J., & 
Wright, S. (1976). Collective coping with stress: Group reactions to fear, anxiety, and 
ambiguity. Journal o
f
 Personality and Social Psychology, 33, 674-679. 
Moscovici, S. (1976). Social in
fluence and social changes. London: Academic Press. 
Moscovici, S. (1980). Toward a theory of conversion behavior. In L. Berkowitz (Ed.), Advances 
in experimental social psychology, Vol. 13. New York: Academic Press. 
Moscovici, S., & Mugny, G. (1983). Minority influence. In P. Paulus (Ed.), Basic group process. 
New York: Springer Verlag. 
Moscovici, S., & Nemeth, CJ. (1974). Minority influence. In CJ. Nemeth (Ed.), Social 
psychology: Classic and contemporary integrations. Chicago: Rand McNally. 
Moya, M.C (1987). Percepcion de es
pisodios y estereotipia sexual. Tesis Doctoral, Servicio de 
Publicaciones, Universidad de Granada. 
Moya, M.C. (1989). Genero, estereotipia y relaciones intergrupales. In P. Ballarin & T. Ortiz 
(Eds.), La mu
jer en Andalucia. Granada: Universidad de Granada. 
Moya, M.C, & Peplau, L.A. (in preparation). 1dentidad de genero y relaciones de pare
ja. 
Mugny, G. (1982). The power o
f
 minorities. London: Academic Press. 


Re
f
erences 
245 
Mugny, G., Kaiser, e., & Papastamou, S. (1983). Influence minontaue, identification et 
relations entre groupes: Etude experimentale autour d'une votation. Cahiers de Psychologie 
Sociale, 19, 1-30. 
Mugny, G., Kaiser, e., Papastamou, S., & Perez, J.A. (1984). Intergroup relations, 
identification and social influence. British Journal o
f
 Social Psychology, 23, 31 7-322. 
Mugny, G., & Papastamou, S. (1982). Minority influence and psycho-social identity. European 
Journal o
f
 Social Psychology, 12, 379-394. 
Mugny, G., & Perez, J.A. (1985). Influence sociale, conflit et identification: Etude 
experimentale autour d'une persuasion 'manquee' lors d'une votation. Cahiers de 
Psychologie Sociale, 6, 1-13. 
Mugny, G., & Perez, J.A. (199Ia). T
he social psychology o
f
 minority influence. Cambridge: 
Cambridge University Press. 
Mugny, G., & Perez, J.A. (199Ib). EI efecto de criptomnesia social. Anthropos: Revista de 
Documentacian Cientifica de la Cultura, 27, 1 10-1 12. 
Mullen, B., Brown, R.J., & Smith, C. (1992). Ingroup bias as a function of salience, relevance 
and status: An integration. European Journal o
f
 Social Psychology, 22, 103-1 12. 
Mullen, B., & Hu, L. (1989). Perceptions of in 
group and outgroup variabilty: A meta-analytic 
integration. Basic and Applied Social Psychology, 10, 233-252. 
Murphy, G.L., & Medin, D.L. (1985). The role of theories in conceptual coherence. 
Psychological Review, 92, 289-3 16. 
Mussolini, B. (1935). Fascism: Doctrine and institutions. Rome: Ardita. 
Nakbi, J.L. (1990). Evolution de I'identite psychosociale feminine et masculine de l'enfance a 
l'age adulte. Cahiers de Psychologie Sociale, 34, 1 -9. 
Nakbi, J.L., & Arnal-Duchemin, M.J. (1987). Dynamique des representations chez les femmes 
et les hommes. Cahiers de Psychologie Sociale, 34, 1-9. 
Nascimento-Schulze, C.M. (1993). Social comparison, group identity and prof
essional identity: 
A study with bank clerks. Revista de Psicologia Social, 8, 69-82. 
Neisser, U. (Ed.). (1987). Concepts and conceptual development: Ecological and intellectual 
f
actors in categorization. Cambridge: Cambridge University Press. 
Nemeth, C. (1986). Differential contributions of majority and minority influence. Psychological 
Review, 93, 1-10, 23-32. 
Nemeth, e., & Staw, B.M. (1989). The tradeoffs of social control and innovation in groups and 
organizations. In L. Berkowitz (Ed.), Advances in experimental social psychology, Vol. 23. 
Orlando, FL: Academic Press. 
Oakes, P.J. (1987). The salience of social categories. In J.e. Turner, M.A. Hogg, P.l. Oakes, 
S.D. Reicher & M.S. Wetherell, Rediscovering the social group: A self
-categorization theory. 
Oxford: Blackwell. 
Oakes, P.J. (1996). The categorization process: Cognition and the group in the social 
psychology of stereotyping. In P. Robinson (Ed.), Social groups and identity: Developing the 
legacy o
f
 Henri T
a
j
f
el. Oxford: Butterworth Heinemann. 
Oakes, P.J., Haslam, S.A., & Reynolds, K.J. (in press). Social categorization and social 
context: An example from stereotype change. In D. Abrams & M.A. Hogg (Eds.), Social 
identity and social cognition. Oxford: Blackwell. 
Oakes, P.J., Haslam, A.S., & Turner, J.e. (1994). Stereotyping and social reality. Oxford: 
Blackwell. 
Oakes, P.J., & Turner, J.C. (\986). Distinctiveness and the salience of social category 
memberships: Is there an automatic perceptual bias towards novelty? European Journal o
f
 
Social Psychology, 16, 325-344. 
Oakes, P.J., & Turner, J.C. (\990). Is limited infonnation processing capacity the cause of 
social stereotyping? In W. Stroebe and M. Hewstone (Eds.), European review o
f
 social 
psychology, Vol. I. Chichester: Wiley. pp. 1 1 1-136. 
Oakes, P.J., Turner, J.e., & Haslam, S.A. (1991). Perceiving people as group members: The 
role of fit in the salience of social categorizations. British Journal o
f
 Social Psychology, 30, 
125-144. 
O'Dea, T.F. (1957). The Mormons. Chicago: University of Chicago Press. 


246 
Social identity: international pers
pectives 
Olden 
quist, A. (1992). Autonomy, social identities, and alienation. In F. Geyer & W.R. Heinz 
(Eds.), Alienation, society, and the individual. New Brunswick, NJ: Transaction Publishers. 
Osgood, C.E. (1962). An alternative to war or surrender. Urbana: University of Illinois Press. 
Ostrom, T.M., & Sedikides, C. (1992). Out-group homogeneity effect in natural and minimal 
groups. Psychological Bulletin, 112, 536-552. 
Paez, D., Arrospide, l., Martinez-Taboada, C., & Ayestaran, S. (1992). Identidad social, 
autoesquematizacion y autoconciencia collectiva: Investigaciones correlacionales en la 
perspectiva de la teona de la identidad social. Revista de Psicologia Social, Monogrti
fico, 3-
20. 
Palmonari, A., Carugati, F., Ricci-Bitti, P., & Sarchielli, G. (1979). Identita imper
f
ette. 
Bologne: II Mulino. 
Palmonari, A., Pombeni, M.L., & Kirchler, E. (1989). Peergroups and evolution of the self­
system in adolescence. European Journal o
f
 Psychology o
f
 Education, 4, 3-15. 
Parelman, A. (1983). Emotional intimacy in marriage: A sex-roles perspective. Ann Arbor, MI: 
UMI Research Press. 
Park, B., & Rothbart, M. (1982). Perception of out-group homogeneity and levels of social 
categorization: Memory for subordinate attributes of in-group and out-group members. 
Journal o
f
 Personality and Social Psychology, 42, 1051-1068. 
Parsons, T. (1968). The position of identity in the general theory of action. In G. Gordon & 
KJ. Gergen (Eds.), The self in social interaction. New York: Wiley. 
Pavelchak, M.A. (1989). Piecemeal and category-based evaluation: An idiographic analysis. 
Journal o
f
 Personality and Social Psychology, 56, 354-363. 
Pawluczuk, B. (1992). Indywiduac
ja systemu norm a toleranc
ja wobec obcych. Unpublished 
research report, Wydzial Psychologii, Uniwersytet Warszawski, Warsaw. 
Peabody, D., & Goldberg, L.R. (1989). Some determinants of factor structures from 
personality-descriptors. Journal o
f
 Personality and Social Psychology, 57, 552-567. 
Pemberton, MJ., Insko, C.A., & Schopler, l. (1996). Memory for and experience of differential 
competitive behavior of individuals and groups. Journal o
f
 Personality and Social 
Psychology, 71, 5, 953-966. 
Peplau, L.A. (1983). Roles and gender. In H.H. Kelley, E. Berscheid, A. Christensen, l. 
Harvey, T.L. Huston, G. Levinger, E. McClintock, A. Peplau & D.R. Peterson (Eds.), Close 
relationshi
ps. New York: W.H. Freeman & Co. 
Peplau, L.A., & Gordon, S.L. (1985). Women and men in love: Gender differences in close 
heterosexual relationships. In V.E. O'Leary, R.K. Under & B.S. Wallston (Eds.), Women, 
gender, and social psychology. Hillsdale, Nl: Erlbaum. 
Perez, l.A., & Mugny, G. (1985a). Categorizacion e influencia minoritaria. Anuario de 
Psicologia. 32, 100-1 16. 
Perez, l.A., & Mugny, G. (1985b). Influencia minoritaria sobre las opiniones frente al aborto y 
los anticonceptivos. Estudios de Psicologia, 23124, 29-54. 
Perez, l.A., & Mugny, G. (I 986a). Efectos paradojicos de la categorizaci6n en la influencia 
minoritaria. Boletin de Psicologia, 12, 65-89. 
Perez, l.A., & Mugny, G. (1986b). Induction experimentale d'une influence minoritaire 
indirecte. Cahiers de Psychologie Sociale, 32, 15-24. 
Perez, l.A., & Mugny, G. (1987). Paradoxical effects of categorization in minority influence: 
When being an out-group is an advantage. European Journal o
f
 Social Psychology, 17, 1 57-
169. 
Perez, l.A., & Mugny, G. (1990). Minority influence, manifest discrimination and latent 
influence. In D. Abrams & M.A. Hogg (Eds.), Social identity theory: Constructive and 
critical advances. Heme! Hempstead: Harvester Wheatsheaf. 
Perez, l.A., Mugny, G., Llavata, E., & Fierres, R. (1993). Paradoxe de la discrimination et 
conflit culture!: Etudes sur Ie racisme. In l.A. Perez, G. Mugny, et aI., I
n
fluences sociales: La 
theorie de l'elaboration du con
flit. Paris: Delachaux et Niestle. 
Perez, l.A., Mugny, G., & Moscovici, S. (1986). Les eff
ets paradoxaux du deni dans l'influence 
sociale. Cahiers de Psychologie Sociale, 32, 1-14. 
Perez, l.A., Mugny, G., & Navarro, E. (1991). EI efecto de la 'paralisis intragrupal': Niveles de 


Re
f
erences 
247 
categorizaci6n del yo y niveles de influencia social. Anthropos: Revista de Documentacion 
Cientifica de la Cultura, 27, 121-132. 
Perez, J.A., Mugny, G., Roux, P., & Butera, F. (1991). Influence via la comparaison et 
influence via la validation. In J.L. Beauvois, R.V. Joule & J.M. Monteil (Eds.), Perspectives 
cognitives et conduites sociales, Vol. 3. Cousset: Delval. 
Pfost, K.S., & Fiore, M. (1990). Pursuit of nontraditional occupations: Fear of success or fear 
of not being chosen? Sex Roles, 23, 1 5-24. 
Phinney, J.S. (1990). Ethnic identity in adolescents and adults: Review of research. 
Psychological Bulletin, 108, 499-514. 
Piaget, J. (1975). L'equilibration de structures cognitives: Probleme central du developpement. 
Paris: Presses Universitaires de France. 
Platow, MJ., McClintock, e.G., & Liebrand, W. (1990). Predicting intergroup fairness and 
ingroup bias in the minimal group paradigm, European Journal o
f
 Social Psychology, 20, 
221-239. 
Pratto, F., & Bargh, J.A. (1991). Stereotyping based on apparently individuating information: 
Trait and global components of sex stereotypes under attention overload. Journal o
f
 
Experimental Social Psychology, 27, 26-47. 
Proshansky, H.M., Fabian, A.K., & Kaminoff, R. (1983). Place identity: Physical world 
socialization of the self. Journal o
f
 Environmental Psychology, 3, 57-83. 
Pryor, J.B., & Ostrom, T.M. (1987). Social cognition theory of group processes. In B. Mullen & 
G.R. Goethals (Eds.), Theories o
f
 group behavior. New York: Springer Verlag. 
Quattrone, G.A. (1986). On the perception of a group's variability. In S.G. Worchel & W.G. 
Austin (Eds.), Psychology o
f
 intergroup relations. Chicago: Nelson-Hall. 
Quevedo, S. (1972). Revindicaci6n urbana y lucha politica. Revista Latinoamericana de 
Estudios Urbanos Regionales, 2, 12-43. 
Rabbie, J.M., & Lodewijkx, H.G.M. (1994). Conflict and interaction: An individual-group 
continuum. In B. Markovsky, K. Heimer & J. O'Brien (Eds.), Advances in group processes, 
Vol. 11. Greenwich, CT: JAI Press. 
Ravaud, J.F., Beaufils, B., & Paicheler, H. (1986). Handicap et integration scolaire. Sciences 
Sociales et Sante, 3-4, 167-194. 
Reicher, S., & Emler, N. (1986). The management of delinquent reputations. In H. Beloff (Ed.), 
Getting into lif
e. London: Methuen. 
Reis, H.T., Nezlek, J., & Wheeler, L. (1980). Physical attractiveness in social interaction. 
Journal o
f
 Personality and Social Psychology, 38, 604-617. 
Reis, H.T., & Wheeler, L. (1991). StUdying social interaction with the Rochester Interaction 
Record. In M.P. Zanna (Ed.), Advances in experimental social psychology, Vol. 24. New 
York: Academic Press. 
Reykowski, J. (1979). Motywac
ja, postawy prospoleczne a osobowoSi:. Warsaw: PWN. 
Reykowski, J. (1989). Dimensions of the development in moral values. In N. Eisenberg, J. 
Reykowski & E. Staub (Eds.), Social and moral values. Hillsdale, NJ: Erlbaum. 
Reynolds, V., Falger, V., & Vine, I. (1987). The sociobiology o
f
 ethnocentrism. London: Croom 
Helm. 
Rips, L.J. (1989). Similarity, typicality, and categorization. In S. Vosniadou & A. Ortony 
(Eds.), Similarity and analogical reasoning. Cambridge: Cambridge University Press. 
Rivera, A.N. (1991). Psicologia y colonizaci6n. In M. Montero (Ed.), Accion y discurso: 
Problemas de psicologia politica en America Latina. Caracas: Eduven. 
Robinson, J.P. (1988). Time diary evidence about the social psychology of 
everyday life. In J 
.E. 
McGrath (Ed.), The social psychology o
f
 time. Newbury Park, CA: Sage. 
Rodriguez, C.P. (1992). Estereotipos de nacionalidad en estudiantes colombianos y 
venezolanaos. Boletin A vepso, 15, 65-74. 
Rogers, C. (1951). Client-centered therapy. Boston: Houghton-Mifflin. 
Rokeach, M. (1973). The nature o
f
 human values. New York: Free Press. 
Rook, D.S. (1987). Social support versus companionship: Effects of life stress, loneliness, and 
evaluations by others. Journal o
f
 Personality and Social Psychology, 52, 1 1 32-1 147. 
Rosch, E. (1973). On the internal structure of perceptual and semantic categories. In T.E. 


248 
Social identity: international perspectives 
Moore (Eds.), Cognitive development and the acquisition o
f
 language. New York: Academic 
Press. 
Rosch, E. (1975). Cognitive reference points. Cognitive Psychology, 7, 532-547. 
Rosch, E. (1978). Principles of categorization. In E. Rosch & B.B. Lloyd (Eds.), Cognition and 
categorization. Hillsdale, NJ: Erlbaum. 
Rosch, E., Mervis, CB., Gray, W.D., Johnson, D.M., & Boyes-Braem, P. (1976). Basic objects 
in natural categorization. Cognitive Psychology, 15, 346-378. 
Rosenberg, S., & Gara, M. (1985). The multiplicity of personal identity. Review o
f
 Personality 
and Social Psychology, 6, 87-1 13. 
Roth, E.M., & Schoben, E.1. (1983). The effect of context on the structure of categories. 
Cognitive Psychology, 15, 346-378. 
Rubin, Z., Peplau, L.A., & Hill, CT. (1981). Loving and leaving: Sex differences in romantic 
attachments. Sex Roles, 7, 821-835. 
Rush, M.C, & Russell, J.E.A. (1988). Leader prototypes and prototype-contingent consensus 
in leader behaviour descriptions. Journal o
f
 Experimental Social Psychology, 24, 88-104. 
Salazar, J.M. (1983a). Bases psicologicas del nacionalismo. Mexico: Trillas. 
Salazar, J.M. (1983b). On the psychological viability of 'Latin-Americanism'. International 
Social Science Journal, 35, 295-308. 
Salazar, J.M. (1989). Niveles de identificacion y estructura cognoscitiva en relacion con el 
latinoamericano. Revista de Psicologia Social, 4, 13-22. 
Salisbury, R.H. (1975). Interest groups. In F.I. Greenstein & N.W. Polsby (Eds.), Handbook o
f
 
political science, Vol. 4. Reading, MA: Addison-Wesley. 
Sanchez-Mazas, M., Perez, J.A., Navarro, E., Mugny, G., & Jovanovic, J. (1993). De la 
paralysie intragroupe au confiit normatif: Etudes dur I'avortement, la contraception et la 
xenophobie. In J.A. Perez & G. Mugny (Eds.), Influences sociales: La tMorie de l'eboration 
du confiit. Paris: Delachaux et Niestle. 
Sande, G.N. (1990). The multifaceted self. In J.M. Olson & M.P. Zanna (Eds.), Self
-in
f
erence 
processes: The Ontario Symposium, Vol. 6. Hillsdale, NJ: Erlbaum. 
Sanders, G.S., & Baron, R.S. (1977). Is social comparison irrelevant for producing choice 
shifts? Journal o
f
 Experimental Social Psychology, 13, 303-314. 
Sarbin, T.R., & Allen, V.L. (1968). Role theory. In G. Lindzey & E. Aronson (Eds.), Handbook 
o
f
 social psychology, Vol. 1 (2nd ed.). Reading, Mass: Addison-Wesley. 
Schachter, S. (1951). Deviation, rejection, and communication. Journal o
f
 Abnormal and Social 
Psychology, 46, 190-207. 
Schopler, J., & Insko, CA. (1992). The discontinuity effect in interpersonal and intergroup 
relations: Generality and mediation. In W. Stroebe & M. Hewstone (eds.), European review 
o
f
 social psychology, Vo!' 1. Chichester: Wiley. 
Schopler, J., Insko, CA., Graetz, K.A., Drigotas, S.M., & Smith, V.A. (1991). The generality 
of the individual-group discontinuity eff
ect: Variations in positivity-negativity of outcomes, 
players' relative power, and magnitude of outcomes. Personality and Social Psychology 
Bulletin, 1 7, 612-624. 
Schopler, J., Insko, CA., Graetz, K.A., Drigotas, S.M., Smith, VA, & Dahl, K. (1993). 
Individual-group discontinuity: Further evidence for mediation by fear and greed. 
Personality and Social Psychology Bulletin, 19, 419-43 1. 
Semin, G.R. (1987). On the relationship between representation of 
theories in psychology and 
ordinary language. In W. Doise & S. Moscovici (Eds.), Current issues in European social 
psychology, Vol. 2. Cambridge: Cambridge University Press. 
Serino, C (1988a). Biases egocentrici e fenomeni di asimmetria nei confronti Se/Altro. Giornale 
Italiano di Psicologia, 15, 213-233. 
Serino, C. (I 
988b). Strategies et structure de la comparaison sociale: Quelques aspects de 
l'asymmetrie soi/autrui dans les relations entre groupes. Cahiers de Psychologie Cognitive, 8, 
627-648. 
Serino, C (1989). Processi di confronto e argomentazione: In G. Bellelli (Ed.), II metodo del 
discorso. Naples: Liguouri. 
Serino, C (1990). La comparaison socia 
Ie et Ie structure des systemes categoriels: Quelques 


Re
f
erences 
249 
fI!flexions sur la prototypicalite du soi. Cahiers Internationaux de Psychologie Socia 
Ie, J, 73-
96. 
Serino, C (1992). From the asymmetry phenomena to the prototypicality of self: Studies on the 
structure of social categories. In L. Arcuri & C Serino (Eds.), Asymmetry phenomena in 
interpersonal comparison: Cognitive and social issues. Naples: Liguori. 
Serino, C (1993). Information processing and self/others comparison: Social-cognitive 
extensions of a non-metrical approach to similarity. In M.C. Hurtig, Pichevin, M.F., & 
M. Piolat (Eds.), Studies on the self and social cognition. London: World Scientific 
Publishing. 
Serino, C (1996). Similarity: Inter
personal and intergroup comparisons in an asymmetrical 
f
rame. Paper presented at the Small Group Meeting on 'Intergroup Relations: Current Work 
and Future Perspectives', Catania, Italy, 26-9 September. 
Serino, C, & Mazzara, B. (1995). Scripts e processi di confronto nella rappresentazione sociale 
della persona handicappata: Una ricerca sperimentale. Giornale Italiano di Psicologia, 22, 
265-292. 
Seton-Watson, H. (1965). Nationalism. old and new. Sydney: Sydney University Press. 
Shaw, M.E. (1976). Group d
ynamics: The psychology o
f
 small group behavior. New York: 
McGraw-Hill. 
Sherif, M. (1936). The psychology o
f
 social norms. New York: Harper & Brothers. 
Sherif, M. (1951). A preliminary experimental study of inter-group relations. In 1.H. Rohrer & 
M. Sherif (Eds.), Social psychology at the crossroads. New York: Harper & Brothers. 
Sherif, M. (1966). In common predicament: Social psychology o
f
 intergroup conflict and 
cooperation. Boston: Houghton Mifflin. 
Sherif, M. (1967). Group conflict and cooperation: T
heir social psychology. London: Routledge 
& Kegan Paul. 
Sherif, M., Harvey, 0.1., White, B.1., Hood, W.R., & Sherif, CW. (1961). I
ntergroup conflict 
and cooperation: The Robbers' Cave experiment. Norman, OK: University Book Exchange. 
Shils, E. (1968). Ideology. In D.L. Sills (Ed.), International encylopedia o
f
 the social sciences. 
Vol. 7. New York: Macmillan and Free Press. 
Sills, D.L. (1958). The volunteers. New York: Free Press. 
Simmel, G. (1955). Conflict: T
he web o
f
 group a
ffiliations. Chicago: Chicago Free Press. 
Simon, B. (1994). The perception of ingroup and outgroup homogeneity: Reintroducing the 
intergroup context. In W. Stroebe and M. Hewstone (Eds.), European review o
f
 social 
psychology. Vol. 3. Chichester: Wiley. pp. 1-3. 
Simon, B., & Brown, R. (1987). Perceived intragroup homogeneity in minority-majority 
contexts. Journal o
f
 Psychology and Social Psychology, 53, 703-71 1 .  
Simon, B., & Hamilton, D.L. (1994). Self stereotyping and social context: The effects of relative 
in-group size and in-group status. Journal o
f
 Personality and Social Psychology, 66, 699-
7 1 1 .  
Singh, A .  (1979). Reliability and validity of self-reported delinquency studies. Psychological 
Reports, 44, 987-993. 
Singleton, R., 1r. (1979). Another look at the conformity explanation of group-induced shift in 
choice. Human Relations, 32, 37-56. 
Smith, A.D. (I 984a). National identity and myths of ethnic descent. Research in Social 
Movements. Con
flict and Change, 7, 95-230. 
Smith, A.D. (1984b). Ethnic myths and ethnic revivals. European Journal o
f
 Sociology, 25, 
283-305. 
Smith, A.D. (199 1). National identity. London: Penguin. 
Smith, P.B., & Bond, M.H. (1993). Social psychology across cultures. New York: Harvester 
Wheatsheaf. 
Snyder, CR., & Fromkin, H.L. (1980). Uniqueness: the human pursuit o
f
 d
iff
erence. New York: 
Plenum Press. 
Snyder, M.L., Kleck, B.E., Strentna, A., & Mentzer, S.l. (1 979). A comparison of egotism, 
negativity, and learned helplessness as explanations for poor performance after unsolvable 
problems. Journal o
f
 Personality and Social Psychology, 40, 24-30. 


250 
Social identity: international pers
pectives 
Snyder, M.L., & Wicklund, R. (1981). Attribute ambiguity. In J.H. Harvey, W. Ickes & R.R. 
Kidd (Eds.), New directions in attribution research. New York: Halsted. 
Sobczyk, K. (1992). Schematy Ja/My/I
nni a spostrzegany d
ystans i ustosunkowania wobec 
narodowosci blizsze
j i dalsze
j. Unpublished master's thesis, Wydzial Psychologii, Universytet 
Warszawski, Warsaw. 
Sokolowska, G. (1992). Tozsamosciowe przes/anki ustosunkowim wobec obcych. Unpublished 
research report, Wydzial Psychologii, Uniwersytet Warszawski, Warsaw. 
Spears, R., Lea, M., & Lee, S. (1990). De-individuation and group polarization in computer­
mediated communication. British Journal o
f
 Social Psychology, 29, 121-134. 
Srull, T.K., & Gaelik, L. (1983). General principles and individual diff
erences in the self as a 
habitual ref
erence point: An examination of self-other judgments of similarity. Social 
Cognition, 2, 108-121. 
Stephenson, G.M. (1984). Interpersonal and intergroup dimensions of bargaining and 
negotiation. In H. Tajfel (Ed.), The social dimension: European developments in social 
psychology, Vol. 2. Cambridge: Cambridge University Press. 
Stroup, H.H. (1945). T
he Jehovah's Witnesses. New York: Russell & Russell. 
Suls, J.M., & Miller, R.L. (Eds.). (1977). Social comparison processes: Theoretical and empirical 
pers
pectives. Washington, DC: Hemisphere. 
Sumner, W.G. (1906). Folkways. New York: Ginn. 
Szuster-Zbrojewicz, A. (1988). Z badan nad zwiazkami pobodienstwalodn;:nosci Ja-Inni z 
gotowoSci̕ do egzocentrycznego angaiowania sil;: w sprawy pozaosobiste. In M. Jarymowicz 
(Ed.), Studia nad spostrzeganiem relac
ji Ja-I
nni: Tozsamosc, ind
ywiduac
ja, przynaleznosc. 
Wroclaw: Ossolineum. 
Szuster-Zbrojewicz, A. (1 990). Formy reprezentacji MY a prze
jawy pozaosobistego 
zaangazowania. Unpublished report G-MEN-9/90, Wydzial Psychologii, Uniwersytet 
Warszawski, Warsaw. 
Tajfel, H. (1959). Quantitative judgment in social perception. British Journal o
f
 Psychology, 50, 
16-29. 
Tajfel, H. (1969a). Cognitive aspects of prejudice. Journal o
f
 Social Issues, 25, 79-97. 
Tajfel, H. (1969b). The formation of national attitudes: A social psychological perspective. In 
M. Sherif & c. Sherif (Eds.), Interdisci
plinary relationshi
ps in the social sciences. Chicago: 
Aldine. 
Tajfel, H. (1970a). Aspects of national and ethnic loyalty. Social Science I
n
f
ormation, 9, 1 19-
144. 
Tajfel, H. (1970b). Experiments in intergroup discrimination. Scientific American, 223, 96-102. 
Tajfel, H. (I 
972a). La categorisation sociale. In S. Moscovici (Ed.), Introduction a la 
psychologie sociale, Vol. 1 .  Paris: Larousse. 
Tajfel, H. (1972b). Experiments in a vacuum. In J. Israel & H. Tajfel (Eds.), T
he context o
f
 
social psychology: A critical assessment (European Monographs in Social Psychology No. 2). 
London: Academic Press. 
Tajf
el, H. (1974). Social identity and intergroup behavior. Social Science Inf
ormation, 13, 65-
93. 
Tajfel, H. (1978a). Interindividual behaviour and intergroup behaviour. In H. Tajfel (Ed.), 
Diff
erentiation between social groups: Studies in the social psychology o
f
 intergroup behaviour. 
London: Academic Press. 
Tajfel, H. (1978b). Social categorization, social identity and social comparison. In H. Tajfel 
(Ed.), Dif
f
erentiation between social groups: Studies in the social psychology o
f
 intergroup 
relations. London: Academic Press. 
Tajfel, H. (I978c). The achievement of group diff
erentiation. In H. Tajfel (Ed.), Diff
erentiation 
between social groups: Studies in the social psychology o
f
 intergroup relations. London: 
Academic Press. 
Tajfel, H. (Ed.). (1978d). Dif
f
erentiation between social groups: Studies in the social psychology 
of intergroup relations. London: Academic Press. 
Tajfel, H. (1978e). The social psychology o
f
 minorities. London: Minority Rights Group. 


Re
f
erences 
251 
Tajfel, H. (1979). Individuals and groups in social psychology. British Journal o
f
 Social 
Psychology, 18, 183-190. 
Tajfel, H. (\981). Human groups and social categories: Studies in social psychology. Cambridge: 
Cambridge University Press. 
Tajfel, H. (1982a). Instrumentality, identity, and social comparison. In H. Tajfel (Ed.), Social 
identity and intergroup relations. Cambridge: Cambridge University Press. 
Tajfel, H. (Ed.). (\982b). Social identity and intergroup relations. Cambridge: Cambridge 
University Press. 
Tajfel, H. (1982c). Social psychology of intergroup relations. In M.R. Rosenzweig & L.R. 
Porter (Eds.), Annual review o
f
 psychology, Vol. 33. Palo Alto, CA: Annual Reviews. 
Tajfel, H., & Billig, M. (\974). Familiarity and categorization in intergroup behavior. Journal 
o
f
 Experimental Social Psychology, 10, 159-170. 
Tajfel, H., Flament, e., Billig, M., & Bundy, R.P. (1971). Social categorization and intergroup 
behaviour. European Journal o
f
 Social Psychology, 1, 149-178. 
Tajfel, H., & Turner, J.C. (\979). An integrative theory of intergroup conflict. In W.G. Austin 
& S. Worchel (Eds.), The social psychology o
f
 intergroup relations. Monterey, CA: Brooks/ 
Cole. 
Tajfel, H., & Turner, J.e. (1986). The social identity theory of intergroup behavior. In S. 
Worchel & W.G. Austin (Eds.), The psychology o
f
 intergroup relations. Chicago: Nelson­
Hall. 
Tajfel, H., & Wilkes, A.L. (1963). Classification and quantitative judgement. British Journal o
f
 
Psychology, 54, 101-1 14. 
Tap, P. (1979). L'identification: Est-elle une alienation de l'identite? In P. Tap (Ed.), Jdentite 
individuelle et personalisation. Toulouse: Privat. 
Tap, P. (1985). Masculin et f
eminin chez I'en
f
ant. Toulouse: Privat. 
Taylor, D.M., & Moghaddam, F.M. (1987). T
heories o
f
 intergroup relations. New York: 
Praeger. 
Toch, H. (\965). The social psychology o
f
 social movements. Indianapolis: Bobbs-Merrill. 
Torregrosa, J.R., & Ramirez, S. (1991). Identidades nacionales y estados plurinacionales: Una 
perspectiva psicosociol6gica. lnteraccion Social, 1, 123-1 36. 
Triandis, H.e. (\ 
988). Collectivism vs. individualism. In G. Verma & e. Bagley (Eds.), Cross­
cultural studies of 
personality, attitudes, and cognition. London: Macmillan. 
Triandis, H.e. (1992). Comment on S. Hinkle, R.J. Brown & P. Ely, Social identity processes: 
some limitations and limiting conditions. Revista de Psicologia Social, Monogra
fico, 1 1 3-
123. 
Triandis, H.e. (1993). Paper for Psychology in the Americas: North-South Exchange, 
Communicaci6n Presentada al Congreso Interamericano, Santiago de Chile. 
Triandis, H.e. (\994). Recherches recentes sur I'individualisme et collectivisme. Cahiers 
1nternationaux de Psychologie Sociale, 23, 14-27. 
Triandis, H.C., Bontempo, R., Betancourt, H., Bond, M., Leung, D., Brenes, A., Georgas, G., 
Hui, e.H., Marin, G., Setiadi, B., Sinha, J.B.P., Verma, J., Spangenberg, J., Touzard, H., & 
Montmollin, G. (1986). The measurement of etic aspects of individualism and collectivism 
across cultures. Australian Journal o
f
 Psychology, 38, 257-267. 
Triandis, H.C., Bontempo, R., Villareal, M.J., Asai, M., & Lucca, N. (\988). Individualism 
and collectivism: Cross-cultural perspectives on self-ingroup relationships. Journal o
f
 
Personality and Social Psychology, 54, 323-338. 
Triandis, H.e., Leung, K., Villareal, M.J., & Clack, F.L. (1985). Allocentric versus idiocentric 
tendencies: Convergent and discriminant validation. Journal o
f
 Research in Personality, 19, 
395-415. 
Triandis, H.C., McCusker, e., & Hui, e.H. (1990). Multimethod probes of individualism and 
collectivism. Journal o
f
 Personality and Social Psychology, 59, 1006-1020. 
Triska, J.F. (Ed.). (\962). Soviet communism: Programs and rules. San Francisco: Chandler. 
Truman, D.B. (\951). The governmental process. New York: Knopf. 
Tuckman, B.W. (1965). Developmental sequence in small groups. Psychological Bulletin, 63, 
384-399. 


252 
Social identity: international pers
pectives 
Tuckman, B.W., & Jensen, M.A.e. (1977). Stages of small group development revisited. Group 
and Organization Studies, 2, 419-427. 
Turner, J.e. (l975a). Social comparison and social identity: Some prospects for intergroup 
behaviour. European Journal o
f
 Social Psychology, 5, 5-34. 
Turner, J.e. (1975b). Social categorization and social comparison in intergroup relations. PhD, 
University of Bristol. 
Turner, J.C. (1978). Social categorization and social discrimination in the minimal group 
paradigm. In H. Tajfel (Ed.), Diff
erentiation between social groups: Studies in the social 
psychology o
f
 intergroup relations. London: Academic Press. 
Turner, J.e. (l98Ia). Towards a cognitive redefinition of the social group. Cahiers de 
Psychologie Cognitive, 1, 93-1 18. 
Turner, J.e. (198Ib). The experimental social psychology of intergroup behavior. In J.C. 
Turner & H. Giles (Eds.), Intergroup behavior. Chicago: University of Chicago Press. 
Turner, J.C. (1982). Towards a cognitive redefinition of the social group. In H. Tajfel (Ed.), 
Social identity and intergroup relations. Cambridge: Cambridge University Press. 
Turner, J.C. ( l983a). Some comments on the measurement of social orientations in the minimal 
group paradigm. European Journal o
f
 Social Psychology, 13, 351-368. 
Turner, J.e. (I 
983b). A second reply to Bornstein, Crum, Wittenbraker, Harring, Insko, and 
Thibaut on the measurement of social orientations. European Journal o
f
 Social Psychology, 
13, 383-387. 
Turner, J.e. (1985). Social categorization and the self-concept: A social cognitive theory of 
group behaviour. In E.J. Lawler (Ed.), Advances in group processes, Vol. 2. Greenwich, CT: 
JAI Press. 
Turner, J.C. (1987). In J.e. Turner, M.A. Hogg, P.J. Oakes, S.D. Reicher & M.S. Wetherell, 
Rediscovering the social group: A self-categorization theory. Oxford: Blackwell. 
Turner, J.e. (1988). Comments on Doise's 'Individual and social identities in intergroup 
relations'. European Journal o
f
 Social Psychology, 18, 1 1 3-1 16. 
Turner, J.C. (1991). Social influence. Milton Keynes: Open University Press. 
Turner, J.e., & Brown, R.I. (1978). Social status, cognitive alternatives and intergroup 
relations. In H. Tajfel (Ed.), Diff
erentiation between social groups: Studies in the social 
psychology o
f
 intergroup relations. London: Academic Press. 
Turner, J.e., Brown, R.J., & Tajfel, H. (1979). Social comparison and group interest in 
intergroup favoritism. European Journal o
f
 Social Psychology, 9, 187-204. 
Turner, J.C., & Giles, H. (Eds.). (1981). Intergroup behaviour. Oxford: Blackwell; Chicago: 
University of Chicago Press. 
Turner, J.e., Hogg, M.A., Oakes, P.I., Reicher, S.D., & Wetherell, M.S. (1987). Rediscovering 
the social group: A self-categorization theory. Oxford: Blackwell. 
Turner, J.e., Hogg, M.A., Oakes, P.I., & Smith, P.M. (1984). Failure and defeat as 
determinants of group cohesiveness. British Journal o
f
 Social Psychology, 23, 97-1 1 1 .  
Turner, J.e., & Oakes, P.J. (1986). The significance of the social identity concept for social 
psychology with ref
erence to individualism, interactionism, and social influence. British 
Journal o
f
 Social Psychology, 25, 237-252. 
Turner, J.e., & Oakes, P.J. (1989). Self-categorization theory and social influence. In P.B. 
Paulus (Ed.), The psychology o
f
 group influence. Hillsdale, NJ: Erlbaum. 
Turner, J.C., Oakes, P.I., Haslam, S.A., & McGarty, e.A. (1994). Self and collective: 
Cognition and social context. Personality and Social Psychology Bulletin, 20, 454-63. 
Turner, J.e., Sachdev, I., & Hogg, M.A. (1983). Social categorization, interpersonal attraction 
and group formation. British Journal o
f
 Social Psychology, 22, 227-239. 
Turner, J.e., Wetherell, M.S., & Hogg, M.A. (1989). Referent informational influence and 
group polarization. British Journal o
f
 Social Psychology, 28, 135- 147. 
Tversky, A. (1977). Features of similarity. Psychological Review, 84, 327-352. 
Vander Zanden, J.W. (1960). The Klan revival. American Journal o
f
 Sociology, 65, 426-459. 
van Knippenberg, A. (1978). Status differences, comparative relevance and intergroup 
differentiation. In H. Tajfel (Ed.), Diff
erentiation between social groups: Studies in the social 
psychology o
f
 intergroup relations. London: Academic Press. 


Re
f
erences 
253 
van Knippenberg, A. (1984). Intergroup diff
erences in group perception. In H. Tajfel (Ed.), The 
social dimension: Developments in European social psychology, 
Vol. 
2. Cambridge: 
Cambridge University Press. 
van Knippenberg, A., & Wilke, H. (1988). Social categorization and attitude change. European 
Journal of Social Psychology, 18, 395-406. 
Verkuyten, M. (1990). Self-esteem and the evaluation of ethnic identity among Turkish and 
Dutch adolescents in the Netherlands. Journal o
f
 Social Psychology, 130, 285-297. 
Volpato, C., Maass, A., Mucchi-Faina, A., & Vitti, E. (1990). Minority influence and social 
categorization. European Journal o
f
 Social Psychology, 20, 1 19-132. 
Wetherell, M.S. (1987). Social identity and group polarization. In J.c. Turner, M.A. Hogg, P.J. 
Oakes, S.D. Reicher & M.S. Wetherell, Rediscovering the social group: A self
-categorization 
theory. Oxford: Blackwell. 
Wheeler, L., Reis, H., & Nezlek, J. (1983). Loneliness, social interaction, and sex roles. Journal 
o
f
 Personality and Social Psychology, 45, 943-953. 
Wilder, D.A. (1977). Perception of groups, size of oppositions, and social influence. Journal o
f
 
Experimental Social Psychology, 13, 253-268. 
Wilder, D.A. (1978). Homogeneity of jurors: The majority's influence depends upon their 
perceived independence. Law and Human Behavior, 2, 363-376. 
Wilder, D.A. (1990). Some determinants of the persuasive power of in-groups and out-groups: 
Organization of information and attribution of independence. Journal o
f
 Personality and 
Social Psychology, 59, 1 202-1213. 
Wilkes, A., & Tajfel, H. (1966). Types de classification et importance du contraste relatif. 
Bulletin du CERP, 15, 77-81 .  
Williams, J.A., & Giles, H .  (1978). The changing status of women in society: An intergroup 
perspective. In H. Tajfel (Ed.), Diff
erentiation between social groups: Studies in the social 
psychology o
f
 intergroup Behaviour. London: Academic Press. 
Williams, R.M. (1970). American society (3rd ed.). New York: Knopf. 
Winter, D.G. (1973). The power motive. New York: Free Press. 
Wish, M. (1976). Comparisons among multidimensional structures of interpersonal relations. 
Multivariate Behavioral Research, 11, 297-324. 
Wish, M., Deutsch, M., & Kaplan, S.J. (1976). Perceived dimensions of interpersonal relations. 
Journal o
f
 Personality and Social Psychology, 33, 409-420. 
Wittenbraker, J. (1983). The etiology o
f
 intergroup discrimination: A developmental approach. 
Unpublished doctoral dissertation, University of North Carolina, Chapel Hill. 
Worchel, S., Coutant-Sassic, D., & Grossman, M. (1992). A developmental approach to group 
dynamics: A model and illustrative research. In S. Worchel, W. Wood & J.A. Simpson 
(Eds.), Group process and productivity. Newbury Park, CA: Sage. 
Worchel,. S., Coutant-Sassic, D., & Wong, F. (1993). Toward a more balanced view of conflict: 
There is a positive side. In S. Worchel & J.A. Simpson (Eds.), Con
flict between people and 
groups. Chicago: Nelson-Hall. 
Worchel, S., Grossman, M., & Coutant-Sassic, D. (1993). Minority influence in the group 
context: How group factors aff
ect when the minority will be influential. In A. Mucchi Faina 
& S. Moscovici (Eds.), Minority influence. Chicago: Nelson-Hall. 
Wright, S.c., Taylor, D.M., & Moghaddam, F.M. (1990). Responding to membership in a 
disadvantaged group: From acceptance to collective protest. Journal o
f
 Personality and 
Social Psychology, 58, 994-1 003. 
Wyer, R.S., & Podeschi, D.M. (1978). The acceptance of generalizations about persons, objects 
and events. In R. Revlin & R. Mayer (Eds.), Human reasoning. Washington, DC: Winston. 
Yamagishi, T. (1995). Social dilemmas. In K.S. Cook, G.A. Fine & J.S. House (Eds.), 
Sociological perspectives on social psychology. Boston: Allyn & Bacon. 
Zander, A. (1968). Group aspirations. In D. Cartwright & A. Zander (Eds.), Group d
ynamics 
(3rd ed.). London: Tavistock; New York: Harper & Row. 
Zander, A., Stotland, E., & Wolfe, D. (1960). Unity of group, identification with group, and 
self-esteem of members. Journal o
f
 Personality, 28, 463-478. 


254 
Social identity: international perspectives 
Zavalloni, M. (1973). Social identity: Perspectives and prospects. Social Science I
n
f
ormation, 
12, 65-9J. 
Zavalloni, M. (1983). Ego-ecology: The study of the interaction between social and personal 
identities. In A. Jacobson-Widdings (Ed.), Identity: Personal and socio-cultural. Stockholm: 
Almqvist & Wicksell International. 
Zavalloni, M., & Louis-Guerin, C. (1984). Identite sociale et conscience: Introduction a /'ego­
ecologique. Toulouse: Privat. 
Zebrowitz, L. (1990). Social perception. Buckingham: Open University Press. 
Zeegers, W. (1988). Andere tijden, andere mensen: De sociale re
presentatie van identiteit. 
Amsterdam: Uitgeveri
j Bert Bakker. 
Zeigarnik, B. (1927). fIber das Behalten von erledigten und unerledigten Handlungen. 
Psychologische Forschung, 9, 1-85. 
Ziller, R. (1964). Individuation and socialization: A theory of assimilation in large 
organizations. Human Relations, 1 7, 341-360. 
Zimbardo, P.G. (1969). The human choice: Individuation, reason, and order versus 
deindividuation, impulse, and chaos. In W.J. Arnold & D. Levine (Eds.), Nebraska 
Symposium on Motivation, Vol. 1 7. Lincoln: University of Nebraska Press. 
Zimbardo, P.G., & Leippe, M.R. (1991). T
he psychology o
f
 attitude change and social influence. 
New York: McGraw-HilI. 


Index 
Abrams, D., 82, 124-41, 154, 1 55, 166, 167, 
169, 177-8, 212, 223 
Abric, I.e., 13 
Adamopoulos, I., 17 
Aebischer, V., 144 
agency, 18, 19 
aggregate groups, 20-1 ,  132 
Allen, V.L., 2, 95, 108 
Allport, F.H., 1 80 
Allport, G.W., 16, 44, 81 
altercentrism, 218-19, 228 
Altman, 1., 1 17 
altruistic rationalization hypothesis, 193-4 
Amancio, L., 22 
Amish society, 96, 102 
Amnesty International, 97-8 
Anastasio, P.A., 1 96 
anchored identities, 20-3 
Anderson, B., 1 20 
Antoni, M., 226 
Apter, D.E., 100 
Ardener, S., 96 
Ardrey, R., 1 17 
Argyle, M., 55 
Arnal-Duchemin, M.J., 15 
Arr6spide, 1.1., 21 1-29 
Asai, M., 201, 213 
Asch, S.E., 36, 81, 104, 151 
assimilation effect, 4 
Asuncion, A.G., 143 
asymmetry phenomena in similarity 
judgments, 37-41, 42 
attraction see interpersonal attraction; social 
attraction 
auto-stereotypy, 7 
Ayestanin, S., 21 1-29 
Bachman, B.A., 196 
Back, K., 126 
Bakan, D., 18 
Ball, P., 169 
Ballachey, E.L., 1 12 
Ballato, S.M., 88 
Banaji, M., 42 
Bargh, I.A., 36 
Baron, R.S., 81 
Barsalou, L.W., 88, 90, 92 
Bar-Tal, D., 93-1 13 
Bar-Tal, Y., 109 
Basques, 1 23 
Beaufils, B., 13 
beliefs, group, 93-1 13 
Bern, D.I., 101, 141 
Bern, S.L., 20 
Billig, M., 24, 32, 105, 1 10, 1 19, 145, 1 52, 
181 
Black Panthers, 107-8 
'black sheep eff
ect', 1 29-31, 150 
Blaine, B., 222 
Boltanski, 1.L., 37 
Bonaldo, E., 170 
Bond, M.H., 219, 225 
Bontempo, R.N., 17, 201 , 213 
Bornstein, G., 183, 184 
Bourhis, R.Y., 218 
Bovarism, 219 
Boyes-Braem, P., 75 
Brewer, M.B., 36, 41, 45, 47, 76, 80, 83, 87-8, 
90, 104, 130, 181, 1 96 
Brito, P.N., 133 
Broadnax, S., 222 
Brosch, M., 109 
Brown, R., 1 32, 1 80 
Brown, R.I., 46, 83, 125, 1 27, 1 55, 167, 168, 
170, 171, 172, 175, 176, 177, 182, 202-3, 
204, 2 1 1 ,  214 
Bruner, I.S., 31, 128 
Bundy, R.P., 24, 105, 145 
Burnstein, E., 81 
Butera, F., 149, 151 
Byrne, D., 45, 47, 154 
Caddick, B., 168, 171-2, 175 
Campbell, D.T., 77, 104, 1 15, 1 19, 180, 
188 
Camps, V., 200 
Cannon, T., 107-8 
Cantor, N., 31-2, 41 
Capozza, D., 170 
Carroll, D.I., 159 


256 
Social identity: international pers
pectives 
Cartwright, D., 98, 125, 126 
Carugati, F., 13 
Carver, C., 226 
categorial diff
erentiation see differentiation, 
intergroup 
categorization, 4-5, 32 
and intergroup competition, 180-97 
recategorization and the discontinuity 
eff
ect, 196-7 
see also self-categorization; self­
categorization theory; social 
categorization 
Cavanni, L., 158 
centrality of group beliefs, 101-2 
Chaiken, S., 188 
Chang, 1.1., 1 59 
CIC theory see social identity theory 
Clack, F.L., 201 
Clark, R.D., 70, 142, 143 
Clemence, A., 13, 25 
Cochrane, S., 82 
Codol, I.-P., I, 9, 13, 14, 27, 28, 34, 35, 47, 
48 
cognitive centrism, 9 
cognitive creativity, 219 
cognitive differentiation eff
ect, 4 
cognitive stereotype effect, 4 
collection groups, 20-1 ,  1 32 
collectivism, 201-2, 203, 204, 205-6, 212-14, 
214, 225, 228 
Comby, L., 214 
common knowledge, in personal identity, 
14-17, 19-20 
communion, 18, 19 
communist ideology, \00 
'comparative fit' principle, 79, 127, 128 
competition and conflict 
intergroup, 5, 66, 67, 68-9, \06-7, 176, 
180-97 
interpersonal, 181-2, 192-3, 213 
concentricity thesis, 121, 122-3 
Condor, S., 41, 163, 172 
confidence in group beliefs, 101, l lO-l l  
Consumer Federation of America, 99 
context, dependence of categorization on, 
75-92 
contrast or cognitive diff
erentiation effect, 
4 
Converse, P.E., I I I  
Cooper, 1., 82 
Cooper-Shaw, L., 84 
cooperation between groups, 68-9, 106-7, 
168, 176 
Cordova, D., 222 
Correia, M.G., 133 
couples, relationships between see 
interpersonal relationships 
Coutant, D., 71 
Coutant-Sassic, D., 57, 69, 71, 72 
covariation hypothesis, xviii, 8-l l ,  26, 27, 30 
Cowan, G., 16 
Crano, W.o., 71 
Crocker, 1., 212, 221 ,  222, 223, 226 
Crook, K.F., 172-3 
Crosby, F., 222 
Crutchfield, R.S., 1 12 
Csepeli, G., 83 
culture, 56, 1 1 7-18, 120-1, 199-229 
Cylwik, A., 51 
Damon, W., 14 
Dann, H.D., 83 
Dardis, G.I., 181 
Darley, I.M., 180 
David, B., 79 
Davidson, B., 82 
Deaux, K., 160 
decay of groups, 61-2, 63, 73 
DeCharms, R., 126 
Deconchy, I.P., 101 
delinquency, adolescent, 22-3 
Delumeau, 1.1., 124 
depersonalization, 7, 77, 84, 91, 120 
Deschamps, I.-C., xviii, I -l l ,  25, 26, 27, 30, 
55, 127, 214 
Deutsch, M., 17, 131, 148 
De Vader, C.L., 88 
deviates, derogation of, 124-5, 132-4, 
1 35-41 
Devos, T., l - l l ,  214, 220 
Diamond, M., 97 
Diehl, M., 181 
diff
erentiation, 27-30 
and groups beliefs, 95, \08 
intergroup, 3, 4, 5-7, 8-l l ,  14, 24, 25-6, 
77, 121-2, 126-41 passim, 199-210, 
203 
intragroup, 14, 25-6, 124-41 
and the meta-contrast principle, 77-80 
and national identity, 120-1 
see also covariation hypothesis; 
prototypicality 
DiMaggio, A., 170 
discontinuity effect, 180-2, 187-97 
dissociation theory, 148-51, 1 52, 153 
Doise, W., 4, 9, 1 3-23, 26, 27, 55, 83, 
127, 132, 142, 143, 147, 195, 213, 
214, 215 
dominant groups, 20, 26, 27, 132, 214, 
228 


dominated groups, 20, 26, 27, 132, 214-20, 
222, 223, 228 
Doob, L.W., 1 17 
doomsday cult, 59 
Doosje, B., 132 
Doukhobors, 96 
Dovidio, J.F., 80, 196 
Drigotas, S.M., 180-97 
Dull, V., 80 
Dumont, L., 199 
Dunbar, R., 120 
Dunning, D., 85 
Durand Delvigne, A., 1 3, 21-2 
Durkheim, E., 180 
Duveen, G., 13 
Eagly, A.H., 132, 188 
Echebarria, A., 122 
ego and the self, 44 
egocentrism, 9, 52 
Eiser, J.R., 25 
Elejabarrieta, F., 122 
Ellemers, N., 1 32, 178, 213 
Ely, P.G., 166-79, 203, 204, 21 1 
Emler, N., 6, 22-3 
emotional experience and individualist! 
collectivist culture, 202 
'entitativity', 180, 183, 203 
Epstein, A.L., 103 
Epstein, S., 44 
Erikson, E., 14, 18, 44 
Etaugh, C., 1 56 
ethnicity, 103, 1 18, 120-1 
ethnocentrism, 1 1 5, 121 
ethnocentrism theory, 1 19 
Etzioni, A., 98 
Fabian, A. 
K., 1 17 
failure of groups, 64, 65 
Falger, V., 1 1 5  
Falomir, J.M., 152 
Farr, R.M., 13 
fascist ideology, 100 
fear-based competition, 194-6 
Fells, J., 122 
Festinger, L., 59, 106, 1 10, 126 
Fierres, R., 1 52 
Fiore, M., 1 57 
Fiske, S.T., 23, 36, 76 
Flament, C., 24, 105, 145 
Forgas, J.P., 158-9 
Forsyth, D.R., 129, 141 
Foti, R.J., 88 
Fox-Cardamone, L., 166-79 
Fraser, S.L., 88 
Index 
Fromkin, H.L., 45, 47 
fundamental group belief, 94, 103-6 
Gachoud, J.P., 142 
Gaelik, L., 28 
Gaertner, S.L., 80, 196 
Gagnon, A., 218 
Gara, M., 164 
Garza, R.T., 223, 224, 227 
gender, 19, 20, 21-2, 26, 143 
Gerard, H.B., 1 10, 126, 148 
Gergen, K.J., 3 
Giles, H., 76, 155, 169 
Gillin, J., 97 
globalization, 1 14, 123 
goals, group, 98-9 
Goethals, G.R., 1 52 
Goffman, E., 165 
Goldberg, L.R., 16, 17 
Gordon, P.c., 28, 32, 48 
Gordon, S.L., 1 56 
Gotab, A., 46 
Gouge, c., 83 
257 
Graduated and Reciprocated Initiatives in 
Tension-reduction (GRIT), 196 
Graetz, K.A., 181 
Gray, W.D., 75 
greed-based competition, 193-4 
Greenwald, A.G., 42, 44 
Grossman, M., 57, 69, 71 
group beliefs, 93- 1 1 3  
group concept, xviii 
group identity, 53-123 
developmental view of, 56-73 
group polarization, 81-3, 91 
group productivity, 60, 63, 67, 68 
Gudkyunst, W., 225 
Guimond, S., 222 
Haberkorn, G., 143 
Halverson, C.F., 16 
Hamilton, D.L., 132 
Hamilton, V.L., 124, 141 
Hampson, S.E., 89 
Hantzi, A., 24 
Hardie, E.A., 84-5 
Hardoin, M., 34, 48 
Harkins, S., 61, 180 
Harring, K., 181, 185, 187, 191 
Harrison, A.A., 156 
Hart, D., 14 
Harter, S., 19 
Harvey, O.J., 54, 107, 196 
Haslam, S.A., 31, 75-92, 127, 128 
Hayes, C.H., 92, 120 


258 
Social identity: international pers
pectives 
Henderson, M., 144 
HERMES project, 200 
Herringer, L.G., 223, 224, 227 
Hertz, F., 1 1 5  
Hewstone, M., 24, 144, 169 
Higgins, E.T., 102 
Hill, C.T., 156 
Hinkle, S., 166-79, 202-3, 204, 21 1-12 
Ho, H., 196 
HolTman, C.D., 16 
HolT
man, M.L., 46 
Hofstede, G., 56, 200 
Hogg, M.A., 76, 79, 82, 83, 84-5, 86, 88, 105, 
1 19, 125-31 passim, 139, 146, 1 54, 1 55, 
162, 166, 167, 1 77-8, 2 1 1  
Holyoak, K.J., 28, 32, 48 
Holzworth, D.W., 84 
Homans, G., 96 
Hood, W.R., 54, 107, 196 
Hopkins, N., 6 
Horwitz, M., 180 
Hostetler, J.A., 96, 102 
Hoyle, R.H., 181, 1 93, 1 95 
Hu, L., 214 
Huguet, P., 149, 1 5 1  
Hui, C.H., 200-1 
Huici, C., 154, 155, 215 
human identity, 31, 77 
Hurtig, M.C., 21, 37-8 
identification stage of group development, 
58-60, 62, 68, 73 
ideology, 99-101 
Imago theory, 18 
INDCOL scale, 201 
independence, 54-6 
individualism, 199-210, 212-14, 214, 225, 
228 
individualization, 35, 215 
individuation stage of 
group development, 
60-1, 63, 73 
influence, 126 
informational, 1 26, 1 34, 1 39-41 
of majorities, 142 
of minorities, 70-2, 73, 147-8 
normative, 1 26, 1 34 
and social categorization, 142-53 
Inoff, G., 16 
Insko, C.A., 180-97, 203, 209 
InstIa, P., 2 1 1 -29 
interest groups, 99 
intergroup relationships, 66-9, 73 
competitive, S, 66, 67, 68-9, 106-7, 176, 
180-97 
cooperative, 68-9, 106-7, 168, 176 
interpersonal relationships as, 1 54, 
1 55-65 
use of new comparison dimensions in, 
166-79 
interpersonal attraction, 1 54 
and group behaviour, 83-4, 85, 91 
interpersonal relationships, S, 6, 55-6 
competition and conflict in, 181-2, 192-3, 
213 
as intergroup relationships, 154, 1 55-65 
Jackson, L.A., 178 
James, W., 2, 18, 44 
Janis, I.L., 126 
Jarymowicz, M., 44-52 
Jaskar, K., 222 
Jehovah's Witness group, 107 
Jensen, M.A.C., 57 
Jodelet, D., I 3  
Johnson, D.M., 75 
Johnston, L., 24 
Jones, E.E., 69, 1 10, 126, 134 
Jovanovic, M., 146 
Jung, C.G., 18 
Kaiser, c., 143, 144 
KaminolT, R., 1 17 
Kaplan, S.J., 17 
Karytowski, J., 45 
Kastersztein, J., 168 
Kedourie, E., 1 15 
Kelley, H.H., 1 54, 183, 187, 192 
Kelly, J., 54 
Kelman, H.C., 120 
Kenny, D.A., 1 29 
Kiesler, C.A. and Kiesler, S.B., 81 
Killian, L.M., 103 
Kinder, D.R., 222 
King, G., 102 
Kirchler, E., 1 5  
Kitayama, S., 201 ,  202 
Kleck, B.E., 1 93 
Klein, M., 1 8  
Kluckholn, c., 97 
Kobuszewska, E., 52 
Kohlberg, L., 46 
Kohn, H., 1 14, 1 1 5 
Kotkin, M., 1 56 
Krech, D., 1 12 
Kruglanski, A.W., 101, 109 
Ku Klux Klan, 98 
Kuhn, M.H., 14 
Lalonde, R.N., 171 
Lane, R.E., 100 


Index 
259 
language, 1 17-18 
Larsen, K., 83 
Lassaline, M.E., 75 
Latane, B., 61, 180 
La Voie, L., 129 
Lea, M., 82 
leadership, 88, 178 
LeBon, G., 180 
L'Ecuyer, R., 14 
Lee, J., 196 
Lee, S., 82 
legitimacy/illegitimacy, 167, 170-2, 175-6 
Legrenzi, P., 149 
Leippe, M.R., 141 
Lemaine, G., 151, 168, 169, 171, 174, 176, 
177 
Leung, K., 201 
Levine, J.M., 57, 125, 126 
Le Vine, R.A., 1 19 
Levy, S.B., 102 
Lewin, K., 54 
Leyens, J.-P., 17, 129, 214, 215 
Lindskold, S., 196 
Linville, P.W., 69 
Linz, J.J., 1 23 
Lippitt, R., 54 
Llavata, E., 152 
Lloyd, B., 13 
Lodewijkx, H.G.M., 1 92 
L6pez-Saez, M., 1 99-210 
Lord, R.G., 88, 90 
Lorenzi-Cioldi, F., 9, 10, 1 3, 1 5, 1 9, 20-1 , 22, 
26, 1 32, 214 
Lott, A.J.and Lott, B.E, 83, 131 
Louis-Gueerin, c., 1 3  
Low, S.M., 1 17 
Lubovitcher Hassidim, 102 
Lucca, N., 201 ,  213 
Luhtanen, R., 212, 221 ,  222, 226 
Lui, L., 80, 87 
Luria, A.R., 27 
Maass, A., 70, 142, 143 
Mackie, D., 82, 143, 144, 146, 2 1 1  
majorities, influence of, 142 
Makris-Botsaris, E., 1 9  
Mann, J., 80, 196 
Maoris, 59 
Maras, P., 176 
March, J.G., 98 
Marcia, J.E., 14 
Mariategui, J.c., 1 1 8  
Markus, H., 1 8, 47, 48, 201 , 202 
Marques, J.M., 124-41, 1 50, 215, 220 
Martin, R., 143, 144 
Martinez-Taboda, M., 21 1-29 
Maslow, A.H., 54 
Mazzara, B., 36 
McAdams, D.P., 18, 19 
McCallum, D.M., 181, 1 87 
McCusker, C., 201 
McDougall, W., 180 
McGarty, C., 77, 79, 82, 86, 90, 91, 92, 127, 
139 
McGrath, J.E., 54 
McGuire, W.J. and McGuire, C.V., 14, 23 
McPartland, T.S., 14 
McPhail, C., 180 
Mead, G.H., 2, 47 
Medin, D.L., 75, 77, 88-9, 89-91, 92 
Mentzer, S.J., 193 
Mervis, C.B., 75 
Messick, D., 21 1 
meta-contrast principle, 77-80, 91, 120-1, 
127, 128, 1 36-7, 138 
Meyer, G., 55, 127 
Miller, N., 196 
Miller, R.L., 47 
minimal group paradigrn, 1 10, 1 82-3 
minorities, 142, 166 
influence of, 70-2, 73, 147-8 
Mischel, W., 31-2, 41 
Mlicki, P.P., 213 
Moghaddam, F.M., 178, 219 
Moise, L.C., 218 
Montei1, J.M., 35 
Montero, M., 218 
Moore, M., 156 
morale, 178 
Morales, J.F., 164, 199-210, 215 
Moreland, R.L., 57, 104, 105, 125, 126 
Mormons, 102 
Morris, W.N., 56 
Moscovici, S., 13, 70, 142, 149 
motivation, 24-5, 28, 178 
Moya, M.C., 1 54-65 
Mucchi-Faina, A., 142 
Mugny, G., 70, 125, 142-53 
Mullen, B., 214 
Multiple Alternative Matrices (MAMs) 
experiments, 183-6 
Murphy, G.L., 89 
Murrell, A., 80, 196 
Mussolini, B., 100 
'myths of descent', 1 18 
Nakbi, J.L., 1 5  
Nascimento-Schulze, C.M., 215, 221 
nation-state, 1 14-15, 1 16, 1 1 8-19, 121 
national identity, 1 14-23 


260 
Social identity: international pers
pectives 
National Rifle Association, 99 
National Right to Life Committee, 99 
Navarro, E., 146 
needs, satisfaction of, 104 
Neisser, D., 75, 92 
Nemeth, c., 70, 72, 149 
Neuberg, S.L., 23, 76 
Nezlek, J., 56 
T
he Night Porter (film), 1 58 
Niznik, J., 1 22 
'normative fit' principle, 79, 91, 127-8 
normative influence, 126, 1 34 
norms, group, 96, 106, 1 33-4 
denotative, 127-9, 1 34, 1 35, 1 36-7 
prescriptive, 129, 134, 1 35, 1 36-7, 138 
Oakes, P.J., 31, 32, 33, 75-92, 1 19, 1 27, 1 28, 
146, 162, 2 1 1  
Odbert, H.S., 16 
O'Dea, T.F., 102 
Oldenquist, A., 55 
Onorato, R., 86 
Ortony, A., 90 
Osgood, C.E., 1 96 
Ostell, A., 83 
Ostrom, T.M., 8, 104 
'outgroup homogeneity effect', 8 
Paez, D., 1 24-41, 2 1 1 -29 
Pages, R., 2 
Paicheler, H., 1 3  
Palmonari, A., 1 3, 1 5  
Papastamou, S., 143, 144 
paralysis, ingroup, 146, 148 
Parelman, A., 1 56 
Park, B., 35, 69 
Parsons, T., 97 
particularization, 32, 35 
Pavelchak, M.A., 23, 36 
Pawluczuk, B., 5 1  
Peabody, D., 1 6  
Pemberton, M.J., 196 
Peplau, L.A., 1 55, 156, 162 
Perez, J.A., 1 25, 142-53 
Perie, M., 85 
persecution, collective, 1 24-5 
personal attraction see interpersonal 
attraction 
personal identity, 24-5, 31, 47-9, 54-6, 76, 
77, 160, 228 
common knowledge in, 14-17, 19-20 
and social identity, 1-1 1 , 213 
as a social representation, 1 3-23 
personal relationships see interpersonal 
relationships 
Personnaz, B., 168 
Pfost, K.S., 157 
Phinney, J.S., 222 
Piaget, J., 46 
Pichevin, M.F., 21, 37-8 
Pinkley, R.L., 181, 1 85, 187, 191, 193, 1 95 
Piolat, M., 37-8 
Pizzigati, K.A.L., 16 
Podeschi, D.M., 32 
polarization, group, 81-3, 9 1  
political participation, 222, 229 
Pombeni, M.L., 15 
positive self-identity/evaluation, 6, 24-5, 
181 
see also self-esteem, personal 
positive social identity, 93, 121-2, 221 
and manifest influence, 148 
use of new dimensions of intergroup 
comparison to attain, 166-79 
see also self
-esteem, collective 
Pratkanis, A.R., 44 
Pratto, F., 36 
primus inter pares (PIP) effect, 27 
see also superior conformity of the self 
productivity, group, 60, 63, 67, 68 
Proshansky, H.M., 1 17 
prototypicality, 30-1, 32, 33-5, 37, 38, 41-2, 
75-92, 127-9, 134 
Pryor, J.B., 104 
Quattrone, G., 69 
Questionnaire o
f
 Social Perception, 48 
Quevedo, S., 212 
Rabbie, J.M., 1 80, 1 82, 1 90 
Ramirez, S., 123 
Rauma, D., 124, 141 
Ravaud, J.F., 13, 14-15 
Raviv, A., 109 
realistic conflict theory, 1 88 
recategorization, and discontinuity eff
ect, 
1 96-7 
referent informational influence, 1 39 
regionalism, 1 16, 1 17, 121 
Reicher, S., 22-3, 76, 1 19, 1 27, 146, 162, 
21 1 
Reis, H.R., 56, 1 96 
relationships see intergroup relationships; 
interpersonal relationships 
Reykowski, J., 45, 46 
Reynolds, Y., 92, 1 1 5  
Ricci Bitti, P., 1 3  
Richard III Society, 99 
Riecken, H.W., 59 
Riley, S., 156 


Index 
261 
Rips, L.J., 90 
Rivera, A.N., 222 
Robalo, E.M ., 130 
Robbers' Cave experiments, 106-7 
Robinson, P.W., 19 
Rocha, S.A., 130 
Rodriguez, C.P., 214 
Rogers, C., 54 
Rokeach, M., 96 
Rook, D.S., 56 
Rosch, E., 30-1, 41, 75, 76, 77, 80, 87 
Rosenberg, S., 164 
Ross, G.F., 168, 170, 171, 175 
Roth, E .M., 89 
Rothbart, M., 35, 69 
Roux, P., 25 
Rubin, Z., 1 56 
Rucker, M.H., 156 
Rush, M.C., 88 
Russell, CT., 107 
Russell, J.E.A., 88 
Rust, M.C., 196 
Rutherford, J.F., 107 
Sachdev, I., 83, 105 
Salazar, J .M., 1 14-23, 219, 224 
Salisbury, R.H., 99 
Sanchez-Mazas, M., 146, 148, 150 
Sande, G.N., 16 
Sanders, G.S., 81 
Sarbin, T.R., 2 
Sarchielli, G., 13 
Schachter, S., 54, 59, 1 26 
Schadron, G., 214 
Schei̖r, M., 226 
schematic traits, 47 
Schoben, E.J., 89 
Schopler, J., 180-97, 203, 209 
Sears, D.O., 222 
Sedikides, C., 8 
self 
and the ego, 44 
fluidity of the, 3 
as object, 44 
and prosocial attachment, 44-6 
social, 44-6 
as subject, 44 
superior conformity of the, 9, 16, 27 
self-categorization, levels of, 14, 31, 36, 
146 
self-categorization theory, 7-8, 30-3, 34, 35, 
65, 93-4, 105 
and idea of depersonalized social 
attraction, 84 
and group polarization, 8 1-3 
prototypicality in, 30-1, 33, 34, 76-81 ,  
85-6, 91, 127-9 
and social influence, 147-8 
self-consciousness, 212, 226-8, 229 
self-definition, 7-8 
self-esteem 
collective, 212, 220-3, 229 
personal, 6, 19, 20, 21 1-12, 220-3, 229 
self-other comparison, 24-42 
see also differentiation; similarity 
relationships 
self-permanence, 3 
self-presentation, 177-8 
self-representation, duality of, 2 
self-stereotyping, 126-7, 1 28, 131 
self structure, 45 
self-we-others schemata, 44-52 
Semin, G.R., 17 
Serino, C., 24-42 
Serooio, R.G., 133 
Seton-Watson, H., 1 1 5  
Sewell, D.R., 88 
Shaw, M.E., 126 
Sherif, CW., 54, 107, 196 
Sherif, M., 5, 54, 69, 81, 83, 106-7, 155, 196, 
197 
Shils, E., 100 
Sills, D.L., 98 
Silver, M., 83 
similarity relationships, 3, 4, 6-8, 14, 27-30, 
33-5, 44-5, 47, 89-90 
asymmetries in, 37-41, 42 
and group beliefs, 95, 108 
intergroup, 14, 14-17, 20-2 
and the meta-contrast principle, 77-80 
metrical approach to, 28-9 
and national identity, 120-1 
see also covariation hypothesis; 
prototypicality 
Simmel, G., 125 
Simon, B., 132 
Simon, H.A., 98 
Sinclair, A., 215 
Singh, A., 22 
Singleton, R., 81 
Smith, A.D., 1 1 5, 1 1 8  
Smith, C., 214 
Smith, J., 47 
Smith, P.B., 219, 225 
Smith, P.M., 83 
Smith, V.A., 181 
Snyder, CR., 45, 47, 193 
Sobczyk, K., 52 
social attraction, and prototypicality, 83-5, 
91 


262 
Social identity: international pers
pectives 
social categorization, 4-8, 105 
context dependence of, 75-92 
deductive effect of, 1 5  
graded structure of, 75-92 
productive character of, 31-2 
and social influence, 142-53 
social comparison see diff
erentiation; self­
other comparison; similarity 
relationships 
social constructivism, and latent influence, 
149 
social control, and intragroup differentiation, 
1 24-41 
social creativity, 166-79 
social identity theory, 4-8, 55, 65, 93, 105, 
1 19, 167-9, 181-6 
social influence see influence 
social loafing, 61 
social representations, in personal identity, 
1 3-23, 213 
social self, 44-6 
sociocentrism, 9 
Sokolowska, G., 52 
Spears, R., 82, 132 
Srull, T.K., 28 
stability/instability, 167, 170, 171, 175-6 
Stahelski, A., 1 87 
state, 1 14-1 5, 1 16, 1 18-19, 1 20 
stateless nationalism, 1 16 
status, 167-8, 170-2, 175-6, 212-23, 228-9 
Staw, B.M., 70 
Stephenson, G.M., 1 57 
stereotypes, 4, 7, 1 5, 2 14, 215 
see also self-stereotyping 
Stewart" 1 32 
Story, A.L., 85 
Stotland, E., 103 
Strenta, A., 193 
Stroup, H.H., 107 
subjective group dynamics, 131-2, 1 39-41 
success of groups, 64-5 
Suls, 1.M., 47 
superior conformity of the self, 9, 16, 27 
supra-nationalism, 1 16, 1 18, 121, 122 
symbolic individual mobility strategy, 219, 
220 
Szuster-Zbrokewicz, A., 51, 52 
t-group movement, 57 
Tajfel, H., xviii, 4, 5, 6, 26, 41, 55, 77, 105, 
1 19, 1 25, 126, 127, 1 3 1 , 145, 1 55, 166, 
168, 169, 181 
on legitimacy/stability of 
groups, 167, 171, 
175-6 
minimal group paradigm, 1 10, 1 82-3 
on minority groups, 103 
on national and ethnic identity, 1 14 
on similarity/difference judgements, 24, 25, 
29 
on temporal aspects of social identity, 163 
Tap, P., 1 3, 19, 46 
Taylor, D.M., 178, 219 
Taylor, L.A., 166-79 
temporal aspects of social identity, 163, 176-7 
territoriality, 1 17, 120-1 
Thibaut, 1.W., 106, 1 54, 183, 192 
Toch, H., 100, 109 
Torregrosa, 1.R., 123 
Tougas, F., 222 
Trans-Species Unlimited, 98-9 
transcendence, and national identity, 120, 121 
Triandis, H.C., 56, 201, 203, 213 
Triska, J.F., 100 
Truman, D.B., 99 
Tuckman, B.W., 57 
Turner, J.e., xviii, 5, 14, 24, 34, 35, 36, 41, 55, 
65, 75-92, 93, 105, 1 19, 1 26, 128, 1 32, 
1 39, 146, 163, 166, 167, 169, 181, 1 84, 
1 85, 203, 21 1 
on comparative fit, 127 
on group polarization, 81, 82, 83 
on influence of groups, 147-8 
on meta-contrast principle, 77, 79, 80, 1 27, 
162 
and the minimal group paradigm, 9, 182 
on normative fit, 1 27-8 
on positive self-identity, 6 
and prototypicality theory, 30, 75-92, 127 
on referent informational influence, 139 
on self-definition, 7-8 
see also self-categorization theory 
Tversky, A., 27-8, 29, 34, 41, 77 
Valencia, J., 122 
values, 96-8, 212, 220-3, 225, 226 
Vander Zanden, J.W., 98 
van Knippenberg, A., 83, 168, 171, 174, 176, 
177, 178, 215 
Vega, L., 199-210 
Verkuyten, M., 1 9  
Villareal, M., 1 22, 201, 2 1 3  
Vine, I., 1 1 5 
Vitti, E., 142 
Volpato, C., 9, 10, 142, 144 
Wattenmaker, W.O., 89 
Wetherell, M.S., 76, 79, 81, 82, 83, 1 19, 127, 
146, 162, 2 1 1  
Wheeler, L., 56, 1 96 
White, B.J., 54, 107 


White, R., 54 
Wicklund, R., 193 
Wilder, D.A., 95, 108, 143, 144, 145-6 
Wilke, H., 83, 178, 215 
Wilkes, A.L., 25, 29, 127 
Williams, J., 1 55, 172 
Williams, K., 61, 180 
Williams, R.M., 97 
Winter, D.G., 18 
Wish, M., 17 
Wisniewski, E.J., 75 
witch-hunting, 124 
Wittenbraker, J., 184 
Wolfe, D., 103 
Wood, G., 69 
Worchel, S., 53-73 
Index 
Worth, L.T., 143 
Wright, S.c., 178, 219 
Wyer, R.S., 32 
Yamagishi, T., 124 
Yzerbyt, V.Y., 129, 214 
Zander, A., 98, 103, 125 
Zavalloni, M., 13, 25, 26 
Zebrowitz, L., 219 
Zeegers, W., 13, 16 
Zeigamik, B., 64 
Ziller, R., 46 
Zimbardo, P.G., 141, 180 
Zuckier, H., 36 
263 


Social Identity, 
Intergroup Conflict, and 
Conflict Reduction,
Volume 3
Richard D. Ashmore
Lee Jussim
David Wilder,
Editors
OXFORD UNIVERSITY PRESS


SOCIAL IDENTITY,
INTERGROUP CONFLICT, and
CONFLICT REDUCTION


RUTGERS SERIES ON SELF AND SOCIAL IDENTITY
Series Editors
Richard D. Ashmore and Lee Jussim
Rutgers, The State University of New Jersey
Editorial Board
Kay Deaux, City University of New York
Kenneth J. Gergen, Swarthmore College
Dorothy C. Holland, University of North Carolina at
Chapel Hill
Daniel M. Ogilvie, Rutgers, The State University of
New Jersey
Suzanne Ouellette, City University of New York
Seymour Rosenberg, Rutgers, The State University of
New Jersey
Diane N. Ruble, New York University
Ronald T
aylor, T
emple University
Peggy A. Thoits, Vanderbilt University
David Wilder, Rutgers, The State University of New
Jersey
Volume 1
Self and Identity: Fundamental Issues
Volume 2
Self, Social Identity, and Physical Health: Interdisciplinary 
Explorations
Volume 3
Social Identity, Intergroup Conflict, and Conflict Reduction


RUTGERS SERIES ON SELF AND SOCIAL IDENTITY
Volume 3
SOCIAL IDENTITY,
INTERGROUP CONFLICT,
and CONFLICT REDUCTION
Edited by
Richard D. Ashmore
Lee Jussim
David Wilder
1
2001


3
Oxford
New York
Athens
Auckland
Bangkok
Bogotá
Buenos Aires
Cape T
own
Chennai Dar es Salaam Delhi Florence Hong Kong Istanbul
Karachi
Kolkata
Kuala Lumpur Madrid Melbourne Mexico City Mumbai
Nairobi
Paris São Paulo Shanghai Singapore T
aipei T
okyo T
oronto Warsaw
and associated companies in
Berlin Ibadan
Copyright © 2001 by Oxford University Press
Published by Oxford University Press, Inc.
198 Madison Avenue, New York, New York 10016
Oxford is a registered trademark of Oxford University Press.
All rights reserved. No part of this publication may be reproduced, 
stored in a retrieval system, or transmitted, in any form or by any means, 
electronic, mechanical, photocopying, recording, or otherwise, 
without the prior permission of Oxford University Press.
Library of Congress Cataloging-in-Publication Data
Social identity, intergroup conflict, and conflict reduction / [edited by] Richard D. Ashmore,
Lee Jussim, and David Wilder.
p. cm.—(Rutgers series on self and social identity ; 3)
Includes bibliographical references and index.
ISBN 0-19-513742-6; 0-19-513743-4 (pbk.)
1. Group identity. 2. Conflict management. 3. Ethnic relations. I. Ashmore, Richard D. II.
Jussim, Lee J. III. Wilder, David, 1949–
IV. Series
HM753.S6197
2001
302.4—dc21
00-053761
9 8 7 6 5 4 3 2 1
Printed in the United States of America
on acid-free paper


Preface to Volume 3
This volume is important to me both professionally and personally: it is a
significant part of my evolving life story. At the same, my co-editors and I
believe that this book provides a start toward an important new perspective
on intergroup violence, a problem that brings extensive material destruc-
tion and widespread human suffering around the globe.
In the spring of 1964, while still an undeclared and uncertain history ma-
jor, I read Gordon Allport’s The Nature of Prejudice for a course entitled,
“Medieval Anti-Semitism.” It was an epiphany—I would become a social
psychologist, and I would seek to understand the causes and cures of inter-
group conflict. That summer I enrolled in my first psychology classes, fin-
ished the psychology major during my senior year, and went on to get my
Ph.D. in social psychology with a specialization in intergroup relations.
When I read Allport and chose my occupation and specialty area, I did
not know Erikson’s seminal work on identity development. As I look back,
though, his ideas about adolescent identity search fit my experience very
well. Becoming a social psychologist and seeking to understand prejudice
and discrimination allowed me to bring together my childhood identifica-
tions (with two parents who were simply good people, with a succession of
outstanding teachers [from Mrs. Cirino in the fourth grade to Professor
Lewis Spitz at Stanford who, to this day, provides the model for how I teach
a large undergraduate class]), my then current concerns and values (I be-
lieved in 1964 that conflict among groups was the greatest problem facing
humans; I felt that we should and could do something about this), and my
future work (becoming a scientist and uncovering why people engage in
group-based harm-doing).


Now it is 2000. With many twists and turns, I today still see myself as an
intergroup relations researcher. And I continue to believe that intergroup
conflict is the major problem facing the world. What is different, one of
those twists and turns, is that I now consider myself also an identity re-
searcher. This volume allows me to bring together, then, my personal life
story, especially my commitment to understanding intergroup conflict, with
my evolving professional self-definition as a social scientist who studies self
and identity.
Although the world today is, in many ways, quite different from that of
almost 40 years ago, group-based violence, aggression, and suppression are
no less in evidence. Although the global cold war has ended, we now ex-
perience a seemingly never-ending succession of small and local “wars.”
These are global in a new way—group-based strife occurs on all continents
and every corner of the world and yet, because of powerful new communi-
cations technologies, the whole world often is involved if only as CNN
viewers.
Since I entered the field in the mid-1960s, the social and behavioral 
sciences have made great strides in seeking to understand and ameliorate
intergroup conflict. At the same time, there are still many gaps in our
knowledge.
My co-editors and I believe that social identity may provide a powerful
perspective for filling in some of these gaps. Social identity has certainly
been implicated in past work on group strife, but these efforts have been
piecemeal and isolated from one another across disciplinary boundaries.
Our goal in this volume was to bring together scientists from diverse social
and behavioral sciences to address the multiple ways that social identity and
intergroup conflict are interrelated. We believe that our contributors have
done an excellent job, on the one hand, of highlighting major current theo-
ries, concepts, and methods and, on the other, of identifying the crucial 
issues that need to be addressed if we are to develop a more complete and
coordinated social identity perspective on intergroup conflict. We hope that
this volume contributes to the development of such a paradigm and that it,
along with other approaches, facilitates social technologies that significantly
reduce the incidence and intensity of people harming others because of
group membership.
Piscataway, New Jersey
Richard D. Ashmore
May 2000
vi
Preface to Volume 3


Acknowledgments
Many individuals and organizations helped us as we planned and carried out
the Third Rutgers Symposium on Self and Social Identity and produced
this volume. We thank you all. We received so much help from so many
sources that we are sure to have forgotten to mention some here, and for
that we apologize.
We first thank the symposium speakers/chapter authors. They were 
chosen because they each have distinguished scientific careers, and that past
behavior predicted their performance in this endeavor. They produced
high-quality symposium presentations and excellent chapters appearing in
this book. In addition, they had to deal with fairly picky editors. We thank
each of you for sticking with us and producing the present polished contri-
butions.
Next, we wish to thank our publisher, Oxford University Press, and our
editors there, Joan Bossert, Philip Laughlin, and Catharine A. Carlin. Joan
has been an invaluable source of guidance and support since she first re-
ceived the proposal for the Rutgers Series, and Philip and Catharine ably 
assisted us as we moved the present volume from an idea, to a conference,
to a book.
There are many people to thank for sharing their ideas about the topic of
social identity and intergroup conflict, for suggesting speakers/authors, and
for providing comments on chapter drafts. We especially wish to thank the
members of our editorial board, who provided valuable feedback on initial
and revised proposals. In addition, the following people provided input
concerning one or more facets of the conference/book: Myron Aronoff,
Mahzarin R. Banaji, James Jones, Louis Kreisberg, B. Ann Bettencourt,


Yael Zerubavel, David Sears, Daniel M. Ogilvie, Jessica Heppen, and Rick
Lau.
This volume, as with its two predecessors, deserves to be part of the Rut-
gers Series on Self and Social Identity. We have been helped at every step of
the way by the faculty, staff, students, and administrators of Rutgers–The
State University of New Jersey. The following members of the Rutgers
University community were particularly helpful: Steven Lione, Thelma
Collins, Deana Guden-Pagnozzi, and the staff of the University Inn and
Conference Center; Stacey Hersh-Ehling of the Rutgers News Service;
Joanne Aguglia of the Douglass College Center; Rosemary Manero and her
colleagues in the Rutgers Dining Service; Marc Beebe, Justine Cardone,
Kathy Maloney, Rae Frank, Maria Brewer, Anne Sokolowski, and Alison
Smith, all of the Department of Psychology, who exhibited patience and
good humor in their extremely efficient handling of day-to-day tasks and
deadline pressures.
Other members of the Rutgers community deserve special thanks. First,
Charles Flaherty, chair of the Department of Psychology, has supported the
Rutgers Series since its inception, and he made departmental staff and other
resources available for our use in putting together the third symposium and
present associated book. Second, we express our gratitude to the faculty and
graduate students of the doctoral program in social psychology. Members
of the Social Psychology Area served as hosts for speakers, helped with
symposium registration, transported speakers to and from airports and train
stations, directed traffic between symposium sessions, and helped conduct
myriad other essential tasks. Third, we thank the following members of the
Rutgers community for their generous financial assistance: Joseph J. Seneca
(Vice President for Academic Affairs), and Richard F. Foley (Dean, Faculty
of Arts and Sciences).
We are also grateful for grants from the National Institutes of Health
(R13MH60129-01; Carolyn Morf, Chief, Personality and Social Cognition
Program) and from the National Science Foundation (SBR-9811485;
Steven J. Breckler, Program Director, Social Psychology) that supported
both the Third Rutgers Symposium and the production of this volume.
viii
Acknowledgments


Contents
Contributors
xi
1
introduction: Social Identity and Intergroup Conflict
3
Lee Jussim, Richard D. Ashmore, and David Wilder
PART I / The Contribution of Individuals’ Identities and the Collective
Identities of Social Groups to Intergroup Conflict
2
Ingroup Identification and Intergroup Conflict: When Does Ingroup Love
Become Outgroup Hate?
17
Marilynn B. Brewer
3
Ethnic Identity, National Identity, and Intergroup Conflict: The 
Significance of Personal Experiences
42
Thomas Hylland Eriksen
PART II / The Contribution of Ethnic and National Identities to 
Political Conflict in the United States
4
The Meaning of American National Identity: Patterns of Ethnic Conflict
and Consensus
71
Jack Citrin, Cara Wong, and Brian Duff


5
Communal and National Identity in a Multiethnic State: A Comparison of
Three Perspectives
101
Jim Sidanius and John R. Petrocik
PART III / The Contribution of Social Identity to Violent Intergroup
Conflict
6
Social and Role Identities and Political Violence: Identity as a Window on
Violence in Northern Ireland
133
Robert W. White
7
Individual and Group Identities in Genocide and Mass Killing
159
Ervin Staub
PART IV / The Role of Social Identity in Reducing Intergroup Conflict
8
The Role of National Identity in Conflict Resolution: Experiences from 
Israeli-Palestinian Problem-Solving Workshops
187
Herbert C. Kelman
9
conclusion: T
oward a Social Identity Framework for 
Intergroup Conflict
213
Richard D. Ashmore, Lee Jussim, David Wilder, and Jessica Heppen
Index of Subjects
251
Index of Names
265
x
Contents


xi
Contributors
Richard D. Ashmore, Department of Psychology, Tillett Hall, Rutgers, The State
University of New Jersey, 53 Avenue E, Piscataway, NJ 08854-8040. E-mail: 
ashmore@rci.rutgers.edu.
Marilynn B. Brewer, Department of Psychology, Ohio State University, 1885 Neil
Avenue, Columbus, OH 43210. E-mail: brewer.64@osu.edu.
Jack Citrin, Department of Political Science, 210 Barrows Hall #1950, University 
of California, Berkeley, Berkeley, CA 94720-1950. E-mail: gojack@uclink.
berkeley.edu.
Brian Duff, Department of Political Science, 210 Barrows Hall #1950, University 
of California, Berkeley, Berkeley, CA 94720-1950. E-mail: duff@socrates.
berkeley.edu.
Thomas Hylland Eriksen, Department of Social Anthropology, University of Oslo,
Moltke Moesvei 31, Eilert Sundts hus, 6th floor, P. O. B. 1091 Blindern, N-0317
OSLO, Norway. E-mail: t.h.eriksen@ima.uio.no.
Jessica Heppen, Department of Psychology, Tillett Hall, Rutgers, The State Uni-
versity of New Jersey, 53 Avenue E, Piscataway, NJ 08854-8040. E-mail: 
jheppen@rci.rutgers.edu.
Lee Jussim, Department of Psychology, Tillett Hall, Rutgers, The State University
of New Jersey, 53 Avenue E, Piscataway, NJ 08854-8040. E-mail: jussim@rci.
rutgers.edu.
Herbert C. Kelman, Psychology Department, Harvard University, Cambridge, MA
02138. E-mail: hck@wjh.harvard.edu.
John R. Petrocik, 210 McAlester Hall, Department of Psychology, University of 
Missouri-Columbia, Columbia, MO 65211. E-mail: PetrocikJ@missouri.edu.


Jim Sidanius, Department of Psychology, 1285 Franz Hall, P. O. Box 951563, 
405 Hilgard Avenue, Los Angeles, CA 90095-1563. E-mail: Sidanius@psych.
ucla.edu.
Ervin Staub, Department of Psychology, University of Massachusetts, Amherst, MA
01003. E-mail: estaub@psych.umass.edu.
Robert W. White, Department of Sociology, Indiana University at Indianapolis, 
Cavanaugh Hall 303, 425 University Boulevard, Indianapolis, IN 46202. E-mail:
rwwhite@iupui.edu.
David Wilder, Department of Psychology, Tillett Hall, Rutgers, The State Uni-
versity of New Jersey, 53 Avenue E, Piscataway, NJ 08854-8040. E-mail:
wilder@psych-b.rutgers.edu.
Cara Wong, Department of Political Science, 210 Barrows Hall #1950, University 
of California, Berkeley, Berkeley, CA 94720-1950. E-mail: cwong@bravo.
berkeley.edu.
xii
Contributors


SOCIAL IDENTITY,
INTERGROUP CONFLICT, and
CONFLICT REDUCTION


This page intentionally left blank 


Lee Jussim
1
Richard D. Ashmore
David Wilder
Introduction
Social Identity and Intergroup Conflict
Intense conflict between nations, ethnic groups, and political groups has
created some of the most enduring and lethal social problems humans have
ever faced. Slavery, conquest, exploitation, war, mass murder, and genocide
are among the most extreme manifestations of intergroup conflict. Such
conflicts remain as much a part of the human condition today as ever. 
Over the last two decades, social and behavioral scientists have increas-
ingly emphasized the role of self and identity in the causes and conse-
quences of intergroup hostility and harm-doing. This is, in part, because
issues of self and identity occur at multiple levels of analysis: individual
within a social context, groups within a society, and nations of the world.
At the level of individual within a social context, ethnic identity may con-
tribute to both in-group bonds and hostility toward other groups (Jones,
1997). At the level of groups within a society, social identity theory and
self-categorization theory emphasize the potential for group-based identi-
ties to foster support for the status quo among higher power and status
groups, and to foster intergroup competition and movements for political
change among lower power and status groups (T
ajfel, 1981; Thoits & Vir-
shup, 1997; Turner, 1987). Finally, at the level of modern nations, identity
is implicated in a variety of conflicts—state violence against internal mi-
norities (e.g., Goldhagen 1996; Staub, 1989), armed uprisings by ethnic
groups against central governments (e.g., Gurr, 1993), and international
conflicts (e.g., Greenfeld & Chirot, 1994; Keegan, 1993). Furthermore,
addressing issues of identity may be necessary to reduce certain conflicts
between groups (e.g., Kelman, 1997; Kriesberg, Northrup, & Thorson,
1989).
3


Although self and identity play a critical role in some (perhaps most) in-
tergroup conflicts, social scientific analyses of that role have thus far been
largely isolated from each other. This is unfortunate because, just as the sci-
entific analysis of conflict has blossomed in recent years, so too has there
been a resurgence in work on self and identity (Ashmore & Jussim, 1997).
Multiple and diverse lines of contemporary social science research and
theory demonstrate that self and identity are central to the understanding
of human thought, feeling, and action. They are also crucial in accounting
for interrelationships between the individual and larger sociocultural insti-
tutions and systems (Holland, 1997; T
ajfel, 1981; Thoits & Virshup, 1997).
Thus, there is much in the areas of self and social identity that might inform
our understanding of intergroup conflict.
The main purposes of this book are to elucidate ways in which: 
• Social identities create and exacerbate intergroup conflicts.
• Intergroup conflict influences social identity.
• Addressing social identity issues might help reduce some inter-
group conflicts. 
T
o achieve these goals, we invited a group of talented social and behav-
ioral scientists to address relationships between intergroup conflict and
identity. First, we asked them to address the locus of social identity. Are so-
cial identities primarily phenomena of individual psychology or are social
identities primarily phenomena that reside in the larger social structure?
That is, do social, group-based identities occur primarily in the minds of in-
dividual people? Or are they embedded in the culture, language, norms, and
social practices of a group of people? Or do social identities emerge from
some combination of individual psychology and socio-cultural discourses,
beliefs, and practices? 
Second, we sought contributors whose work addresses different types of
intergroup conflict. Although developing a taxonomy of the dimensions on
which intergroup conflicts may vary is beyond the scope of this book, it is
clear that social identities may be involved in a great variety of intergroup
conflicts. For example, whether it is “identity politics” in American elec-
tions, ethnic cleansing in Yugoslavia, or riots targeting Muslims in India,
group-based identities often play a critical role in maintaining, and some-
times exacerbating, intergroup conflicts. Although addressing all types of
intergroup conflicts is beyond the scope of any single book, we sought con-
tributions that would address a wide range of conflicts: conflicts played out
within normal legal structures; conflicts involving illegal acts of aggression;
conflicts over symbolic and ideological issues; and conflicts over concrete
resources and power.
Although each contributor did not address the role of identity in this full
range of conflicts, the contributions addressed many different types of con-
flicts (rather than focus on a single type of conflict—e.g., genocide). There-
4
Introduction


fore, the first pair of contributions (by Brewer and Eriksen) focuses on 
identifying broad principles or common themes underlying a range of in-
tergroup conflicts. The second pair (by Citrin, Wong, and Duff, and Sida-
nius and Petrocik) focuses on the extent to which ethnic identity in the
United States remains strong and predicts strength of attachment to U.S.
political institutions and culture. The third pair of contributions (by White
and Staub) addresses the role of social identity in lethal violence by and
against the state. The seventh chapter (by Kelman) addresses the role of 
social identity in reducing intergroup conflict. 
In setting the stage for the chapters to follow, we address the following
questions: What do we mean by social identity? What do we mean by con-
flict? What is the organizational scheme for this volume? What specific
topics do the chapters address?
What Do We Mean by Social Identity?
Self and identity are not simple concepts with widely agreed upon mean-
ings. As we pointed out in the first volume in this series (Ashmore & Jussim,
1997), the terms have been used in a bewildering diversity of ways. Some-
times, the same word or phrase is used by different writers to refer to differ-
ent concepts; sometimes different words or terms are used to refer to what
appears to be the same construct. T
o make matters worse, self and identity
are both evolving concepts, which have meant different things in different
historical epochs (Baumeister, 1997; Danziger, 1997). Self and identity are
also doubly contested concepts today. First, the nature and meaning of group-
based identities, such as national, ethnic, or gender-based identities, are 
often contested in the context of sociocultural discourses (for example, con-
sider the potentially strong reactions, both pro and con, that the statement
“My country right or wrong” sometimes evokes). Second, the meaning and
nature of self and identity as social scientific constructs are contested among
researchers from different traditions (Danziger, 1997; Holland, 1997). So-
cial constructivists, for example, often argue against the possibility of even
studying self and identity objectively (Holland, 1997), whereas many psy-
chologists and sociologists implicitly assume that self and identity can be
studied objectively (Jussim & Ashmore, 1997). 
This volume, however, does not attempt to address the multiplicity of
controversies surrounding self and identity (though see Ashmore & Jussim,
1997, which does just this). This volume focuses on social identity (and 
its role in intergroup conflict). Social identity is a narrower and more spe-
cific subset of the phenomena included under the general rubric of “self 
and identity.” As such, it may be at least somewhat easier to introduce the
term in a way that will meet with some agreement across disciplines and
perspectives.
Social Identity and Intergroup Conflict
5


Within psychology and sociology, one of the clearest and most common
definitions of social identity was provided by T
ajfel (1981, p. 255): “that part
of an individual’s self-concept which derives from his knowledge of his
membership in a social group (or groups) together with the value and emo-
tional significance attached to that membership.” Thus, according to T
ajfel,
social identity has two components: belief that one belongs to a group (e.g.,
“I am an American”) and the importance of that group membership to one’s
self (e.g., “and I am damn proud to be a citizen of the greatest country on
earth”). 
We have not insisted that the contributors use this definition; they were
free to define social identity in the manner they saw as most appropriate to
their work. Nonetheless, five of the seven contributors cite and use T
ajfel’s
definition. Limitations to this definition, however, were implicitly high-
lighted by the two authors (Eriksen, Kelman) who did not use it. T
ajfel’s 
definition is almost purely individualistic, focusing exclusively on how the
individual thinks and feels about group memberships. Both Eriksen and
Kelman suggest that ethnic and national identities are not purely individu-
alistic—that such identities emerge within specific sociocultural contexts.
Sociocultural discourses, national myths, and intergroup relations are all es-
sential to the creation of any particular ethnic or national identity, so that,
according to Eriksen and Kelman, social identity resides at least partly
within one’s national or cultural community, rather than exclusively within
the individual.
What Do We Mean by Intergroup Conflict?
Conflict, of course, can occur between almost any type of groups (e.g.,
sports teams, corporations, university departments, nations, political par-
ties, ethnic groups). The chapters in this book, however, primarily (but not
exclusively) emphasize conflict between ethnic, racial, religious, and na-
tional groups. This emphasis reflects some of the dominant concerns within
the social sciences. Those concerns, we suspect, have largely been moti-
vated by a longstanding desire to understand some of humanity’s deepest
and most tragic social ills—war, genocide, mass murder, and institutional-
ized oppression.
Within social psychology, conflict is often defined as some sort of incom-
patibility of goals, beliefs, attitudes, and/or behavior (e.g., Myers, 1999).
The nature of the conflicts addressed in this volume are many and varied.
Some address economic and political discrimination within a nation; others
address how such discrimination can flare up into murderous intergroup 
violence. Five chapters focus on specific, real-world conflicts. Eriksen ad-
dresses conflicts between Serbs and Bosnians; Fijians and ethnic Indians in
Fiji; Hindus and Muslims in India. Citrin et al. and Sidanius and Petrocik
address conflicts between white Americans and ethnic minorities in
6
Introduction


America. White discusses the multi-sided conflict in Northern Ireland, and
Kelman describes his own workshops focusing on reducing the conflict be-
tween Israelis and Palestinians. 
Three chapters (Brewer, Eriksen, Staub) seek to provide broad, general
insights into the intersection of social identity and intergroup conflict.
Brewer develops a theoretical perspective on the steps leading to progres-
sively greater intergroup hostility. Eriksen derives general insights into in-
tergroup hostility on the basis of his analysis of three real-world conflicts.
Staub focuses specifically on the origins of genocide and mass murder. 
The starting point for understanding intergroup conflict often varies
across the different social science disciplines. Within social psychology,
conflict of some type or another is often addressed, implicitly or explicitly,
in theories and approaches to intergroup relations (e.g., Allport, 1954;
Oakes, Haslam, & Turner, 1994; Sherif, 1958; T
ajfel, 1981). The goal of
such research is typically to provide broadly applicable general explanations
of phenomena such as bias, prejudice, and discrimination. Social psycho-
logical theories customarily provide individual-level explanations for conflict
and the method of choice is, in general, laboratory experimentation. The
starting point for many such perspectives, therefore, is the accumulated ex-
perimental evidence addressing issues such as stereotypes and prejudice.
This is in stark contrast to many of the other social sciences, such as an-
thropology, political science, and sociology. Work in these disciplines often
focuses on understanding the history, sources, and nature of particular in-
tergroup conflicts. When general principles are sought, however, the start-
ing point is usually real-world conflicts between nations, ethnic groups, or
religious groups. There is often more emphasis on organizational, institu-
tional, and sociocultural levels of explanation. The methods of choice tend
to be surveys, interviews, ethnographies, archives, and other sources for his-
torical analysis. Consequently, the literatures addressing intergroup conflict
in the different social sciences have developed largely independently, and
often may appear to be talking past one another. One purpose of this book,
therefore, was to begin to redress this tendency by drawing contributors
from several social sciences (anthropology, political science, psychology,
and sociology). 
Themes Underlying This Book
Several themes provide the framework for the specific topics addressed by
the volume: (1) identity issues are important to diverse intergroup conflicts;
(2) conflict influences social identity; (3) social identity can contribute to in-
tergroup conflict; (4) addressing issues of social identity may be useful or
even necessary for reducing some conflicts; and (5) to understand how iden-
tity figures in intergroup strife, it is necessary to consider multiple levels of
analysis.
Social Identity and Intergroup Conflict
7


Identity Issues Are Important to Diverse Intergroup Conflicts
Regardless of whether intergroup conflicts are played out peacefully in cul-
tural and political arenas, or become violent and murderous, social identity
is usually involved. On the peaceful side, consider the various conflicts
within the United States over issues such as affirmative action, immigration,
and official use of the English language. These conflicts often seem to in-
volve social identities—with policy preferences and attitudes often varying
among different ethnic groups in such a manner as to enhance the power,
status, or privileges of one’s own group.
Or consider the more brutal side: the various incidents of mass murder
and genocide that have occurred throughout human history. In many cases,
extreme nationalist movements link a deep commitment to a national or
ethnic identity to hatred of a minority group—a pattern that can be found
in the Turkish genocide of the Armenians, the German genocide of the
Jews, and in Serbian mass murder of Bosnians.
Conflict May Influence Social Identity
People are usually affected, sometimes deeply, by intergroup conflicts. Cul-
tural stereotypes and prejudices can lead members of stigmatized groups to
disidentify with and disengage from tasks in which they are most heavily
stigmatized (Steele, 1997). Some African-American youths, for example,
disengage from academic achievement because negative intellectual stereo-
types about African Americans are too threatening and because high
achievement may be seen as “acting white” (Fordham & Ogbu, 1986;
Steele, 1997).
Furthermore, groups that have once been victimized by conquest, en-
slavement, or genocide may, in turn, undergo identity changes that render
them more vulnerable to committing atrocities against other groups in the
future. At least sometimes, a “we better get them before they get us” atti-
tude becomes a critical component of a national or ethnic identity. When
such an attitude takes hold on a large scale, the society may become less
likely to consider making peace with enemies, more likely to seek violent
solutions to conflicts, and more likely to commit genocide (du Preez, 1994;
Goldhagen, 1996; Staub, 1989).
Social Identity May Contribute to Intergroup Conflict
Intergroup violence can be fostered by social identities as well as stem from
group-based self-interest (the notion of realistic group conflict [Ashmore &
Del Boca, 1976]). For example, the recent Serbian-Kosovar conflict was
started over a material issue: who, Serbs or Kosovars, would have political
authority over Kosovo? But one of the reasons this was important to the
Serbian government was the historical linkage of Kosovo with Serbian
8
Introduction


identity (it was the location of a 600-year-old stand by Serbs against invad-
ing Turks).
Addressing Issues of Social Identity May Be Useful for Reducing Some Conflicts
At the same time, social identity can assist in alleviating intergroup conflict.
For example, Sherif’s (1958) classic work on “superordinate goals” demon-
strated that members of two groups in conflict can work toward a shared
goal. Doing so can involve a new sense of “us,” which likely contributes 
to the lessened hostility that “superordinate goals” produce (cf. Gaertner,
Dovidio, Anastasio, Bachman, & Rust, 1993).
To Understand How Identity Figures in Intergroup Strife, It Is Necessary 
to Consider Multiple Levels of Analysis
The incorporation of the final organizing variable—level of analysis—
underscores that the links of identity to intergroup conflict cannot be re-
duced either to the characteristics of individuals or to the qualities of
groups. Fortunately, however, different social and behavioral sciences tend
to emphasize different levels of analysis. Thus, it was necessary that con-
tributors to this volume come from diverse academic disciplines that cover
the full continuum—from those that focus primarily on the individual (as
most psychologists do), to others that emphasize understanding the rela-
tionship of individuals to institutions and social structure within a society
(as do many sociologists and political scientists), to still others that investi-
gate multiple societies (as do many anthropologists). 
The Chapters in This Volume
Three sections, each including two chapters, address different aspects of re-
lations between identity and conflict: (1) general analyses of how social
identities contribute to conflicts between groups and how conflict can influ-
ence social identity; (2) analyses of the extent to which ethnic identities 
influence degree of patriotism and attachment to mainstream American po-
litical culture and institutions; and (3) perspectives on how social identity
contributes to lethal intergroup violence. The fourth topic is conflict reso-
lution, which is addressed in a chapter that describes a specific conflict re-
duction intervention and the role of social identity in this intervention.
How Social Identities Contribute to Intergroup Conflict
The first section analyzes the role of identity, as both a group- and indi-
vidual-level variable, in contributing to intergroup conflict. Ethnic and na-
tional rivalries have a long history of leading to intergroup conflicts with
Social Identity and Intergroup Conflict
9


horrendous consequences. In the extreme, ethnic identity has led to mass
murder (e.g., in the former Yugoslavia). National identities, especially in the
form of aggressive nationalism, have contributed to genocides and global
war (e.g., the Nazi Holocaust and World War II). Ethnic identities have
also contributed to slavery, discrimination, and exploitation. 
In chapter 2, Brewer reviews the psychological research on social 
identity and intergroup strife. Social psychologists have long studied how
individuals’ social identities and group memberships influence their reac-
tions to, and perceptions of, others. Much work has addressed the role of
individual-level stereotypes, prejudice, and discrimination in intergroup
conflict. In addition, two related theories—social identity theory and self-
categorization theory (T
ajfel, 1981; Turner, 1987) have explicitly identified
ways in which individuals’ group memberships and social identities may
lead to intergroup strife. Brewer integrates existing theoretical perspectives
on social identity with social cognition research on stereotypes and preju-
dice and with her own work on optimal distinctiveness into a new perspec-
tive. Her analysis suggests that intergroup harm-doing is the end result of 
a series of progressively more hostile steps that include the psychological
categorization of people into groups, ingroup trust and preference, per-
ceived moral superiority of the ingroup over outgroups, and antagonism to-
ward outgroups. She argues that although progression from each step to the
next is not inevitable, the occurrence of each step increases the likelihood
that the next will occur.
In chapter 3, Eriksen addresses this topic by focusing on three case stud-
ies of the role of identity in intergroup conflict: the Serbian-Bosnian con-
flict; the Hindutva movement in India; and relations between ethnic Indians
and Fijians on Fiji. In all three cases, Eriksen describes how conceptions of
“us” and “them” (social identities and intergroup relations) are influenced
by societal/cultural phenomena such as the existing power structure, na-
tionalistic leaders, national myths, and modernization. He then discusses
how these socially and historically constructed views of “us” and “them” in
turn contribute to intergroup conflict. 
ToWhat Extent Do Ethnic Identities Influence Attachment 
to the Mainstream Political Culture of the United States?
The second section of the book focuses specifically on relations between
ethnicity and attachment to mainstream American political culture. T
o what
extent have different ethnic groups adopted mainstream American values
and become strongly attached to American political institutions? T
o what
extent does a strong ethnic identity undermine attachment to American 
political culture? T
wo chapters address these issues.
In chapter 4, Citrin, Wong, and Duff assess ethnic similarities and dif-
ferences regarding key, symbolic aspects of American identity. Levels of
chauvinism, patriotism, and pride in America are strikingly similar among
10
Introduction


blacks, whites, Asians, and Hispanics. Personal commitment to an Ameri-
can identity serves to bind Americans to the United States and to bring di-
verse groups together in providing a common identity, and, to some extent,
a common set of interests. 
However, alongside this common sense of attachment to America, many
people, especially members of minority groups, express a strong sense of 
attachment to their ethnic groups. Among whites, furthermore, a strong at-
tachment to a white racial identity seems to foster intergroup conflict.
Specifically, such whites tend to be most resistant to policies aimed at im-
proving the social and political conditions facing Blacks. In contrast, strong
ethnic attachment among minority groups did not seem to foster inter-
group hostility.
In chapter 5, Sidanius and Petrocik compare three general perspectives—
melting pot, pluralist, and social dominance—regarding relations between
ethnic identity and national identity among African-American, white, La-
tino, Asian, and Jewish United States citizens, and among resident aliens liv-
ing in the United States They find that: (1) ethnic identities remain strong
for many people; (2) ethnocentrism exists among all groups they studied; 
(3) differences in patriotism between whites and the non-black ethnic groups
is small or nonexistent; but (4) differences in patriotism between African
Americans and whites are larger. Sidanius and Petrocik interpret these results
as most consistent with the social dominance model, which predicts that the
strongest support for a society’s culture and institutions will be found among
those groups that benefit most from that culture and those institutions.
Social Identity and Lethal Intergroup Conflict
The third section of the book is devoted to the role of national and ethnic
identity in the instigation of deadly violence. Intergroup hostility, even
when intense, does not always erupt into violence. And even when violent,
conflicts do not always erupt into mass killing, ethnic cleansing, or geno-
cide. Nonetheless, systematic and widespread state-sponsored violence is
clearly one of the ultimate manifestations of conflict between groups. From
biblical stories of conquest, to the Romans’ treatment of foes and rebellious
peoples (e.g., Carthage, Israel), the Crusades, the Turkish slaughter of Ar-
menians, the German Holocaust, the Cambodian genocide, and widespread
killings in Rwanda and Bosnia, mass murder and genocide have emerged as
gruesome recurrent themes of intergroup relations. Thus, two chapters ad-
dress the role of social identity in lethal intergroup conflicts.
In chapter 6, White presents an analysis of the ongoing hostilities in
Northern Ireland. In tracing the historical roots of the existing conflict, he
identifies three primary perpetrators of violence: the Irish Republican para-
militaries, Protestant paramilitaries, and the state’s security forces. He ana-
lyzes the role of religious, political, and national identities in perpetuating
the conflict and the violence. He concludes that: (1) social identities play a
Social Identity and Intergroup Conflict
11


crucial role in maintaining the conflict; and (2) very different types of iden-
tities underlie each group’s involvement. IRA members often view them-
selves as working for freedom and to overcome the oppression of the Irish
by the British; Protestant paramilitary members view themselves as fighting
for the defense of their religious/ethnic group; and British political leaders
view themselves, at least according to their public statements, as unbiased
peacekeepers (a view that White argues is not completely consistent with
their actions there). 
In chapter 7, Staub presents a social psychological analysis of factors
leading to, and preventing, genocide. He describes the social and historical
conditions that often predispose a society toward genocidal violence. In so
doing, he highlights the various ways in which self and social identity are in-
volved in the processes of mass killing. For example, he concludes that “in-
jured” social identities—the belief that one’s group has been unfairly vic-
timized in the past and, therefore, is at greater risk of being victimized in
the future—is one major precursor of genocide (“we have to get them be-
fore they get us”). Staub also discusses the types of interventions that seem
to avert genocide. 
Social Identity Approaches to Reducing Intergroup Conflict
In chapter 8, Kelman describes his extensive experience running conflict-
resolution workshops for Palestinians and Israelis. Through these work-
shops, it has become clear that each side tends to have components of a na-
tional identity that seem to reduce the conflict to a zero-sum game. Indeed,
even acknowledging the identity and existence of the other group has often
been viewed as a threat to one’s own identity and existence as a nation.
However, Kelman also points out that, although the core identity of both
peoples may be non-negotiable, there are many peripheral aspects of those
identities that have emerged from these workshops as negotiable. Kelman’s
chapter, therefore, describes the role of identity in both contributing to the
Israeli-Palestinian conflict and in seeking solutions to that conflict.
Conclusion
Research on identity and on conflict have long histories within the social
and behavioral sciences. Until relatively recently, however, research on rela-
tions between identity and conflict has been less common. In the last 20
years, researchers in anthropology, political science, psychology, and soci-
ology have begun to study how group identifications contribute to conflict,
how intergroup conflict can influence identity, and, most recently, how
identities might be harnessed to reduce intergroup hostility. This volume
does not summarize all such existing work. The following chapters do pre-
sent a sophisticated and detailed analysis of some of the most fundamental
12
Introduction


issues involved in understanding identity and conflict. Chapters address
conflict between ethnic, racial, religious, and national groups; they address
ethnic differences in national attachment; political conflicts over territory;
and genocidal conflicts. As the chapters take stock of existing work on iden-
tity and conflict within several social scientific disciplines, we suspect that
both their diversity and some similarities in underlying themes will stand
out.
References
Allport, G. (1954). The nature of prejudice. Cambridge, MA: Addison-Wesley.
Ashmore, R. D., & Del Boca, F. K. (1976). Psychological approaches to under-
standing intergroup conflict. In P. Katz (Ed.), Towards the elimination of racism
(pp. 73–123). Elmsford, NY: Pergamon Press.
Ashmore, R. D., & Jussim, L. (Eds.) (1997). Self and identity: Fundamental issues
(Rutgers Series on Self and Social Identity, Volume 1). New York: Oxford Uni-
versity Press.
Baumeister, R. F. (1997). The self and society: Changes, problems, and opportuni-
ties. In R. D. Ashmore, & L. Jussim, (Eds.), Self and identity: Fundamental issues
(Rutgers Series on Self and Social Identity, Volume 1; pp. 191–217). New York:
Oxford University Press.
Danziger, K. (1997). The historical formation of selves. In R. D. Ashmore, & L. Jus-
sim, (Eds.), Self and identity: Fundamental issues (Rutgers Series on Self and Social
Identity, Volume 1; pp. 137–159). New York: Oxford University Press.
du Preez, P. (1994). Genocide: The psychology of mass murder. Southampton, UK: 
Bowerdean Publishing Co.
Fordham, S., & Ogbu, J. U. (1986). Black students’ school success: Coping with the
“burden of acting White.” Urban Review, 18, 176–206.
Gaertner, S. L., Dovidio, J. F., Anastasio, P. A., Bachman, B. A., & Rust, M. C.
(1993). The common ingroup identity model: Recategorization and the reduc-
tion of intergroup bias. In W. Stroebe & M. Hewstone (Eds.), European review of
social psychology, 4, pp. 1–26, London: Wiley.
Goldhagen, D. (1996). Hitler’s willing executioners: Ordinary Germans and the holo-
caust. New York: Alfred A. Knopf.
Greenfeld, L., & Chirot, D. (1994). Nationalism and aggression. Theory and Society,
23, 79–130.
Gurr, T. R. (1993). Why minorities rebel: A global analysis of communal mobiliza-
tion and conflict since 1945. International Political Science Review, 14, 161–201.
Holland, D. (1997). Selves as cultured: As told by an anthropologist who lacks a
soul. In R. D. Ashmore, & L. Jussim, (Eds.), Self and identity: Fundamental Issues
(Rutgers Series on Self and Social Identity, Volume 1, pp. 160–190). New York:
Oxford University Press.
Jones, J. (1997). Prejudice and Racism (second edition). New York: McGraw-Hill.
Jussim, L., & Ashmore, R. D. (1997). Fundamental issues in the study of self and
identity—contrasts, contexts, and conflicts. In R. D. Ashmore, & L. Jussim,
(Eds.), Self and identity: Fundamental issues (Rutgers Series on Self and Social
Identity, Volume 1, pp. 218–230). New York: Oxford University Press.
Social Identity and Intergroup Conflict
13


Keegan, J. (1993). A history of warfare. New York: Alfred A. Knopf.
Kelman, H. C. (1997). Some determinants of the Oslo breakthrough. Negotiaion
Journal, 13, 327–340.
Kreisburg, L., Northrup, T. A. & Thorson, S. J. (1989). Intractable conflicts and
their transformation. Dynamics of Identity, 4, 55–82. 
Myers, D. G. (1999). Social psychology (fifth edition). New York: McGraw-Hill.
Oakes, P. J., Haslam, S. A., & Turner, J. C. (1994). Stereotyping and social reality.
Cambridge, MA: Blackwell.
Reich, W. A. (1997). An expanded role for identity in the analysis of intergroup con-
flict. Unpublished manuscript.
Sherif, M. (1958). Superordinate goals in the reduction of intergroup conflict.
American Journal of Sociology, 63, 349–363.
Staub, E. (1989). The roots of evil: The origins of genocide and other group violence. Cam-
bridge: Cambridge University Press.
Steele, C. (1997). A threat in the air: How stereotypes shape intellectual identity and
performance. American Psychologist, 52, 613–629. 
T
ajfel, H. (1981). Social identity and intergroup relations. London: Cambridge Univer-
sity Press.
Thoits, P. A., & Virshup, L. K. (1997). Me’s and we’s: Forms and functions of social
identities. In R. D. Ashmore, & L. Jussim, (Eds.), Self and identity: Fundamental
issues (Rutgers Series on Self and Social Identity, Volume 1; pp 106–133). New
York: Oxford University Press.
Turner, J. C. (1987). Rediscovering the social group: A self-categorization theory. New
York: Basil Blackwell.
14
Introduction


PART I
THE CONTRIBUTION OF 
INDIVIDUALS’ IDENTITIES
AND THE COLLECTIVE 
IDENTITIES OF SOCIAL
GROUPS TO INTERGROUP
CONFLICT


This page intentionally left blank 


Marilynn B. Brewer
2
Ingroup Identification and 
Intergroup Conflict
When Does Ingroup Love Become Outgroup Hate?
Social psychologists’ interest in social identity and ingroup bias arose in the
context of research on intergroup relations (T
ajfel &Turner, 1986; Brewer &
Brown, 1998). As a consequence, there has been a coupling of the study of in-
group favoritism and outgroup prejudice, the implicit assumption being that
the more we know about why individuals become attached to their ingroups,
the more we will understand why they come to dislike or derogate outgroups.
The present chapter seeks to challenge that assumption by examining more
carefully the nature of the relationship between ingroup love and outgroup
hate. The basic thesis is that the formation of ingroups and ingroup identifi-
cation arises independently of attitudes toward outgroups. Attachment to in-
groups is presumed here to be the primary process, fundamental to indi-
vidual survival and well-being. Ingroup identification thus precedes
outgroup hostility and intergroup conflict but is not by itself a sufficient ex-
planation for such conflict. Instead, as this chapter will argue, we need to
trace the steps that lead from ingroup formation to outgroup hostility, speci-
fying the conditions under which the two become entangled.
Sumner’s Theory of Ethnocentrism
An early theory of the interrelationship between ingroup attachment and
identification and outgroup antagonism was implicit in the definition of
“ethnocentrism” when the term was introduced by Sumner (1906). Ethno-
centrism was described by Sumner as a universal characteristic of human
social groups whereby
17


a differentiation arises between ourselves, the we-group, or in-group, and
everybody else, or the others-group, out-groups. The insiders in a we-group
are in a relation of peace, order, law, government, and industry, to each other.
Their relation to all outsiders, or others-groups, is one of war and plunder.
. . . Ethnocentrism is the technical name for this view of things in which
one’s own group is the center of everything, and all others are scaled and rated
with reference to it. . . . Each group nourishes its own pride and vanity,
boasts itself superior, exalts its own divinities, and looks with contempt on
outsiders. . . . (pp. 12–13)
The features characteristic of Sumner’s ethnocentrism syndrome include
negative attitudes toward outgroups as well as positive feelings and evalua-
tions of the ingroup. Further, an explicit negative correlation between in-
group and outgroup attitudes is postulated: the greater the attachment and
solidarity within the ingroup, the greater the hostility and contempt di-
rected toward outgroups. This hypothesized relationship between ingroup
love and outgroup hate derived from Sumner’s functional theory of the ori-
gins of social groups and intergroup conflict:
The relation of comradeship and peace in the we-group and that of hostility
and war towards others-groups are correlative to each other. The exigencies
of war with outsiders are what make peace inside, lest internal discord should
weaken the we-group for war. . . . Thus war and peace have reacted on each
other and developed each other, one within the group, the other in the inter-
group situation. . . . Sentiments are produced to correspond. Loyalty to the
ingroup, sacrifice for it, hatred and contempt for outsiders, brotherhood
within, warlikeness without—all group together, common products of the same
situation. (1906, pp. 12–13, emphasis added)
The last 20 years of social psychological research on ingroup bias and in-
tergroup relations have cast doubt on the validity of the functional relation-
ship between ingroup and outgroup attitudes and behavior presumed by
Sumner. Results of the laboratory experiments using the “minimal group
paradigm” designed by T
ajfel and his colleagues (T
ajfel, Billig, Bundy, & Fla-
ment, 1971) explicitly challenged the assumption that differential treatment
of ingroup and outgroup members derives from cooperative interdepen-
dence within groups and competitive relations between groups. In the mini-
mal group experiments, mere categorization of persons into differentiated
social groupings has proved to be sufficient to give rise to preferential dis-
crimination in favor of ingroup over outgroup members, in the absence of
any differential interdependence (T
ajfel et al., 1971; Turner, 1978; Brewer,
1979; Diehl, 1990). Field studies assessing the relationship between identifi-
cation and attachment to ingroups and discrimination against outgroups
have also failed to find any straightforward negative correlation between in-
group positivity and intergroup attitudes (Hinkle & Brown, 1990). For ex-
ample, in a study of the reciprocal attitudes among 30 ethnic groups in East
Africa, Brewer and Campbell (1976) found that almost all of the groups ex-
hibited systematic differential positive evaluation of the ingroup over all out-
18
The Contribution of Identities to Intergroup Conflict


groups on dimensions such as trustworthiness, obedience, friendliness, and
honesty, but the correlation between this measure of ingroup positivity and
distancing from outgroups was essentially .00 across the 30 groups (p. 85).
Thus, the question of when and how ingroup identification is related to out-
group hostility and intergroup conflict requires further investigation.
Decomposing Ethnocentrism: Ingroup, Intergroup, and 
Outgroup Orientations
Sumner’s definition of ethnocentrism cited at the beginning of this chapter
contains four separate propositions. The four distinguishable elements can
be characterized as follows:
1. Human social groups are organized into discrete ingroup-out-
group categories (the social categorization principle).
2. Individuals value their ingroups positively and maintain positive, 
cooperative relationships with members of the ingroup (the in-
group positivity principle).
3. Ingroup positivity is enhanced by social comparison with out-
groups in which ingroup attributes and outcomes are evaluated as
better than or superior to those of outgroups (the intergroup compar-
ison principle).
4. Relationships between ingroup and outgroups are characterized by 
antagonism, conflict, and mutual contempt (the outgroup hostility
principle).
In contrast to Sumner’s (1906) expectation that these four elements nec-
essarily cohere into a pattern that is universally characteristic of intergroup
relations, a review of relevant social psychological research suggests that the
components can be distinguished both empirically and conceptually. How-
ever, the components are not entirely independent features of ethnocentric
attitudes, either. Instead, I propose that the elements represent a systematic
progression along a continuum of possible relationships between ingroup
formation and intergroup behavior in which each element in the progres-
sion provides a necessary but not sufficient condition for occurrence of the
subsequent elements. Within this system, the first two elements are proba-
bly universal characteristics of human social groups (as Sumner postulated),
but the third and fourth elements require additional social-structural and
motivational conditions that are not inherent in the processes of group for-
mation itself.
Social Categorization:The Basic Substrate
Virtually all current theories of social identity and intergroup relations rec-
ognize social categorization as the basic cognitive process underlying all in-
Ingroup Identification and Intergroup Conflict
19


tergroup phenomena (e.g, Allport, 1954; T
ajfel, 1969). Categorization (a)
partitions the multidimensional variability among human beings into dis-
crete subsets, accompanied by (b) accentuation of perceived intracategory
similarities and intercategory differences (Doise, Deschamps, & Meyer,
1978; T
ajfel, 1969; T
ajfel & Wilkes, 1963). The basic processes of catego-
rization and category accentuation are presumed to be the same whether we
are talking about individuals’ partitioning of the world of physical objects
and events or of the social world. Individuals learn to classify objects as
functionally interchangeable and develop concepts that distinguish mem-
bers of one category from those of another as a fundamental tool for nego-
tiating the physical and social environment. 
The assimilation and contrast processes underlying category formation
and accentuation are independent of any self-involvement in the category
system. Just as one can acquire a concept of “apple,” which may include a
representation of the prototypic member of the category, a theory of the
“essence” of appleness, and an understanding of the variability among ex-
emplars of the category, so one can acquire similar conceptualizations of so-
cial objects such as “librarian” or “African American.” Stereotypic represen-
tations (prototypes) of social categories are a natural consequence of
categorization and concept formation processes, motivated by the need for
cognitive simplification and social structure. As with any other category
concept, whether social category stereotypes are positive or negative (or
mixed) is a function of what has been learned about the content of the cate-
gory representation.
What makes social categorization different from nonsocial categoriza-
tion systems, however, is the potential for self-involvement in that social
perceivers are themselves members of some social categories (ingroups) and
not others. Thus, an additional categorical distinction between ingroup and
outgroups may be superimposed onto particular social classifications. In the
absence of the engagement of this ingroup-outgroup distinction, social 
categories are merely cognitive concepts that help individuals to make sense
of their social world and to direct their behavior toward other individuals.
When social categories are also recognized as social ingroups, however, ad-
ditional processes associated with self-categorization (Turner, Hogg, Oakes,
Reicher, & Wetherell, 1987) are engaged that lay the groundwork for fur-
ther social differentiation. Categorization of the self as an ingroup member
entails assimilation of the self to the ingroup category prototype and en-
hanced similarity to other ingroup members. Self-categorization provides
the cognitive substrate for attachment to ingroups and differentiation from
outgroups—the first step toward ingroup bias and discrimination. 
Ingroup Positivity:The Causes and Consequences of Social Identification
Differentiation between ingroups and outgroups does not necessarily carry
evaluative implications, and acknowledging one’s ingroup membership is
20
The Contribution of Identities to Intergroup Conflict


not equivalent to identification with that social group. This is clear from
T
ajfel’s (1981) definition of social identity as “that part of an individual’s
self-concept which derives from his knowledge of his membership of a so-
cial group . . . together with the value and emotional significance attached
to that membership” (p. 255; emphasis added). Social identification, then,
entails affective and evaluative processes that are above and beyond mere
cognitive classification of the self and others into a shared social category.
The affective significance of social identification arises from the felt attach-
ment between the self and the ingroup as a whole. 
Social identification represents the extent to which the ingroup has been
incorporated into the sense of self, and at the same time, that the self is ex-
perienced as an integral part of the ingroup. With high levels of social iden-
tification, the group’s outcomes and welfare become closely connected to
one’s own sense of well-being (Brewer, 1991). According to social identity
theory, it is this engagement of the self that accounts for the positive valua-
tion of the ingroup and positive orientations toward fellow ingroup mem-
bers. Self-worth is both projected onto and derived from positive ingroup
evaluation. The distinction between ingroup membership and ingroup
identification implies that social identification is not an automatic by-
product of categorization. Identification implies some motivation to define
oneself in terms of that group membership and to achieve and maintain 
inclusion in the ingroup category. Since identification entails some sacrifice
of an autonomous self-concept, the question arises as to why individuals
would attach some measure of their self-worth and well-being to the fate of
a collective. 
One explanation that has been offered is that social identity and ingroup
positivity are engaged in the interests of enhanced self-esteem (Turner,
1975; T
ajfel, 1978). But group outcomes are not always positive or self-
aggrandizing, and identification often entails assimilation of group failures
as well as successes (Brewer & Weber, 1994; Crocker & Major, 1989; Hirt,
Zillmann, Erickson, & Kennedy, 1992; Turner, Hogg, Turner, & Smith,
1984). Overall, results of research on the causal role of self-esteem in social
identification are far from convincing (Abrams & Hogg, 1988; Rubin &
Hewstone, 1998). In fact, there is good reason to believe that attachment of
the self to the group (or other objects) leads to positive valuation of the
group, rather than vice versa (Greenwald & Banaji, 1995). 
If strength of ingroup identification is largely independent of ingroup
status, then motives other than self-enhancement must be implicated in se-
lecting and maintaining social identities. An alternative motivational expla-
nation is provided by optimal distinctiveness theory (Brewer, 1991). The
theory postulates that human beings have two powerful social motives: a
need for inclusion that motivates assimilation of the self into large, imper-
sonal social collectives, and an opposing need for differentiation that is satis-
fied by distinguishing the self from others. As opposing motives, the two
needs hold each other in check. When a person feels isolated or detached
Ingroup Identification and Intergroup Conflict
21


from any larger social collective, the drive for inclusion is aroused; on the
other hand, immersion in an excessively large or undefined social collective
activates the search for differentiation and distinctiveness. 
Identification with distinctive social groups emerges from the processes
involved in satisfying both needs simultaneously. Groups that are exclusive
rather than inclusive, groups that have clear categorical boundaries—satisfy
the need for inclusion (intragroup assimilation) at the same time that they
provide a basis for satisfying the need for differentiation (intergroup con-
trast). Clear ingroup boundaries serve to secure both inclusion and exclu-
sion. According to optimal distinctiveness theory, ingroups that meet these
simultaneous conditions become an integral part of the individual’s sense of
self and the basis for a secure and stable self-concept. Group identification
in this sense involves “transformations in the definition of self and the basis
for self-evaluation. When the definition of self changes, the meaning of
self-interest and self-serving motivation also changes accordingly” (Brewer,
1991, p. 476). 
Optimal distinctiveness theory proposes that social identification is the
product of the search for inclusion and differentiation, rather than a conse-
quence of the search for self-esteem. Nonetheless, although the desire for
self-enhancement may not be a primary cause of ingroup identification,
identification may well lead to a motivation to view the ingroup in the most
favorable possible terms. Drawing on Heider’s (1958) principles of cogni-
tive balance, the connection between self and ingroup constitutes a unit re-
lationship. Assuming that affect toward the self is primarily one of postive
regard, positive attitudes toward the group creates balance between unit
and sentiment relations. There is empirical evidence that positivity biases
associated with ingroup identification arise automatically and without
awareness. Perdue, Dovidio, Gurtman, and Tyler (1990), for instance,
demonstrated that ingroup signifiers (terms such as “we,” “us,” and “our”)
automatically activate positive evaluative responses, as indicated by differ-
ential reaction time to recognize subsequent words as “good” (vs. “bad”).
Consistency motives are also implicated in relationships between atti-
tudes and behavior; positive evaluations and expectations give rise to trust
and cooperative behavior that in turn justify positive feelings and future ex-
pectancies. Thus, the process of assimilation between self and others that
underlies ingroup formation instigates a “benign cycle” of positive affect
and behavior.
The emphasis at this stage on positive affect and evaluations directed to-
ward the ingroup is not devoid of implications for intergroup relations.
Positive attitudes and expectations about ingroup members, perceived simi-
larity, and cooperative orientation are all limited to those who share mem-
bership in the common ingroup. The extension of positive self-regard to
others stops at the boundary between ingroup and outgroups; attitudes to-
ward those outside the boundary are, at best, characterized by indifference.
Differential indifference is not without its consequences. When positive re-
22
The Contribution of Identities to Intergroup Conflict


gard, empathy, and cooperation are extended to some individuals but not
others on the basis of category membership, an initial form of intergroup
discrimination is evident (Brewer, 1979, 1996). Compared to ingroupers,
outgroupers are less likely to be helped in ambiguous circumstances (Frey
& Gaertner, 1986; Gaertner & Dovidio, 1977; Gaertner, Dovidio, & John-
son, 1982), more likely to be seen as provoking aggression (Baron, 1979;
Rogers & Prentice-Dunn, 1981), less likely to receive the benefit of the
doubt in attributions for negative behaviors (Weber, 1994), and likely to be
seen as less deserving of public welfare (Pettigrew & Meertens, 1995).
Studies of ethnic and racial prejudice in the United States and Europe
confirm that the essence of “subtle racism” is not the presence of strong
negative attitudes toward minority outgroups but the absence of positive sen-
timents toward those groups (e.g., Dovidio & Gaertner, 1993; Pettigrew &
Meertens, 1995; Stangor, Sullivan, & Ford, 1991, Study 1). As Gaertner,
Dovidio, Anastasio, Bachmann, & Rust. (1993) put it, “Initially it was 
assumed that in situations in which blacks were treated differently than
whites, whites discriminated against blacks, and thus reflected subtle, indi-
rectly expressed negative racial attitudes. Alternatively, it is possible that, 
at some fundamental level, aversive racism reflects a pro-white (i.e., pro-
ingroup) rather than the solely anti-black bias that was originally pro-
posed” (p. 4).
If social groups differ in access to resources and power, the benefits of in-
group favoritism accrue to members of some groups more than others and
contribute to substantial differences in outcomes between the groups as a
whole. Ultimately, many forms of discrimination may develop not because
outgroups are hated, but because positive affects such as admiration, sympa-
thy, and trust are reserved for the ingroup and withheld from outgroups
(Pettigrew & Meertens, 1995). Differential positivity toward ingroups over
outgroups also sets the stage for more pernicious forms of intergroup 
discrimination.
Intergroup Comparison: Social Competition for Positive Distinctiveness
Ingroup positivity implies a sort of self-absorption whereby group members
attend to the concerns and welfare of their own group with disregard for
the status or outcomes of those outside the ingroup boundary. The orienta-
tion is simply “us” versus “not us,” with little need to differentiate among
outgroups or to take outgroup outcomes into account while pursuing in-
group goals. Satisfaction with the ingroup’s well-being is evaluated against
internal standards and expectations, independent of the status of those out-
side the group.
The idea that assessments of goodness and satisfaction can be made in an
absolute sense against clear internal standards may hold true for some di-
mensions of evaluation or some conditions of group living. In studies of the
individual self-concept, Neidenthal and Beike (1997) demonstrated that
Ingroup Identification and Intergroup Conflict
23


some self-evaluations are affected by comparative context, but others re-
main stable and independent of comparative information. A parallel distinc-
tion may hold for ingroup self-evaluations. Groups that are physically or
psychologically isolated from outsiders, for instance, may measure success
in terms of basic sustenance or survival in the face of environmental hard-
ships, or in terms of adherence to dogma or beliefs and practices that are
not shared by those outside the group. 
In most cases, however, standards of evaluation are not so clear-cut, and
knowledge of the status of outgroups becomes relevant to assessing the state
of welfare of the ingroup. The conditions of social comparison at the inter-
group level are parallel to those for interpersonal comparison in the pursuit
of self-evaluation as outlined by Festinger (1954). The need for social com-
parison is aroused when there is uncertainty about one’s standing on some
dimension of self-evaluation, uncertainty that can be resolved by comparing
one’s own position to that of relevant others. The relevance requirement is
particularly interesting because it implies that the self and other (or ingroup
and outgroup) must be perceived as similar in some sense (i.e, as subject to
the same values or sharing a common frame of reference for evaluation) in
order to be useful targets for comparative appraisal. Thus, when ingroup
evaluation is dependent upon intergroup comparison, it is necessary to
identify specific outgroups for comparison and to distinguish between out-
groups that are relevant to ingroup self-regard and those that are not. 
Comparison (whether interpersonal or intergroup) is not inherently
competitive. This depends on what motives social comparison is serving
(Wood, 1989). Comparison can be undertaken for purposes of objective
self-appraisal or to motivate self-correction and improvement. When as-
sessments are relative and evaluative, however, the better the other is judged
to be, the worse the self-evaluation. In this case, ingroup favoritism takes a
subtle shift from motivated perceptions that “we are good” to perceptions
that “we are better.” When all of the groups in a social milieu are jockeying
for this type of ingroup advantage, ingroup-outgroup relations are charac-
terized by what Turner (1975) referred to as “social competition” for posi-
tive identity. Combined with motivations for self-enhancement, then, social
comparison becomes social competition and the pursuit of positive self-
regard can be achieved only at the expense of the other. 
In this case, competition is not necessarily inherent in the availability of
positive outcomes but is created by the evaluation that is attached to out-
comes for the ingroup and outgroup. If resources (economic and social 
benefits) are valued in an absolute sense, increases in ingroup welfare can be
achieved even if the outgroup is also benefitting. However, when valuation
(e.g., subjective states such as status) is relative, the welfare of the ingroup
and the outgroup is psychologically a zero-sum distribution. We cannot 
improve our position or sense of well-being unless the outgroup is doing 
less well than we are. When resources are also attached to intergroup differ-
ences in status and power, this zero-sum mentality is exacerbated. Ingroup
24
The Contribution of Identities to Intergroup Conflict


bias motivated by a desire to establish or maintain positive distinctiveness 
of the ingroup relative to the outgroup is a second form of intergroup dis-
crimination that can be differentiated, at least conceptually, from simple 
ingroup favoritism (Brewer, 1996). 
Interestingly, the more similar groups are in their values and aspirations,
the more acute the intergroup social competition. When the ingroup and
outgroup are arrayed along a single dimension of evaluation that all groups
regard as important and relevant to collective self-esteem, comparative ap-
praisals are inevitably competitive. In a context of self-enhancement mo-
tives, outgroup achievements are threatening to ingroup position and the
closer the outgroup, the greater the threat (T
esser, 1988). Social competi-
tion and intergroup threat are diluted when the outgroup is very remote or
distant from the ingroup on the dimension of evaluation. Ingroups that en-
joy very high status often display largess or magnanimity, rather than in-
group discrimination, toward outgroups that are much lower down on the
status ladder. Conversely, low-status groups often exhibit outgroup posi-
tivity in evaluations that are related to the status distinction, conceding out-
group superiority on dimensions where the status difference is clear and
large (Hinkle & Brown, 1990; Mullen, Brown, & Smith, 1992). 
Either of these forms of outgroup favoritism, however, is qualified by
strategies that serve ingroup positivity overall. Whereas low-status groups
exhibit outgroup positivity on status-relevant evaluations, they generally
display greater ingroup bias (compared to their high-status counterparts)
on dimensions unrelated to the status differential (Brewer & Campbell,
1976; Brewer, Manzi, & Shaw, 1993; Mullen, Brown, & Smith, 1992). Simi-
larly, outgroup largess on the part of high-status groups quickly reverts to
ingroup favoritism when the status gap between ingroup and outgroup is
perceived to be closing (Sachdev & Bourhis, 1991; Sidanius, 1993).
In a pluralistic context where multiple dimensions of evaluation are possi-
ble, social competition for positive distinctiveness may be more diffused and
less intense. In research on individual self-evaluation maintenance it has
been demonstrated that individuals feel threatened by a friend’s achieve-
ments when the domain of performance is highly important and relevant to
their own self-evaluation, but a friend’s accomplishments are celebrated
when the domain is not self-important (T
esser, 1988). Similar effects have
been observed in research on ingroup bias and intergroup discrimination.
Mummendey and Schreiber (1983) found that when individuals are given the
opportunity to choose dimensions of evaluation, they evaluate ingroup prod-
ucts more favorably than outgroups on dimensions chosen by the ingroup
but rate the outgroup positively on other dimensions. In general, in-group
favoritism is highest on dimensions of evaluation that are important to the
ingroup, but is mitigated for dimensions that are important to the outgroup
but relatively unimportant to the ingroup (Mummendey & Schreiber, 1984;
Mummendey & Simon, 1989). 
In interpreting these findings, Mummendey and Schreiber (1983) sug-
Ingroup Identification and Intergroup Conflict
25


gest that social competition occurs only when there is no alternative for es-
tablishing one’s own positive identity than at the expense of the outgroup.
When positive identity is possible for both groups at the same time, judg-
ments are more influenced by fairness than by ingroup bias. Thus, positive
distinctiveness does not necessitate across-the-board favoritism of ingroup
over outgroups. Ingroup distinctiveness at the expense of outgroups obtains
only when valuation is relative and on dimensions that are particularly im-
portant to the group’s collective self-worth.
Outgroup Antagonism and Intergroup Aggression
Intergroup comparison and social competition provide an initial connection
between valuing the ingroup and devaluing relevant outgroups. In the lit-
erature on social motives (McClintock, 1972), however, a clear distinction is
made between competition—the motivation to seek relative gain for the in-
group over others—and aggression—the motivation to harm the other as an
end in itself. Following this distinction, we can distinguish between inter-
group discrimination that is based solely on ingroup favoritism, and dis-
crimination or prejudice that entails an active component of outgroup dero-
gation and aggression (Levin & Sidanius, 1999; Struch & Schwartz, 1989).
Aggression directed against an outgroup may originate in the service of
protection or enhancement of the ingroup, but it may also become an end
in itself—outgroup hate that has become dissociated from ingroup en-
hancement or desire to benefit the ingroup.1
When outcomes are objectively (or perceived to be) zero-sum, it be-
comes difficult to pin down the locus of bias (Brewer, 1979). If ingroup out-
comes improve only as outgroup benefits decrease, it is not clear whether
the motivation for discriminatory behavior is to benefit the ingroup (i.e.,
discrimination for us) or to disadvantage the outgroup (discrimination
against them). Still, the experimental literature on intergroup discrimina-
tion provides evidence that the primary motivation is to benefit the ingroup
rather than harm the outgroup. In studies where evaluations of the ingroup
can be assessed independently of evaluations of the outgroup, conditions
that increase ingroup identification and favoritism generally involve en-
hancement of ingroup evaluations rather than decreased evaluation or dero-
gation of outgroups (e.g., Bettencourt & Dorr, 1998; Brewer, 1979; Brewer,
Manzi, & Shaw, 1993). 
More direct evidence is provided by research on the asymmetry of posi-
tive and negative ingroup bias (Mummendey & Otten, 1998). Although the
initial experiments by T
ajfel et al. (1971) claimed that ingroup bias in the
minimal intergroup situation occurred for allocations of both positive and
negative outcomes, subsequent research suggests that when the outcomes
to be distributed are negative or harmful, the usual intergroup discrimina-
tion may be lessened or may disappear all together. Hewstone, Fincham,
and Jaspars (1981) modified the normal paradigm by asking group members
26
The Contribution of Identities to Intergroup Conflict


to subtract money from ingroup and outgroup recipients. Although some
ingroup bias was observed in this context, the levels were lower than those
obtained with the standard allocation matrices. Mummendey et al. (1992)
extended this finding by asking participants to distribute different durations
of an unpleasantly high-pitched tone to ingroup and outgroup members.
With harm as an outcome, ingroup bias was generally eliminated and
strategies to equalize or minimize the amount of aversive stimulation were
observed.
Mummendey and Otten (1998) have suggested that the positive-negative
asymmetry in intergroup discrimination is explained by normative con-
staints that make it more difficult to justify differential allocations that harm
others directly than just benefit some more than others. As long as everyone
obtains some benefit, it is relatively easy to find justification for why the in-
group might be entitled to benefit more than the outgroup. There is even
evidence that such ingroup benefiting is considered normative in its own
right (Blanz, Mummendey, & Otten, 1997; Platow, O’Connell, Shave, &
Hanning, 1995). Thus, allocations that favor the ingroup over the outgroup
may be motivated primarily by intentions to be a “good” ingroup member,
rather than by any sentiments against the outgroup. 
Inflicting harm, however, is not sanctioned so directly. It is more difficult
to justify protecting the ingroup if it comes at the expense of inflicting harm
on others (Mummendey & Otten, 1998). Thus, even when the ingroup may
be disadvantaged, intergroup discrimination is constrained when there 
is no direct motivation to harm or derogate the outgroup. These are the
constraints that are lifted when outgroups are viewed with hatred or con-
tempt—emotions that justify outgroup harm above and beyond ingroup
benefit. Yet outside of the laboratory, the world is full of evidence of the 
existence of such self-justified harm toward outgroups, even at great costs
to the ingroup. A desire for positive distinctiveness can account for prefer-
ential hiring of ingroup members over outgroup members and for selective
investment in projects that benefit one’s own group over others. But wars 
of conquest, pograms, and ethnic cleansing require explanation that goes
beyond that of achieving positive distinctiveness for the ingroup.
Social identity theory provides an adequate framework for understand-
ing ingroup bias and intergroup discrimination of the positive types de-
scribed earlier. Assimilation of the self to the ingroup accounts for ingroup
positivity, and from there, the extension to relative enhancement of the in-
group over outgroups under conditions of intergroup comparison is not a
big motivational leap. The story becomes more complicated, however,
when one tries to use the same framework to account for more virulent out-
group hate and intergroup hostility. T
o justify aggression against outgroups
in the interest of the ingroup, the very existence of the outgroup, or its
goals and values, must be seen as a threat to the maintenance of the ingroup
and to one’s own social identity. Thus, understanding the relationship be-
tween ingroup identification and outgroup hostility requires understanding
Ingroup Identification and Intergroup Conflict
27


how the interests of the ingroup and those of the outgroup come to be 
perceived as in conflict. For that purpose, we need to reconsider the funda-
mental motivations underlying ingroup attachment and the path from in-
group identity to outgroup threat.
From Ingroup Formation to Intergroup Conflict: Another Look
Realistic competition over scarce vital resources provides one explanation for
perceived conflict of interests. When resources for survival or flourishing are
scarce and groups are locked into negative interdependence (competing for
the same resources), the link between ingroup survival and destruction of the
outgroup is direct. But analyses of warfare and intergroup violence call into
question the necessary role of realistic, objective conflicts of interest in many
cases of intergroup conflict (e.g., Sears & Funk, 1991). Perceived threat is of-
ten subjective and symbolic, and overt conflict often appears to be the prod-
uct of prior intergroup antagonism rather than its cause.
The thesis that is being developed here gives prior status to ingroup 
categorization and identification for understanding the origins of intergroup
behavior. This view assumes that ingroups are formed for reasons indepen-
dent of intergroup conflict and competition over resources. The social and
psychological functions served by ingroup formation and identification pro-
vide the necessary conditions for collective action—either intragroup or 
intergroup. Hence, ingroup attachment and loyalty lay the groundwork for
intergroup conflict, but outgroup antagonism is not a necessary extension of
ingroup positivity and enhancement. Explicating the relationship between
social identity and intergroup conflict requires a fuller understanding, first,
of the functions of ingroup identification and, then, of the additional condi-
tions that lead to a connection between ingroup attachment and outgroup
hostility.
Ingroup Formation: The Boundaries of Cooperation and Trust
If ingroup-outgroup distinctions do not always involve intense (or even
mild) competition or conflict over scarce resources, there is need for a
theory of the evolution of social groups that does not depend on intergroup
conflict per se. Such a theory starts from the recognition that group living
represents the fundamental survival strategy that characterizes the human
species. In the course of our evolutionary history, humans abandoned most
of the physical characteristics and instincts that make possible survival and
reproduction as isolated individuals or pairs of individuals in favor of other
advantages that require cooperative interdependence with others in order
to survive in a broad range of physical environments. In other words, as a
species we have evolved cooperation rather than strength and social learn-
ing rather than instinct as basic adaptations.
28
The Contribution of Identities to Intergroup Conflict


The result is that, as a species, human beings are characterized by obliga-
tory interdependence (Brewer, 1997; Caporael, 1997). For long-term survival,
we must be willing to rely on others for information, aid, and shared re-
sources, and we must be willing to give information and aid and to share 
resources with others. At the individual level, the potential benefits (receiv-
ing resources from others) and costs (giving resources to others) of mutual
cooperation go hand-in-hand and set natural limits on cooperative interde-
pendence. The decision to cooperate (to expend resources to another’s 
benefit) is a dilemma of trust since the ultimate benefits depend on every-
one else’s willingness to do the same (Brewer, 1981). A cooperative system
requires that trust dominate over distrust. But indiscriminate trust (or indis-
criminate altruism) is not an effective individual strategy; cooperation must
be contingent on the probability that others will cooperate as well.
Social categorization and clear group boundaries provide one mechanism
for achieving the benefits of cooperative interdependence without the risk of
excessive costs. Ingroup membership is a form of contingent altruism. By
limiting aid to mutually acknowledged ingroup members, total costs and
risks of nonreciprocation can be contained (see T
akagi, 1996, for a related ar-
gument). Thus, ingroups can be defined as bounded communities of mutual
trust and obligation that delimit mutual interdependence and cooperation. 
An important aspect of this mutual trust is that it is depersonalized
(Brewer, 1981), extended to any member of the ingroup whether personally
related or not. Psychologically, expectations of cooperation and security
promote positive attraction toward other ingroup members and motivate
adherance to ingroup norms of appearance and behavior that ensure one
will be recognized as a good or legitimate ingroup member. Symbols and
behaviors that differentiate the ingroup from local outgroups become par-
ticularly important here, to reduce the risk that ingroup benefits will be 
inadvertently extended to outgroup members, and to assure that oneself 
is recognized as a member of the ingroup and entitled to those benefits. As-
similation within and differentiation between groups is thus mutually rein-
forcing, along with ethnocentric preference for ingroup interactions and 
institutions.
The optimal distinctiveness model of social identity (Brewer, 1991) was
based on this evolutionary perspective on the functions of ingroup for-
mation and differentiation. The theory assumes that psychological me-
chanisms at the individual level (opposing needs for differentiation and 
inclusion) co-evolved with structural requirements at the group level for
successful group functioning and coordination (size and boundedness). The
relative strength of inclusion and differentiation needs at the individual
level interacts with group properties such as size and permeability to de-
termine social identification and ingroup attachment and loyalty. The psy-
chology of assimilation and differentiation limits the extent to which strong
social identification can be indefinitely extended to highly inclusive, super-
ordinate social groups or categories. 
Ingroup Identification and Intergroup Conflict
29


Thus, one implication of optimal distinctiveness theory is that ingroup
loyalty, and its concomitant depersonalized trust and cooperation, is most
effectively engaged by relatively small, distinctive groups or social cate-
gories. Groups that satisfy these needs are those that provide the necessary
boundaries on mutual obligation and trust. Optimal group identities be-
come essential to a secure self-concept. Once group identification has been
achieved, maintaining a secure sense of inclusion and boundedness becomes
tantamount to protecting one’s own existence. When optimal inclusion
within a distinctive ingroup is not achieved, the sense of self is threatened
and vulnerable.
Intergroup Differentiation and Outgroup Stereotypes
A consequence of ingroup identification and intergroup boundaries is that
individuals modify their social behavior depending on whether they are in-
teracting with ingroup or outgroup members. Ingroup behavior is governed
by norms and sanctions that reinforce expectations of mutual cooperation
and trustworthiness. Depersonalized trust is supported by implicit under-
standings that ingroup members will monitor the behavior and interactions
of other group members, sanctioning deviations from group expectations
about appropriate ingroup attitudes and behavior. Thus, shared ingroup
membership may be taken as prima facie evidence that other members of the
group will live by the codes of conduct that bind them as a group (Kramer,
Brewer, & Hanna, 1996).
Just as there is a realistic basis for ethnocentric trust of ingroups, differ-
ences in norms and sanctions applied to ingroup behavior compared to be-
havior in interactions with outgroup members provide a realistic basis for
outgroup distrust and negative stereotypes. At the same time that groups
promote trust and coopration within, they caution wariness and constraint
in intergroup interactions. Thus, even in the absence of overt conflict be-
tween groups, the differentiation between ingroup and outgroup behavior
creates a kind of self-fulfilling prophecy in the realm of intergroup percep-
tions. As LeVine and Campbell (1972) put it, “if most or all groups are, in
fact, ethnocentric, then it becomes an ‘accurate’ stereotype to accuse an
outgroup of some aspect of ethnocentrism” (p. 173). 
Combined with the accentuation principle that exaggerates perceived
differences between social categories, this leads to a set of “universal stereo-
types” to characterize ingroup-outgroup differences. Whereas “we” are
trustworthy, peaceful, moral, loyal, and reliable, “they” are clannish, exclu-
sive, and potentially treacherous. What is particularly interesting about this
pattern of stereotypes is that the same behaviors that are interpreted as rea-
sonable caution on the part of the ingroup in dealings with outgroup mem-
bers become interpreted as “clannishness” and indicators of potential
treachery when exhibited by outgroupers toward the ingroup. 
Evidence for this hypothesized universal difference in ingroup versus out-
30
The Contribution of Identities to Intergroup Conflict


group stereotypes was provided by data from our survey of intergroup per-
ceptions in East Africa (Brewer & Campbell, 1976). Respondents in the 
survey were presented with a lengthy list of character traits (both positive 
and negative) and asked to indicate which groups were most likely to exhibit
each trait. When responses to this trait list were factor analyzed, the first 
factor that emerged was a bipolar evaluative factor interpretable as trustwor-
thiness versus untrustworthiness. Positive loadings on this factor were asso-
ciated with traits such as “peaceful,” “obedient,” “honest,” “gentle,” and
“friendly.” Negative loadings were obtained for traits such as “quarrelsome,”
“disobedient,” “dishonest with others,” “cruel,” and “hot-tempered.” An 
index of positive regard was constructed by taking the frequency with which
a group was mentioned in connection with positive traits on this factor and
subtracting the frequency of mentions in connection with the negative traits.
Comparing ingroup and outgroup ratings on this index revealed that all 30
groups in the survey rated their own group more positively than the average
rating given to outgroups. Further, 27 of the groups rated the ingroup higher
than any outgroup in the sample (Brewer & Campbell, 1976, p. 79). Although
the assignment of negative traits such as “quarrelsome,” “dishonest,” and
“cruel” varied considerably across outgroups, attributions of positive traits
such as “peaceful,” “honest,” and “friendly” were almost universally reserved
for the ingroup.
The socialization experiences that lead to differential expectations and
comfort in ingroup dealings compared to dealings with outgroups may be
generalized in the form of a generic schema that is brought to all intergroup
contexts. Whenever ingroup-outgroup distinctions become salient, the in-
tergroup schema is activated. Just such a generic schema was suggested by
T
ajfel in his initial explanations for the minimal intergroup discrimination
findings (T
ajfel et al., 1971). At that time, T
ajfel (1970) speculated about a
“generic norm” of intergroup behavior that prescribed outgroup competi-
tion. More recently, Insko and Schopler (1987) also postulated such a group
schema as one explanation for their robust finding that in interdepen-
dent situations, people are consistently more competitive in intergroup ex-
changes than when facing the same situation in an interpersonal context—a
finding that is known as the “individual-group discontinuity effect” (Insko
et al., 1987, 1990). A primary component of the group schema, as devel-
oped by Insko and his colleagues, is schema-based distrust—the learned be-
lief or expectation that intergroup relations are inherently competitive and
therefore the outgroup cannot be trusted and the ingroup’s welfare must be
protected (see also Fiske & Ruscher, 1993). In an experimental test of the
schema-based mistrust hypothesis, Insko et al. (1990) found that group dis-
cussion preceding an intergroup prisoner’s dilemma game was characterized
by frequent references to the potential untrustworthiness of the other
group, and there was a strong negative correlation between the amount of
recorded distrust during discussion and later cooperation in the PDG play.
Further, groups more than individuals chose to withdraw from play when
Ingroup Identification and Intergroup Conflict
31


that option was made available, where withdrawal serves as an indicator of
fear that the other player will prefer competition to cooperation.
The Role of Emotions in Intergroup Contexts
When distrust leads groups to compete in an effort to protect themselves
from anticipated outgroup defection, a self-fulfilling cycle of competitive-
ness is set in motion (Kelley & Stahelski, 1970). What is of importance here
is the idea that this cycle is the product of a misattribution process whereby
universal ethnocentric behavior (intragroup trust and lack of trust in inter-
group encounters) is used to infer malevolent intent on the part of out-
groups. Similar misattribution processes may be at work in interpreting and
evaluating intergroup differences more generally. If individuals feel discom-
fort and unease in interactions with outgroups, those feelings may well be
attributed to intergroup differences in values, attitudes, and cultural prac-
tices. Hence, negative affect becomes associated with intergroup differences
and, through a process of misattribution of arousal, intergroup anxiety is
transformed into more virulent intergroup emotions of fear, hatred, or dis-
gust (Smith, 1993; Stephan & Stephan, 2000). It is this emotional compo-
nent that is postulated here to be the critical ingredient that turns inter-
group comparison into intergroup antagonism.
One perspective on the etiology of intergroup emotions can be derived
from the theory of optimal distinctiveness (Brewer, 1991) described earlier.
The theory holds that strong ingroup identification is associated with meet-
ing needs for secure inclusion (though similarity to the ingroup) and secure
differentiation (through intergroup distinctions). From this perspective,
there are two fundamental sources of potential threat to a person’s social
identity. One threat comes from feelings of marginalization from the 
desired ingroup—loss of a secure sense of inclusion and fit. One way of cop-
ing with this type of threat is to reaffirm support for ingroup positivity 
and moral superiority. However, hypervaluation of the ingroup alone may
actually increase the distance between the self and the ingroup ideal. Dero-
gation of outgroups, however, increases the contrast between ingroup 
and outgroup and, indirectly, the relatively similarity between the self 
and the ingroup by comparison. Thus, threats to inclusion are predicted 
to heighten feelings of moral superiority, intolerance of difference, and con-
comitant emotions of contempt and disgust toward relevant outgroups.
The emotions associated with moral superiority may justify some nega-
tive discrimination against outgroups, but do not necessarily lead directly to
hostility or conflict. In various contexts, groups have managed to live in a
state of mutual contempt over long periods without going to war over their
differences. The emotions of contempt and disgust are associated with
avoidance rather than attack, so intergroup peace may be maintained
through segregation and mutual avoidance. However, social changes that
give rise to the prospect of close contact, integration, or influence from the
32
The Contribution of Identities to Intergroup Conflict


outgroup engage the second form of threat to social identity—the threat of
loss of distinctiveness, accompanied by feelings of invasion. The emotions
engendered by this form of threat are fear and anger—emotions that propel
action against the outgroup rather than mere avoidance. This analysis sug-
gests that the combined emotions of contempt (engendered by moral supe-
riority in the service of insecure identity) and anger (engendered by fear of
invasion and loss of distinctiveness) provide the potent ingredients that are
sufficient to kindle hatred, expulsion, and even ethnic cleansing.
Some Moderating Conditions
The preceding analysis indicates that there are a number of steps by which
the effects of social categorization and ingroup-outgroup differentiation
may provide a fertile ground for conflict and hate. As the steps progress, the
line is crossed between the absence of trust and the presence of active dis-
trust, or between noncooperation and overt conflict. In this final section I
consider a few of the conditions that may exacerbate or moderate the
process whereby maintaining ingroup integrity and loyalty paves the way to
outgroup hate and hostility.
Common Goals.
The presence of superordinate goals or common threat is
widely believed to provide the conditions necessary for intergroup coopera-
tion and reduction of conflict (e.g., Sherif , 1966). This belief is an extrapo-
lation of the general finding that intragroup solidarity is increased in the face
of shared threat or common challenge. It may be true that loosely knit in-
groups become more cohesive and less subject to internal factioning when
they can be rallied to the demands of achieving a common goal. The dy-
namics of interdependence are quite different, however, in the case of
highly differentiated social groups. Among members of the same ingroup,
engaging the sense of trust necessary for cooperative collective action is es-
sentially nonproblematic. In an intergroup context, however, perceived in-
terdependence and the need for cooperative interaction make salient the
absence of mutual trust. Without the mechanism of depersonalized trust
based on common identity, the risk of exploited cooperation looms large
and distrust dominates over trust in the decision structure. It is for this rea-
son that I have argued elsewhere (Brewer, 2000) that the anticipation of posi-
tive interdependence with an outgroup, brought on by perceptions of com-
mon goals or common threat, actually promotes intergroup conflict and
hostility. When negative evaluations of the outgroup such as contempt are
also already present, common threat in particular may promote scapegoat-
ing and blame rather than mutual cooperation.
Perceived positive interdependence with the outgroup also threatens 
intergroup differentiation. T
o the extent that feelings of secure inclusion,
ingroup loyalty, and optimal identity are dependent upon the clarity of in-
group boundaries and intergroup distinctions, shared experiences and co-
Ingroup Identification and Intergroup Conflict
33


operation with the outgroup threaten the basis for social identification. Par-
ticularly for individuals who are exclusively vested in a single group identity,
the threat of lost distinctiveness may override the pursuit of superordinate
goals and lead to resistance to cooperation (collaboration) even at the cost
of ingroup self-interest.
Power Politics. Moral superiority, distrust of outgroups, and social compari-
son are all processes that emerge from ingroup maintenance and favoritism
and can lead to hostility and conflict between groups even in the absence of
realistic conflict over material resources or power. When groups are politi-
cal entities, however, these processes may be exacerbated through deliber-
ate manipulation by group leaders in the interests of mobilizing collective
action to secure or maintain political power. Social category differentiation
provides the fault lines in any social system that can be exploited for politi-
cal purposes. When trust is ingroup-based, it is easy to fear control by out-
siders; perceived common threat from outgroups increases ingroup cohe-
sion and loyalty; appeals to ingroup interests have greater legitimacy than
appeals to personal self-interest. Thus politicization—an important mecha-
nism of social change—can be added to the factors that may contribute to a
correlation between ingroup love and outgroup hate.
Culture. In addition to psychological needs, there are social structural and
cultural factors that may promote the intensity of ingroup attachment and
loyalty and enhance the distance between ingroup and outgroups. When
the salience and strength of intragroup interdependence and mutual obliga-
tion is increased, the importance of maintaining group boundaries is also
increased. Hence we might expect that in collectivist societies ingroup-
outgroup distinctions and distrust of outgroups would be higher than in 
individualistic societies where social interdependence is less emphasized.
Indeed, findings from cross-cultural studies of ingroup bias in collectivist
and individualist societies supports this prediction (Triandis, 1995).
Social Structural Complexity. A strong relationship between intense ingroup
favoritism and outgroup antagonism might also be expected in highly seg-
mented societies that are differentiated along a single primary categoriza-
tion, such as ethnicity or religion. And this would be especially true if the
categorization is dichotomous, dividing the society into two significant sub-
groups. Such segmentation promotes social comparison and perceptions of
conflict of interest that give rise to negative attitudes toward outgroups and
high potential for conflict. By contrast, the potential for intergroup conflict
may be reduced in societies that are more complex and differentiated along
multiple dimensions that are not perfectly correlated. Such a complex social
structure gives rise to cross-cutting category distinctions that mean that, at
the individual level, a person may be attached to one ingroup by virtue of
ethnic heritage, to another by religion, to yet another based on occupation,
34
The Contribution of Identities to Intergroup Conflict


or region of residence, and so forth. With this profusion of social identities
some individuals will be fellow ingroup members on one category distinc-
tion but outgroupers on another. Such cross-cutting ingroup-outgroup dis-
tinctions reduce the intensity of the individual’s dependence on any particu-
lar ingroup for meeting psychological needs for inclusion, thereby reducing
the potential for polarizing loyalties along any single cleavage or group dis-
tinction and perhaps increasing tolerance for outgroups in general.
Summary and Concluding Perspectives
The main point of this chapter has been to argue that any relationship be-
tween ingroup identification and outgroup hostility is progressive and con-
tingent rather than necessary or inevitable. Ingroup formation arises from
processes of social categorization and the need for cooperative interdepen-
dence within groups. At the individual level, identification with ingroups 
is motivated by the need to belong and to participate in bounded coopera-
tive social units. The bounded nature of ingroup attachment means that
trust and cooperation are limited to ingroup relations and withheld from 
outgroups.
Once the self is attached to a distinctive ingroup, additional motives to
achieve positive valuation, maintain secure inclusion, and protect ingroup
boundaries are engaged. It is at this point that outgroups may become rele-
vant to the maintenance or enhancement of ingroup identity. Social compe-
tition for positive valuation shifts initial indifference about outgroup out-
comes to concern for the relative position of the ingroup over outgroups.
Under these conditions, benefiting the ingroup comes at the expense of
outgroup benefits, and positive outcomes for the outgroup may arouse re-
sentment and antagonism if the comparison threatens important bases of
ingroup evaluation. This is the first step in connecting ingroup identity to
negative outgroup attitudes and underlies many forms of social discrimina-
tion and intergroup competition for social status.
At a more extreme point in the progression from ingroup identity to in-
tergroup conflict, perceived threats to ingroup inclusion or differentiation
from outgroups promote perceived conflict between ingroup identity and
the very existence of the outgroup or its goals and values. This is the point
at which preservation of the ingroup justifies direct aggression against 
the outgroup and underlies conflicts over political control, dominance, and
wars of annihilation. Such intense outgroup hostility is not a necessary or
inevitable concomitant of strong ingroup identity, but the basic processes of
social categorization that give rise to differential treatment of ingroup and
outgroup lay the groundwork for distrust and vulnerability that make this
escalation possible.
This analysis of the relationship between ingroup identity and outgroup
attitudes has a number of implications for research on intergroup relations
Ingroup Identification and Intergroup Conflict
35


and the prevention of the more virulent forms of intergroup conflict. First,
it implies that the study of ingroup bias, prejudice, and discrimination
should make a clearer distinction between discriminatory attitudes and be-
havior that are based on ingroup preference and positivity and discrimina-
tion that is based on antagonism and hostility toward outgroups. This is not
to suggest that ingroup-based prejudices are entirely benign or acceptable.
Indeed, these are the prejudices that are most likely to be subtle, uncon-
scious, and pervasive and thus more difficult to detect and address than
“old-fashioned” outgroup hatred. But understanding the motives underly-
ing discrimination based on group membership leads to different prescrip-
tions for preventing or reducing the discriminatory outcomes.
When differential cooperation, positivity, and trust toward others is
based on shared ingroup membership, discrimination can best be reduced
by conditions that foster greater inclusiveness, extending the boundaries of
the ingroup to include former outgroup members. This is the basic ratio-
nale for the common ingroup identity model for reducing intergroup bias
through recategorization of ingroup and outgroup into a superordinate 
category identity (Gaertner et al., 1993). This is most likely to be successful
when symbols of superordinate category membership are highly salient and
individual needs for differentiation are not aroused.
When intergroup attitudes and relations have moved into the realm of
outgroup hate or overt conflict, however, the prospect of superordinate
common group identity may constitute a threat rather than a solution
(Brewer, 2000). When outgroup attitudes are characterized simply by lack
of trust rather than active distrust, there is opportunity for capitalizing on
common interests and shared identities. But when intense distrust has al-
ready developed, common group identities are likely to be seen as threats
(or opportunities) for domination and absorption. In this case, the prescrip-
tion for conflict reduction may first require protection of intergroup
boundaries and distinctive identities.
From this analysis there is an inevitable tension between promoting co-
operative interdependence and shared interests at high levels of inclusive-
ness and the need to maintain and recognize intergroup differentiation and
distinctiveness. The negative effects of divisiveness will be most virulent
when individuals have very limited options for meeting their needs for se-
cure, optimal group identification. When social cleavages are defined along
a single dimension of differentiation, vulnerability to threats to inclusion
and differentiation is very high, and maintenance of ingroup identity is
most likely to be achieved at the cost of high levels of distrust and antago-
nism toward outgroups. When societies are more complex, when individu-
als have multiple avenues for meeting needs for inclusion and distinctive-
ness, intergroup comparisons along any particular division are likely to be
less competitive or critical to member identities. Under these conditions,
the positive benefits of distinctive ingroup identities may be attained with-
out the costs of intergroup hostility and conflict.
36
The Contribution of Identities to Intergroup Conflict


Note
1. Note, we are dealing here with prejudice and hostility at the collective level.
The hatred that a specific individual bears toward a particular type of people is not
at issue unless that hatred is grounded in norms and attitudes shared by the ingroup
as a whole.
References
Abrams, D., & Hogg, M. A. (1988). Comments on the motivational status of self-
esteem in social identity and intergroup discrimination. European Journal of Social
Psychology, 18, 317–334.
Allport, G. W. (1954). The nature of prejudice. Cambridge, MA: Addison-Wesley.
Baron, R. A. (1979). Effects of victim pain cues, victim’s race, and level of prior insti-
gations upon physical aggression. Journal of Applied Social Psychology, 9, 110–114.
Bettencourt, B. A., & Dorr, N. (1998). Cooperative interaction and intergroup bias:
Effects of numerical representation and cross-cut role assignment. Personality and
Social Psychology Bulletin, 24, 1276–1293.
Blanz, M., Mummendey, A., & Otten, S. (1997). Normative evaluations and fre-
quency expectations regarding positive and negative outcome alloctions between
groups. European Journal of Social Psychology, 27, 165–176.
Brewer, M. B. (1972). (Series editor). HRAFLEX Ethnocentrism Interview Series.
New Haven, CN: Human Relations Area Files.
Brewer, M. B. (1979). In-group bias in the minimal intergroup situation: A cognitive
motivational analysis. Psychological Bulletin, 86, 307–324.
Brewer, M. B. (1981). Ethnocentrism and its role in intergroup trust. In M. Brewer
& B. Collins (Eds.), Scientific inquiry in the social sciences (pp. 214–231). San Fran-
cisco: Jossey-Bass.
Brewer, M. B. (1991). The social self: On being the same and different at the same
time. Personality and Social Psychology Bulletin, 17, 475–482.
Brewer, M. B. (1996). In-group favoritism: The subtle side of intergroup discrimi-
nation. In D. Messick & A. T
enbrunsel (Eds.), Codes of conduct: Behavioral research
into business ethics (pp. 160–170). New York: Russell Sage Foundation.
Brewer, M. B. (1997). On the social origins of human nature. In C. McGarty & S. A.
Haslam (Eds.), The message of social psychology (pp. 54–62). Oxford: Blackwell.
Brewer, M. B. (2000). Superordinate goals versus superordinate identity as bases of
intergroup cooperation. In D.Capozza & R. Brown (Eds.), Social identity processes:
Trends in theory and research (pp. 117–132). London. Sage.
Brewer, M. B., & Brown, R. (1998). Intergroup relations. In D. Gilbert, S. Fiske, 
& D. Lindzey (Eds.), The handbook of social psychology. Fourth Edition (Vol. 1, 
pp. 99–142). Boston: McGraw-Hill.
Brewer, M. B., & Campbell, D. T. (1976). Ethnocentrism and intergroup attitudes: East
African evidence. Beverly Hills, CA: Sage.
Brewer, M. B., Manzi, J., & Shaw, J. S. (1993). In-group identification as a function
of depersonalization, distinctiveness, and status. Psychological Science, 4, 88–92.
Brewer, M. B., & Weber, J. G. (1994). Self-evaluation effects of interpersonal versus
intergroup social comparison. Journal of Personality and Social Psychology, 66,
268–275.
Ingroup Identification and Intergroup Conflict
37


Caporael, L. R. (1997). The evolution of truly social cognition: The core configura-
tions model. Personality and Social Psychology Review, 1, 276–298.
Crocker, J., & Major, B. (1989). Social stigma and self-esteem: The self-protective
properties of stigma. Psychological Review, 96, 608–630.
Deschamps, J-C., & Brown, R. (1983). Superordinate goals and intergroup conflict.
British Journal of Social Psychology, 22, 189–195.
Diehl, M. (1990). The minimal group paradigm: Theoretical explanations and em-
pirical findings. In W. Stroebe & M. Hewstone (Eds.), European review of social
psychology (Vol. 1, pp. 263–292). Chichester, England: Wiley.
Doise, W., Deschamps, J-C., & Meyer, G. (1978). The accentuation of intra-
category similarities. In H. T
ajfel (Ed.), Differentiation between social groups
(pp. 159–170). London: Academic Press.
Dovidio, J. F., & Gaertner, S. L. (1993). Stereotypes and evaluative intergroup 
bias. In D. Mackie & D. Hamilton (Eds.), Affect, cognition, and stereotyping
(pp. 167–193). San Diego, CA: Academic Press.
Festinger, L. (1954). A theory of social comparison proccesses. Human Relations, 7,
271–282.
Fiske, S. T., & Ruscher, J. (1993). Negative interdependence and prejudcie: Whence
the affect? In D. Mackie & D. Hamilton (Eds.), Affect, cognition, and stereotyping
(pp. 239–268). San Diego, CA: Academic Press.
Frey, D., & Gaertner, S. L. (1986). Helping and the avoidance of inappropriate in-
terracial behavior: A strategy than can perpetuate a non-prejudiced self-image.
Journal of Personality and Social Psychology, 50, 1083–1090.
Gaertner, S. L., & Dovidio, J. F. (1977). The subtlety of white racism, arousal, and
helping behavior. Journal of Personality and Social Psychology, 35, 691–707.
Gaertner, S. L., Dovidio, J. F., & Johnson, G. (1982). Race of victim, non-respon-
sive bystanders, and helping behavior. Journal of Social Psychology, 117, 69–77.
Gaertner, S. L. , Dovidio, J. F., Anastasio, P., Bachman, B., & Rust, M. (1993). The
common ingroup identity model: Recategorization and the reduction of inter-
group bias. In W. Stroebe & M. Hewstone (Eds.), European review of social psy-
chology (Vol. 4, pp. 1–26). Chichester, England: Wiley.
Greenwald, A. G., & Banaji, M. R. (1995). Implicit social cognition: Attitudes, self-
esteem, and stereotypes. Psychological Review, 102, 4–27.
Heider, F. (1958). The psychology of interpersonal relations. New York: Wiley.
Hewstone, M., Fincham, F., & Jaspars, J. (1981). Social categorization and similarity
in intergroup behaviour: A replication with “penalties.” European Journal of 
Social Psychology, 11, 101–107.
Hinkle, S., & Brown, R. (1990). Intergroup comparisons and social identity: Some
links and lacunae. In D. Abrams & M. Hogg (Eds.), Social identity theory: Con-
struction and critical advances (pp. 48–70). London: Harvester Wheatsheaf.
Hirt, E. R., Zillmann, D., Erickson, G., & Kennedy, C. (1992). Costs and benefits of
allegiance: Changes in fans’ self-ascribed competencies after team victory versus
defeat. Journal of Personality and Social Psychology, 63, 724–738.
Insko, C. A., & Schopler, J. (1987). Categorization, competition and collectivity. In
C. Hendrick (Ed.), Personality and Social Psychology Review. Vol. 8: Group Processes
(pp. 213–251). Beverly Hills, CA: Sage.
Insko, C. A., Pinkley, R., Hoyle, R., Dalton, B., Hong, G., Slim, R., Landry, P.,
Holton, B., Ruffin, P., & Thibaut, J. (1987). Individual versus group discon-
38
The Contribution of Identities to Intergroup Conflict


tinuity: The role of intergroup contact. Journal of Experimental Social Psychology,
23, 250–267.
Insko, C. A., Schopler, J., Hoyle, R., Dardis, G., & Graetz, K. (1990). Individual-
group discontinuity as a function of fear and greed. Journal of Personality and 
Social Psychology, 58, 68–79.
Kelley, H. H., & Stahelski, A. J. (1970). The social interaction basis of cooperators’
and competitors’ beliefs about others. Journal of Personality and Social Psychology,
16, 66–91.
Kramer, R. M., Brewer, M. B., & Hanna, B. A. (1996). Collective trust and collec-
tive action: The decision to trust as a social decision. In R. Kramer & T. Tyler
(Eds.), Trust in organizations (pp. 357–389). Thousand Oaks, CA: Sage.
Levin, S., & Sidanius, J. (1999). Social dominance and social identity in the United
States and Israel: Ingroup favoritism or outgroup derogation? Political Psychology,
20, 99–126.
LeVine, R. A. & Campbell, D. T. (1972). Ethnocentrism: Theories of conflict, ethnic 
attitudes and group behavior. New York: Wiley.
McClintock, C. (1972). Social motivation: A set of propositions. Behavioral Science,
17, 438–454.
Mullen, B., Brown, R., & Smith, C. (1992). Ingroup bias as a function of salience,
relevance, and status: An integration. European Journal of Social Psychology, 22,
103–122.
Mummendey, A., & Otten, S. (1998). Positive-negative asymmetry in social dis-
crimination. In W. Stroebe & M. Hewstone (Eds.), European review of social psy-
chology (Vol. 9, pp. 107–143). Chichester, England: Wiley.
Mummendey, A., & Schreiber, H. (1983). Better or just different? Positive social
identity by discrimination against or by differentiation from outgroups. European
Journal of Social Psychology, 13, 389–397.
Mummendey A., & Schrieber, H. (1984). ‘Different’ just means ‘better’: Some obvi-
ous and some hidden pathways to in-group favoritism. British Journal of Social
Psychology, 23, 363–368.
Mummendey, A., & Simon, B. (1989). Better or different? III: The impact of impor-
tance of comparison dimension and relative ingroup size upon intergroup dis-
crimination. British Journal of Social Psychology, 28, 1–16.
Mummendey, A., Simon, B., Dietze, C., Grunert., M., Haeger, G., Kessler, S.,
Lettgen, S., & Schaferhoff, S. (1992). Categorization is not enough: Intergroup
discrimination in negative outcome allocations. Journal of Experimental Social 
Psychology, 28, 125–144.
Neidenthal, P. M., & Beike, D. (1997). Interrelated and isolated self-concepts. Per-
sonality and Social Psychology Review, 1, 106–128.
Perdue, C. W., Dovidio, J. F., Gurtman, M. B., & Tyler, R. B. (1990). “Us” and
“Them”: Social categorization and the process of intergroup bias. Journal of Per-
sonality and Social Psychology, 59, 475–486.
Pettigrew, T. F., & Meertens, R. W. (1995). Subtle and blatant prejudice in Western
Europe. European Journal of Social Psychology, 25, 57–75.
Platow, M., O’Connell, A., Shave, R., & Hanning, P. (1995). Social evaluations of
fair and unfair allocations in interpersonal and intergroup situations. British Jour-
nal of Social Psychology, 34, 363–381.
Rogers, R. W., & Prentice-Dunn, S. (1981). Deindividuation and anger-mediated
Ingroup Identification and Intergroup Conflict
39


interracial aggression: Unmasking regressive racism. Journal of Personality and 
Social Psychology, 41, 63–73.
Rubin, M., & Hewstone, M. (1998). Social identity theory’s self-esteem hypothesis:
A review and some suggestions for clarification. Personality and Social Psychology
Review, 2, 40–62.
Sachdev, I., & Bourhis, R. (1991). Power and status differentials in minority and ma-
jority group relations. European Journal of Social Psychology, 21, 1–24.
Sears, D. O., & Funk, C. L. (1991). The role of self-interest in social and political
attitudes. In M. Zanna (Ed.), Advances in experimental social psychology (Vol. 24, 
pp. 2–92). New York: Academic Press.
Sherif, M. (1966). In common predicament: Social psychology of intergroup conflict and 
cooperation. New York: Houghton Mifflin.
Sidanius, J. (1993). The psychology of group conflict and the dynamics of op-
pression: A social dominance perspective. In S. Iyengar & W. McGuire (Eds.),
Explorations in political psychology (pp. 183–219). Durham, NC: Duke University
Press.
Smith, E. R. (1993). Social identity and social emotions: T
oward new conceptualiza-
tions of prejudice. In D. Mackie & D. Hamilton (Eds.), Affect, cognition, and
stereotyping (pp. 297–315). San Diego, CA: Academic Press.
Stangor, C., Sullivan, & Ford, T. E. (1991). Affective and cognitive determinants of
prejudice. Social Cognition, 9, 359–380.
Stephan, W., & Stephan, C. (2000). An integrated threat theory of prejudice. In 
S. Oskamp (Ed.), Reducing prejudice and discrimination (pp. 23–45). Mahwah, NJ:
Lawrence Erlbaum.
Struch, N., & Schwartz, S. H. (1989). Intergroup aggression: Its predictors and 
distinctness from in-group bias. Journal of Personality and Social Psychology, 56,
364–373.
Sumner, W. G. (1906). Folkways. New York: Ginn.
T
ajfel, H. (1969). Cognitive aspects of prejudice. Journal of Social Issues, 25, 79–97.
T
ajfel, H. (1970). Experiments in intergroup discrimination. Scientific American,
223(2), 96–102.
T
ajfel, H. (1978). Differentiation between social groups: Studies in the social psychology of
intergroup relations. London: Academic Press.
T
ajfel, H. (1981). Human groups and social categories. Cambridge: Cambridge Univer-
sity Press.
T
ajfel, H., Billig, M., Bundy, R., & Flament, C. (1971). Social categorization and 
intergroup behaviour. European Journal of Social Psychology, 1, 149–178.
T
ajfel, H., & Turner, J. C. (1986). The social identity theory of intergroup behavior.
In S. Worchel & W. Austin (Eds.), Psychology of intergroup relations (pp. 7-24).
Chicago: Nelson-Hall.
T
ajfel, H., & Wilkes (1963). Classification and quantitative judgement. British Jour-
nal of Psychology, 54, 101–114.
T
akagi, E. (1996). The generalized exchange perspective on the evolution of altru-
ism. In W. Liebrand & D. Messick (Eds.), Frontiers in social dilemma research
(pp. 311–336). Berlin: Springer-Verlag.
T
esser, A. (1988). T
oward a self-evaluation maintenance model of social behavior. 
In L. Berkowitz (Ed.), Advances in experimental social psychology (Vol. 21, pp. 181–
227). San Diego, CA: Academic Press.
Triandis, H. C. (1995). Individualism and collectivism. Boulder, CO: Westview Press.
40
The Contribution of Identities to Intergroup Conflict


Turner, J. C. (1975). Social comparison and social identity: Some prospects for 
intergroup behaviour. European Journal of Social Psychology, 5, 5–34.
Turner, J. C. (1978). Social categorization and social discrimination in the minimal
group paradigm. In H. T
ajfel (Ed.), Differentiation between social groups (pp. 101–
140). London: Academic Press.
Turner, J. C. , Hogg, M., Turner, P
., & Smith, P
. (1984). Failure and defeat as deter-
minants of group cohesiveness. British Journal of Social Psychology, 23, 97–111.
Turner, J. C. , Hogg, M., Oakes, P., Reicher, S., & Wetherell, M. (1987). Rediscover-
ing the social group: A self-categorization theory. Oxford: Blackwell.
Weber, J. G. (1994). The nature of ethnocentric attribution bias: Ingroup protection
or enhancement? Journal of Experimental Social Psychology, 30, 482–504.
Wood, J. V. (1989). Theory and research concerning social comparisons of personal
attributes. Psychological Bulletin, 106, 231–248.
Ingroup Identification and Intergroup Conflict
41


Thomas Hylland Eriksen
3
Ethnic Identity, National Identity,
and Intergroup Conflict
The Significance of Personal Experiences
There exists by now an enormous literature on contemporary ethnicity and
nationalism in its various forms. It deals with a wide range of phenomena—
North American multiculturalism and indigenous rights movements, post-
Soviet ethnonationalism in Central and Eastern Europe, urban minority
dilemmas and Islamic revivalism in Western Europe, indigenista movements
in Latin America, and processes of political fission and fusion in contempo-
rary Africa. The analytical focus of this chapter will be on the concept of
identity. The naïve question to be asked at the outset is: what is it about
identity politics that makes it such a formidable force in the contemporary
world? Some important issues will have to be omitted—notably the rela-
tionships among identity politics, globalization, and reflexive modernity
(e.g., Bauman, 1993; Friedman, 1994; Giddens, 1994). Instead, what is of-
fered amounts to an anthropological perspective on the relationship be-
tween personal identity and political identity, using conflicts based on iden-
tity politics as empirical examples. 
This chapter sets out to do three things. First, a brief overview of the
standard social anthropological perspective on the politics of identity is pro-
vided. Identity politics should be taken to mean political ideology, organiza-
tion, and action that openly represents the interests of designated groups
based on “essential” characteristics such as ethnic origin or religion, and
whose legitimacy lies in the support of important segments of such groups.
Membership in such groups is generally ascribed, unlike membership in
other political groups (socialists, liberals, trade unions, etc.). Second, by way
of examples from India, Fiji, and Yugoslavia, parallels and differences be-
tween some such conflicts are highlighted. The examples are chosen mainly
42


for their differences: what they have in common is, apart from their tempo-
ral location in the 1990s, apparently only that they are based on ideologies
of culture and identity. The Yugoslav conflict has been extremely brutal and
tragic; the Fijian one has involved only a few casualties, but has led to im-
portant constitutional changes; while the Indian conflict, although occa-
sionally bursting into violence, is largely contained within the framework of
institutional Indian politics. If it can persuasively be argued that these con-
flicts have important features in common, it is likely that those features will
also be present in other settings where identity politics has a major impact.
The third and final part of the chapter draws on classic political anthropol-
ogy in order to suggest some general features of the relationship between
personal identity and politics.
The Anthropology of Identity
The last decades of the 20th century saw a dramatic reconceptualization 
of core concepts, including culture and society, within the social sciences.
Until the 1960s, the close overlap between culture and ethnicity, or even
culture and nationhood, was generally taken for granted in the scholarly
community. During the past 30 years, however, hardly a single serious con-
tribution to the field has failed to point out that there is no one-to-one rela-
tionship between culture and ethnicity (the seminal text here is Barth,
1969); that cultural differences cut across ethnic boundaries; and that ethnic
identity is based on socially sanctioned notions of cultural differences, not
“real” ones. While ethnic identity should be taken to refer to a notion of
shared ancestry (a kind of fictive kinship), culture refers to shared represen-
tations, norms, and practices. One can have deep ethnic differences without
correspondingly important cultural differences (as in the Bosnian example
below); and one can have cultural variation without ethnic boundaries (as,
for example, between the English middle class and the English working
class).
Several recent debates in anthropology and neighboring disciplines pull
in the same direction: away from notions of integrated societies or cultures
toward a vision of a more fragmented, paradoxical, and ambiguous world.
The currently bustling academic industry around the notion of globaliza-
tion (see Featherstone, 1990, for an early, influential contribution) repre-
sents an empirically oriented take on these issues, focusing on the largely
technology-driven processes that contribute to increasing contact across
boundaries and diminished importance of space. This focus on unbounded
processes rather than isolated communities has contributed to a reconcep-
tualization of the social, which is radically opposed to that of classic Durk-
heimian sociology and anthropology; where flux, movement, and change
become the rule and not the exception in social life (Strathern, 1991; Han-
nerz, 1992; Lash & Urry, 1994). 
Ethnic Identity, National Identity, and Intergroup Conflict
43


Particularly in North America, the classic concept of culture has been
used for what it is worth in domestic identity politics, leading in some cases
to controversial policies of multiculturalism, where individuals have been
endowed with special rights in accordance with their ethnic origins. Critics
might point out that multiculturalism in some of its versions resembles
apartheid; also, that by positing a simple one-to-one relationship between
ethnic origin and culture, it not only encouraged a “disuniting of America”
(Schlesinger, 1992) but also contributed to reifying misleading notions of
culture seen as the commonalities of a bounded set of individuals, like so
many nationalisms writ small. 
A further disruptive tendency has been the so-called postcolonial move-
ment in literary studies, spilling into anthropology and other disciplines,
that has raised the question of who has the right to identify whom; a stan-
dard text in this field of discourse is Edward Said’s Orientalism (1978), al-
though Frantz Fanon developed similar insights two decades earlier. Said
and others argued, briefly, that ethnocentrism was deeply embedded in
Western scholarship dealing with non-Western peoples. Postcolonial critics
also tend to call attention to the multiplicity of voices (an academic cliché
by the late 1990s) present in any society and the general unwillingness of
academic researchers to give all of them the attention they deserve. 
T
wo related debates defined the field for many years. First, there was the
controversy over primordialism and instrumentalism. Was ethnic identity
“primordial”—that is, profoundly rooted in, and generative of, collective
experiences; or did it arise as an ad hoc supplement to political strategies?
An early, powerful defense of the instrumentalist view was represented in
Abner Cohen’s work on urban ethnicity in Africa (1969, 1974), showing the
conscious manipulation of kinship and cultural symbols by political entre-
preneurs seeking political gain. This perspective is still used with consider-
able success in studies of identity politics. Who, then, were (or are) the pri-
mordialists? Clifford Geertz is often associated with this view, arguing as 
he does along hermeneutic lines that cultural systems are more or less self-
sustaining and are thus not subject to the willful manipulation of individuals
(Geertz, 1973), a perspective he retained when writing about nationhood 
in the Third World (Geertz, 1967). Typically, however, ethnicity studies
were—and are—instrumentalist in their basic orientation (Rex, 1997). 
The second debate, usually framed as the opposition between construc-
tivism and essentialism, concerns the question of whether ethnic or national
communities are created more or less consciously, or whether they grow or-
ganically, as it were, out of preexisting cultural communities. In nationalism
studies the most highly profiled antagonists regarding this issue have been
the late Ernest Gellner (1983, 1997) and Anthony D. Smith. Smith (1986,
1991) has developed an intermediate position in arguing the importance of
preexisting ethnies for the development of nationalism while acknowledging
its essential modernity. Gellner, on the other hand, champions the view that
nations are entirely modern creations—the progeny of industrialism and
44
The Contribution of Identities to Intergroup Conflict


the state—that more or less fraudulently invent their past to gain a sem-
blance of antiquity and deep roots; his final statement on the issue is repro-
duced in the posthumously published Nationalism (Gellner, 1997) under the
heading “Do Nations Have Navels?”, a pun on the biblical enigma relating
to Adam’s navel. Regarding definitions, there are also important differences
between theorists. While Gellner holds nations to be ethnic groups who ei-
ther control a state or who have leaders who wish to do so, Anderson (1983)
sees no necessary link between the abstract “imagined community” of the
nation and particular ethnic groups; indeed, several of the main examples in
his famous book on “imagined communities,” including the Philippines and
Indonesia, are multiethnic countries. Yet others have distinguished between
ethnic nations and “civic” ones (Smith, 1991). There is nonetheless general
agreement that nations are by definition linked with states, whether they
are based on a common ethnic identity or not.
In anthropology at least, the recent shift toward the study of identities
rather than cultures has entailed an intense focus on conscious agency and
reflexivity; and for many anthropologists, essentialism and primordialism
appear as dated as pre-Darwinian biology. In addition, there seems to be
good political sense in discarding the old, static view of culture, which is be-
ing used for many political purposes that are difficult to endorse by acade-
mics committed to democratic values, ranging from the Balkan war to dis-
crimination against ethnic minorities in Western Europe. Further, this is an
age when the informants talk back. It could, perhaps, be said that a main
purpose of an earlier anthropology consisted of identifying other cultures.
Representatives of these so-called other cultures are now perfectly able to
identify themselves, which leaves the scholars either out of a job or with a
new mission—that is, to identify their identifications—in other words, to
study reflexive identity politics. 
Not so many years ago, anthropology was still a discipline fueled by a
programmatic love of cultural variation for its own sake, and anthropolo-
gists involved in advocacy tended to defend indigenous peoples’ or other
minorities’ traditional ways of life against the onslaught of modernity. A
main tendency in recent years has, on the contrary, consisted of decon-
structing instrumentalist uses of notions of authenticity and tradition, and
showing not only that the internal variation within a group is much greater
than one would expect but also that traditionalist ideologies are, paradoxi-
cally, direct results of modernization (e.g., Roosens, 1989). This theoretical
shift is a very significant one. It offers a method for investigating extremely
well the strategic action, politics of symbols, and contemporary processes of
identity politics within a uniform comparative framework. 
In other words, cracks in the edifice of mainstream social and cultural
anthropology, some of them directly inspired by events well beyond the
confines of academia, have led to a widespread reconceptualization of soci-
ety and culture. Reification and essentialism have become central terms of
denunciation; multiple voices, situational identification, and cultural flows
Ethnic Identity, National Identity, and Intergroup Conflict
45


are some of the key words delineating the current intellectual agenda. It has
become difficult if not impossible to talk of, say, Nuer culture, Hopi cul-
ture, Dutch culture, and so on, since such terms immediately invite critical
questions of whose Nuer culture, Hopi culture, and so on, intimating that
there are an infinite number of versions of each culture, none of which is
more “true” than the others (Holland, 1997). Ethnicity and nationalism,
then, become the political reifications or constructions of a particular au-
thorized version of a culture, freezing that which naturally flows, erecting
artificial boundaries where they did not exist before, trimming and shaping
the past to fit present needs, and inventing traditions where no organic tra-
ditions exist, or are not adequate, to ensure a sense of continuity with the
past. 
A new kind of political responsibility has entered academia in acute ways
during the last decades. Academic or semiacademic statements about na-
tions, ethnic groups, or cultures may now immediately be picked up, or as-
similated more or less subconsciously, by ideologists and politicians wishing
to build their reputation on national chauvinism, ethnic antagonism, enemy
images, and so on. The liberal academic establishment thus wags a warning
finger at those who dare to talk of culture as the cause of conflicts, shaking
their heads sadly over those lost souls who have not yet heeded the words of
leading theorists such as Barth (1969) and Gellner (1983), criticizing those
who do not realize that culture is chimerical and fleeting, and that reified
culture is a dangerous tool. It is, thus, not only intellectually correct but
also politically correct to reject all forms of essentialism. 
The current scholarly orthodoxy on ethnicity and the politics of identity
can be summed up as follows:
• Although ethnicity is widely believed to express cultural differ-
ences, there is a variable and complex relationship between eth-
nicity and culture; and there is certainly no one-to-one relationship
between ethnic differences and cultural ones. 
• Ethnicity is a property of a relationship between two or several
groups, not a property of a group; it exists between and not within
groups. 
• Ethnicity is the enduring and systematic communication of cultural
differences between groups considering themselves to be distinct. It
appears whenever cultural differences are made relevant in social
interaction, and it should thus be studied at the level of social life,
not at the level of symbolic culture. 
• Ethnicity is thus relational and also situational: the ethnic character
of a social encounter is contingent on the situation. It is not, in
other words, inherent.
This instrumentalist framing of ethnicity, which may appear simply as a
set of methodological guidelines, has firm, although usually untheorized,
philosophical foundations and is, as I have tried to show elsewhere (Eriksen,
46
The Contribution of Identities to Intergroup Conflict


1998b), deeply embedded in empiricist thought. I shall argue that this ap-
proach, notwithstanding its strengths, is limited in overemphasising choice
and strategy (instrumental aspects) when analyzing identity politics. As a re-
sult, the self is taken for granted (A.P. Cohen, 1994), and it is therefore not
shown how it can be possible to mobilize particular aspects of personal
identity for antagonistic identity politics. Yet, anthropology as a discipline is
in a privileged situation to study the dynamics of identity politics, precisely
because of its focus on the ongoing flow of social interaction.
Although an enormous amount of anthropological research has been
carried out on ethnicity and nationalism since around 1970, surprisingly
few studies have dealt with violent conflicts and conflict resolution (see,
however, T
ambiah, 1994; Turton, 1997). The dominant approaches to eth-
nicity have been instrumentalist (with a focus on politics) or constructivist
(with a focus on ideology), and research questions have concentrated on the
establishment and reproduction of ethnically incorporated groups, not on
the circumstances under which ethnicity may become politically less impor-
tant. While my examples (below) and the ensuing discussion will indicate
the fruitfulness of these approaches, it is also necessary to point out the
need for a phenomenological understanding of social identity, which sees it
as emerging from experiences, not as a mere construct of ideology. In this, I
follow scholars such as A. P. Cohen (1994) and Jenkins (1996), who have
called for an anthropology of identity that does not concentrate exclusively
on its political and ideological aspects, but also strives to understand the
self. 
As noted by Holland (1997), anthropologists are generally associated
with a culturalist view of the self, arguing its cultural specificity as against
psychologists, who have been more prone to a universalist view—the self as
something proper not to particular cultures but to humanity, with universal
characteristics lurking below a thin veneer of culture. For the purposes of
the present argument, it is not necessary to take a stance on this contro-
versy, partly since it can be presupposed that modernity creates a particular
kind of selves with important shared characteristics everywhere (Giddens,
1991), but also because this chapter restricts its scope to the relationships
among personal experiences, ideology, and political mobilization. As the 
examples will, it is hoped, show, the similarities are more striking than the
differences here.
Culture and the Breakup of Yugoslavia
No other recent ethnic conflict has been more intensely studied, discussed,
and moralized over than the breakup of Yugoslavia in 1991. This was 
followed by three major wars and a number of smaller skirmishes, and the
situation in many parts of ex-Yugoslavia remains unstable and tense. In 
Europe, the outbreak of war in Yugoslavia has been interpreted by hun-
Ethnic Identity, National Identity, and Intergroup Conflict
47


dreds of commentators. In the press, it was occasionally argued, along 
sociobiological or Hobbesian lines, that humans are driven by aggressive
instincts that emerge when the social fabric falls apart, in this case orga-
nized as factions ultimately based on kinship. Another view, popular in the
European nationalist right, implied that ethnic conflict was inevitable when
different groups are forcibly integrated into one state. Most scholars have,
on the contrary, tended to focus on the cultural logic of feuding in Balkan
society, the deep economic crisis underlying the conflict, or the rise of 
Serbian supremacy during the 1980s, looking for contradictions within Yu-
goslav society rather than into human nature for an explanation of the con-
flict. A widespread view nevertheless sees the cultural differences between
the constituent groups as a basic cause of the conflict (cf., the influential
analysis by Ignatieff, 1994, or Huntington’s controversial model, 1996). 
Yugoslavia was a state that came into being twice: after the First World
War and after the Second World War. The first Yugoslav republic (1918–
41) was for all practical purposes the Serb monarchy writ large, and was rid-
dled with continuous internal tension. Croats and Slovenes reluctantly 
supported King Alexander’s regime, seeing it as a possible defense against
Italian, Austrian, and Hungarian aggressors. It was a precarious state, peri-
odically dictatorial, that had been on the verge of collapse several times 
before it fell apart as Germany invaded the country in 1941. While Serbs
generally resisted the Germans, Croats collaborated and saw the German
intervention as an opportunity to create their own state. 
Before the formation of the second Yugoslav republic in 1945, both of
the largest constituent groups suffered large-scale massacres—Serbs at the
hands of Croatian fascists (Ustasa) in 1941 and Croats killed by Serbian
communists (Partisans) in 1945. The new Yugoslav state (1945–91) was a
nonaligned socialist federation led by a pro-Serbian Croat, Josip Broz Tito,
until his death in 1980. Ethnicity was officially declared a nonissue in so-
cialist Yugoslavia. This official blindness stems from the Marxist view that
class is a more objective and more authentic vessel of social identity than
ethnicity or nationality, which were officially seen as expressions of false
consciousness. This does not, however, imply that Yugoslav policies were
particularly repressive regarding expressions of cultural distinctiveness sym-
bolizing ethnic identity; on the contrary, in this area Yugoslavia was more
liberal than many Western European countries whose leadership feared
separatism and social fission. Ethnic identity was seen as politically irrele-
vant, and partly for this reason, the use of various languages and the prac-
tice of different religions were tolerated in civil society. It is true that the
merging of Serbian and Croatian into one language, Serbo-Croatian, in 
the 1950s signaled an attempt at building a unitary Yugoslav identity, but
the two languages were so closely related that few, except Croat intellectu-
als reacting against the relegation of specific Croat variants as “dialect,”
seem to have taken offense (Schöpflin, 1993). Albanian remained an official
language in Kosovo, as did Slovene in Slovenia. 
48
The Contribution of Identities to Intergroup Conflict


There were nearly twice as many Serbs as Croats in Yugoslavia, and
twice as many Croats as Slovenes or Muslims. Regarding the territorial 
dimension, Slovenes were—and are—largely confined to the nearly mo-
noethnic republic of Slovenia. Both Croatia and Serbia had large minorities
of Serbs and Croats, respectively, as well as Gypsies, while Serbia also in-
cluded nearly a million Albanians (in Kosovo) and smaller numbers of Hun-
garians and others. In Macedonia, most of the population were (and are)
Slav-speaking Macedonians, while the Montenegrins of Montenegro are
culturally close to the Serbs. Bosnia-Herzegovina was the most throroughly
mixed republic, with roughly equal numbers of Serbs, Croats, and Bosnian
Muslims, often living in mixed areas. As is well known, the conflicts of the
1990s have modified this picture somewhat, creating large monoethnic 
territories in formerly mixed areas.
Although the ruling Communist party seems to have believed that a
common Yugoslav identity would eventually supersede the national identi-
ties based on ethnic membership, ethnic identity remained strong in most
parts of the country throughout the postwar era. There were nevertheless
important exceptions, particularly in cities such as Belgrade, Sarajevo, and
Zagreb, where many people increasingly identified themselves primarily as
Yugoslavs and where mixed marriages were common.
Ethnic identities did, in other words, not disappear during the existence
of Yugoslavia. In some urban areas they were arguably weakened, but it
could be—and has been—argued that the nonethnic character of Yugoslav
politics actually led to its strengthening as a vehicle for the political opposi-
tion and made it possible for Serbs to gain control over the armed forces
and state bureaucracy: since political ethnicity officially did not exist (only
cultural ethnicity did), there were no institutionalized ways of preventing
one group from dominating the public sector. 
The wars in the former Yugoslavia have bequeathed to the world the 
neologism ethnic cleansing. It is nevertheless easy to show that the conflicts
involving Serbs, Croats, Bosnian Muslims, Slovenes, and Albanians were
never conflicts over the right to assert one’s ethnic or cultural identity, but
were based on competing claims to rights such as employment, welfare, and
political influence. What needs to be explained is the fact that the conflicts
over these resources were framed in ethnic terms rather than being seen as,
say, regional, class-based, or even ideological. 
The relevant questions are therefore: what is the stuff of ethnic identity
in the former Yugoslavia; in which ways do the groups differ from one an-
other, and why did group allegiances turn out to be so strong? Bosnia-
Herzegovina may be considered as an example. There are three large ethnic
groups inhabiting Bosnia: Serbs, Croats, and Muslims. The main difference
between the groups is religion—Serbs are Orthodox and Croats are
Catholics. (Ironically, religious fervor was not particularly widespread in
prewar Bosnia.) They all have common origins: Slav immigration into 
Illyria from the north took place between c. a.d. 400 and 700, and the cul-
Ethnic Identity, National Identity, and Intergroup Conflict
49


tural differences between Croats and Serbs are perhaps comparable to those
between Norwegians and Swedes. The “objective” difference between
Bosnian Christians and Bosnian Muslims, further, has been compared to
the difference between English Protestants and English Catholics (Cornell
& Hartmann, 1998). Unlike the impression sometimes given in Serbian and
Croatian propaganda, Bosnian Muslims are not the descendants of alien 
invaders, but of locally residing converts. Although each group has its nu-
merical stronghold, many Bosnian regions and villages were mixed before
the war. This implies, among other things, that they went to school to-
gether, worked together, and took part in various leisure activities together.
A Serbian villager in Bosnia had more in common, culturally speaking, with
a Muslim co-villager than with a Serb from Belgrade. This would hold true
of both dialect spoken and way of life in general. However, since religion
turned out to be the central marker of collective identity in the Bosnian
conflict, the effective boundary was drawn not between villagers and city-
dwellers, but between religious categories.
The boundaries between the groups may seem arbitrary. However, 
the large, “national” groups are clearly embedded in smaller, local networks
based on kinship and informal interaction, as well as being culturally
founded in religious schisms, collective myths or memories of treason and
resistance under Ottoman rule, massacres, deception, and humiliations. 
Although it is tempting to argue that any so-called cultural trait can be 
exploited in the formation of national or ethnic groups, it is obvious that
not just anything will do. Nothing comes out of nothing, and strong collec-
tive identities—such as the ones revealed during the war in Bosnia—are 
always embedded in personal experiences. In one of the most detailed 
accounts of ethnicity at the village level in prewar Bosnia, Bringa (1996)
shows that although cultural differences between the groups were perhaps
negligible, and although relations among Serbs, Muslims, and Croats might
be cordial at the local level, there were nevertheless important social prac-
tices of affiliation that created boundaries between them—not in the cos-
mopolitan Sarajevo middle class embraced by Western commentators, per-
haps, but elsewhere. Intermarriage was rare, the close informal networks 
of friends tended to be monoethnic, and the discrete groups maintained 
different, sometimes conflicting myths of origin. The intimate sphere, in
other words, seems to have been largely monoethnic and by this token,
Bosnia was a plural society in the classic sense (Furnivall, 1948); the public
arenas were shared, but the private ones were discrete. 
One may choose not to speak of such features of social reality and every-
day life in terms of “culture,” but they are no more “invented” than any
other social fact. People do not choose their relatives, they cannot choose to
do away with their childhood and everything they learned at a tender age.
These are aspects of identity that are not chosen, that are incorporated and
implicit. People relate to them as reflexive agents, but they do so within
limitations that are not chosen. Such limitations form the objective founda-
50
The Contribution of Identities to Intergroup Conflict


tions of social identification. When analysts such as Cornell and Hartmann
(1998) argue that ethnic identities before the war were weak, sometimes so-
cially irrelevant, and in many cases ambiguous (many had parents from dif-
ferent groups), they refer to particular segments of society—a main example
is the Yugoslav basketball team—whose members were active participants in
pan-Yugoslav contexts. 
The maintenance of ethnic boundaries in socialization and the private
sphere reveals a main cause of the failure of Yugoslav social engineering in
doing away with ethnic identification. It does not explain the outbreak of
war in the early 1990s, but it indicates why the groups that emerged were 
so strong, and why they were based on ethnicity (seen as fictive kinship) 
instead of, say, class or region. Their foundation must be sought not in the
biology of kinship, as some might want to argue, but in the phenomenology
of social experience, the raw material of personal identity. This argument
will be elaborated after an examination of two very different examples of 
intergroup conflict in polyethnic societies.
The Fijian Coup-D’etat
The Pacific island-state of Fiji, located at the ethnographic crossroads of
Melanesia and Polynesia, is perhaps less heterogenous than Yugoslavia, but
it is scarcely less ethnically divided. Its population of about 800,000 is
largely composed of two ethnic categories: Fijians and Indians. The Fijians
are indigenous, largely Christian, speak a Polynesian language, and make
up slightly less than half the population. The Indians are uprooted “over-
seas Indians” whose ancestors were brought to Fiji during colonialism un-
der the British indentureship system described, probably a trifle too grimly,
as “a new system of slavery” by Tinker (1974). They are overwhelmingly
Hindus (with a Muslim minority), speak a locally modified dialect of Hindi,
and were slightly more numerous than the indigenous Fijians until the po-
litical changes in the late 1980s leading to mass migration of Indo-Fijians.
The small minorities of Europeans and Chinese are politically insignificant,
but the economically powerful Europeans, representing the former colonial
regime, have in no small measure shaped the Fijian public sphere, notably
through establishing English as the national lingua franca.
The relatively brief history of democratic Fijian politics up to 1987 has
been described with the metaphors of balance and power sharing (Premdas,
1993; Kelly, 1998). While Indians in practice wielded disproportionate eco-
nomic power, it was tacitly agreed that Fijians should be politically para-
mount. However, there were indications that the “equilibrium” model was
under severe stress in the early- to mid-1980s, and in the 1987 elections, a
coalition supported by most Indians and only a few Fijians won. Although
the new prime minister was a Fijian, many saw his government as a vehicle
for Indian communal politics. In May and September 1987, the military
Ethnic Identity, National Identity, and Intergroup Conflict
51


seized power through two successive coups-d’etat, explicitly doing so to
protect “native” Fijian interests. In the period following the coups, thou-
sands of Indians emigrated. 
The tension between Fijians and Indians had been evident throughout
Fijian history. Unlike the situation in Yugoslavia and especially Bosnia, 
nobody would question the view that there are deep cultural differences 
between Indians and Fijians. Their languages, cultural traditions, religions,
and gender relations differ markedly—indeed, one of their recent ethnogra-
phers has poetically described striking, and culturally potent, differences 
in body language between Fijians and Indians (Williksen-Bakker, 1991).
There seems to be little informal interaction between the groups, most
rural areas are dominated by one or the other, and intermarriage has always
been nearly nonexistent. The effective separation of Fijians from Indians
has always been much deeper than that obtaining between the major groups
(or “nationalities”) in rural Bosnia after the Second World War. Until 1987,
policies of compromise had nonetheless ensured political stability and had
made Fiji a remarkably liberal and relaxed society.
Invoking the slogan “Fiji for Fijians” (which had been launched before
the 1977 elections), the writers of the new constitution, promulgated in
1990, ensured continued indigenous Fijian dominance of the political
sphere, by according that group disproportionate representation in parlia-
ment, ruling that only Fijians can become prime ministers, and giving 
Fijians preferential treatment in other areas as well, such as religion. In ad-
dition, Fijians are guaranteed control over most of the arable land, about 82
percent of which has been communally owned by Fijian kin groups since
the beginning of colonialism in the 19th century. 
The conflict culminating in the military coups can, at one level of analy-
sis, be seen as a clear case of group competition. Indians have done better
economically than Fijians. Ironically, this may partly be explained through
the British colonial policy of indirect rule relating to Fijians, who were al-
lowed to retain important traditional institutions, such as chieftainship and
the rudiments of a caste system, making them in consequence unprepared
to compete with Europeans and Indians in a capitalist economy later. The
Indians, by contrast, had shed important aspects of their traditional social
organization in the process of migration, and were accustomed to economic
individualism through the indentureship system whereby they were made
to work on the European-owned plantations. 
The demographic growth rate among Indians has been higher than
among Fijians (just as the Muslims in Bosnia were more prolific than the
other groups), and there was generally a growing sentiment among Fijian
leaders that they were becoming a minority in the land of their ancestors.
The inegalitarian measures introduced by the military regime, discriminat-
ing between categories of citizens on ethnic grounds, were condemned by
the international community, but less strongly than one might have ex-
pected in a different setting. What is remarkable about the Fijian case is the
52
The Contribution of Identities to Intergroup Conflict


nativist quality of the supremacist rhetoric—how they brought “sons-of-
the-soil” arguments to bear on a national legislation, creating, in effect, a
two-tier society where non-Fijians were relegated to the status of second-
class citizens. They argued that their culture, like that of Maoris in New
Zealand and Aborigines in Australia, was threatened with marginalization
from outside forces. In contrast, nobody in countries like Mauritius,
Trinidad and T
obago, and Guyana—which are in many ways similar to Fiji
but lack substantial indigenous populations—would have been able to in-
voke arguments of cultural authenticity and preservation of traditional cul-
tures in a bid to introduce differential treatment for different ethnic groups
(see Eriksen, 1992, 1998a, for details on Trinidad and Mauritius).
It may well be asked whether contemporary Fijians, being Methodists and
proficient English speakers, are any less culturally uprooted than, say, Trini-
dadians of African descent or, for that matter, Fijians of Indian descent. The
question of cultural authenticity is outside the scope of this chapter; let me
now make a few more pertinent points relating to this example.
Fiji had developed an informal formula for interethnic accommodation
where the largest ethnic groups divided societal sectors between them. In
addition there were—and still are, at least to some extent—developing
fields of shared meaning and cross-cutting alignments, such as the common
use of English as a national language and a shared educational system. Nev-
ertheless the segregation between the groups in both social and cultural do-
mains is more striking to the outside observer than tendencies toward as-
similation. Conflict avoidance would thus have to rely chiefly on group
compromise rather than the development of a hybridized, shared identity—
an option that has occasionally been proposed by politicians and intellectu-
als in other insular, postcolonial plantation societies such as Mauritius and
Trinidad. Finally, owing to historical circumstances and cultural differences,
Indians and Fijians have participated in different ways and have succeeded
to varying degrees in the modern sectors of politics and the economy. It
could indeed be argued that processes of modernization in Fiji, far from 
reducing cultural differences, have deepened them, at least at the socially
operational level. Unlike in Bosnia, it is possible to refer to differences in
local organization, cosmology, and traditional economical practices when
accounting for the ethnic conflict in Fiji, which is nevertheless much less 
violent than the Bosnian one. This is a reminder, against cultural determin-
ists à la Huntington (1996), of the relative unimportance of cultural differ-
ences for ethnic conflict—and it also indicates one of the main strengths of
the constructivist-instrumentalist perspective on identity politics: cultural
differences do not in themselves lead to intergroup conflict, but are invoked
strategically to mobilize support. At the same time, it must be conceded
that the differences in life-worlds and personal identities in Fiji, as in
Bosnia, explain why the political cleavages were given ethnic expression.
The differences were already there before they were exploited for particular
political ends.
Ethnic Identity, National Identity, and Intergroup Conflict
53


Hindutva: An Apparent Anomaly in Contemporary Indian Society
My third and final example differs from the two previous ones in significant
ways. India is a tough case for any scholar trying to develop a general theory
of ethnicity or nationalism, and with few exceptions, it does not figure in
general introductory texts on the field. India is hardly a state based on cul-
tural similarity or even equality in the Western sense; it is a country with
deeply embedded hierarchies and a very considerable degree of internal cul-
tural variation. Its population of nearly a billion is divided by language, reli-
gion, caste, and culture, and it has often been argued that India is culturally
more complex than continents such as sub-Saharan Africa or Europe. Al-
though 80 percent of the population are Hindus in one meaning of the
word or another, India also has the second-largest Muslim population in the
world (after Indonesia) and more Christians than all the Scandinavian
countries put together. Since independence (and partition) in 1947, India
has been defined in Gandhian-Nehruvian terms as a secular, federal country
using English and Hindi as national languages, but with another dozen or
so official regional languages.
Since the early 1980s—but particularly forcefully during the 1990s—a
formerly marginal political movement has steadily increased its influence in
India, culminating in its victories in the successive general elections of 1998
and 1999. This is the movement often referred to as hindutva, meaning
roughly “Hindu-ness,” which rallies behind slogans to the effect that India
should be redefined as a Hindu country. The hindutva movement, led by an
organization called the RSS (Rashtriya Swayamsevak Sangh) began mod-
estly in the interwar years, and its more recent parliamentary wing, the BJP
(Bharatiya Janata Party, “The Indian People’s Party”) is now in power not
only federally, in New Delhi, but in several of the states as well.
The rhetoric of hindutva is strongly reminiscent of European ethnic na-
tionalism. It invokes ancient myths of bitter defeats and noble sacrifices, re-
framing them to fit a contemporary political scene. It quotes liberally from
19th-century poets and sacred texts, and it redefines history to make the
past conform to a redefined present. It advocates a return to the roots, con-
demns Westernization and its adverse moral effects on the young, praises
the family as the key institution of society, and seeks to promote the vision
of India as a hindu rashtra—a Hindu nation. While the late Rajiv Gandhi al-
lowed himself to be photographed wearing a Lacoste shirt and khaki shorts,
BJP leaders always wear traditional Indian clothes. The main enemy image
is nevertheless not the West but Islam, which is depicted as a martial and
cruel religion alien to the subcontinent, and Indian Muslims (the descen-
dants of converts, like Bosnian Muslims) are represented partly as traitors to
Hinduism, partly as foreign invaders. The demolition of a mosque in the
northern town of Ayodhya in December 1992, the ensuing riots in several
Indian cities, and the call for the rebuilding of a Hindu temple allegedly de-
54
The Contribution of Identities to Intergroup Conflict


stroyed by a Mughal ruler four centuries ago marked a climax of sorts in
this respect (see van der Veer, 1994, for details).
The phenomenal rise of this traditionalist movement is a result of several
connected processes of sociocultural change or modernization. First, the
very notion of hindutva, Hindu-ness, is a modern one. Hinduism is not a
“religion of the Book.” It is an noncentralized religion with scores of holy
scriptures, thousands of avatars (incarnations of divinities), and very many
ways of worshiping them. The idea of the Hindu identity as an imagined
community based on cultural similarity is alien to Hinduism as such, which
is a religion based on complementarity, difference, and hierarchy. Regard-
ing political Hinduism, some Indian commentators actually speak of a
Semitization of Hinduism whereby it takes on structural characteristics from
the great religions of West Asia. 
Second, the hindutva movement is explicitly modeled on European na-
tionalism—some early hindutva ideologists were even warm admirers of
Hitler—which has been, for 150 years, an attempt to reconcile change and
continuity by talking of roots and traditions in a situation of industrializa-
tion and urbanization. This is obvious in hindutva practice, whereby issues
regarding national anthems, dress, and foreign foods are given prominence,
while profound social changes continue to affect everyday life as before.
There is a clear connection between the rise of the BJP and the liberaliza-
tion of the Indian economy, the rise of a substantial new middle class with 
a strong consumerist orientation and the rapid spread of new mass media
including the Rupert Murdoch-controlled Star TV network. While liber-
alization of this kind stimulates consumerism (perceived as Westernization),
it also indirectly boosts traditionalism since the new patterns of consump-
tion and the new media scene may indicate that cherished traditions are 
under threat. 
Third, the “contagious” influence from political Islam is obvious; hin-
dutva is the assertion of Hindu identity as opposed to Muslim identity both in
Pakistan and in India itself. Doubly ironic, hindutva has double origins in
European romanticism and West Asian political Islam. When its first ide-
ologist, Dr. Veer Savarkar, wrote in the 1920s that “Hindutva is not the
same thing as Hinduism,” he was therefore right, but not for the reasons he
believed. Savarkar saw hindutva as a wide-ranging social movement emanat-
ing from Hindu faith and practices, while a more historically correct ac-
count sees it as the result of cultural diffusion from Europe and West Asia.
Fourth, and perhaps most important, the hindutva movement can be
seen as a reaction against a growing egalitarianism in Indian society. Al-
ready in the 1950s, policies attempting to improve the conditions of the
“Untouchables,” the lowest castes, were introduced, and during the 1990s,
very radical measures have been proposed to this effect—and in some cases
carried out. About half of India’s population are now defined as being either
Dalits (“Untouchables”), tribals, low-caste people, or “OBCs” (Other Back-
Ethnic Identity, National Identity, and Intergroup Conflict
55


ward Classes), and in theory, all of 49.5 percent of jobs in the public sectors
should be reserved for these groups, following the recommendations of the
government-appointed Mandal commission. Since the early 1990s, this
principle has been enforced in many areas. Naturally, many members of the
“twice-born,” upper castes feel their inherited privileges eroding away, and
hindutva is largely a movement representing the interests of the disen-
chanted upper castes. It is for the most part a reaction against the move-
ment toward greater equality in Indian society. Although hindutva seems to
promote equality among Hindus, an implication of its traditionalist Hin-
duism is the reinvigoration of the caste system, which in effect benefits only
the “twice-born” castes. 
This analysis of hindutva must by necessity be a superficial one (see
Hansen, 1999, for a full treatment). It must be remarked, however, that in-
terreligious marriages (and, indeed, intercaste marriages) are rare outside
certain elite groups. Casual interaction between Hindus and Muslims is far
from unusual, but as in the Bosnian and Fijian cases, the intimate (family)
spheres, as well as personal networks of close friends, rarely cross religious
boundaries. Social classification in India is nonetheless complex, and as will
be indicated later, the Hindu-Muslim divide is only one of several possible
social dichotomies—unlike in Bosnia and Fiji, where religious or ethnic
contrasts tend to be paramount.
Some Comparisons
Previous sections have outlined three contemporary conflicts involving col-
lective identity as a political resource. The differences are obvious; the 
focus here will therefore be on the similarities.
The conflicts have three important sociological features in common.
First, there is in all three cases competition over scarce resources. As Horowitz
(1985) and many others writing about group conflict in contemporary soci-
eties have shown, such conflicts invariably involve perceptions of scarcity
and struggles to retain or attain hegemony or equality. Successful mobiliza-
tion on the basis of collective identities presupposes a widespread belief that
resources are unequally distributed along group lines. “Resources” should
be interpreted in the widest sense possible, and could in principle be taken
to mean economic wealth or political power, recognition, or symbolic
power—although what is usually at stake is either economic or political re-
sources. This feature is easy to identify in all three examples described
above: Fijians and Indians compete over relative political and economic
power; the constituent groups of Bosnia compete over political power
and/or sovereignty; hindutva is an attempt to defend the political and eco-
nomic interests of “Hindus” in secular India.
Second, modernization actualizes differences and triggers conflict. With the
integration of formerly discrete groups into shared economic and political
56
The Contribution of Identities to Intergroup Conflict


systems, inequalities are made visible, as comparison between the groups
becomes possible. In a certain sense, ethnicity can be described as the
process of making cultural differences comparable, and to that extent, it is a
modern phenomenon. The Fijian example, where the increasing integra-
tion of Fijians into the modern sphere made it apparent that Indians were
doing better economically, illustrates this point. In India, the rise of the
Dalit movement struggling for recognition and equal rights on behalf 
of “Untouchable” groups is an expression of the modern value of equality,
and the counterreaction from the Hindu right is an attempt to stop egali-
tarianism from spreading, as well as reflecting—almost with the accuracy of
a mirror image—symbolic competition with Muslims within and (espe-
cially) outside India. The Bosnian example, admittedly, seems less straight-
forward, as socialist Yugoslavia was in many ways no less modern than its
successor countries (some would indeed argue that at least at the level of
ideology, it was infinitely more modern than them). What is clear, and
which also holds true for other Eastern and Central European countries, 
is that the sudden introduction of liberal political rights and a capitalist
economy around 1990—core characteristics of non-socialist moderniza-
tion—created a new dimension of comparison between individuals and new
arenas of competition. 
Third, the groups are largely self-recruiting. Intermarriage is rare in all three
cases (excepting urban Yugoslavia). Although biological self-reproduction is
by no means necessary for a strong collective identity to come about, it
should be kept in mind that kinship remains an important organizing princi-
ple for most societies in the world, and a lot of what passes for ethnicity at the
local level is really kinship. Kinship has an important social dimension in ad-
dition to its symbolic side, which is highlighted in ideologies of fictive or
metaphoric kinship. Symbolic boundaries are never effective unless under-
pinned by social organization. 
Further, there are several important ideological similarities. First, at the
level of ideology, cultural similarity overrules social equality. Ethnic national-
ism in Yugoslavia, political Hinduism in India, and the “sons-of-the-soil”
rhetoric of Fiji all depict the ingroup as homogeneous, as people “of the
same kind.” Internal differences are undercommunicated, and moreover,
in the wider political context, equality values are discarded for ostensible
cultural reasons. (Although it could be argued that hindutva is a Trojan
horse concealing upper-caste interests with all-Hindu rhetoric, the point 
is that it stresses the commonalities of all Hindus irrespective of caste or
language.)
Second, images of past suffering and injustice are invoked. Serbs bemoan the
defeat at the hands of the Turks in Kosovo in 1389; Hindu leaders have
taken great pains to depict Mughal (Muslim) rule in India from the 1500s as
bloody and authoritarian; and indigenous Fijian leaders compare their
plight to that of other indigenous peoples who have suffered foreign inva-
sions. Violence targeting the descendants of the invaders can therefore be
Ethnic Identity, National Identity, and Intergroup Conflict
57


framed as legitimate revenge. Even hindutva leaders, who claim to represent
80 percent of India’s population, complain that Hinduism is under siege and
needs to defend itself with all means available. 
Third, the political symbolism and rhetoric evokes personal experiences. This is
perhaps the most important ideological feature of identity politics in gen-
eral. Using myths, cultural symbols, and kinship terminology in addressing
their supporters, promoters of identity politics try to downplay the differ-
ence between personal experiences and group history. In this way, it be-
comes perfectly sensible for a Serb to talk about the legendary battle of
Kosovo in the first person (“We lost in 1389”), and the logic of revenge is
extended to include metaphorical kin, in many cases millions of people.
The intimate experiences associated with locality and family are thereby
projected onto a national screen. This general feature of social integration
has been noted by Handelman (1990), analyzing national rituals, and much
earlier in Turner’s (1967) studies of ritual among the Ndembu of Zambia. In
showing that rituals have both an instrumental and an emotional (or sen-
sory) dimension—one socially integrating, the other metaphorical and per-
sonally meaningful—Turner actually made a point crucial to the present
analysis—namely, that loyalty to a larger collectivity (such as a tribe or a na-
tion) is contingent on its imagery being personally meaningful. 
Fourth, first-comers are contrasted with invaders. Although this ideological
feature is by no means universal in identity politics, it tends to be invoked
whenever possible, and in the process, historical facts are frequently
stretched. In Fiji, the Fijian population—although genetically a Polynesian-
Melanesian mix—has a strong case here, although it is less obvious that
Indo-Fijians can be immigrants to a country in which they were born, and
therefore legitimately deprived of equal rights. Regarding Bosnia and India,
as mentioned above, there is nothing to suggest that the ancestors of Mus-
lims in the respective countries were more recent arrivals than the ancestors
of Christians or Hindus, although Islam is a relatively recent import. What
is interesting here is how the varying depth of cultural genealogies (“roots”)
is used to justify differential treatment. The historical location of the self
along the dimensions of descent and place is thereby invested with political
significance.
Fifth and finally, the actual social complexity in society is reduced to a set of
simple contrasts. As Adolf Hitler wrote in Mein Kampf, the truly national
leader concentrates the attention of his people on one enemy at the time.
Since cross-cutting ties reduce the chances of violent conflict, the collective
identity must be based on relatively unambiguous criteria (such as place, 
religion, mother-tongue, kinship). Again, internal differences are under-
communicated in the act of delineating boundaries toward the demon-
ized Other. This mechanism is familiar from a wide range of interethnic 
situations, from social classification in Zambian mining towns (Epstein,
1992) to Norwegian–Sami relations in sub-Arctic Scandinavia (Eidheim,
1971), the Sinhalese–T
amil conflict (Kapferer, 1988) and Quebecois nation-
58
The Contribution of Identities to Intergroup Conflict


alism (Handler, 1988): the Other is reduced to a minimal set of “traits,” and
so is the collective Self.
These similarities do not necessarily indicate that there are universal
mechanisms linking personal selves and larger collectivities, but they do
suggest that there is a universal “grammar” common to contemporary iden-
tity politics everywhere. In the final sections of this chapter this argument
will be pursued slightly further, and it will be suggested that universal con-
nections between the self and the collective exist, which must be under-
stood not only to account for traditional societies but to make sense of the
present. Far from being an “atavistic” or “primitive” counterreaction to
globalization or modernization, identity politics is a special case of some-
thing more general—namely, collective identity anchored in personal 
experiences.
Where Is the Identity of Identity Politics?
Social scientists have proposed many typologies of ethnic conflict, dividing
the groups involved into categories such as majority, minority, irridentist,
and separatist, using variables such as division of labor, relative political
power, and historical intergroup relations as criteria of classification. In my
view, this kind of exercise can at best generate a limited understanding of
the dynamics of group conflict. T
o begin with, the very adjective “ethnic” is
hardly appropriate to describe all conflicts based on identity politics. Indian
Hindus are not an ethnic group in any meaningful sense, and it is a matter
of definition whether Serbs, Croats, and Muslims in Bosnia should be con-
sidered ethnic groups (they have shared origins only a few centuries back).
Many contemporary conflicts displaying some or all of the features listed
above cannot be seen as ethnic. I note a few African examples: The Su-
danese civil war is partly fought over religion (northeners are Muslims try-
ing to Islamicize the south), partly over culture and language—neither
northern nor southern Sudanese are ethnic groups. Hutus and Tutsis in
Rwanda and Burundi, like the constituent groups of Bosnia, are culturally
very close; they speak the same language and have the same religion. The
Somali civil war presents an even more puzzling case, as Somalia is one of
the few sub-Saharan states that are truly ethnically homogeneous and so far
the only one that seems to have relinquished the trappings of statehood
completely, having dissolved into warring clans (an intermediate level of so-
cial organization, between the family and the ethnic group) since the early
1990s. T
o the northwest of Somalia, one of the great forgotten wars of
Africa is being fought over a contested border area between Eritrea and
Ethiopia. Eritrea, which seceded from Ethiopia in 1991, has never been
based on religion or ethnic identity, but has a vague legitimacy as a histori-
cal nation in the brief period of Italian colonialism before the Second
World War. The current war is being fought between Tigrinya speakers on
Ethnic Identity, National Identity, and Intergroup Conflict
59


both sides of the borders, who are united through religion, language, cus-
toms, history, and even kinship ties, but they are no less bitterly divided po-
litically. This conflict in turn creates its strange bedfellows in the alliance
between Ethiopian Tigrinyas and Amhara speakers from the highlands,
who are traditional enemies.
In other words, the concept “ethnic conflict” is misleading, whether it is
used to classify phenomena or to explain hostilities. Several of the alterna-
tive terms one might consider are, however, no less misleading: “Cultural
conflict” will clearly not do, as it is obviously not what is usually thought of
as cultural differences that lie at the heart of the conflicts. At the village
level, even Hindus and Muslims in India hold many of the same beliefs and
worship in similar ways. The low-intensity conflict in Fiji involves groups
that are by any criterion more culturally different than, say, the Bosnian
groups. All the conflicts considered here are over resources perceived as
scarce: territory, political power, economic gain, employment, recogni-
tion—rights in a wide sense. What they have in common is their successful
appeal to collective identities perceived locally as imperative and primor-
dial, identities associated with a deep moral commitment, whether ethnic
(based on notions of kinship and descent), regional (based on place), or reli-
gious (based on beliefs and forms of worship). For these reasons, the term
“identity politics” is preferable as a generic term for all such political move-
ments, whether nonviolent or violent.
This final section will therefore amount to an attempt to unravel the
identity of identity politics. What is it that makes it so powerful? What is
the “identity” that such political movements can draw upon? 
Benedict Anderson proposes an answer in the introduction to his seminal
Imagined Communities (1983), where he points out that nationalism has
more in common with phenomena such as religion and kinship than with
ideologies like liberalism and socialism. He argues that nationalism (and,
one might add, any form of identity politics) expropriates personal identity,
transforming intimate experiences into the raw material of politics. I owe
my existence to my parents, and by metonymical extension they represent
the larger, abstract collective. I harbor tender feelings for my childhood,
which by extension becomes my group’s glorious and tragic history. I feel
attached to the place where I grew up, which was not just any arbitrary
place but the nation (or, as the case might be, the sacred land of Hinduism,
the traditional territory of the Fijians, the tormented country of the brave,
but sadly misunderstood Serbs). Indeed, this argument can profitably be
seen as echoing Turner’s aforementioned argument on the instrumental and
emotional dimensions (or “poles”) of ritual. In both cases, the integrative
strength of the imagined community (be it a tribe or a nation) depends on
its ability to mobilize emotions proper to the intimate sphere of kinship and
personal experience. 
The conditions for this transformation to take place—the move from an
interpersonally anchored identity to an abstract national, ethnic, or reli-
60
The Contribution of Identities to Intergroup Conflict


gious identity—are usually tantamount to certain general conditions of
modernity (cf. Gellner, 1997). It is through school and mass media that
people are taught to identify with an abstract, mythically rooted community
of people “of the same kind.” Through the replacement of traditional
economies with an abstract labor market, they become participants in a
large-scale system of subsistence. Through the implementation of a bureau-
cratic system of political management, their allegiances are at least partly
moved from the concrete to the abstract community. 
It is important to remember, as theorists of nationalism and ethnicity
have pointed out time and again, that identification is relational, situational,
and flexible, and that each person carries a number of potential identities,
only a few of which become socially significant, making a difference in
everyday life. Even fewer gain political importance, forming the basis of
power struggles and group competition. This is not, however, to say that
collective identities can be created out of thin air. They have to be con-
nected, in credible ways, to people’s personal experiences. These experi-
ences in turn are flexible—not only historians but everybody else as well 
selects and interprets events to make a particular kind of sense of the past—
but not indefinitely so. Regarding our main examples, in Fiji virtually no-
body doubts whether he or she is Indian or Fijian, and politics—whether
based on compromise or conflict—will have to take this into account for the
foreseeable future. In Yugoslavia after the breakup, cross-cutting ties and
cultural hybridity were undercommunicated. Cosmopolitanism was in-
creasingly seen as a suspect, unpatriotic attitude, and people of mixed an-
cestry were forced to choose a bounded, unambiguous identity: they had 
to select past experiences that made them either Serbian, Croatian, or 
Muslim—more or less like the proverbial North African mule, who speaks
incessantly about his uncle, the horse, but never mentions his father, the
donkey. In India, finally, some of the strongest scholarly arguments against
the lasting influence of hindutva have actually pointed toward people’s per-
sonal experiences (Frøystad, 1999). Since Indian everyday life is still perme-
ated by caste distinctions, and caste continues to define the very fabric of
social integration, these scholars argue that hindutva—the idea that all Hin-
dus have something profound in common—is so counterintuitive to most
Indians that it can unite Hindus only as long as the enemy image of Mus-
lims can be kept ablaze. 
All the basic components of political identity familiar from classic politi-
cal anthropology can be identified in contemporary identity politics: it is
based on a sometimes ambiguous mix of kinship and locality; it has well de-
veloped myths of origin and myths of past suffering; and it distinguishes
clearly between “us” and “them.” The main difference between, say, the no-
madic Nuer society studied by E. E. Evans-Pritchard (1940) in the 1930s
and Serbian (or Croatian) nationalism today is probably that of scale: while
the Nuer rarely imagined themselves as members of larger groups than the
clan, a Serb in Vojvodina can readily identify him or herself with a Serb in
Ethnic Identity, National Identity, and Intergroup Conflict
61


Kosovo. The act of transformation from personal, concrete social experi-
ences to the abstract community is naturally much more demanding in a
large-scale society than in a village-based one, hence the importance of
modern institutions of communication, economic transactions, and political
rule for the growth of abstract communities.
Some Final Lessons from Political Anthropology
Having long ago abandoned the early ambition of becoming “a natural sci-
ence of society,” social and cultural anthropology has for decades been re-
luctant to formulate lawlike propositions about the functioning of society.
The constructivist turn of recent years seems to confirm that contemporary
anthropology is less concerned with absolute truths than with the analysis
of local cultural constructions. This need not be so, and the study of current
identity politics may illustrate the power of comparative anthropology in
generating general hypotheses.
Early instrumentalist research on ethnic groups, particularly in Africa (as
in A. Cohen, 1974), searched for the logic of group cohesion, which they
assumed to be roughly the same everywhere. The related, actor-based per-
spective developed by Barth (1969) and his colleagues assumed the logic of
action to be quite universal—people act to maximize benefits. Later analy-
ses of the constructedness of ethnic and national ideologies (the seminal
text is Hobsbawm and Ranger, 1983; see Chapman, Malcolm, McDonald,
& T
onkin, 1989, for an overview) also emphasize universal characteristics of
a particular kind of societal formation (the modern state) and its relation-
ship to group identities based on notions of culture. The canonical texts on
nationalism (such as Gellner, 1983; Anderson, 1983; Smith, 1991) also have
clear universalist ambitions. As this chapter has made clear, these ap-
proaches have obvious strengths, but they need to be supplemented by de-
tailed research on the experiential world of the everyday—the Lebenswelt
(life-world) of the actors. A renewed focus on the informal, intimate, and
often noninstrumental dimensions of everyday life reveals that terms such
as “ethnicity” by themselves explain little. The parallels between a sup-
porter of the BJP and a supporter of Serbian supremacists should not be lo-
cated to their respective “ethnic” identities or “civilizational” membership,
but to the fact that their everyday life, social networks, and personal obliga-
tions connect them to particular groups that may be exploited politically,
given the right circumstances. It should also be kept in mind that class poli-
tics can sometimes be a form of identity politics (Shore, 1993), which can
profitably be understood along the same lines as ethnic or religious identity
politics. The cause of group allegiance lies in the everyday, not in the over-
arching ideology.
In order to complete this analysis, it is necessary to go a few decades
back, to classic political anthropology, in order to see how the perspectives
62
The Contribution of Identities to Intergroup Conflict


developed earlier in the chapter can be enriched by the work of previous
generations. 
The integration of persons into groups can be described as the work of
an inverted refrigerator: the function of a refrigerator is to generate inward
coldness, but in order to do so it more or less inadvertently, as a side effect,
creates outward warmth. Conversely, groups form to create warmth for
their members, but they necessarily create some outward coldness in order
to be able to do so. Under particular circumstances, the outward coldness is
more readily perceptible than the inward warmth. A sociological principle
originally formulated by Georg Simmel, known as “Simmel’s Rule,” simply
states that the internal cohesion of a group is contingent on the strength of exter-
nal pressure. This principle may explain why group integration generally is
so much stronger in small groups, especially if they are oppressed, than in
large ones—why, for example, Scots seem to have fewer difficulties defining
who they are than do the English. 
An interesting corollary of Simmel’s Rule is the fact that what kind of
group emerges depends on where the perceived pressure comes from. Both
gender-based and class-based social movements have periodically been suc-
cessful, given that the perceived threat was seen, not as alien religion or for-
eign ethnic groups, but as male supremacy and ruling classes, respectively.
In accordance with this, some inhabitants of Sarajevo during the war felt
that the conflict was really an urban–rural one, since city-dwellers had a lot
in common, irrespective of religion, that they did not share with rural peo-
ple. Strong opposition groups in India, similarly, argue against a view of 
Indian politics as divided between Hindu communalists and liberals because
they see the main problem of Indian society as one of poverty and distribu-
tion of resources, to which neither of the parties seems to give priority. In
Fiji, finally, the immediate reason for the 1987 coup was the establishment
of a government of national unity promising to address issues of social wel-
fare and economic development rather than intergroup issues. In other
words, redefinitions of societal cleavages are entirely possible insofar as they
do not contradict people’s everyday experiences too obviously. 
In the course of this chapter, cross-cutting ties and conflicting loyalties
have already been mentioned as mitigating forces in situations of inter-
group conflict. Phrased within the terminology used here, one might say
that shared experiences across boundaries reduce the risk of conflict. In
Max Gluckman’s reinterpretation of Evans-Pritchard’s Nuer material from
the 1930s (Gluckman, 1956–1982), this point was made forcefully. The
Nuer were organized along kinship lines across villages, but they were also
locally integrated in villages. The women married out of the village and the
lineage, so that everybody had affines (in-laws) in other villages. Further-
more, men were tied to nonrelatives through trade, initiation rituals, and
friendship. All of these factors led to a reduction in the incidence of vio-
lence among the Nuer. In contemporary identity politics, it can easily be
seen—and has been remarked above—how political leaders emphasize in-
Ethnic Identity, National Identity, and Intergroup Conflict
63


ward similarity and outward boundaries in order to reduce the potentially
mitigating impact of cross-cutting ties. They are true to people’s everyday
life, but try to emphasize certain experiences at the expense of others (in-
ward solidarity and similarity, outward conflict and difference).
Another, even more time-honored principle from political anthropology
is the twin notions of fusion and fission in tribal societies. When the sole
organizing principle for a group lacking hierarchies and formal political of-
fice is kinship, there are limits to the group’s growth; at a certain point, it
splits into two. Without such a fission, internal conflicts would soon be-
come overwhelming given the simple social organization of such societies,
and the effects would be destructive. Fusion of discrete groups has also been
studied extensively, but in many acephalous societies it is seasonal (nomadic
groups fuse in the dry season or in winter) and fragile.
A more dynamic view of contraction and expansion of tribal groups 
was developed, especially by Africanists, from the 1940s (Fortes & Evans-
Pritchard, 1940). In studies of feuding and political competition, they
showed how two or several local groups that might be periodically involved
in mutual feuding united temporarily when faced with an external enemy.
This form of organization, described as segmentary by Evans-Pritchard, fol-
lows the proverb often cited in recent years to explain the logic of the So-
mali civil war: “It’s me against my brother, my brother and I against our
cousins, and our cousins, my brother and I against everybody else.” A form
of segmentary logic is apparent in politics nearly everywhere; a distinguish-
ing mark of modern nation-building has nevertheless been its attempt to
channel loyalties away from various subnational levels of identity in order to
monopolize the political loyalty of individual citizens. 
The segmentary logic creates a fluid, relational political organization
that, in its pure form, is impracticable in modern state societies given their
requirements for stability, centralized power, and reified systems of political
representation. This does not, however, mean that segmentary identi-
fication does not continue to exist, and one of the causes of oppositional
identity politics in modern nation-states is their not providing subnational
identity groups appropriate political arenas, thereby encouraging counter-
reactions in the form of identity politics directed against the state. 
The formation of identity-based political groups generally entails both
an expansion and a contraction of the focus for identification. At the time of
the breakout of conflict in Bosnia, the federal or even state level was in-
creasingly seen as irrelevant—the process was one of fission. At the same
time, internal conflicts and schisms within each constituent group were
minimized, and as a result each group became more coherent and united
than before. In the cases of Fiji and India, this is even more obvious: among
Fijians, rivalry between chiefs and clans has diminished in importance as 
Fijian politics has grown increasingly ethnic; similarly, rifts within the In-
dian population on the basis of regional origin, which could formerly lead
to Indian subgroups supporting Fijian-dominated governments, have be-
64
The Contribution of Identities to Intergroup Conflict


come much less important since the military coup. In the Indian case, the
very idea of hindutva implies an enormous expansion of the ingroup for
Hindus. Trying to bridge differences based on language, caste, region, and
culture, hindutva tries to create a morally committing all-Indian Hindu
identity based on symbolic equality. This is, in the Indian context, a very
radical move. Simultaneously, Indian citizenship and (overarching, supra-
religious) national identity become less important since the federal Indian
state includes millions of non-Hindus, who are depicted as internal enemies
by hindutva spokespersons.
In all the examples considered here, group segmentation at a higher
level, and the ensuing formation of imagined communities larger than the
locality but smaller than the state, is immediately related to the need for a
firm boundary in a situation of conflict within the state. The alternative
identity of national citizenship, which encompasses the other as well, no
longer functions. Interestingly, contemporary identity politics is very simi-
lar to nationalism—for example, in its appeal to mythical foundations, its
abstract postulation of similarity and equality, its rejection of segmentary
identity formation, and its attempt to reduce a world of many small differ-
ences to a world of only a few, major ones. In many cases, it is more success-
ful than nationalism, particularly in postcolonial, multiethnic states. This 
is not only because identity politicians promise its adherents that they will
win zero-sum games against political competitors, but also because they 
are able to represent themselves as natural extensions of people’s personal,
experience-based identities. 
A challenge for modern states, thus, consists in coping with the fact that
personal identity can be exploited politically not only by the state itself but
by others as well, not because the self is infinitely multifaceted (it is not),
but because the experiences and relationships that make up the self can be
expanded symbolically in several, often conflicting ways. Processes of seg-
mentary fusion and fission, the formation of different kinds of groups in-
volving overlapping personnel owing to the functioning of Simmel’s Rule,
group-based antagonism, and competition: these ways of expressing politi-
cal interests, underpinned by shared meaning within the ingroup, are no
more eradicated by modernity than is personal identity. The challenge thus
consists in laying the foundations for “a sense of belonging to a community
larger than each of the particular groups in question” (Laclau, 1995, p. 105),
and this can be done only by first acknowledging both the richness and the
variability of personal identities.
In the face of violent identity politics, “ethnic cleansing,” and the strong
attention to “roots” and historically based identities (ungenerously de-
scribed as “the narcissism of small differences” by Michael Ignatieff) char-
acterizing many societies in recent years, it is not surprising that intellectu-
als have recently tried to think essentialism away, emphasising the endlessly
flexible and fluid character of human identification. A typical expression of
this position can be found in a recent text by the influential sociologist 
Ethnic Identity, National Identity, and Intergroup Conflict
65


Zygmunt Bauman, where he states: “If the modern ‘problem of identity’ is
how to construct an identity and keep it solid and stable, the postmodern
‘problem of identity’ is primarily how to avoid fixation and keep the options
open” (Bauman, 1996, p. 18).
In today’s world it may be ethically imperative to endorse Bauman’s posi-
tion, but it is equally important to keep in mind that humans are not free-
floating signifiers, and no amount of benevolent intentions will be able to
change people’s life-worlds overnight. Rather than trying to think them
away, it is necessary to understand them and come to terms with their 
enduring power. Notwithstanding globalization and the universalization 
of modernity, cultural differences continue to exist, within and between
places, within and between nations and ethnic groups. It is also, however,
doubtlessly true that carbon can be turned into graphite as well as dia-
monds, and the ways in which cultural differences become socially relevant
vary importantly. But to pretend they do not exist outside ethnic and na-
tionalist ideologies would be intellectually indefensible; people’s personal
experiences are the very raw material of such ideologies. Here lies an im-
portant limitation in constructivist models of identity. Collective identities
are constructed, consciously or not, but nothing comes out of nothing. In
locating the universal not in the workings of identity politics (it changes
historically and varies geographically), nor in the eternal sovereignty of 
the state (the same objection applies), but rather in the social life-worlds in
that individuals make sense of the world, we may have found a basis for
comparison that will outlive academic fads and contemporary politics. 
Acknowledgments
The author would like to express his thanks to the editors and three anonymous
readers for their critical readings and useful comments on an earlier version of the
chapter.
References
Anderson, B. (1983). Imagined communities: Reflections on the origin and spread of na-
tionalism. London: Verso.
Barth, F. (Ed). (1969). Ethnic groups and boundaries. Oslo: Scandinavian University
Press.
Bauman, Z. (1993). Postmodern ethics. Cambridge: Polity.
Bauman, Z. (1996) From pilgrim to tourist; or a short history of identity. In S. Hall
& P. Du Gay (Eds)., Questions of cultural identity (pp. 18–36). London: Sage.
Bringa, T. (1996). Being Muslim the Bosnian way: Identity and community in a central
Bosnian village. Princeton, NJ: Princeton University Press.
Chapman, M., McDonald, M., & T
onkin, E. (1989). Introduction—history and so-
cial anthropology. In E. T
onkin, M. McDonald, & M. Chapman (Eds.), History
and ethnicity (pp. 1–21). London: Routledge.
66
The Contribution of Identities to Intergroup Conflict


Cohen, A. (1969). Custom and politics in urban Africa: A study of Hausa migrants in a
Yoruba town. London: Routledge & Kegan Paul.
Cohen, A. (1974). Two-dimensional man: An essay on power and symbolism in complex 
society. London: Routledge & Kegan Paul.
Cohen, A. P. (1994). Self consciousness: An alternative anthropology of identity. London:
Routledge.
Cornell, S., & Hartmann, D. (1998). Ethnicity and race: Making identities in a chang-
ing world. London: Pine Forge Press.
Eidheim, H. (1971). Aspects of the Lappish minority situation. Oslo: Scandinavian Uni-
versity Press. 
Epstein, A. L. (1992). Scenes from African urban life: Collected Copperbelt essays. Edin-
burgh: Edinburgh University Press.
Eriksen, T. H. (1992). Us and them in modern societies: Ethnicity and nationalism in
Trinidad, Mauritius and beyond. Oslo: Scandinavian University Press.
Eriksen, T. H. (1998a). Common denominators: Ethnicity, nationalism and compromise in
Mauritius. Oxford: Berg.
Eriksen, T. H. (1998b). Culture and ethnicity: A second look. Lecture at 6th SIEF
conference, Amsterdam, 21 April 1998.
Evans-Pritchard, E. E. (1940). The Nuer. Oxford: Clarendon.
Featherstone, M. (Ed). (1990). Global culture. London: Sage.
Fortes, M., & Evans-Pritchard, E. E. (Eds.). (1940). African political systems. London:
Oxford University Press.
Friedman, J. (1994). Global culture and local process. London: Sage.
Frøystad, K. (1999). Personal communication.
Furnivall, J. S. (1948). Colonial policy and practice: A comparative study of Burma and
Netherlands India. New York: New York University Press.
Geertz, C. (Ed.). (1967). Old societies and new states: The quest for modernity in Africa
and Asia. New York: The Free Press.
Geertz, C. (1973). The interpretation of cultures. New York: Basic Books.
Gellner, E. (1983). Nations and nationalism. Oxford: Blackwell.
Gellner, E. (1997). Nationalism. London: Weidenfeld & Nicolson.
Giddens, A. (1991). Modernity and self-identity. Cambridge: Polity.
Giddens, A. (1994). Beyond left and right. Cambridge: Polity.
Gluckman, M. (1956/1982). Custom and conflict in Africa. Oxford: Blackwell.
Handelman, D. (1990). Models and mirrors: Towards an anthropology of public events.
Cambridge: Cambridge University Press.
Handler, R. (1988). Nationalism and the politics of culture in Quebec. Madison: Wiscon-
sin University Press.
Hannerz, U. (1992). Cultural complexity: Studies in the social organization of meaning.
New York: Columbia University Press.
Hansen, T. B. (1999). The saffron wave: Democracy and Hindu nationalism in modern
India. Princeton, NJ: Princeton University Press.
Hobsbawm, E., & Ranger, T. (Eds.) (1983). The invention of tradition. Cambridge:
Cambridge University Press.
Holland, D. (1997). Selves as cultured: As told by an anthropologist who lacks 
a soul. In R. Ashmore & L. Jussim (Eds.), Self and identity: Fundamental issues
(pp. 160–190). Oxford: Oxford University Press.
Horowitz, D. L. (1985). Ethnic groups in conflict. Berkeley: University of California
Press.
Ethnic Identity, National Identity, and Intergroup Conflict
67


Huntington, S. (1996). The clash of civilizations and the remaking of a world order. New
York: Simon and Schuster.
Ignatieff, M. (1994). Blood and belonging: Journeys into the new nationalism. London:
Vintage.
Jenkins, R. (1996). Social identity. London: Routledge.
Kapferer, B. (1988). Legends of people, myths of state. Baltimore, MD: Smithsonian 
Institution Press.
Kelly, J. D. (1998). Aspiring to minority and other tactics against violence. In D. C.
Gladney (Ed.), Making majorities: Constituting the nation in Japan, Korea, China,
Malaysia, Fiji, Turkey and the United States (pp. 173–197). Stanford, CA: Stanford
University Press.
Laclau, E. (1995). Universalism, particularism and the question of identity. In 
J. Ratchmann (Ed.), The identity in question. London: Routledge.
Lash, S., & Urry, J. (1994). Economies of signs and space. London: Sage.
Premdas, R. (1993). Balance and ethnic conflict in Fiji. In J. McGarry & B. O’Leary
(Eds.), The politics of ethnic conflict regulation (pp. 251–274). London: Routledge.
Rex, J. (1997). The nature of ethnicity in the project of migration. In M. Guibernau
& J. Rex (Eds.), The ethnicity reader: Nationalism, multiculturalism and migration.
Cambridge: Polity.
Roosens, E. E. (1989). Creating ethnicity. London: Sage.
Said, E. W. (1978). Orientalism. New York: Pantheon.
Schlesinger, A. M. Jr. (1992). The disuniting of America. New York: Norton.
Schöpflin, G. (1993). The rise and fall of Yugoslavia. In J. McGarry & B. O’Leary
(Eds.), The politics of ethnic conflict regulation (pp. 172–203). London: Routledge.
Shore, C. (1993). Ethnicity as revolutionary strategy: Communist identity construc-
tion in Italy. In S. Macdonald (Ed.), Inside European identities (pp. 27–53). Oxford:
Berg.
Smith, A. D. (1986). The ethnic origins of nations. Oxford: Blackwell.
Smith, A. (1991). National identity. Harmondsworth: Penguin.
Strathern, M. (1991). Partial connections. Savage, MD: Rowman and Littlefield.
T
ambiah, S. J. (1989). The politics of ethnicity. American Ethnologist, 16, 335–349.
Tinker, H. (1974). A new system of slavery: The export of Indian labour overseas 1830–
1920. Oxford: Oxford University Press.
Turner, V. (1967). The forest of symbols: Aspects of Ndembu ritual. Ithaca, NY: Cornell
University Press.
Turton, D. (Ed.) (1997). War and ethnicity: Global connections and local violence. Wood-
bridge, Suffolk: University of Rochester Press.
van der Veer, P. (1994). Religious nationalism: Hindus and Muslims in India. Berkeley:
University of California Press.
Williksen-Bakker, S. (1991). Fijians in business: A study in the transformation of
symbols. Doctoral dissertation, University of Oslo.
68
The Contribution of Identities to Intergroup Conflict


PART II
THE CONTRIBUTION 
OF ETHNIC AND 
NATIONAL IDENTITIES 
TO POLITICAL CONFLICT 
IN THE UNITED STATES


This page intentionally left blank 


Jack Citrin
4
Cara Wong
Brian Duff
The Meaning of American 
National Identity
Patterns of Ethnic Conflict and Consensus
National identity is one among many often co-existing and overlapping so-
cial identities, including territorial, racial, religious, linguistic, and gender
identities. The psychological task for individuals is to order and integrate
their different collective “selves” (Ashmore & Jussim, 1997). The political
task for government is to balance the need for national unity with the com-
peting claims of other group identities. In modern society, emphasizing loy-
alty to the “nation” and making this the essence of the individual’s political
self-definition emerged as the dominant way to boost social solidarity
(Greenfield & Chirot, 1994). As an immigrant nation, the United States has
always faced the problem of coping with ethnic diversity; the motto e
pluribus unum expresses the desire for a strong sense of common American
identity without indicating the proper balance between the national “one”
and the ethnic “many.”
The contemporary meaning of American national identity is the focus of
this chapter. Immigration and differences in fertility rates have changed the
ethnic and religious composition of the United States (Farley, 1996;
Warner, 1993). Compared to 50 years ago, the country is more diverse, with
many more people today from Latin America, Asia, and Africa. Moreover,
this demographic change has occurred in the context of technological trans-
formations eroding national sovereignty in economic life and of the emer-
gence of intellectual trends legitimating the primacy of ethnicity in political
life.
Scholars disagree about whether globalization and multiculturalism have
weakened Americans’ sense of attachment to the “nation,” particularly
among minority groups (Reich, 1992; Barber, 1995; Hollinger, 1997;
71


Wolfe, 1998; de la Garza, Falcon, & Garcia 1996). There also are conflict-
ing views about the implications of a stronger sense of ethnic, as opposed to
national, identity. Some scholars (Raz, 1994; Young, 1990) regard ethnic
pride as a source of individual self-esteem among cultural minorities that
contributes to the achievement of their group’s goals. Others (Schlesinger,
1992; Miller, 1995) claim that the emphasis on ethnic distinctiveness (in-
group favoritism) inevitably weakens common bonds and intensifies group
conflict (outgroup hostility), raising the specter of cultural and political
Balkanization. 
This chapter employs survey research to analyze the meaning and politi-
cal consequences of American national identity in the context of the demo-
graphic, economic, and intellectual changes noted above. The conceptu-
alization of national identity formulated here draws on social identity
theory (T
ajfel & Turner, 1986; Thoits & Virshup, 1997) to encompass both
self-categorization (or identification as) and affect (or identification with).
However, it goes beyond merely assessing the individual’s positive (or nega-
tive) feelings about the nation to determine the normative foundations of
their sense of American identity. 
The content of American identity is the set of ideas and sentiments that
form the conceptual framework of nationhood. At the cultural level, this
refers to a particular collective representation of the nation indicating the
subjective criteria for membership in the national community (Citrin, Haas,
Muste, & Reingold, 1994). A given conception of national identity thus in-
cludes normative assumptions about how members of different ethnic and
cultural groups should relate to one another. For example, must minority
groups assimilate to a dominant language, religion, or political outlook in
order to achieve full acceptance as fellow-nationals, or can a nation accom-
modate loyalty to a variety of cultural traditions? Historical experience sug-
gests that there is no single answer to this question, so it is important to
conceive of national identity in a way that recognizes the possibility of sev-
eral conceptions of “Americanness” vying for popular support. 
The next section of the chapter provides a more detailed account of the
tripartite definition of national identity that distinguishes among the di-
mensions of self-categorization, affect, and normative content. The second
section summarizes the content of competing ideological conceptions of
American identity, with particular attention to their normative assumptions
regarding the articulation of strong ethnic identities. The analysis of survey
data, presented in the third section, addresses three main questions: 
1. What is the degree of national versus ethnic identification in
American public opinion, as assessed by several different measures
of self-categorization and affective support? 
2. What is the pattern of consensus and cleavage among white, black,
Hispanic, and Asian respondents? Specifically, do minority ethnic
groups have a weaker sense of national identity than whites, and
72
The Contribution of Ethnic and National Identities to Political Conflict


are there ethnic group differences in how people order their na-
tional and ethnic identities? 
3. What is the impact of national and ethnic identities on policy pref-
erences? Drawing on symbolic politics theory (Sears, 1993), the 
final section of the chapter tests hypotheses about the relative in-
fluence of these identifications on preferences regarding foreign
policy, racial policies, and “multicultural” issues such as immigra-
tion and group representation. 
These analyses shed light on the impact of a superordinate “American”
identity on feelings about “outgroups,” both domestic and foreign, and pro-
vide a basis for speculating about the implications of different patterns of
national and ethnic identity for conflict and cooperation among America’s
racial and ethnic groups.
The Dimensions of Social Identity
Identity is a slippery concept. It is an assertion of both sameness and differ-
ence. One answers the question “What is your identity?” by naming who
one is like. This entails drawing boundaries: one is the same as some others
and different from everyone else. For this reason, a social identity both in-
tegrates and divides. 
The first step in measuring a social identity is to determine the basis of
self-categorization; one identifies the characteristic(s) shared by those with
whom one psychologically belongs. We each possess multiple potential 
social identities whose degree of overlap and whose relative significance 
for our self-concept and behavior may vary (Thoits & Virshup, 1997).
Moreover, while self-categorization may be the initial step in the formation
of a psychologically meaningful social identity, the range of one’s options is
finite. Biology, the life cycle, social structure, the economy, and the govern-
ment create the categories that demarcate social identities with potential
political consequences. Thus, some social identities are optional and others
imposed. Furthermore, the borders dividing identity groups are permeable,
but not entirely open. It is easy to be a Bulls fan and then to “exit” psycho-
logically for the Celtics when Michael Jordan retires. It is harder to shed
one’s nationality or ethnicity and almost impossible to change one’s race or
sex. Finally, even when one can adopt a particular social identity by self-
categorization (Turner, Hogg, Oakes, Reicker, & Blackwell, 1987), this
choice frequently must be affirmed by the agreement of others in the group
that you do indeed possess the criteria for membership. The daughter of
Chinese immigrants may call herself an American, yet might be told by 
fellow-citizens of European origin that she does not share their national
identity.
Self-categorization refers to “identification as,” connoting a perceived
The Meaning of American National Identity
73


self-location in a group, and should be distinguished from “identification
with,” which indicates positive affect toward others in the group and must be
measured separately. Social identity theory (T
ajfel, 1978) posits that mere
awareness of belonging to a group engenders positive feelings about the
group and a tendency to act on behalf of other group members, even if they
are unknown to one personally. In the same vein, Miller (1995) has argued
that the moral value of a strong sense of national identity is that it fosters
diffuse feelings of sympathy and obligation toward fellow citizens. 
In addition, there is a difference between membership in a group and psy-
chological attachment to it. One can identify with members of a minority
group, agreeing that they deserve more access to jobs or political power,
without categorizing oneself as a member of that group. This example un-
derscores the potential significance of identities based on shared values
rather than sociological similarities and the need to determine the condi-
tions under which “identification as” and “identification with” are closely
connected. 
“Identification with” entails learning about a group’s defining customs,
expectations, and values, and making them one’s own. These common val-
ues or ideas define the third component of a social (e.g., national) identity—
its content. The specific content of a group identity is socially constructed, in
the sense that consensual decisions shape and reinforce the normative crite-
ria for membership. Since these criteria are vulnerable to challenge and
subject to change, it is important to assess the extent of their legitimacy. 
Still, social identities are not automatically political. They are politicized
when feelings of identification with a group are combined with a belief in
advancing its goals through collective action (Miller, Gurin, Gurin, &
Malanchuk, 1981). During the British Raj, Indians undoubtedly differenti-
ated themselves cognitively from the English. The emergence of an ide-
ology and organized movement demanding change in the cultural, eco-
nomic, and legal status of Indians infused this social identity with political
content. 
The Affective Dimension: Patriotism versus Chauvinism
A nation denotes a group of people seeking or possessing a common home-
land. Nationalism as a doctrine asserts that a group of people sharing char-
acteristics that differentiate them, in their own minds, from others should be
politically autonomous. The affective dimension of national identity (iden-
tification with) refers to feelings of closeness to and pride in one’s country
and its symbols. Many scholars go further, however, and distinguish love of
one’s own country from a sense of superiority to other countries (Schaar,
1981; Staub, 1997; Kosterman & Feshbach, 1989; Sullivan, Fried, & Dietz,
1992; Connor, 1993; T
aylor 1995). They invoke the concept of patriotism
to refer to the positive emotion of love for one’s own people and homeland
74
The Contribution of Ethnic and National Identities to Political Conflict


and use “nationalism” pejoratively, not technically, to designate arrogance
and contempt toward other countries. Writing in this vein, for example,
George Orwell described patriotism as defensive and nationalism as aggres-
sive. For Schaar (1981, p. 285), nationalism is a perversion of patriotism,
“its bloody brother.” Psychologists developing separate measures for these
concepts tend to give the label “nationalism” or “blind patriotism” to name
disdain for other countries and a drive to dominate them (Bar-T
al, 1993;
Kosterman and Feshbach, 1989; Staub, 1997). 
Drawing on social identity theory’s distinction between ingroup fa-
voritism and outgroup hostility clearly is useful when studying national and
ethnic identity. The emotive use of the term “nationalism,” however, is un-
necessarily confusing. “Chauvinism,” not nationalism per se, is the term
used here to refer to an extreme and bounded loyalty, the belief in one’s
country’s superiority, whether it is right or wrong. Whether patriotism, an
ideology of mutual affection among those with a common sense of national
consciousness (Kelman, 1997, p. 170), and chauvinism empirically slide into
each other, such that caring for one’s own typically is associated with hos-
tility toward others, is a controversial issue to be explored below (see also
Brewer, this volume). 
Normative Conceptions of American Identity
History shows that the attributes used to define the idea of a “people” or
“nation” are numerous and malleable (Greenfield, 1992). Historians distin-
guish “civic” nationalism, where criteria for belonging are identical to citi-
zenship and so, in principle at least, open to all, from “ethnic” nationalism,
where membership is based on a sense of shared blood (Connor, 1993).
There is a modern tendency to endorse the principle of ethnicity as the le-
gitimate basis for political autonomy, but there are prominent counterex-
amples of multiethnic states such as the United States (Greenfield, 1992).
As Max Weber put it, nationality ultimately is based on the principle of ter-
ritory, ethnicity on lineage. 
Whatever the specific foundations of a sense of national identity, nation-
alism, in its neutral, technical meaning, implies that membership in the na-
tion is the citizen’s overriding group loyalty, taking precedence over other
available foci of affiliation, including ethnicity, in circumstances where they
conflict. If a strong national identity means consciousness of special bonds
to one’s fellow-nationals, so that they become, at least symbolically, a large
extended family, then constructing a nation may involve psychological
losses as well as gains. The development of a national identity may require
diminishing or abandoning one’s attachment to other groups, including
one’s ethnic group. 
The third dimension of our conception of national identity refers to its
normative content. This means the particular set of ideas about what makes
The Meaning of American National Identity
75


the nation distinctive—ideas about its members, its core values and goals,
the territory it ought to occupy, and its relations to other nations (Citrin et
al., 1994). In a multiethnic society, these ideas lay down the nature of le-
gitimate commitments to subnational communities of descent. In contem-
porary American politics, liberalism, nativism, and multiculturalism are 
alternative political theories about how society should be organized and
how national identity should be defined. Each proposes a different solution
for how properly to balance national and ethnic identities, with conflicting
implications for policymakers. The analysis below therefore assesses the 
degree of popular support for these competing ideas. 
The liberal conception of American identity is “civic,” not “ethnic.”
Commitment to the national “creed” of democracy and individualism
(Huntington, 1981; Gleason, 1980) is what makes one an American. Since
belonging to the nation is equated not with shared blood but with common
beliefs and customs (Gleason, 1980; Lind, 1995), anyone, regardless of an-
cestry, can become American through adherence to the dominant set of
ideals, which include equality of opportunity and respect. The liberal image
of nationality is ethnically inclusive, in principle, if not always in practice
(Smith, 1997). 
When it comes to ethnic diversity, the liberal conception of national
identity is optimistic about the ability of contemporary American society to
assimilate newcomers. This means that ethnic Balkanization is not a serious
threat, so there is no reason to extirpate the traditions of new immigrants as
attempted by the Americanization program in the 1920s. Over time, the
largely Asian and Hispanic newcomers will blend into the cultural main-
stream, just like their European predecessors. Thus, people are free to
honor their ethnic heritage if they wish, without undermining America’s
ability to create one people out of many. Still, according to the liberal con-
ception of national identity, when the claims of nationality and ethnicity
conflict, the former should take precedence. In psychological terms, liberal-
ism rejects the notion of primordial identities, the idea that differences
among communities of descent are fundamental and enduring, exercising an
inherent dominion over the individual’s political conduct. 
Rogers Smith (1997) documents the persistent gap between the liberal
theory of national identity and more nativist political practices. From the
beginning of the United States, the official definition of “Americanness” ex-
cluded blacks and Native Americans and consigned women to a lesser role.
Then, in the 19th and 20th centuries, restrictive immigration and natural-
ization laws discriminated on a racialist basis. Asians were denied citizen-
ship, Mexicans were repatriated, and in 1924, immigration was limited to a
small number of people of Northern European origin. 
The nativist response to ethnic diversity (Higham, 1988) is to insist on
cultural conformity. If there is to be immigration, the nation’s policy should
favor the admission of people who already are familiar with American po-
litical culture (Brimelow, 1995). Once here, newcomers should undergo a
76
The Contribution of Ethnic and National Identities to Political Conflict


program of indoctrination that cleanses them of their traditional loyalties
and imparts knowledge of cultural practices. Nativism thus gives an ethno-
centric cast to American national identity. Here, nationality does not simply
trump one’s ethnicity; rather, nationality fuses with a particular Anglo-
Saxon ethnicity. 
Scholars disagree about the extent and the voluntary nature of cultural
assimilation of immigrants to America (Gerstle, 1997; Hollinger, 1997;
Gordon, 1964; Glazer & Moynihan, 1959). While the passage of the 1964
Civil Rights Act and the 1965 Immigration and Nationality Act marked the
ascendancy of the liberal conception of citizenship, in previous decades na-
tivism and racism strongly influenced public policy. Gerstle (1997, p. 556)
argues that government support for nativism and the end of mass immigra-
tion suppressed the hyphenated American identities that had thrived earlier
and narrowed the range of acceptable cultural behavior. Even so, nativism
never entirely succeeded (Glazer & Moynihan, 1959); every immigrant did
not truly “melt” into a single, unchanging political culture. 
Multiculturalism is an alternative ideological response to the presence of
ethnic diversity in America. The “soft” or “liberal” version of this concep-
tion of national identity emphasizes the virtues of cultural pluralism and
calls on the government to help ensure tolerance toward minorities (Miller,
1995; Appiah, 1994; Raz, 1994). “Hard” or “radical” multiculturalism goes
further, conceiving of the nation as a confederation of ethnic groups with
equal rights and construing ethnicity as the preferred basis of one’s political
identity (Citrin, Sears, Muste, & Wong, forthcoming; Sears, Citrin, Chele-
den, & van Laar, 1999). 
Unlike liberalism, then, multiculturalism assumes that one’s ethnic iden-
tity is essential to one’s personal dignity and self-realization (Raz, 1994).
This conception of national identity does hold that differences among com-
munities of descent are basic and persisting. Since no culture is superior to
any other, none should be privileged in a multiethnic polity (Young, 1990;
Goldberg, 1994). T
o assure that minority cultures survive and flourish—
something that is viewed as indispensable for the well-being of their mem-
bers—“hard” multiculturalism therefore demands group rights and govern-
ment efforts to preserve minority cultures (Okin, 1997). Without such
support, it is argued, the standing of minority ethnic groups inevitably will
decline in the face of the cultural and economic power of the numerical 
majority, and members of minority ethnic groups will lose the core of their
social identity. 
Multiculturalism emphasizes ethnic consciousness and pride, but pro-
vides no apparent basis for social solidarity among America’s ethnic groups
save the principle of mutual tolerance (Raz, 1994). Because of this, critics
(Schlesinger, 1992; Kateb, 1994) warn that elevating the psychological
salience of ethnic identity, as Yugoslavia’s tragic recent history illustrates,
will increase group conflict or make it difficult to mobilize citizens on be-
half of national goals. In the vocabulary of social identity theory, they claim
The Meaning of American National Identity
77


that legitimating ingroup favoritism will lead inexorably to hostility and ag-
gression toward outgroups, including other ethnic groups within the
American national community.
Research Questions and Method
Discussion about the consequences of changing the balance between na-
tional and ethnic identities often proceeds in an empirical vacuum, without
much reference to the prevailing attitudes and beliefs of the American pub-
lic. This chapter employs survey data to present evidence about the follow-
ing questions: What is the level of patriotism in the United States, and does
it differ substantially across ethnic groups? Do white, black, Hispanic, and
Asian Americans differ in how they order their national and ethnic identi-
ties or in how they conceive of membership in the national community?
And does one’s particular conception of national identity influence opinions
about government policies?
Psychological theories regarding intergroup relations differ in their pre-
dictions about the preeminence of ethnicity in how people define themselves
politically (Sidanius, Levin, Rabinowitz, & Frederico, 1999; Sears et al.,
1999). Our own expectations are guided by symbolic politics theory (Sears,
1993; Sears et al., 1999). This approach assumes that individuals possess sta-
ble predispositions established in early socialization (e.g., through parents or
mass media) and reinforced by later experience (e.g., through education or
work). Ethnicity has no inherent priority in identify formation. Ethnic iden-
tification is one potentially significant predisposition, national attachment 
is another. The relative strength of these identities varies according to one’s
socialization experiences, and their influence on current behavior depends
upon the nature of environmental cues. For example, feelings of national
identity are likely to be evoked by international terrorist activities because
these raise the salience of one’s status as an American, but not by proposals to
increase the tax on tobacco.
By emphasizing the role of social learning, the symbolic politics perspec-
tive is compatible with either ethnic antagonism or harmony in multiethnic
societies. In the American context, it appears that national pride and a sense
of the country’s exceptionalism are strongly socialized attitudes, and this 
socialization may mitigate ethnic conflict by integrating diverse groups in an
overarching identity (Horowitz, 1985). All the country’s ethnic groups are
exposed to patriotic norms and the country’s liberal political tradition (Hartz,
1955; Lipset, 1996) that transmits belief in individual rights, not group
rights. The potency of these cultural messages would predict only minor 
ethnic differences in national identity and support for multiculturalism. 
Symbolic politics theory conceives of national and ethnic identities as
structures of valenced beliefs stored in memory (Lau & Sears, 1986). Which
of a person’s political predispositions influences her political choices partly
78
The Contribution of Ethnic and National Identities to Political Conflict


depends on the nature of the stimuli she encounters, such as events, poli-
cies, and candidates’ messages (Zaller, 1992). When a particular attitude
such as national pride is evoked, the theoretically predicted response is to
act in a way that reinforces one’s preexisting orientation. Accordingly, sub-
jective conceptions of national identity are hypothesized to govern opinions
about specific issues, such as foreign policy, immigration, or teaching
American history, which evoke symbols of nationhood, but not about “neu-
tral” issues, such as protecting endangered species.
The survey method relies on self-reports as the basis for classifying social
identities. The main body of evidence here comes from the 1996 General
Social Survey (GSS) of the American public aged 18 and older conducted
by the National Opinion Research Center (Davis & Smith, 1997). This sur-
vey employed a split-sample design and the results are based on the 1,367
respondents of the total sample of 2,094 who were asked questions about
national identity. In part because the 1996 GSS national survey included
very few Hispanic and Asian respondents, comparable evidence from the
1994 and 1995 Los Angeles County Social Surveys (LACSS), conducted by
UCLA’s Institute for Social Science Research (Sears et al., 1999) also is re-
ported. The Los Angeles data have special interest since they delineate the
degree of ethnic consensus and conflict in an extraordinarily diverse local
community where the salience of multiculturalism is a political constant.
The 1994 LACSS sample included 279 whites, 231 blacks, 264 Latinos, and
47 Asian respondents, and the corresponding numbers in the 1995 sample
were 259, 71, 202, and 47.
The 1996 GSS included a variety of items regarding beliefs and feelings
about America. From this set the analyses in this chapter use items covering
the different aspects of patriotism and chauvinism, as well as other ques-
tions assessing subjective beliefs about the normative criteria for American
nationality. Other questions about the respondent’s level of pride in distinct
features of American society and history were used to provide convergent
evidence regarding national attachment. In addition, questions concerning
one’s political self-categorization and one’s level of identification with one’s
ethnic group provided evidence about the balance between national and
ethnic identities.
Public Attachment to America
The present survey evidence shows clearly that most Americans retain a
strong, positive sense of national identity. T
able 4.1 presents the responses
of the 1996 GSS national sample, broken down by ethnicity, to an array of
relevant items, with the relatively few “don’t know” answers excluded. (For
the full wording of the questions, see Davis & Smith, 1997). T
able 4.1
groups these questions to reflect the tripartite definition of national identity
outlined above and to indicate the composition of multi-item attitude
The Meaning of American National Identity
79


table 4.1 Beliefs about American National Identity (1996 GSS)
T
otal
Whites
Blacks
Hispanics
(1367)
(1018)
(156)
(61)
Political Self-Categorization
Think of self mainly as “just American” in
social and political issues*
90%
96%
66%
79%
Affective Attachment to America
Patriotism scale items:**
How important is being an American to you?
(% saying ‘most important thing in their life’)
46
47
49
26
How close do you feel to America? (% saying 
‘close’ or ‘very close’)
81
83
69
83
Agree ‘I would rather be a citizen of America 
than any other country’
91
92
89
79
Pride in Aspects of America
% saying ‘very proud’ or ‘somewhat proud’
The way democracy works
83
85
77
81
Its political influence in the world
80
81
77
75
Its history
88
92
69
81
Its fair and equal treatment of all groups in 
society
57
60
42
51
Chauvinism scale items:
Agree ‘World would be a better place if people 
from other countries were more like Americans’
40
40
47
47
Agree ‘America is a better country than most 
other countries’
81
83
74
73
Agree ‘America should follow its own interests, 
even if this leads to conflicts with other nations’
44
44
50
39
Agree ‘People should support their country even 
when it’s in the wrong’
32
30
41
33
Normative Conceptions of American Identity
% Saying very or fairly important for a True 
American:
Nativism scale items:
T
o have been born in America
69
67
82
42
T
o be a Christian
54
50
77
50
T
o have lived in America for most of one’s life
73
71
84
77
Assimilationism scale items:
T
o be able to speak English
93
94
91
85
T
o feel American
87
89
81
72
T
o have American citizenship
93
92
93
93
Multiculturalism
Agree ‘It is impossible for people who do not 
share American customs and traditions to 
become fully American’
35
34
47
34
Agree ‘Ethnic minorities should be given 
govt assistance to preserve their customs 
and traditions’
17
11
44
45
Think it is better if groups adapt and blend into 
larger society rather than maintain their 
distinct customs and traditions
58
61
56
50
*This question was not asked in the 1996 GSS.  The data we report is from the 1994 GSS.
**The items listed under Patriotism, as well as those under the headings Chauvinism, Assimilationism, and
Nativism, are the items that were used in constructing these scales for the analyses appearing in other tables.


scales. The Political Self-Categorization item asks respondents to consider
which identity is primary to them when they think about political and social
issues. The Patriotism and Chauvinism items assess the affective dimension
of national identity. These items are similar in content to other measures of
psychological attachment to the nation (Kosterman & Feshbach, 1989;
Staub, 1997). Another set of “pride” questions taps the respondent’s sense
of pride in regards to specific aspects of American society. The “true Ameri-
can” questions capture normative conceptions of national identity and are
modeled on an earlier Traditional Americanism scale (Citrin, Reingold, &
Green, 1990). In addition, several items assess support for the maintenance
of diverse cultures within the political community, a core principle of the
multiculturalist conception of American identity. 
Clearly, table 4.1 shows that patriotic sentiment remains pervasive in the
American public. One poll is just a snapshot in time, of course, and the level
of positive affect one records will vary with the response categories used to
measure the underlying attitude. Nevertheless, the consistency in expres-
sions of pride and belonging is impressive. When asked to rate the impor-
tance of being an American on a 10-point scale, fully 46 percent of the 1996
GSS sample said it was “the most important thing in their life.” In response
to similar questions, 81 percent said they felt “close to America” and 91 per-
cent agreed that they would “rather be a citizen of America than any other
country.” Responses to the items in the “Pride” subheading in table 4.1
show that this generalized feeling of attachment carries over to judgments
about most specific features of national experience. For example, 83 percent
of the 1996 GSS sample expressed pride in “the way democracy works in
America.” However, many fewer, 57 percent, said they were proud of the
country’s record in treating all groups in society equally. Clearly, one can
have a strong sense of patriotism without believing that the nation is per-
fect. 
Nor does everyone extend pride in America to a conviction in the coun-
try’s superiority. While 81 percent of the 1996 national sample did agree
that America is a better country than most others, just 40 percent thought
the world would be a better place if people from other nations were more
like Americans. Only 32 percent agreed that one should support their coun-
try ”even when it is wrong,” compared to 50 percent who disagreed. Chau-
vinism, in the sense of a latent hostility toward other nations, is present, but
the dominant outlook seems to be just a preference for one’s own people
and place. 
The pattern of answers to the questions about what is important for 
being “truly American” reinforces this conclusion. Each of the attributes
listed was deemed either “very” or “fairly” important, suggesting that most
Americans believe that there is a distinct national identity. The more inclu-
sive political criteria—American citizenship (87 percent) and simply feeling
American (93 percent)—were chosen more often than the ascriptive quali-
ties of being native born (69 percent) or Christian (54 percent). The ability
The Meaning of American National Identity
81


to speak English (93 percent) might appear as an ethnocentric criterion, but
this is an achieved rather than an ascribed trait and research indicates that
within two generations most immigrants do learn the country’s common
language (Portes, 1996). In fact, language minorities and foreign-born resi-
dents are as likely as white or black Americans to emphasize the linguistic
criterion for national identity (Citrin et al., 1990). 
Multiculturalism asserts the need to maintain minority cultures within
America and regards the symbols of assimilation with suspicion, if not out-
right hostility. The 1996 GSS asked respondents whether they believed it is
“better for a country if different racial or ethnic groups maintain their dis-
tinct customs and traditions” or if they “adapt and blend into the larger so-
ciety.” Of those who expressed a definite opinion, 58 percent said that
“blending in” would be better. (In the case of this question, however, fully
27 percent of the respondents did not express a definite opinion, suggesting
substantial ambivalence about the choice as posed. When the respondents
without a definite opinion are included, 43 percent of the total sample are
in favor of “blending in,” while 31 percent favor the maintenance of cultural
diversity). Even fewer, only 17 percent of the GSS sample, agreed with the
position of many advocates of multiculturalism that ethnic minorities
should receive government support to preserve their traditions.
At the same time, there was considerable confidence about the possibility
of assimilation. More respondents (42 percent) disagreed than concurred
(35 percent) with the proposition that “it is impossible for people who do
not share American customs and traditions to become fully American.” Per-
haps because of this optimism, many Americans may seem to feel that as-
similation and maintaining connections to one’s ethnic heritage are mutu-
ally compatible. Public opinion endorses the importance of a common
national identity without insisting on a forced march to cultural conformity.
For example, a 1994 GSS poll found that only 26 percent of the public said
that ”ethnic history is getting too much attention in the public schools,” an-
other indication of the tendency to view the existence of a unifying Ameri-
can culture as compatible with respect for pluralism (Merelman, Streich, &
Martin, 1998). 
Using exploratory and confirmatory factor analysis, we combined the
items listed in table 4.1 to construct Patriotism (alpha = .64) and Chauvinism
(alpha = .66) indices and summary measures of Nativist (alpha = .79) and 
Assimilationist (alpha = .64) conceptions of Americanism. Although there is
statistical evidence for treating these dimensions as distinct, clearly they are
interrelated. Nativism and Assimilationism correlate .58 (Pearson’s r) in the
1996 GSS sample, and Patriotism and Chauvinism also are associated 
(r = .37). An interesting note is that Assimilationism was more strongly re-
lated to Patriotism (r = .47) than to Chauvinism (r = .38), whereas the obverse
was true for Nativism (r = .29 and .45, respectively), a pattern pointing to a
“psycho-logical” bond between an exclusionary definition of American iden-
tity and a sense of superiority to other peoples. However, beliefs about the
82
The Contribution of Ethnic and National Identities to Political Conflict


value of cultural diversity are not strongly related to these variants of positive
attachment to the nation. Responses to the question about whether or not
groups should blend into the mainstream culture correlated positively, but
weakly, with the Patriotism (.13), Chauvinism (.13), Nativism (.07), and As-
similationism (.13) scales. Admittedly, these correlations may be attenuated
owing to the use of just a single-item to measure support for assimilation into
the mainstream culture and due to the low variance on the scales. 
Normative Consensus or Ethnic Conflict? 
Although strong feelings of national attachment are the norm, the possi-
bility of ethnic conflict in outlook remains. The nativist prediction that 
diversity erodes national unity assumes that minority groups are less com-
mitted to the idea of a common American identity. The multiculturalist
perspective yields a similar expectation. If ethnicity is the dominant crite-
rion of one’s social identity, then minority group members should be less
likely than those in the dominant ethnic group to view the entire nation as
“theirs” and, therefore, to be less likely to express the usual forms of pride
and affection. 
Despite frequent claims that there is a vast racial divide in American
opinion (Kinder & Sanders, 1996), our surveys indicate that ethnic differ-
ences are slight when it comes to patriotic sentiment (table 4.1). There is a
strong similarity in the outlook of blacks and whites regarding the personal
importance of American citizenship. Blacks (69 percent) are somewhat less
likely to say they feel “very close” or “close” to America than whites (83
percent) or Hispanics (83 percent), but, again, the overwhelming majority
of all three groups express a strong sense of attachment to the nation. In the
1994 LACSS, 79 percent of the whites and 73 percent of the black respon-
dents said they were “extremely” or “very” proud to be an American.
The opinions of Hispanics point to a continuing connection to their
countries of origin among recent immigrants. As table 4.1 shows, they are
less likely than either blacks or whites to say being an American is impor-
tant to them, more willing to say they would move to another country, and
less likely to say that being born in America or feeling American is an im-
portant characteristic of being a true American. The same general pattern
exists among the Hispanic and Asian respondents in the Los Angeles data
(reported in full in Sears et al., 1999). Only 44 percent of the Hispanic re-
spondents and 58 percent of the Asians in the Los Angeles study felt “ex-
tremely” or “very” proud to be an American, but these differences from
white opinion are largely because many of these respondents are recent ar-
rivals in the country and are not yet citizens. The findings of the Latino
National Political Survey (de la Garza, Falcon, & Garcia, 1996, p. 346) also
indicate that after controlling for demographic characteristics, including
length of time in the United States, Mexican Americans “express patriotism
The Meaning of American National Identity
83


at levels equal to or higher than do Anglos.” Somewhat surprisingly, the
Los Angeles survey found virtually no ethnic differences on the question of
whether “it is better if groups change so they blend into the larger society
rather than maintain their distinct cultures.” 
The pervasive patriotic and assimilative tendencies among all ethnic
groups frays, however, when one moves from the realm of ideals to judging
historical experience. In the 1996 GSS survey, as reported in table 4.1,
blacks and, to a lesser extent, Hispanics were less likely than whites to ex-
press pride in America’s history, in the way its democracy works, and in the
treatment of all social groups. Members of minority groups also are less
likely than whites to agree that America has provided “people of my ethnic
group a fair opportunity to get ahead in life” (LACSS) and more likely to
agree that the government should give special assistance to ethnic minori-
ties (GSS). Minority groups, then, clearly display a heightened sensitivity to
their collective status, largely agreeing with whites about the goal of e
pluribus unum, but, as previous research also found (Hochschild, 1995), 
asserting more often that discrimination and unequal opportunities for 
minority groups are ongoing problems in the United States.
The normative conceptions of what it means to be a “true American”
also are very similar in all ethnic groups. In this regard, blacks are somewhat
more likely to endorse what we have termed ascriptive or nativist criteria of
identity. They are significantly more likely (by the chi-square test) than ei-
ther white or Hispanic respondents in the 1996 GSS to agree that being
born in America, living in America for most of one’s life, and being a Chris-
tian are “very important” qualities of national identity. In an important
sense, these opinions express the visceral quality of their American roots for
the black respondents. African Americans constitute the oldest group of
nonwhite immigrants in the country, despite the forcible nature of their ar-
rival. Thus, the symbolic meaning of their invocation of nativist criteria for
national identity might well be a positive statement about their own “true
American-ness” rather than prejudice toward outsiders. 
Symbolic politics theory emphasizes the role of socialization in the for-
mation of social identities and would predict that common exposure to the
patriotic themes so dominant in American life would engender high levels
of national attachment across all ethnic groups over time. Of course, so-
cialization is ongoing and multifaceted, and there generally are group dif-
ferences in prior learning and current experience that have an impact on
political predispositions (Kinder & Sears, 1985). T
able 4.2 summarizes the
ethnic group differences in feelings of national identity by comparing their
mean scores on the multi-item measures of Patriotism, Chauvinism, Na-
tivism, and Assimilationism Indices, with each variable scored from 0 (low)
to 1 (high). The results show that in all three ethnic groups, positive attach-
ment to American nationality prevails. Nevertheless, ANOVA indicates that
several of the ethnic differences that do appear around the highly favorable
norm are statistically significant. Blacks do have lower mean scores than
84
The Contribution of Ethnic and National Identities to Political Conflict


whites on Patriotism, though higher scores on Nativism. Hispanics also
have lower Patriotism scores than whites, as well as less support for the As-
similationist conception of national identity. Whether one should empha-
size the element of consensus or these minor and spotty differences is a
matter of debate (compare Sidanius, this volume). But whatever one makes
of the divergences among whites, blacks, and Hispanics, clearly the glass of
patriotism is much more than half full in every group. The real ethnic di-
vide in American politics emerges on specific policies bearing directly on
the status of minority groups, not on conceptions of national identity. 
The results of a multiple regression analysis in which race, ethnicity,
gender, age, income, education, region, and immigrant generation were
employed as predictors of the Patriotism, Chauvinism, Nativism, and As-
similationism scales, respectively, confirm this conclusion. (For space rea-
sons, the full results are not reported here, but will be provided on request.)
With the imposition of statistical controls, there is no “race effect” on Pa-
triotism or Chauvinism scores. Blacks remain significantly more likely than
whites to endorse the nativist conception of American identity, largely be-
cause of their greater support for Christianity as a legitimate defining crite-
rion of nationality. This suggests that another social identity, religion, is im-
portant to understanding ethnicity, national identity, and intergroup
conflict in America. 
More generally, there is a striking similarity in the demographic under-
pinnings of the four measures of national attachment: the elderly, those
with lower levels of education, and those residing in the South were signifi-
cantly more likely to have high, “pro-America” scale scores. Finally, first-
generation immigrants were significantly less likely to have high scores on
the Patriotism, Nativism, and Assimilationism indices. However, neither
second- nor third-generation Americans differ from those with an even
The Meaning of American National Identity
85
table 4.2 Ethnic Differences in National Identity1 (1996 GSS)
Mean Score on a 0 to 1 Scale
T
otal
Whites
Blacks
Hispanics
Affective attachment to America
Patriotism
.80
.81
.76
.75
Chauvinism
.59
.59
.62
.56
Normative conceptions of American identity
Assimilationism
.86
.86
.85
.79
Nativism
.64
.62
.79
.60
1. The four attitude measures are additive indices created by summing responses to the several groups of
items designated in table 4.1.  The response options for each item were first recoded to range from zero to
one.  After summation, index scores were also recoded to range from zero to one with higher scores indicat-
ing higher levels of patriotism, chauvinism, etc.  Reliabilities (Chronbach’s alpha) for the respective scales
are patriotism (.64), chauvinism (.66), assimilationism (.64), and nativism (.79).


longer family history in the country, once again suggesting that socializa-
tion into a sense of national pride and belief in a distinctive American iden-
tity does not take long. 
National, Hyphenated, and Ethnic Identities
Since citizens can have multiple loyalties (or identities), it is important to
assess the intensity of ethnic and national identifications separately.
Whether these attachments are complementary or competitive is an em-
pirical matter. The 1996 GSS asked respondents separate questions about
how close they felt to the United States and to their own racial or ethnic
group. A cross-tabulation of answers to these items show that 63 percent of
the white, 65 percent of the black, and 66 percent of the Hispanic respon-
dents expressed feelings of closeness to both the nation and their own ethnic
group. Among whites, only 11 percent felt close attachment to their own
ethnic group, but not to the country as a whole, whereas 24 percent of the
black and 14 percent of the Hispanic respondents fell into this category.
This suggests a stronger salience of an ethnic identity among blacks, with
the possible implication that they would be more likely than whites to 
privilege ethnicity over nationality when trade-offs must be made. Still, the
dominant result is that a large majority among both races expresses af-
finity to both identity groups. Subjectively, national and ethnic identities in
the United States tend to be experienced as complementary rather than
competing.
T
able 4.1 reports the responses to the political self-categorization ques-
tion: “When you think of social and political issues, do you think of yourself
mainly as a particular ethnic, racial, or nationality group, or do you think of
yourself mainly as just an American?” (The data come from the 1994, not
the 1996 GSS poll, which did not include the item). This question wording
calls for a ranking of identities and explicitly places the choice in a political
context. Of the white respondents, 96 percent answered “just an American,”
compared to 66 percent of the black and 79 percent of the Hispanic respon-
dents (see table 4.1). The uniformity of opinion among whites is consistent
with the view of the various groups of European immigrants having
“melted,” so that their most salient political identity is no longer cultural or
religious, but simply a superordinate identity as Americans (Alba, 1990;
Waters, 1990; Roediger, 1994). 
The 1994 and 1995 LACSS studies asked a similar self-categorization
question with equivalent results. In the pooled samples, 93 percent of the
white respondents gave the overarching “just an American” self-designa-
tion. Blacks (69 percent), Hispanics (70 percent), and Asians (71 percent)
were significantly less likely (by the chi-square test) than whites to say they
thought of themselves as just Americans. When forced to choose, however,
a large majority in all three minority ethnic groups opted for a common na-
86
The Contribution of Ethnic and National Identities to Political Conflict


tional identity, not ethnic particularity, as their first choice. Moreover,
among Hispanics and Asians, those who categorized themselves in terms of
ethnicity tended to be first-generation immigrants, suggesting, once again,
that the absorption of a sense of national identity is the normal outcome of
socialization in this country. 
The 1995 LACSS allowed respondents to say whether they identified as
just Americans or as both Americans and as members of an ethnic, racial, or
nationality group. As reported in table 4.3, only 17 percent of the whites
opted for a dual or hyphenated identity. But, when given this opportunity to
choose, a majority in all three minority groups preferred to categorize
themselves as both Americans and members of an ethnic group and not as
exclusively “American.” Unfortunately, there are no national data to repli-
cate this result, but it is consistent with other small-scale studies testing so-
cial identity theory that suggest the relatively greater salience of an ethnic
identity for members of smaller, easily demarcated, homogeneous, and 
disadvantaged groups (Prentice & Miller, 1999; Brewer, this volume). The
historically dominant group in a multiethnic state often fuses rather than
hyphenates their ethnic and national identities (Connor, 1993). It is beyond
the scope of this chapter to explore the antecedents of these alternative self-
categorizations by minority group members. Among immigrant groups, it
appears that length of residence in the United States and speaking English
diminish the likelihood of identifying oneself primarily in ethnic terms 
(de la Garza et al., 1996), but these factors could not account for the varied
responses of blacks. 
Identities and Policy Preferences
The political relevance of social identities rests partly on how they influence
attitudes and behavior toward one’s own and other groups. A strong sense
of group identity is expected to promote conduct that favors one’s fellow
group members, even in a “minimal group” (T
ajfel, 1978). But since there
are multiple dimensions of self-identification, how one defines the “I” as
“we” when a political decision arises is likely to be significant. Symbolic
politics theory holds that broad attitudes, including feelings of national and
ethnic identity, are more likely to determine preferences on specific issues
when those predispositions are central to an individual’s self-definition and
when they are cued by stimuli associated with the objects of these attitudes
(Sears, 1993; Zaller, 1992). For example, a strong sense of Jewish identity is
more likely to govern opinions about American policy toward Israel or a
mandatory school prayer than the future of the spotted owl. In addition, as
shown above, ethnic identity may indeed be more significant for the con-
struction of the social self among some groups than others. 
Nationalism is an integrating ideology that overrides the claims of less
comprehensive group loyalties. In fact, liberal theorists (T
amir, 1997;
The Meaning of American National Identity
87


table 4.3 Political Self-Categorization: American, Dual, or Ethnic Identity (1995 LACSS)
Whites
Blacks
Hispanics
Asians
(n = 254)
(70)
(174)
(46)
Just an American Identity
78%
16%
11%
4%
Both American and ethnic identity
17
54
59
59
Ethnic identity
5
30
31
37
Whites
Blacks
Hispanics
Asians
US born 
Not US born 
US born 
Not US born
US born 
Not US born 
US born 
Not US born
(n = 223) 
(31)
(68)
na
(59)
(115) 
(8)
(38)
Just an American identity 
80%
65%
16%
na
24%
4% 
13%
3%
Both American & ethnic identity 
17
16
54
na
53
62 
63
58
Ethnic identity 
3
19
29
na
24
34 
25
40


Hollinger, 1997; Miller, 1995) justify holding on to a national identity on
the grounds that this sentiment generates feelings of special obligation to
and responsibility for the other members of one’s national community. If
this surmise is correct, then people should be less likely to feel a sense of
duty to members of groups who define themselves as somehow outside or
different from the national “family.” Accordingly, since multiculturalism
validates the maintenance of group differences, the stronger one’s attach-
ment to an overarching American national identity, the more likely one
should be to oppose recognition and support for groups emphasizing their
cultural distinctiveness. Alternatively, if the majority group views ethnic mi-
norities as disadvantaged members of a community with a shared identity,
feelings of national attachment could accommodate, even facilitate, support
for policies aimed at helping them. 
In considering the political implications of ethnic consciousness, Tyler,
Lind, Ohbuchi, and Sugawara (1998) found that assimilated and “bicultur-
alist” members of minority groups (that is, those who emphasize their
American identity and dual identity, respectively) were equally likely to ex-
press trust in government, to regard existing institutions as fair, and to com-
ply with government policy. Both groups were significantly more likely
than “separatist” respondents from the same ethnic group (those emphasiz-
ing their ethnic identity) to express these attitudes. The present data show
that the relative salience of an ethnic as opposed to national self-identifica-
tion affects how people evaluate multiculturalist principles (see table 4.4).
Minority group members for whom ethnic identification takes priority over
national identity generally are less likely than those who say they view poli-
tics mainly as an American to express pride in America and to agree that
their group has had a fair chance to get ahead, and are more likely to believe
that groups should maintain their own distinct cultures, to deny that politi-
cal organizations based on ethnicity promote separatism, and to think that
people are best represented by leaders with the same ethnic background.
(See table 4.4 for results of chi-square tests for statistical significance.) 
T
able 4.4 also suggests that the association between an ethnic self-identi-
fication and support for multiculturalism is stronger and more systematic
among blacks than among either Hispanics or Asians. For example, among
blacks, 61 percent of those who identified themselves as “just American” felt
it is better for minority groups to blend into the larger society rather than
maintain their distinct cultures, compared to only 25 percent of those who
defined themselves primarily as members of an ethnic group. Among His-
panics, the equivalent figures were 60 percent and 54 percent, a far smaller
difference, and among Asians, they were 57 percent and 48 percent, respec-
tively. The future trend in how the different minority groups balance their
ethnic and national identities thus has clear implications for the pattern of
intergroup conflict over issues of cultural recognition and representation.
Ordinary least-squared regression analysis was employed to delineate the
connections between the affective and normative dimensions of national
The Meaning of American National Identity
89


table 4.4 The Impact of Political Self-Categorization on Beliefs about the American National Community (1994 LACSS)
Blacks
Hispanics
Asians
Just
Ethnic
Just
Ethnic
Just
Ethnic
American
Identity
American
Identity
American
Identity
(n = 156)
(68)
(182)
(66)
(36)
(8)
Feel extremely or very proud to be an American
81%
54%
53%
22%
61%
33%
Agree that American society has provided people of my ethnic
50
22*
70
63
69
63
group a fair opportunity to get ahead in life
Think it is better if groups change so that they blend into the larger
61
25
60
54
57
48#
society (as in the idea of a melting pot) rather than maintain their
distinct cultures1
Agree that political organizations based on race or ethnicity
70
57#
71
52*
70
38#
promote separatism and make it hard for all of us to live together
Leader background doesn’t matter for representation
68
40*
67
50*
60
38
Chi-square test of differences between “just American” and “ethnic identity” respondents within each racial/ethnic group:
*p<.05
#p<.10
1The percentages in this row represent pooled data from the 1994 and 1995 LACSS.


table 4.5 The Relationship between Attitudes T
oward National Identity  and Policy Preferences (1996 GSS)
Affective Attachment
Normative Conceptions
(3)
(6)
Patriotism
Assimilationism
(1)
(2)
& Chauvinism
(4)
(5)
& Nativism
Patriotism
Chauvinism
(patrsm) (chauv)
Assimilationism
Nativism
(assim) (nativ)
Policy Preferences
Multiculturalism
(A) Need to share American customs
.05
.38**
–.08
.42**
.17**
.27**
–.07
.29**
(B) Support melting pot as ideal
.20*
.24**
.18
.17
.25**
.01
.33**
–.10
(C) Reduce immigration levels
.09
.18**
.06
.17**
.26**
.17**
.21**
.09*
(D) No cultural benefit from immigrants
.04
.20**
–.01
.21**
.11**
.15**
–.01
.16**
Racial Policies
(E) Oppose affirmative action
.02
–.01
.01
.00
–.05
–.03
–.05
–.01
(F) Oppose govt. help for blacks
.06
.14*
.03
.13
.12*
.06
.08
.04
(G) Oppose more spending on blacks
.17
.06
.14
.00
–.01
.06
.03
.05
Domestic spending
(H) Oppose social spending
.03
.03
–.03
.05
.03
.00
.05
–.02
Foreign Policy
(I) Support more spending on defense
.23*
.31**
.12
.25*
.19*
.13*
.10
.10
(J) Support protectionism
.18**
.42**
.06
.40**
.29**
.23**
.16**
.16**
Entries are unstandardized regression coefficients for national identity variables in equations with controls for race, ethnicity, age, education, income, gender, region (South), generation, 
ideology, and party identification.
*p<.05
**p<.01


identity on the one hand and policy preferences on the other. The statistical
model estimates the effects of the Patriotism, Chauvinism, Assimilation,
and Nativism Scales, (columns 1, 2, 4, and 5 in table 4.5), respectively, on 
issues related to cultural diversity (rows A–D), racial issues (rows E–G), 
domestic spending (row H), and foreign policy (rows I–J). The equations
included race, ethnicity, age, education, income, gender, region, immigrant
generation, partisan affiliation, and liberal-conservative self-identification
as control variables. Owing to the small number of Hispanic respondents in
the 1996 GSS sample, this analysis is confined to whites and blacks. For
simplicity of presentation, the table omits the effects for the control vari-
ables and reports only the unstandardized regression coefficients for the
measures of national identity (Patriotism, Chauvinism, Assimilationism,
and Nativism). All variables were recoded to vary between 0 and 1, and the
dependent variables all are scored in the “nationalist” or “conservative” 
direction. Accordingly, the positive coefficient of .24 for the association 
between Chauvinism and the Melting Pot item (column 2, row B) indicates
that believing that America is superior to other countries is strongly related
to the idea that ethnic minorities in the United States should blend into the
mainstream culture. 
T
able 4.5 also reports the results of models that simultaneously included
both Patriotism and Nationalism Scales (column 3) and the Assimilationism
and Nativism Scales (column 6) as predictors. These equations were de-
signed to provide a tentative estimate of the relative degree of association
between each of these affective orientations and normative conceptions of
national identity. The results (which are reviewed by comparing the coeffi-
cients across columns on a row-by-row basis) do provide some evidence of
tension between a strong sense of national identity and multiculturalism.
Although the respondent’s race has no relationship to these multiculturalist
policy positions in the multivariate model, three of the four indicators 
of national identity predict support for the belief that it is impossible for
people who do not share American customs to become fully American (row
A). With regard to the question asking whether ethnic and racial groups
should blend into mainstream society, Patriotism, Chauvinism, and Assimi-
lationism had statistically significant relationships with giving the melting-
pot response (row B). Immigration is the major source of ethnic diversity
and new claimants for cultural recognition in America, and Chauvinism,
Assimilationism, and Nativism, but not Patriotism, had statistically signifi-
cant relationships with support for reducing the level of immigration (row
C) and disagreement that immigrants have a positive cultural impact on
America (row D). 
When Patriotism and Chauvinism are included as predictors simultane-
ously, the more aggressive outlook tends to have the stronger association
with opposition to increased cultural diversity in the United States. And
when the effects of Assimilationism and Nativism are compared, it is the
latter more exclusionary attitude that generally appears to have the stronger
92
The Contribution of Ethnic and National Identities to Political Conflict


negative relationship to support for multiculturalism. An ethnocentric con-
ception of American national identity, therefore, may intensify internal po-
litical conflict by hardening resistance to the demands of cultural minorities
for recognition and support. 
T
able 4.5 also reports on opinions about government efforts targeted at
helping blacks (rows E–G) and about spending on domestic social programs
(measured by a Social Spending Index combining beliefs about whether the
government should spend more on health, education, welfare, and the envi-
ronment—row H). The results are a virtual mirror image of the earlier
findings about the factors influencing opinions about cultural diversity and
immigration. In the case of the domestic policy issues, race is strongly asso-
ciated with preferences, with blacks consistently more favorable toward
governmental activism. But none of the four measures of national identity
has a statistically significant relationship with scores on the Social Spending
Index, opinions about affirmative action, or attitudes about whether the
government should spend more to help blacks. The only coefficients in the
array that attain conventional levels of statistical significance are the weak
associations of Chauvinism and Assimilationism with opposition to special
efforts to improve the living conditions of blacks. The results of these
analyses indicate that generalized feelings of national attachment neither
boost support for assisting one’s fellow citizens through government spend-
ing nor enhance opposition to policies, such as affirmative action, aimed at
assisting blacks or other ethnic minorities. 
This may not be surprising, since the controversies over domestic spend-
ing and racial policy tend not to be framed in terms that engage attitudes
toward the nation’s integrity or power. As hypothesized above, the impact
of social identities (and other predispositions) on specific policy preferences
should vary across political domains. Feelings of national attachment
should be strongly engaged only when the specific issues clearly touch on
the values of American power, sovereignty, or purity. The data presented in
table 4.5 consistently support this proposition. In addition to their effects
on beliefs about cultural diversity and immigration, all four measures of na-
tional attachment had statistically significant effects in predicting opinions
about whether military spending should be increased and on scores on an
American Protectionism Index (alpha = .65), constructed by summing an-
swers to items about limiting foreign imports, prohibiting foreigners from
buying land in America, and requiring television stations to give preference
to American programs. 
A final regression analysis, reported in table 4.6, explored potential dif-
ferences in the connections of ethnic as opposed to national identities on
racial policy preferences. Here, responses to both the question “How close
do you feel to the United States?” and the question “How close do you feel
toward your own racial or ethnic group?” were included as predictors of
opinions regarding affirmative action, support for government efforts to
help blacks, and government spending for blacks. In this analysis, the vari-
The Meaning of American National Identity
93


able coding and controls are identical to those reported for table 4.5. How-
ever, to distinguish between the impact of ethnic identifications among
whites and blacks, these equations were estimated separately for whites and
blacks. (Because of the paucity of Hispanic respondents, the analysis could
not be replicated for that group.) 
T
o highlight the results, table 4.6 reports only the coefficients for na-
tional attachment and ethnic closeness. Confirming the results reported
above, positive affect for the nation as a whole has no impact one way or the
other on the racial policies queried. There are obviously significant racial
differences in policy preferences, but among blacks, feeling close to one’s
own ethnic group had no great additional influence on support for policies
aimed at improving their collective status. The fear of some political com-
mentators that black racial identification leads to a highly distinctive politi-
cal outlook seems overstated. 
Among whites, however, a stronger sense of ethnic identity does have a
significant effect across the three policies; white identification seems consis-
tently to increase opposition to government policies designed to help
blacks. Given our concern about the impact of ethnic identification on in-
tergroup conflict, these results are sobering. “Identity politics,” by which
we mean the tendency to judge issues and events in terms of how they affect
the standing of one’s own group, can juxtapose one’s ethnic and national
94
The Contribution of Ethnic and National Identities to Political Conflict
table 4.6 National Attachment and Ethnic Identity as Predictors of Opinion 
on Race-T
argeted Policies (1996 GSS)
Oppose
Oppose Special
Oppose More
Affirmative
Govt. Help
Spending on
Action
for Blacks
Blacks
Among White Respondents
Close to own ethnic group (close =1)
.07**
.05*
.08*
(white identification)
Close to the US
–.02
–.02
–.03
(national attachment)
Sample size (n)
543
546
347
Among Black Respondents
Close to own ethnic group (close =1)
.12
–.05
.09
(black identification)
Close to the US
–.11
.16
.01
(national attachment)
Sample size (n)
72
74
50
Entries are unstandardized regression coefficients for national and ethnic identity variables in equations with
controls for age, education, income, gender, region (South), generation, ideology, and party identification.
*p<.05
**p<.01


“selves.” While minorities may be more likely to express a strong ethnic
identity, counterintuitively, it appears that the effects of ethnic identification
are more prominent among whites. In a disturbing way, the results in table
4.6 are a reminder that the evocation of white consciousness, whether
through nativist mobilization or through reaction to the assertion of ethnic
pride by other groups, can lead to divisiveness and prejudice. 
Conclusion
A large body of survey evidence confirms that a positive sense of national
identity is pervasive in the United States, as is the belief that there is a
unique American identity. In terms of affective attachment, patriotism—the
sense that America is “best for me”—was more widespread than chauvin-
ism—the sense that this country is inherently superior to all others. Signifi-
cantly, too, we found no evidence of a deep ethnic divide in feelings about
American nationality. The slightly lower level of patriotism among blacks,
as a group, than whites is less compelling than the fact that the large major-
ity of all ethnic groups express pride in and closeness to their country and
its symbols. In addition, there is a broad consensus among all ethnic groups
regarding the main criteria for American identity. 
The strong sense of attachment to an American identity among ethnic
minorities should calm fears about the consequences of the nation’s in-
creased ethnic diversity. Nearly as much as whites, minority respondents in
our surveys rejected the idea of organizing political life along ethnic lines.
Furthermore, most of them do not express their primary political identity
in purely ethnic terms. 
These results do not fit easily into an image of public attitudes as driven
primarily by interest-based conflict between dominant and subordinate eth-
nic groups. The symbolic politics perspective outlined here presumes that
attitudes and values are acquired through a process of reinforcement, and
that the role of purely instrumental motivation often is secondary (Sears &
Funk, 1990). The survey evidence suggests that members of all ethnic
groups may be attracted to an American identity because of its traditional
liberal premises. In both California and national surveys (Citrin et al., 1990,
1994), the criteria of a “true American” chosen as important most often by
all four major ethnic groups were egalitarianism and tolerance (“treating
people of all races equally” and “respecting people’s freedom to say what
they want”).
The staying power of this liberal conception of American identity may
help explain why many minority respondents viewed attachment to the
country and a sense of closeness to their own ethnic group as complemen-
tary rather than competitive. The results presented above show that mi-
nority group members frequently conceive of themselves in dual terms,
defining themselves as both Americans and part of a distinctive racial or
The Meaning of American National Identity
95


ethnic group. Whites, by contrast, generally see themselves as “just Ameri-
cans”; for them, national and ethnic identities are merged. What is critical,
both psychologically and politically, is how the dual identity—national and
ethnic—is accommodated intrapersonally. That is, when does the con-
sciousness of being simultaneously American and black lead to a sense of
ambivalence that pits the self against either the nation or the group, and
how is this conflict among loyalties resolved? The present data reinforce
the familiar point that blacks identify with American ideals, not American
realities. As James Jones suggests (personal communication), their dual
identity reflects recognition of the persistence of what Myrdal (1944) called
the “American dilemma.”
In the political domain, a strong sense of national identity consistently
was related to opposition to liberal immigration policies and official support
for maintaining minority cultures. There is, then, an ideological tension be-
tween multiculturalism and nationalism. Strong feelings of national attach-
ment also were associated with approval for increasing America’s military
strength, as well as a willingness to take quite drastic measures to reduce the
potential impact of foreign interests and influences. These issues, of course,
focus on the boundaries between citizens and strangers, a distinction that is
less relevant for opinion formation when it comes to domestic spending or
racial policies. 
The strength of national attachment was unrelated to how whites felt
about policies targeted at helping blacks, their fellow-citizens. However, a
strong ethnic identification among whites was associated with greater an-
tagonism to affirmative action and more government spending to assist
blacks. In this instance, ingroup favoritism does seem to lead to greater 
inter-group conflict. 
The portrait of normative consensus among the general public seems
strangely discordant with the seeming prevalence of ethnic competition in
American political life. The division of political offices, government con-
tracts, and other benefits among ethnic groups is closely watched by ac-
tivists. Racial issues deeply divide the country’s two political parties. Given
all this and more, what accounts for the failure of the surveys to detect in-
tense disagreements between the “majority” and “minority” groups when it
comes to attitudes about national identity?
One answer is that the existence of a gulf between the views of political
activists and ordinary citizens is not uncommon. Another possibility is that
the current approach to measuring national and ethnic identifications is in-
adequate. Surveys may pay too much attention to abstract norms and too
little to responses to distributive conflicts or actual social encounters among
members of different groups. In this regard, researchers should frame issues
in a way that asks respondents to make trade-offs among their multiple
identities when it comes to allocating costs and benefits. This would enable
a more comprehensive assessment of intragroup, as well as intergroup, con-
flict in how these choices are made. 
96
The Contribution of Ethnic and National Identities to Political Conflict


Another important issue concerns the enduring quality of the tendency
of recent immigrants to prefer a dual or hyphenated political identity. The
present data indicate that a developing sense of patriotism and a sense of
identity as an American are part of the process of acculturation that His-
panic and Asian groups are experiencing. More generally, today’s immi-
grants, like their predecessors, become “desocialized from their native cus-
toms over time, even if the emotional significance of attachment to the
group persists” (Glazer & Moynihan, 1975, p. 8). Because of this continu-
ing affective tie, appeals to ethnicity still can be an effective strategy for mo-
bilizing political support for the pursuit of group or individual interests
(Horowitz, 1985). The utility of ethnicity in political combat, then, may 
reinforce the tendency of political leaders to emphasize the salience of this
aspect of identity and strive to sustain its psychological significance within
their ethnic group. 
In addressing the unresolved issue concerning the conditions under
which hostility toward outgroups accompanies ingroup pride, Gaertner,
Dovidio, Nier, Ward, & Banker (1999) argue for the efficacy of a superordi-
nate common identity in mitigating intergroup conflict. Applied to the
larger realm of American national politics, the question posed by social
identity theory becomes whether an enhanced sense of ethnic identification
among minorities leads toward satisfactory conflict resolutions or toward
entrenched Balkanization. Clearly, the mobilization of group consciousness
among the disadvantaged can force movement toward less inequality and
more social justice. On the other hand, the enhanced expression of ethnic
identification in one group tends to be emulated, often provoking resistance
to change and the hardening of group boundaries. Whatever the direction
of future change in public attitudes, the evidence summarized here points to
the advantages of a common sense of American identity founded on the 
realization of equal citizenship. 
References
Alba, R. D. 1990. Ethnic identity: The transformation of white America. New Haven:
Yale University Press.
Appiah, K. A. (1994). Identity against culture: Understandings of multiculturalism.
Berkeley, CA: Doreen B. T
ownsend Center for the Humanities.
Ashmore, R. D., & Jussim, L. (1997). Self and identity: Fundamental issues. New York:
Oxford University Press. 
Barber, B. (1995). Jihad vs. McWorld. New York: Times Books.
Bar-tal, D. (1993). Patriotism as fundamental beliefs of group members. Politics and
the Individual, 3, 45–62.
Brimelow, P. (1995). Alien nation. New York: Random House.
Citrin, J., Sears D., Muste C., & Wong, C. (Forthcoming). Multiculturalism in
American public opinion. British Journal of Political Science.
Citrin J., Haas E. B., Muste C., & Reingold B. (1994). Is American nationalism
The Meaning of American National Identity
97


changing?: Implications for foreign policy. International Studies Quarterly, 38,
1–31.
Citrin J., Reingold B. & Green. D. P. (1990). American identity and the politics of
ethnic change. Journal of Politics, 52, 1124–1154.
Connor, W. (1993). Beyond reason: The nature of the ethno-national bond. Ethnic
and Racial Studies, 163, 373–389.
Davis J., & Smith. T. (1997). The general social survey codebook. Chicago: National
Opinion Research Center.
de la Garza, R. O., Falcon A., Garcia C. (1996). Will the real Americans please stand
up: Anglo and Mexican-American support of core American political values.
American Journal of Political Science, 40, (2), 335–350.
Farley, R. (1996). The new American reality : Who we are, how we got here, where we are
going. New York: Russell Sage Foundation.
Gaertner, S. L., Dovidio J.F., Nier J. A., Ward C. M., & Banker, B. S. (1999). Across
cultural divides: The value of superordinate identity. In D. Prentice & D. Miller.
(Eds.), Cultural divides: The social psychology of cultural contact (pp. 173–212). New
York: Russell Sage Foundation.
Gerstle, G. (1997). Liberty, coercion, and the making of Americans.” Journal of
American History, 84, 524–558.
Glazer, N. & Moynihan D. P. (1963). Beyond the melting pot. Cambridge: MIT Press.
Glazer, N., & Moynihan. D. P. (1975). Ethnicity: Theory and experience. Cambridge,
MA: Harvard University Press.
Gleason, P. (1980). American identity and Americanization. In S. Thernstrom, A.
Orlov, & O. Handlin (Eds.), Harvard Encyclopedia of American Ethnic Groups
(pp 31–58). Cambridge, MA: Harvard University Press. 
Goldberg, D. T. (1994). Introduction: Multicultural conditions. In D. T. Goldberg,
(Ed.), Multiculturalism: A critical reader (pp. 1–41). Cambridge, MA: Blackwell.
Gordon, M. (1964). Assimilation in American life: The role of race, religion, and national
origins. New York: Oxford University Press.
Greenfeld, L. (1992). Nationalism : Five roads to modernity. Cambridge, MA: Harvard
University Press.
Greenfield, L., & Chiriot, D. (1994). Nationalism and aggression. Theory and Soci-
ety, 23, 79–130.
Hartz, L. (1955). The liberal tradition in America: An interpretation of American political
thought since the revolution. New York: Harcourt, Brace.
Higham, J. (1988). Strangers in the land: Patterns of American nativism, 1860–1925.
2nd ed. New Brunswick, NJ: Rutgers University Press.
Hochschild, J. L. (1995). Facing up to the American dream: Race, class, and the soul of the
nation. Princeton, NJ: Princeton University Press. 
Hollinger, D. (1995). Postethnic America. New York: Basic Books.
Hollinger, D. A. (1997). National solidarity at the end of the twentieth century: 
Reflections on the United States and liberal nationalism. Journal of American 
History, 84, 559–569.
Horowitz, D. (1985). Ethnic groups in conflict. Berkeley: University of California
Press.
Huntington, S. P. (1981). American politics: The promise of disharmony. Cambridge,
MA: Belknap Press.
Kateb, G. (1994). Notes on pluralism. Social Research, 61, 511–537. 
98
The Contribution of Ethnic and National Identities to Political Conflict


Kelman, H. (1997). Nationalism, patriotism, and national identity: Social-psycho-
logical dimensions. In D. Bar-tal & E. Staub (Eds.), Patriotism (pp. 165–189).
Chicago: Nelson-Hall.
Kinder, D. R., & Sanders, L. M. (1996). Divided by color: Racial politics and democratic
ideals. Chicago: University of Chicago Press.
Kinder, D., & Sears, D. O. (1985). Public opinion and political action. In 
G. Lindzey & E. Aronson (Eds.), Handbook of Social Psychology. 3rd ed., vol 2., 
(pp. 659-741) Reading, MA: Random House.
Kosterman, R., & Feshbach, S. (1989). T
owards a measure of patriotic and national-
istic attitudes. Political Psychology, 10, 257–274.
Lau, R., & Sears, D. (Eds.). (1986). Political cognition : The 19th Annual Carnegie
Symposium on Cognition.. Hillsdale, NJ: Lawrence Erlbaum Associates.
Lind, M. (1995). The new nationalism and the fourth American revolution. New York:
The Free Press.
Lipset, S. M. (1996). American exceptionalism: A double-edged sword. New York: W. W.
Norton. 
Merelman, R. M., Streich, G. & Martin, P. (1998). Unity and diversity in American
political culture: An exploratory study of the national conversation on American
pluralism and identity. Political Psychology, 19, 781–808.
Miller, A. H., Gurin, P., Gurin G., & Malanchuk O. (1981). Group consciousness
and political participation. American Journal of Political Science, 25, 494–511.
Miller, D. (1995). On nationality. Oxford: Clarendon Press.
Myrdal, G. (1944). An American dilemma: The Negro problem and modern democracy.
New York, London: Harper & Brothers.
Okin, S. M. (1997). Is multiculturalism bad for women? Boston Review, 22, 5.
Portes, A. (Ed.) (1996). The new second generation. New York: Russell Sage Foundation.
Prentice, D. A. & Miller, D. T. (Eds.) (1999). Cultural divides: Understanding and
overcoming group conflict. New York: Russell Sage Foundation.
Raz, J. (1994). Multiculturalism: A liberal perspective. Dissent, 41, 67–79.
Reich, R. B. (1992). The work of nations: Preparing ourselves for 21st century capitalism.
New York: Vintage Books.
Roediger, D. R. (1994). Towards the abolition of whiteness: Essays on race, politics, and
working class history. London: Verso.
Schlesinger, A. M. Jr. (1992). The disuniting of America: Reflections on a multicultural
society. New York: Norton.
Schaar, J. H. (1981). Legitimacy in the modern state. New Brunswick, NJ: Transaction
Books.
Sears, D. O. (1993). Symbolic politics. In S. Iyengar & W. J. McGuire, (Eds.), Explo-
rations in political Psychology (pp. 113–149). Durham: Duke University Press.
Sears, D. O., Citrin J., Cheleden S., & Van Laar C. (1999). Cultural diversity and
multicultural politics: Is ethnic Balkanization psychologically inevitable? In D.
Prentice & D. Miller, (Eds.), Cultural divides: The social psychology of cultural con-
tact. (pp. 35–79) New York: Russell Sage Foundation. 
Sears, D. O., & Funk, C. L. (1990). Self-interest in Americans’ political opinions. 
In J. J. Mansbridge, (Ed.), Beyond self-interest (pp. 147–170) (1999). Chicago:
University of Chicago Press.
Sidanius J., Levin S., Rabinowitz J. L., & Frederico C. M. (1990). Peering into the
jaws of the beast: The integrative dynamics of social identity, symbolic racism
The Meaning of American National Identity
99


and social dominiance. In D. A. Prentice & D. T. Miller, (eds.), Cultural divides:
Understanding and overcoming group conflict (pp. 80-132) New York: Russell Sage
Foundation. 1999.
Smith, R. M. (1997). Civic ideals: Conflicting visions of citizenship in U.S. history. New
Haven: Yale University Press.
Staub, E. (1997). Blind versus constructive patriotism: Moving from embeddedness
in the group to critical loyalty and action. In D. Bar-tal & E. Staub, (Eds.), Patri-
otism (pp. 213–228).Chicago: Nelson Hall.
Sullivan, J., Fried A., & Dietz M. (1992). Patriotism, politics, and the presidential
election of 1988. American Journal of Political Science, 36, 200–234.
T
ajfel, H. (1978). Social categorization, social identity and social comparison. In H.
T
ajfel (Ed.), Differentiation between social groups (pp. 61–76). London: Academic
Press. 
T
ajfel, H., & Turner J. (1986). The social identity theory of intergroup behavior. 
In S. Wrochel & W. Austin, (Eds.), Psychology of intergroup relations. Chicago:
Nelson-Hall.
T
aylor, C. (1995). Cross purposes: The liberal-communitarian debate. In Philosophi-
cal arguments. (pp. 181–203). Cambridge, MA: Harvard University Press.
Thoits, P. A., & Virshup, L.K. (1997). Me’s and we’s: Forms and functions of social
identities.” In R. D. Ashmore & L. Jussim, (Eds.), Self and identity: Fundamental
issues. New York: Oxford University Press. 
Turner, J. C., Hogg, M.A. Oakes, P. J. Reicher, S. D. & Blackwell, M. S. (1987). 
Rediscovering the social group: A self-categorization theory. Oxford: Basil Blackwell.
T
yler, T., Lind E. A., Ohbuchi K., & Sugawara, I. (1998). Conflict with outsiders:
Disputing within and across cultural boundaries. Personality & Social Psychology
Bulletin, 24, 137–147.
Warner, R. S. (1993). Work in progress: T
oward a new paradigm for the sociological
study of religion in the U.S. American Journal of Sociology, 98, 1044–1093.
Waters, M. (1990). Ethnic options: Choosing identities in America. Berkeley: University
of California Press. 
Wolfe, A. (1998). One nation after all. New York: Viking.
Young, I. M. (1990). Justice and the politics of difference. Princeton, NJ: Princeton
University Press.
Zaller, J. (1992). The nature and origins of mass opinion. New York: Cambridge Uni-
versity Press. 
100
The Contribution of Ethnic and National Identities to Political Conflict


Jim Sidanius
5
John R. Petrocik
Communal and National Identity
in a Multiethnic State
A Comparison of Three Perspectives
As recent episodes of “ethnic cleansing” in Rwanda, Bosnia, and Kosovo
make painfully clear, some multiethnic states seem as inclined to interethnic
violence at the dawn of the 21st century as they were at the start of the 20th.
Indeed, the national mobilization of a 50-year-long cold war may have tem-
porarily restrained hostilities that lost none of their underlying virulence.
Grievances that were set aside by a need to face a common enemy have now
become priority issues for organized interests, parties, and public policy.
Furthermore, ethnic conflict is not peculiar to nation-states mobilized into
the stand-off between the cold war alliances, but appears quite generalized
across the globe (see Gurr & Harff, 1994).
In many cases, intergroup conflict within multiethnic states turns on the
compatibility of subgroup identities and commitments to the nation-state.
Thus, one wonders if Israel can rely only upon Jewish citizens? Can a
Gypsy also be Hungarian? Are Finns of Swedish extraction loyal to Fin-
land? Can Francophones be as interested in the welfare of Canada as En-
glish speakers? Can a Turkish immigrant (and his children) be as German as
an ethnic German? Can Chinese, Filipino, and Mexican immigrants also be
good Americans? The national loyalty question arises because “social ho-
mogeneity and political consensus are regarded as prerequisites for, or fac-
tors strongly conducive to, stable democracies” (Lijphart, 1977, p. 1). These
questions are crucial because the modern nation-state is predominantly 
(if not universally) organized around ethnic definitions of citizenship
(Horowitz, 1985). Yet few nations are homogeneous and, as migration in-
creases, ethnic diversity creates more opportunities for conflict within them
along ethnic, racial, religious, and linguistic divides (hereafter referred to
101


collectively as “communal” differences). A better understanding of the
manner in which people are able to negotiate loyalty to their ethnic particu-
larisms while adopting or retaining the civic culture and loyalty demanded
by the nation-state seems as worthwhile today as it ever was in the past.
This is especially true given that it is the institutions of the state that pro-
vide the principal framework within which communal differences are re-
solved. This chapter examines the relationship between ethnic and national
identities within the United States.
Diversity as a Source of Conflict
While we recognize a history of conflict among communal groups, the
ethos of American society treats such behavior as aberrant, largely idiosyn-
cratic individual choices. Communal conflict is seen as a “breakdown” of a
normal condition. We argue the contrary: communal conflict is a manifes-
tation of a systematic tendency for: (1) groups to form social identities
around communal differences, (2) there to be an inverse relationship be-
tween the strength of communal and common identities, and (3) both mate-
rial and symbolic differences between groups to become occasions for con-
flict. The thesis underlying the following analysis is that social diversity
shapes social identities that provide material and symbolic reasons for com-
munal conflict in the United States. Further, the thesis asserts that conflict
is an unavoidable, systematic by-product of social diversity itself. Material
interests that coincide with social identities (especially when they are redis-
tributive) will make communal conflict more common and severe. Institu-
tions can moderate its severity. A civic and social culture can be constructed
that pronounces communal conflict illegitimate, thereby reducing it to a
minimum. But it will always be present because social identities will always
be salient to some fraction of the society, and any material interest that co-
incides with a communal distinction will insert that communal difference
into the conflict because of the sensitivity of people to communal distinc-
tions (Lijphart, 1977; Powell, 1982).
Ethnic, Racial, and National Loyalty
This chapter analyzes ethnic conflict by focusing on certain relationships
between communal identity and national loyalties. We have two “nation-
ality” groups—Hispanic Americans and Asian Americans. The first is asso-
ciated with the countries of Latin America (principally Mexico), even
though many are the progeny of families that have been in the United
States since the early 18th century. Although some Asian Americans also
come from families that are long-term residents of the United States, most
are recent immigrants. Consequently both fit into the conditions that are
102
The Contribution of Ethnic and National Identities to Political Conflict


presumed to create the “divided loyalty” often attributed to identifiable
subnational communities. 
African Americans have a singular history. Even though many can trace
their time in the United States to the 17th century, slavery and segregation
have made them strangers in the land. Africa may not exert a sense of iden-
tity for many, but the separation from the larger society that was forced on
them for almost three centuries created a group experience that is, we be-
lieve, functionally equivalent to being placed on the margins of the society. 
Specifically, this chapter examines the relationship between national loy-
alty and a preference for one’s ethnic and “racial” group. Surprisingly little
empirical research has been directed at this issue. Within the American
context, most thinking and research directed at this question falls into two
categories: the classical “melting pot” model of incorporation and an alter-
native “pluralist” perspective, with the latter giving rise to multicultural 
visions of the social order. 
The rarely examined assumption of these images of American society is
that there is no inherent conflict between subgroup and national identity. A
few attempts to examine the relationship have usually reported a compati-
bility, sometimes an easy one, between these identities (de la Garza, Falcon,
& Garcia, 1996). However, this assumption and the empirical evidence in
favor of it flies in the face of emerging work on social dominance, which
finds incompatibility between subgroup and national social identities. In
general, this work suggests that national identity is generally associated with
communal identity for one group and the exclusion of others (Sidanius,
1993; Sidanius, Feshbach, Levin, & Pratto, 1997; Sidanius & Pratto, 1999).
The purpose of this chapter is to further examine the linkage between com-
munal identity and national loyalties in the United States.
The United States is a hard case because disaffection is limited and disaf-
fected groups are few in number. The extreme group cleavages that assure a
negative association between commitment to common and subgroup iden-
tities are absent in the United States. Moreover, the prevailing civic culture
emphasizes the compatibility of subgroup and common identities and loy-
alty. Yet, as the following data show, such negative relationships actually ex-
ist. The nuances of this relationship and their implications for multiethnic
states are elaborated below.
A Thematic History of American Identity and Ethnic Diversity
The earliest comparative studies of mass politics quickly established that
conflict in even the most well-ordered democratic societies closely followed
communal divisions. Religious, ethnic, racial, linguistic, and regional differ-
ences shaped social and political debate more strongly than most other dis-
tinctions (class, for example; see Lipset & Rokkan, 1967; Horowitz, 1985;
Lijphart, 1979; Petrocik, 1981, 1998). The history of the United States fol-
Communal and National Identity in a Multiethnic State
103


lows this pattern (see, e.g., Kleppner, 1970; Kelley, 1979). Communal con-
flict was fueled by cycles of immigration that transformed the nation from a
population of largely English Protestants (with some dissenter religions) to
one that included continental nationalities, Catholics, and Jews. Making
even the most modest assumption about fertility patterns, it seems likely
that immigrants and first-generation Americans (who could be expected to
be more like their immigrant parents than the children of “old stock”
Americans) represented upwards of a third of the population of the country
by 1900.
American politics was affected by this diversity from the earliest days of
the republic. The Alien Act of 1798, for example, was passed by the declin-
ing Federalists to limit the ability of immigrants to vote in the hope that the
growth of the Democratic-Republicans would be slowed. Their lack of suc-
cess did not forestall many further attempts to limit the impact of immigra-
tion and population diversity on an originally English, Protestant society.
The Irish, and Catholics in general, were the first to redefine what America
could represent; central and north Europeans came next; southern and east-
ern Europeans were the major group for about 60 years in the late 1800s
and early 1900s; immigration from Asia during this same period was impor-
tant in the west. In different ways and at different times, this population di-
versification led to new political parties and legislation to limit the number
of immigrants and their influence on American life. 
The surge in immigration in the late 1980s—the immigration rate was
2.3 per thousand in 1980 but 6.1 per thousand in 1990—reawakened latent
sensitivity to the new “strangers in our midst.” This reawakening has been
especially pronounced in California, where the number of new immigrants
with a non-European “racial” and ethnic heritage is very large. The new
immigrants, like their predecessors, are perceived by the native population
to be substantially different, raising familiar questions about how they fit
into the society, where their loyalties and commitments lie, what changes
should be made to accommodate them, and what costs they will impose on
the society. School boards, zoning boards, city councils, utilities regulators,
local and state elections, and presidential politics have all been forums for
conflicts between Americans and the new immigrants. Some of the disputes
are material, turning on taxes and social welfare costs; others are symbolic
(e.g., Chinese residents of the San Gabriel Valley objected to a switch in the
telephone area code from 818 to 626 because of the numerological signifi-
cance of the numbers for the Chinese). Unresolved problems between
white and black Americans have acquired renewed significance with the 
rising tide of immigration.
Americanization and the Melting Pot
Specific disputes between natives and immigrants have always been con-
flicts over the identity of the national community: about what values, view-
104
The Contribution of Ethnic and National Identities to Political Conflict


points, and interests were compatible, and which could not be abided be-
cause they served and promoted loyalties that were hostile to a common
American identity. It rested on the expectation that different groups em-
braced significantly different values, beliefs, and perceptions. A de facto
policy of “Americanization” was the first way in which the nation dealt with
diversity. Everybody was expected to adopt the language, dress, customs,
beliefs, and loyalties of the dominant English culture that defined the
United States. By the end of the 19th century, the polity replaced “Ameri-
canization” with the “melting pot.” 
The melting pot replaced the expectation of submersion in the preexisting
English Protestant culture with a preference (not necessarily universal) for a
continuous reformulation of what constituted American culture. Immi-
grants, in the melting pot formulation, had valuable things to offer America,
and this value was to be recognized as American society incorporated the lan-
guage, cuisine, social sensibilities, and economic and political values of the
immigrant cultures. This reformulation of American culture, coupled with
marriage across the lines of national origin and religion, became the alterna-
tive resolution of ethnic diversity and the suspicions it produced. It resolved
(at least intellectually) the struggle over whether immigrants were real Amer-
icans by loosening (if not severing) the connection between WASP and
American. The result was a continually changing American culture and a
continually changing American ethnicity still expressing the basic theme
“from many, one” (see e.g., Salins, 1997). Restrictions on European (and es-
pecially German) nationality groups during World War I, immigration lim-
its after 1920, and “Americanization” campaigns (largely tied to learning En-
glish and shedding the most obvious habits and custom of the immigrant
culture) through World War II produced substantial assimilation and largely
eliminated the salience of nationality and ethnic differences by the middle
1960s (Walch, 1994; Freedman, 1996).
The Unmelted Pot of Pluralism
A more widespread recognition of the pluralism of American society is of-
ten attributed to the reemergence of a white ethnic identity in response to
the civil rights movement (Greeley, 1971, 1976; Novak, 1971), but, as Mil-
ton Gordon (1964) demonstrated almost four decades ago, and others con-
firmed (e.g., Wolfinger, 1965; Parenti, 1967), resistance to melting-pot
pressures was strong and seems to be to the present day (Alba, 1990). While
the notion of the melting pot has been the popular metaphor for most of
this century, it co-existed with a “pluralist” image of “salad bowls,” “quilts,”
or “glorious mosaics” to explain how one could be an American while 
retaining a distinct “racial” or ethnic identity.
While there is variation in the precise manner in which the term “plural-
ism” is understood, the central idea is that subgroup identities are retained.
In a pluralist society, communal segments defined by religion, language, na-
Communal and National Identity in a Multiethnic State
105


tional origin, or “race” remain salient to individuals. Catholics recognize
Protestants; the progeny of Italian immigrants see “Yankees” and Polish
Americans; blacks and whites are aware of a “racial” divide; Louisiana Ca-
juns recognize differences between themselves and their Protestant neigh-
bors. While the pluralist image expects social views and policy preferences
to vary by ethnic and “racial” segment, there is no notion that the differ-
ences are unbridgeable because the pluralist political culture expects in
dividuals in the subgroups to feel an identity and loyalty to the national
community, and accept a set of obligations to it. Some formulations of plu-
ralism expect positive supra-communal effects from the communal identity.
Rather than contributing to “disuniting” and fractionation (Schlesinger,
1992), these distinct ethnic loyalties are thought to co-exist and even pro-
mote a common national identity through the communal identity (Powell,
1995; de la Garza et al., 1996).
The Melting Pot, Pluralism, and Conflict 
A melting-pot, immigrant society expects virtually no communal conflict
because the melting pot society has virtually no communal divisions around
which material or symbolic grievances can be politicized. Pluralism, al-
though it has a strong expectation of communal conflict and expects groups
to be sensitive to how social resources (e.g., income, education, employ-
ment opportunities, and social standing) are distributed, assumes a political
framework within which affected groups have different advantages and are
able to bring them to bear in ways that ensure that their core interests are
promoted. (Dahl, 1961, offers the classic exposition; but also see Polsby,
1963, for a significant elaboration of the basic theory that guided virtually
all research on the topic.) Every turn to government may energize inter-
group conflicts because such collective action enforces contributions that
some groups will not prefer to make, especially when the outcome is redis-
tributive. In the textbook model of political pluralism, all groups participate
equally and have a similar positive regard for the basic institutions that
structure the political competition because no group is permanently ex-
cluded or permanent losers to any other groups.
Diversity, Social Dominance, and Communal Conflict
Central to the pluralist argument is the notion that all ethnic groups be re-
garded as co-equal partners in the pursuit of the American dream. However,
in contrast to this vision, there are a cluster of theories of intergroup rela-
tions that suggest that “co-equality” never describes relationships among
salient groups within multiethnic societies. These theoretical models have
been referred to as group dominance models, where social dominance
theory is the most explicit and recent statement of this general position (see
106
The Contribution of Ethnic and National Identities to Political Conflict


Sidanius & Pratt. 1999; for related models see Blumer, 1960; Jackman,
1994). These group dominance models suggest that societies tend to be 
organized as group-based hierarchies, with dominant groups enjoying dis-
proportionate power, prestige, and privilege. This group-based hierarchy
applies to both democratic and nondemocratic states, despite differing dis-
courses concerning inclusivity or pluralism. 
Social dominance theory, as it has been developed by Sidanius and Pratto
and their colleagues (see, e.g., Pratto, Sidanius, Stallworth, & Malle, 1994;
Sidanius, 1993; Sidanius & Pratto, 1999) shares with pluralism an expecta-
tion that communal identities will not be submerged into a larger national
consciousness, but, contrary to pluralism, it expects some groups to be relatively
consistent losers in the struggle. The foundation for this expectation is social
dominance theory’s assertion that social systems are structured as group-
based social hierarchies with one or few groups at the top of the hierarchy,
with all others—sometimes in a carefully ordered fashion—arrayed below
them. The mechanisms and processes that establish and maintain the ex-
pected social hierarchy can be informal—group-based associational pat-
terns that create and limit opportunities—or formal—political and eco-
nomic institutions that promote the interest of dominant segments (see
especially Sidanius & Pratto, 1999). 
In time, institutions can become identified with specific groups rather
than as common institutions of the society, and considerable political con-
flict can turn on perceptions of who is served by them. Examples of this in
the United States are easy to observe. Social welfare agencies are popularly
identified with minority ethnic groups. White Americans often resist finan-
cial pleas from urban school districts because the school systems in big cities
are identified with minority children. And while police forces are staffed
with “officer friendlies” for whites, they are commonly regarded as hostile
and dangerous by African Americans and other minorities (see Sidanius &
Pratto, 1999, chap. 8). 
Exclusionary Patriotism
Not only do such group dominance models have much to say about how
positive and negative social value will be distributed across the group-based
social hierarchy, but these models also have implications for the general in-
terface between national and group identity. Since public institutions are
closely linked to dominant groups, evaluations of them should vary with the
salience of an individual’s identification with the dominant as compared to
the subordinate groups. Specific intergroup differences in evaluations of
political figures, the police, or particular government agencies are well
known. The interest here is in demonstrating a more general relationship
between communal identity and a common national identity. This can have
several manifestations, but our data allow us to examine only one aspect: re-
gard for the nation or patriotism.
Communal and National Identity in a Multiethnic State
107


We expect the dynamics of communal conflict in an environment of 
social dominance to create exclusionary patriotism: a situation in which 
communal and national identities are supportive and positively correlated
among socially dominant groups, but in conflict among those in subordi-
nate communal groups. Specifically, we expect the following: 
1. The prominence of the society’s public institutions in maintaining
communal hierarchies will cause dominant communal groups to be
more identified with the nation’s symbols.
2. We expect a differential relationship between the intensity of com-
munal identities and a common national identity across groups.
Among dominate groups we expect communal and national identi-
ties to be fused and positively correlated; among subordinate
groups we expect them to be at odds and negatively correlated.
3. We expect the negative relationship to be moderated by the experi-
ences of the subordinate group. That is, those further down the
communal hierarchy are likely to see more of a conflict between
their communal identity and national symbols that are associated
with the communal identity of the dominant group. The subjective
social distance among the groups predicts some of the effect. Of
course, a real history of significant conflict between the dominant
and subordinate group would lead to even stronger effects. 
This relationship should be observable even though “exclusionary pa-
triotism” is specifically contrary to the American civic culture. The key
variable is not the avowed norms but the actual operation of the society. 
Because the society functions in a way that allows and even encourages a
perception that the political order is more closely identified with European-
Americans than it is with other major racial and ethnic groups, the dynam-
ics of exclusionary patriotism will shape the relationship between subgroup
identities and attitudes toward American national identity.
Empirical Implications of the Interface Between 
National and Ethnic Attachment
The melting-pot, pluralist, and group dominance models of American soci-
ety lead to a different set of expectations concerning the interface between
national and communal identity. This chapter will explore three expecta-
tions for each model. 
• The Salience of Ethnic Identity. The melting-pot model predicts that
ethnic or “racial” identification will be a trivial social category for
most Americans. Both the pluralist and group dominance perspec-
tives predict that these social identities will remain important and
salient social identities. The salience of ethnicity should express it-
108
The Contribution of Ethnic and National Identities to Political Conflict


self not only in terms of social identity but also in terms of a gener-
alized ethnocentric bias against Americans from other ethnic
groups.
• Patriotism and Ethnicity. Both the melting-pot and pluralist models
predict that patriotic attachment to the United States will be essen-
tially equal across ethnic groups. In contrast, and for reasons ex-
plained above, the group dominance model predicts that dominant
European-Americans will have significantly higher levels of patrio-
tism than will members of subordinate groups (e.g., Asian, Latino,
or African Americans).
• The Interface Between National and Communal Identity. The melting-
pot and pluralist models predict no correlation between patriotism
and communal identity. The melting-pot model predicts no corre-
lation because it expects essentially no ethnocentrism in the society.
The variance in ethnocentrism is trivial and idiosyncratic. The plu-
ralist model of the society predicts no correlation between commu-
nal identity and regard for national institutions because the wins
and losses of one’s ethnic group are not predetermined by an inher-
ently subordinate position to dominant groups who use public in-
stitutions to enforce their preferences. Indeed, the results of de la
Garza et al. (1996) even suggest that ethnic identity can actually
promote a stronger sense of regard for the nation. However, social
dominance theory (Sidanius et al., 1997) expects the correlation 
between national and communal identity to vary across the social
status continuum. It should be positive among dominants (e.g., 
European-Americans) and negative among subordinates (e.g.,
African-Americans). T
able 5.1 summarizes the expected relation-
ships between ethnic and national attachment for each model.
Thus far, we know of only two studies that have attempted to explore the
relative plausibilities of these conflicting models (but see also Citrin et al.,
this volume). Using a nationally representative sample of Mexican Ameri-
cans, de la Garza and his colleagues (1996) found clear empirical support
for the pluralist perspective. They found that Mexican Americans who were
strongly attached to their ethnic heritage were no less patriotic than 
Mexican-Americans with weak ethnic attachments. Furthermore, their data
showed virtually no relationship between the strength of this ethnic identity
and core American values such as individualism. Even more striking was
their finding that Mexican Americans appeared slightly more patriotic than
native-born white Americans. 
A later study by Sidanius and colleagues (1997) found only partial 
support for the de la Garza findings. When comparing the relationships 
between ethnic and national identity between European- and African-
Americans, the results showed clear support for the group dominance per-
spective in two respects. First, European-Americans showed significantly
Communal and National Identity in a Multiethnic State
109


higher levels of patriotism than African-Americans. Second, there was clear
evidence of asymmetry in the correlation between patriotism and commu-
nal identity across dominant and subordinate ethnic groups. Among domi-
nants (i.e., whites), ethnocentrism was positively correlated with patriotism,
while among subordinates (i.e., blacks) ethnocentrism was negatively corre-
lated with patriotism, as the social dominance model expects. While the
dominant-subordinate asymmetry was quite consistent for the contrast be-
tween European- and African Americans; it was less consistent with respect
to the contrast between European and Latino Americans. Moreover, while
the asymmetry held among a sample of European- and Latino-American
college students, it did not hold among a national probability sample of
white and Latino adults (see Sidanius et. al., 1997). In Sidanius et al.’s
(1997) analysis of an American national probability sample, the results were
essentially consistent with the de la Garza et al. (1996) findings.
How might we explain the somewhat conflicting findings between the de
la Garza et al. (1996) and the Sidanius et al. (1997) studies? First and most
obviously, the de la Garza et al. study only compared whites and Latinos,
whereas the Sidanius et al. study also included blacks. Second, even the
slightly different findings with respect to Latinos might have something to
do with major differences in political context. In recent years, immigration
from Mexico and Central America has become a salient and contentious po-
litical issue in California, where the Sidanius et al (1997) data were col-
lected. It is quite possible that Southern California’s highly racialized public
discourse concerning illegal Latino immigration (e.g., Proposition 187) in-
duced Latinos to feel under attack and as political outsiders as African
Americans have felt for the bulk of American history. Finally, we think that
their analysis does not measure their key conceptual variable very well. The
acculturation variable, the centerpiece of their analysis, is only a weak ap-
110
The Contribution of Ethnic and National Identities to Political Conflict
table 5.1 Hypotheses Derivable from the Melting-Pot, Pluralism, and Group
Dominance Models
Theoretical Model
Hypotheses
Melting-pot
Pluralism
Group Dominance
1. Ethnicity remains strong 
No
Yes
Yes
social identity
2. Dominants will be more 
No
No
Yes
patriotic than subordinates
3. There will be a correlation 
No
No, or positive 
Yes. Positive 
between ethnocentrism and 
correlation
correlation among 
patriotic attachment to
dominant, negative 
the nation
correlation among 
subordinates


proximation of subgroup attachment. Their measure categorized individu-
als by place of birth and language facility: U.S.-born Mexican Americans
who were English speakers were regarded as the most acculturated while
foreign-born Spanish speakers were classed as the least acculturated. Accul-
turation, however, is not the relevant variable. Although there is probably a
relationship between the individual’s level of acculturation and the strength
of his or her preference for Mexicans over non-Mexican-Americans, there
is no reason to assume that it is strong. It may, in fact, be extremely weak.
Among European Americans (the Irish, Italians, Poles, etc.), identity with
national origin is poorly related to rates of endogamy, the observance of
ethnic customs, and facility with the immigrant language (Alba, 1990). A
priori, there is no compelling reason to expect Mexican Americans to be 
different. 
T
o be sure, the null results also appear with the one approximate mea-
sure of identity used in the analysis. Respondents classed as Latinos were
those who, when asked their “racial” identity, chose the option “something
else” rather than white. This group, slightly less than half of their sample,
clearly thought of themselves in ethnic terms. In our experience, those who
self-describe as Latinos are more ethnically identified than those who do
not use this appellation—and this subgroup was not less patriotic or less in-
dividualistic in their study. However, an unwillingness of the respondent to
choose the category “white” as their ethnic or racial group is a weak mea-
sure of identity, and the null results probably should not be surprising.
Consider that it is fairly common for white Americans to respond with a na-
tional origin distinction when asked their nationality, but there is tremen-
dous variability in the strength of attachment to the national origin of
Americans who call themselves Italian, Polish, or Lithuanian when report-
ing their nationality. This variance almost certainly exists among Mexican
Americans. It is at least reasonable to suspect that it does, and the weak rela-
tionship de la Garza and his colleagues report for the measure may only re-
flect the indeterminate property of the categorization.
The finding that Mexican Americans are more patriotic and individualis-
tic than native-born whites is even less persuasive. As de la Garza and his
colleagues note, the whites with whom they draw the contrast were avail-
able for the analysis only because they lived in areas with large Mexican-
American populations. These whites are almost certainly not representative
of the white population as a whole, given the patterns of residential segre-
gation that exist in the United States. We would expect such whites to be
poorer and less educated. Their social class is likely to orient them toward a
preference for government action (the exact opposite of the “individualism”
expected of the median American in the de la Garza et. al. formulation).1
The social values that permit these whites to live comfortably in Mexican-
American neighborhoods may also have negative implications for expres-
sions of patriotism, depressing, or even reversing, therefore, the expected
relationship between it and nativity and ethnicity.
Communal and National Identity in a Multiethnic State
111


Method and Design
We used five different data sets in these analyses and a total of 7,400 
respondents. 
Data set 1: UCLA Student Sample
This data set consisted of 725 randomly sampled UCLA undergraduates
collected in the fall of 1993. The sampling frame stratified a list of all regis-
tered students into four “ethnic” strata (whites, blacks, Latinos, and Asians),
and randomly sampled from each stratum. The subjects were enticed to
participate by the offer of four $50 prizes. The ethnic breakdown of all of
the samples is found in table 5.2. The students were given a questionnaire
primarily assessing their attitudes toward patriotism and their affect toward
various American ethnic groups. However, unlike the Sidanius et. al. (1997)
study, in this chapter we will explore the responses of resident-aliens as well.
Patriotism Measures.
Kosterman and Feshbach (1989) and Sidanius et al.
(1997) have shown that patriotism could be well measured by an eight-item
scale consisting of items such as: “I am proud to be an American,” and
“Every time I hear the national anthem, I feel strongly moved.” Not only
did this scale have a high degree of face validity, but it also had a fairly high
level of reliability (with a Cronbach a = .91). Because the patriotism mea-
sures were assessed using different scales across the different studies, in or-
der to aid in cross-study comparisons, we transformed all of the patriotism
measures to the same 0.00 to 1.00 scale, where 0 represents the lowest pos-
sible score a respondent can get and 1.00 represents the highest possible
score a respondent can get. 
Ethnocentrism. While communal identity can be regarded as expressing the
general salience of one’s “racial,” ethnic, or religious identity, we will use
the term “ethnocentrism” to refer to preference for communal ingroups over
communal outgroups. The ethnocentrism measures were operationalized
by use of thermometer rating scales. The respondents were asked to indi-
cate their degree of positive or negative feeling toward each of four ethnic
groups using a seven-point rating scale (7–very positive, 4–neutral, 1–very
negative). The groups were whites, African Americans, Latinos, and Asians.
T
wo different types of ethnocentrism scores were computed: (1) general-
ized ethnocentrism and (2) hierarchical ethnocentrism. Generalized ethno-
centrism (GEI) was used to measure the respondents’ degree of ethnocen-
trism against the “generalized other,” regardless of whether this generalized
other had relatively high or low ethnic status. GEI was computed simply by
taking the thermometer rating given to one’s ethnic ingroup minus the
thermometer rating given to the average ethnic outgroup. Hierarchical eth-
nocentrism (HEI), on the other hand, was designed to be sensitive to the 
112
The Contribution of Ethnic and National Identities to Political Conflict


hierarchical nature of the status differences among the ethnic groups. Thus,
HEI was measured slightly differently, depending upon whether one was a
member of a dominant or subordinate ethnic group. For whites, hierarchi-
cal ethnocentrism (HEI) was defined in precisely the same way as for gener-
alized ethnocentrism (GEI). However, consistent with the notion of sub-
ordination to a dominant social group, HEI for the each of the three
subordinates groups was defined as the positive rating given to one’s ethnic
ingroup minus the rating given to higher-status whites. 
Data Set 2: Los Angeles County Social Survey (LACSS)
These data are drawn from two years (1997 and 1998) of the Los Angeles
County Social Survey (LACSS). The LACSS is a countywide random-digit
dial telephone survey of adults living in households conducted by the Com-
puter Assisted T
elephone Interviewing unit of UCLA’s Survey Research
Center. The LACSS employs a 12 call-back procedure, systematically vary-
ing the day of the week and the time of day, before dropping any numbers
from the sample. A total of 1,232 respondents were interviewed (see table
5.2).
Patriotism Measures. T
wo items were used to operationalize patriotism: (1)
“I find the sight of the American flag very moving,” and (2) “I have great
love for the United States.” The questions were answered on a four-point
scale from “4–strongly agree” to “1–strongly disagree.” The reliability of
the scale was considered adequate (Cronbach a = .68).
Ethnocentrism.
Generalized and hierarchical ethnocentrism were defined
by use of affect differentials in the same manner as used in Data set 1, the
UCLA student sample. 
Communal and National Identity in a Multiethnic State
113
table 5.2 Data Sets and Ethnic Groups
Samples
Data Set 1
Data Set 2
Data Set 3
Data Set 4
Data Set 5
Ethnic
UCLA Students
LACSS
NES
NES
GSS
Group
1993
1997–98
1992
1996
1996
T
otal
Whites
153
539
1,905
1,325
1,015
4,937
Blacks
113
139
312
191
170
925
Latinos
98
221
207
149
54
729
Asians
72
68
28
22
10
200
Jews
32
—
—
—
—
32
Immigrants
257
265
—
—
55
577
T
otal
725
1,232
2,452
1,687
1,304
7,400


Ethnic Grievances. Another series of questions assessed respondents’ “sense
of ethnic grievance”—that is, the extent to which their ethnic or racial group
has been disadvantaged by those in power. These questions assessed whether
respondents: (1) believed that their group should have more influence in gov-
ernment; (2) believed that their group was often held back by other people;
(3) believed that the government helps particular groups and does not pay
enough attention to people in general; (4) believed that public officials from
their own group would be more likely to do things respondents approved of;
(5) would not vote for someone who was not likely to help their group; (6)
oppose tax increases that help groups other than their own; and (7) would
vote for a candidate from their own group, even if that candidate did not
agree with them on every issue. All questions except the first two used an
agree or disagree format (which also included options for don’t know and re-
fused, and neither agree or disagree if this was volunteered by the respon-
dent). The response options for the first question were should have more in-
fluence, less influence, or just about right (as well as don’t know and refused).
The response options for the second question were often held back, some-
times held back, and treated as well as anybody else (as well as don’t know and
refused). For this question, both held back and sometimes held back were
coded as “agreement” (i.e., agreeing that their group has been held back).
Data Set 3: National Election Study—1992 (NES)
The third data set was a national probability sample of Americans collected
as part of the 1992 National Election Study (NES) conducted by the Cen-
ter for Political Studies at the Institute for Social Research at the University
of Michigan. Altogether, the 1992 study interviewed 2,452 respondents and
the analysis is limited to the four major ethnic groups that are the focus of
this study (i.e., whites, African Americans, Latinos, and Asians; see table 5.2
for details.
Patriotism. Patriotism was defined by two questions that had a content that
was virtually identical to the patriotism questions used in data set 2: (1)
“When you see the American flag flying does it make you feel extremely
good, very good, somewhat good, or not very good?” (2) “How strong is
your love for your country? Extremely strong, very strong, somewhat
strong or not very strong?” (Cronbach a = .78). 
Ethnocentrism. Ethnocentrism was defined in precisely the same way in data
sets 2 and 3.
Data Set 4: National Election Study—1996 (NES)
The fourth data set was the 1996 National Election Study (NES), also con-
ducted by the Center for Political Studies at the Institute for Social Re-
114
The Contribution of Ethnic and National Identities to Political Conflict


search at the University of Michigan (N = 1,687; see table 5.2). This data
set did not contain measures of patriotism, but it did contain feeling ther-
mometers that were used to compute ethnocentrism in the same way as in
data sets 2 and 3. The 1996 NES permitted a replication of the communal
identity patterns observed in the 1992 study, and a check against a one-time
pattern.
Dataset 5: General Social Survey—1996 (GSS)
This data set (N = 1,304 ) was used to replicate earlier findings and examine
dimensions of national identity beyond what was possible with the patrio-
tism index used in the 1992 NES study (data set 3) and the LACSS samples
(data set 2). While the dimension of patriotism, or love of the nation, is a
central dimension of national attachment, it is clearly only one of several
different possible ways in which citizens can be psychologically attached to
the nation. Previous research (Kosterman & Feshbach, 1989; Sidanius et
al., 1997; Hofstetter, Feierabend, & Klicperova-Baker, 1999) has shown
that there are at least two important dimensions of national attachment: pa-
triotism (love of country and its symbols) and nationalism (the desire that
one’s country dominate other countries). Previous research also suggests
that while dominants are significantly more patriotic than subordinates,
they are not more nationalistic than subordinates (see Sidanius et al., 1997). 
The 1996 GSS permitted a further exploration (as well as replication) of
the communal-national trade-off because it contained 34 questions con-
cerning Americans’ feelings of national attachment. While the 34 national
attachment items formed a reasonable coherent and homogeneous scale
(Cronbach a = .84), examination of the items revealed distinctly different
types of national attachment within this scale. T
o further explore this issue
and before any further analyses were attempted, we first factor-analyzed
these items using an alpha factor extraction approach. While there were as
many as 10 factors with eigen values greater than 1.00, the scree test sug-
gested that only four of these factors were meaningful. Altogether these
four factors accounted for approximately 37 percent of the total variance.
The factors were interpreted as: 
1. Patriotism or pride in one’s nation (51 percent of the common vari-
ance) was primarily defined by nine items such as: “Proud of the
way democracy works in America,” “Proud of America’s economic
achievements,” “Proud of America’s political influence in the
world,” Proud of America’s scientific and technical achievements.”
2. Nationalism (20 percent of common variance) essentially embraced
a chauvinistic and belligerent attitude toward other nations and
peoples (see also Bar-tal & Staub, 1997). The items most strongly
defining this dimension were: “TV should give preference to
American films,” “Foreigners should not be allowed to buy land,”
Communal and National Identity in a Multiethnic State
115


“America should follow its own interests,” “People should support
their country even when the country is wrong,” and “People would
be better off if they were more like Americans.”
3. Unwillingness to leave America (16 percent of common variance) pri-
marily expressed people’s unwillingness to move away from
America (e.g., “Willing to move outside of America,” “Willing to
move outside of North America”).
4. Importance of American identity (13 percent of common variance), as
distinct from patriotism or pride in one’s nation, expressed people’s
sense of social identity as being American. The dimension was pri-
marily defined by items such as: “It is important to have lived in
America for life,” “It is important to have been born in America,”
“It is important to have American citizenship.”
Factor scores were computed for each respondent on each factor. Not
surprisingly, these factor scores were positively correlated, with the
strongest correlation found between “Patriotism” and “American identity”
(r = .42), and the weakest between “nationalism” and “unwilling to leave
America.” ( r = .10).
The 1996 GSS included one question that we used to assess the salience
of ethnic identification. This question read “Compared to people who are
not (your ethnicity), would you say that you feel much closer to people who
are (your ethnicity), somewhat closer to them, or not closer at all?” The
term “your ethnicity” did not actually appear—it was always replaced by the
respondent’s own ethnicity (e.g., if they said they were Asian, it read
“Asian”; if they said they were “Latino” etc.).
Results
The alternative models of American society posit ethnic differences (or in
the case of the melting-pot model, no ethnic differences) in: the salience 
of ethnic identity, group levels of patriotism, and the relationship between
patriotism and ethnocentrism.
Issue 1: The Continued Salience of Ethnic Identity
Within the melting-pot model, ethnic identification should not be a salient
social identity. We explored the validity of this assumption by examining: 
(1) the degree to which American citizens also think of themselves in
“racial”/ethnic terms; (2) ethnocentric bias against other American ethnic
groups, and (3) people’s political grievances as a function of their own 
ethnic/racial ingroups. First T
able 5.3 shows ethnicity to be a salient social
identity for most Americans. For example, some 76 percent of European-
Americans in the nation as a whole feel close (34 percent are very close) to
116
The Contribution of Ethnic and National Identities to Political Conflict


other whites. Among, African, Latino and Asian Americans, this percentage
is even greater at 89 percent, 81 percent, and 84 percent, respectively. The
same general pattern was found for Americans in Los Angeles County. How-
ever, not surprisingly and in both cases, the salience of ethnicity varied sig-
nificantly as a function of ethnic group membership. In both cases ethnicity
was more salient for the ethnic minorities than for European Americans (na-
tional data set: X2(6) = 81.71, p < .001; LA dataset X2(6) = 61.75, p < .001).
Second, using the Los Angeles and national probability samples, we ex-
amined the generalized ethnocentrism scores of each ethnic group. If one’s
ethnic/racial particularism is no longer a salient feature of people’s social
identity, as a strict reading of the melting-pot model would expect, then
Americans from one particular ethnic or “racial” group should not exhibit
ethnocentric bias against other American “racial”/ethnic groups. However,
as the data in table 5.4 show, all four major American ethnic/racial groups
display an ethnocentric bias against other groups. While European Ameri-
cans in Los Angeles appeared to be slightly less ethnocentric than members
of the other major ethnic groups (the thermometer difference was approxi-
mately 3 degrees compared to differences of 8, 10 , and 8 degrees among
blacks, Hispanics, and Asians, respectively), the level of general ethnocen-
tric bias among European Americans in the nation as a whole appeared to
be no more or less extreme than among other major ethnic groups in the
nation as a whole (i.e., an 11 degree average difference compared to an av-
erage difference of 17 degrees for African Americans, 10 degrees for Lati-
nos, and 8 degrees for Asians). This very broad and generalized level of eth-
nocentric bias is consistent with the thrust of social identity theory (T
ajfel &
Turner, 1986).
Finally, ethnic group membership also has political relevance and politi-
cal overtones. Using the 1997 LACSS, we asked respondents from each of
the major ethnic groups a series of questions concerning whether they felt
Communal and National Identity in a Multiethnic State
117
table 5.3 Feelings of Closeness to One’s Own Ethnic/Racial Group (Citizens
Only)
Whites
Blacks
Hispanics
Asians
USA
LA
USA
LA
USA
LA
USA
LA
Respondent feels:
Very close
34%
27%
70%
58%
47%
43%
46%
37%
Some closeness
42
52
19
39
34
40
38
41
Not close
24
21
11
10
19
14
17
22
T
otal
100% 100%
100% 100%
100% 100%
100% 100%
Number of cases
975
559
166
140
343
432
14
95
Source: The U.S. national data are drawn from the 1996 General Social Survey. The Los Angeles data are
from the 1997 Los Angeles County Social Survey.


that members of their own ethnic group were being fairly treated by the po-
litical system, whether public officials from their own ethnic group would
treat them more fairly, and whether they would vote for a member of their
own group, even if they disagreed with the respondent (see table 5.5). For
example, while members of the ethnic minorities are significantly more
likely than whites to feel that their own ethnic group needs more influence
in the political system (X2(3) = 15.82, p < .01), a majority of all groups feel
that their ethnic ingroups need more political influence. Similarly, a ma-
jority of all ethnic groups believe that some groups get too much help and
not enough attention is paid to what is good for all people. The only issue
where one’s ethnic identity does not seem to play a major role concerns re-
spondents’ willingness to vote for a member of one’s own ethnic group,
even when the respondent disagrees with the candidate. T
able 5.5 shows
that a distinct minority of respondents is willing to vote strictly on the basis
of ethnicity, varying from 23 percent for blacks to 6 percent for Asians.
Altogether, ethnicity remains an important and salient social identity
with clear political overtones. In general, the data give little support to the
melting-pot image, at least to the extent that the melting pot assumes that
Americans are blended and substantially unaware of their “racial” and eth-
nic differences. T
able 5.5 also makes clear that the pluralist notion of con-
tented collaboration among competing groups is at least contested. Ethnic
minorities are particularly likely to believe that they suffer at the hands of
others because of their communal identity and that members of their group
are likely to be more sympathetic and protective.
118
The Contribution of Ethnic and National Identities to Political Conflict
table 5.4 Thermometer Differences Among Ethnic/Racial Groups (Citizens
Only)
Whites
Blacks
Hispanics
Asians
USA
LA
USA
LA
USA
LA
USA
LA
Rating of:
Whites
71
81
73
79
71
75
67
74
Blacks
62
77
85
87
69
65
63
66
Hispanics
60
78
70
79
78
79
59
66
Asians
58
79
61
78
63
66
71
77
Own group
71
81
85
87
78
79
71
77
Average of
other groups
60
78
68
79
68
69
63
69
Difference
+11
+3
+17
+8
+10
+10
+8
+8
Number of cases
3151
559
495
140
343
432
61
95
Source: The U.S. national data are drawn from the 1992 and 1996 National Election Study surveys of those
years. The Los Angeles data are from the 1997 and 1998 Los Angeles County Social Survey.


Issue 2: Ethnicity and Patriotism
Our second question concerns whether the degree of patriotic attachment
to the nation is the same across all major ethnic groups. While both the
melting-pot and pluralist models would expect no differences among the
groups, the group dominance model of American society expects consider-
able differences. Dominant European Americans should have significantly
higher levels of patriotism than Hispanics, Asians, and African Americans.
In addition, since African Americans have long been, and remain, the group
at the very bottom of America’s ethnic hierarchy, the group dominance ap-
proach would expect the mean level of patriotism among African Americans
to be relatively low and no greater than that found among resident-aliens,
or people barred from participation in the political system. T
o explore these
questions, we compared patriotism as a function of ethnicity across four of
the five data sets (UCLA students, Los Angeles County residents, and two
representative samples from the nation as a whole).
The results in table 5.6 are inconsistent with both the melting-pot and
the pluralism models. Contradicting the melting-pot and the pluralist per-
spectives on America, patriotic attachment is not uniform across the ethnic
status hierarchy. However, the differences are not substantially large. More-
over, the absolute scores indicate that all groups have relatively high patrio-
tism scores. The bulk of the intergroup difference, reflected in the correla-
tion ratios (varying between .14 and .35), is due to the relatively low
Communal and National Identity in a Multiethnic State
119
table 5.5 Sense of Grievance Among Those Identifying with an Ethnic/Racial
Group (Citizens Only)
White
Black
Hispanic
Asian
Percent agreeing:
Believe that ones group needs more 
influence
58
86
71
67
Believe that he/she has been “held back” 
because of their ethnic/racial identity
27
77
88
67
Believe that some groups are helped too 
much and not enough attention is paid 
to what is good for all people
54
55
59
67
Believe that public officials from own 
group are more likely to do preferred 
things
30
41
59
56
Oppose programs that help some group 
but not mine
47
46
44
33
Would not vote for a candidate if he/she 
would not help people like respondent
55
73
61
28
Would vote for a candidate from my group 
even if I disagreed about some things
21
23
17
6
Number of cases
151
53
162
46
Source: The 1997 Los Angeles County Social Survey.


patriotism scores of the African Americans. But even their scores are not
particularly low, except in the UCLA student data set. In an absolute sense,
then, no group is substantially alienated from the American polity, but there
is a clear tendency for patriotism to be highest among European Americans
and unequivocally lower among America’s most subordinate group, African
Americans.2 
These differences survive our tests for the confounding effects of other
social attitudes. The test used demographic information such as gender, age
and socioeconomic status as proxies for confounding effects that might be
correlated to a general sense of disaffection resulting from material disad-
vantages tied to social status. The analysis was done with multiple regres-
sion that controlled for gender, age, and SES while testing for the partial 
effect of a binary coded ethnic group membership variable. In all of the re-
gression analyses, European-American ethnicity was treated as the contrast
category. Use of this coding scheme had the distinct advantage of providing
a direct test of the patriotism differences between the dominant ethnic
120
The Contribution of Ethnic and National Identities to Political Conflict
table 5.6 Patriotic Attachment to America as a Function of Ethnicity Across Four
Samples.
Ethnic Group
Resident Alien in US
More
than 5
Less than
Whites
Asians
Latinos
Jews
Blacks
yrs
5 yrs
Sample 1 (UCLA Students, 1993): F(6,703) = 16.64, p<.001, h = .35
Mean patriotism
.693
.628
.617
.555
.432
.585
.477
N
153
72
98
32
106
219
30
SD
.23
.20
.23
.23
.24
.20
.22
Sample 2 (L.A. County, 1997–98): F(4,1203) = 9.53, p<.001, h = .18
Mean patriotism
.849
.806
.818
—
.757
.763*
—
N
530
66
217
—
138
257
—
SD
.20
.19
.22
—
.25
.23
—
Sample 3 (NES-1992): F(3,2206) = 66.47, p<10–10, h = .29
Mean patriotism
.794
.712
.789
—
.589
—
—
N
1,736
26
167
—
281
—
—
SD
.22
.22
.22
—
.29
—
—
Sample 5 (GSS-1996): F(3,1204) = 8.72, p<.01, h = .14
Mean patriotism
.531
—
.505
—
.488
.480*
—
N
973
—
48
—
38
49
—
SD
.11
—
.13
—
.13
.16
—
*Indicates that length of residency for this alien group is unknown.


group (i.e., whites) and each of the subordinate ethnic groups, net of the 
effects of all other factors.
The results of these analyses for the four different data sets are found in
table 5.7. The table reports only the coefficients for the ethnic group con-
trasts (i.e., white vs. ethnic minority). The coefficients for gender, age, and
various SES measures are not individually reported because they are only
proxies for unexamined control variables that have no theoretical impor-
tance.3 The tables show a moderate degree of consistency across data sets:
after considering the effects of the major demographic factors, the patrio-
tism differences among ethnic groups in table 5.6 persist in table 5.7. This
was especially true among the college students. Without exception, the
contrasts between socially dominant whites and all of the subordinates
groups showed whites to have significantly higher patriotism scores, with,
as before, the most powerful contrast being blacks versus whites (t = –8.60, p
< .001, b = –.39). The contrasts between whites on the one hand, and Asian
and Latino Americans on the other hand, tended not to be significant
within the other general population samples. 
African Americans, though exhibiting relatively high absolute levels of pa-
triotic attachment and attaching relatively high importance to their Ameri-
can identity (see figure 5.1), express less patriotic sentiment than any of the
other ethnic groups, and they are more likely to exhibit the exclusionary con-
flict between their communal identity and national sentiment. Indeed,
African Americans tended to be no more patriotic than resident-aliens (see
table 5.6). In the student data set, the patriotism scores among African Amer-
icans were significantly lower than those found among resident-aliens in the
United States for more than five years (Scheffé post-hoc, p < .001). The effect
is apparent even after controlling for demographic differences between
groups: European Americans have a significantly greater level of patriotic 
attachment than resident-aliens (data set 1: t = –4.79, p < .001; t = –4.54, 
p < .001; dataset 2: t = –2.02 , p < .05; dataset 5: t = –1.83, p < .05). African-
Communal and National Identity in a Multiethnic State
121
table 5.7 Patriotism and Ethnocentrism Among Dominant and Subordinate-
Groups
European- vs. 
European- vs. 
European- vs. 
African-Americans
Hispanic-Americans
Asian-Americans
B
b
B
b
B
b
UCLA
–.25**
–.39**
–.06*
–.10*
–.08*
–.10*
LACSS 1997/1998
–.07**
–.11**
.01
.01
–.01
–.01
NES 1992
–.22**
–.30**
–.00
–.00
–.00
–.02
GSS 1996
–.04**
–.10**
.00
.00
NA
NA
Note: Numbers under the columns labeled “B” are standardized coefficients from a regression that controls
for gender, age, and socio-economic status. Numbers in the “b” columns are the unstandardized coefficients.
**p<.05; *p = .05


Americans, controlling for demographic differences, are never significantly
more patriotic than resident-aliens. Among university students, African
Americans had significantly lower patriotism scores than aliens residing in
the United States five years or longer (data set 1: –t = 5.20, p < .001).
Issue 3: Patriotism and Communal Identity
Both the melting-pot and pluralist models of American society expect eth-
nocentrism and national attachment to be uncorrelated. This expectation is
not met in these data. Feelings about the nation are related to communal
identity. As one would infer from social dominance theory, this relationship
is asymmetrical across the social status hierarchy. Ethnocentrism and na-
tional assessments are positively correlated for whites but negatively related
for others, especially African Americans.
The following analysis regressed patriotism on the hierarchical ethnocen-
trism (HEI) index within each major ethnic group within the data sets
where we had appropriate measures (data sets 1, 2, and 3). The results were
reasonably consistent across all three data sets, but were, again, most con-
sistent among the university students (see table 5.8). For example, the more
white students were ethnocentric against African Americans, the more pa-
triotic they were (b = .37, p < .05; see table 5.8). The same general tendency
122
The Contribution of Ethnic and National Identities to Political Conflict
figure 5.1 Importance of American Identity as a Function of Ethnicity
Whites
Latinos
Resident-aliens
Blacks
Ethnicity
Importance of American Identity


held with respect to whites’ HEI against both Latinos and Asians (i.e., b =
.44, p < .01, and b = .41, p < .05, respectively). Among subordinates, on the
other hand, the exact opposite trend was found. Without exception, ethno-
centrism among subordinates was negatively related to patriotism. Thus,
while the regression of patriotism upon ethnocentrism (HEI) among whites
against Latinos was significantly positive (i.e., b = .44, p <. 01), the corre-
sponding regression of patriotism on HEI against whites among Latinos
was negative (i.e., b = -.49, p <.01). T
o test whether or not these asymmetries
in the relationships between ethnocentrism and patriotism were statistically
significant across the status continuum, we used simple slopes analyses to
test for interaction between ethnic group and ethnocentrism (see Aiken &
West, 1993). As can be seen for the student dataset in table 5.8, these slope
differences were significant in all cases (i.e., F = 19.63, p < .001, F = 20.80, p
< .001, F = 17.78, p < .001 for the white vs. black, white vs. Latino, and
white vs. Asian contrasts, respectively).
The intergroup asymmetries were substantially weaker in the general
population but generally consistent with expectations. Only two results are
patently contrary to the exclusionary patriotism hypothesis: in the Los An-
Communal and National Identity in a Multiethnic State
123
table 5.8 Patriotism Regressed upon Hierarchical Ethnocentrism Within Domi-
nant and Subordinate Groups (Entries are unstandardized regression coefficients.)
Ethnocentrism Index among 
Ethnocentrism Index among
Interaction
Dominant Group (i.e., Whites)
Subordinate Groups
Statistics
Among
Among
Among
Among Whites
Blacks
Latinos
Asians
F-statistics
White-
White-
White-
Black-
Latino-
Asian-
for Slope
Black
Latino
Asian
Whites
White
White
Differences
UCLA Sample
.37*
–.47***
19.63**
.44**
–.49***
20.80**
.41*
–.51*
17.78**
Los Angeles Sample 1997–98
–.00n.s.
–.09n.s.
0.20n.s.
–.00n.s.
–.19*
2.05n.s.
–.20**
–.10n.s.
4.94*
NES 1992 Sample
.15**
–.21**
30.09***
.16**
–.12n.s.
8.64**
.16**
–.19n.s.
0.73n.s.
*p<.05, **p<.01 ***p<.001


geles sample white ethnocentrism toward blacks and Hispanics is unrelated
to patriotism. Every other coefficient is properly signed. The insignificant
difference between whites and Asians in the 1992 NES is a product of a
small number of cases; however, the numerical value of the differences is
substantial. Altogether, six of the nine interactions showed that the slope of
the relation between ethnocentrism and patriotism differed between domi-
nants and subordinates. For whites, seven of nine correlations between eth-
nocentrism and patriotism were positive and significant. For the other eth-
nic groups, five of nine correlations between ethnocentrism and patriotism
were negative and significant (see table 5.8).
Summary and Discussion
This paper offered three basic ways of thinking about the linkage between
communal identity and national attachment in American society: (1) the
traditional American melting-pot approach, (2) the pluralist model, and (3)
the group dominance approach. In an effort to assess the relative validity 
of these different ways of thinking about the intersection of national and
ethnic identity, we explored three issues (see table 5.1). These issues con-
cerned: (1) the salience of ethnic identity and the presence of ethnocentrism
against other groups of Americans, (2) the relationship between patriotic at-
tachment to the nation and ethnic group membership, and (3) the nature of
the interface between communal identity and patriotic attachment.
The results are mixed, but on balance more supportive of expectations
proffered by social dominance theory. Ethnic subgroup identities remain a
salient social identity for each of the major groups we examined and, consis-
tent with the expectations of social identity theory (T
ajfel & Turner, 1986),
average Americans display clear ethnocentric bias against other ethnic
groups of American citizens. Further, and inconsistent with both the melt-
ing-pot and pluralism perspectives, patriotic attachment to the nation was
generally not uniform across ethnic groups. European Americans have
stronger patriotic attachment than those from subordinate communal
groups. This tendency was particularly strong and consistent with respect
to the white/black contrast. There was not a single instance in which this
white/black contrast failed to be statistically reliable.
Finally, the pluralism image of the society was contradicted by the corre-
lation between ethnic and national attachment across ethnic groups. Within
the pluralism model of American society, communal attachments should
have no implication for national attachment. That is not the case, especially
for black Americans. Here, the data were clearly more consistent with the
group dominance thesis than with the pluralism perspective. The clear
asymmetry in the relationship between communal identity and patriotism
across the social status continuum can be regarded as a special case of ideo-
124
The Contribution of Ethnic and National Identities to Political Conflict


logical asymmetry discussed by social dominance theory and demonstrated in
a number of studies (Fang, Sidanius, & Pratto, 1998; Levin & Sidanius,
1999; Levin, Sidanius, Rabinowitz, & Federico, 1998; Sidanius, Pratto, &
Rabinowitz, 1994; Sidanius & Pratto, 1999).4 The robustness of these basic
asymmetrical findings is supported not only by their consistency with work
generated by social dominance research but also by recent work outside of
this tradition (Hofstetter et. al, 1999). 
The distinctive relationships observed for black Americans raises the
likelihood that the degree to which communal and national identity are
connected depends upon contextual and situational factors that cannot be
systematically considered in this chapter. For example, the degree to which
the group dominance perspective accurately describes intergroup status re-
lationships may well be the critical variable that links communal and extra-
communal identities. African Americans, almost uniquely, have a history of
subordination that makes it easy and reasonable to associate the institutions
of the nation-state (and sentiments such as traditional patriotism) with
abuse and oppression. In such an environment it is simple to understand
why they, relative to Hispanics, Asians, or any other group of immigrants,
are inclined to see a conflict between their communal identity and their
sentiments about the nation. Many other groups, including European im-
migrants in their first decades in this country, also suffered (Freedman,
1996), but the severity of the oppression of American blacks is relatively
unique. 
The issue that policymakers and interpreters of the American social or-
der need to address is whether the patterns characteristic of white/black
communal and national sentiments can be approximated symbolically—in
the absence of the distinctive history of blacks. We suspect that it can. Peter
Skerry’s (1993) analysis of Mexican Americans in San Antonio and Los An-
geles offers affirmative evidence on this point. He persuasively argues that
the ethnic politics practiced by Mexican Americans in San Antonio replicate
the pluralist experiences of European immigrants, with the result that San
Antonio Mexican Americans participate in and feel committed to the insti-
tutions of the society. In Los Angeles, by contrast, Skerry found an ethos
that defined Mexican Americans as oppressed, a mind-set that created an
insurgent political style, and a dramatically less supportive view of the soci-
ety’s institutions. In brief, in objectively similar conditions, an ideologically
shaped political ethos in Los Angeles created among L.A. Mexican Ameri-
cans a subordinate group perspective that established the conditions for ex-
clusionary patriotism. We suspect, although the data to demonstrate it are
lacking, that the strong asymmetry in the UCLA student data set is a reflec-
tion of the “oppressed minority” ethos of Hispanic and Asian politics in the
region. This speculation suggests that future research in this area should ex-
plore the material and perceptual situational and contextual factors associ-
ated with the linkage of communal and national identity.
Communal and National Identity in a Multiethnic State
125


Notes
1. We have doubts about the appropriateness of measuring individualism with
the questions used in their study. They are, more appropriately, viewed as mea-
sures of some attitude about the government’s responsibility for the social welfare
of the population. This is almost certainly related to an individualism-communi-
tarian attitude, but it is a related orientation, not individualism itself. The failure
of the three questions to form a reliable scale (de la Garza, 1996, p. 342) is a fur-
ther indicator of the weakness of the questions as indicators of some underlying
social orientation.
2. However, as mentioned before, unlike the other data sets examined, the 1996
GSS data set contained a much richer array of national attachment questions. These
additional questions afforded us the opportunity to examine the connection between
ethnic and national attachment across additional dimensions of national attachment.
Thus, besides the dimension of patriotism, we were also able to identify three addi-
tional dimensions of national attachment: (1) nationalism—i.e., an aggressive and
dominance-oriented stance toward other nations; (2) unwillingness to leave America,
and (3) importance of American identity. Most interestingly, the ethnic differences with
regard to these three additional dimensions of national attachment were very differ-
ent from those found for patriotism. First, there were no significant ethnic groups
differences with respect to nationalism. Second, while there was an overall difference
with respect to Unwillingness to leave America (i.e., F(3,1249) = 10.68, p < .001, h =
.16), inspection of Scheffé post-hoc comparisons showed that this difference was es-
sentially due to the fact that resident-aliens were significantly less opposed to leav-
ing the United States than all other American groups—not surprisingly.
The one national attachment dimension where the group differences were quite
different from previous results was importance of American identity. Not only did the
different ethnic groups show significant overall differences here (F(3,1290) = 13.68, p
< .001; h = .18), but the nature of these differences were qualitatively different from
those found with respect to patriotism (see figure 5.1). As seen in figure 5.1, while
one should expect resident-aliens to have the lowest level of American identity,
given what we have already seen, it is not immediately obvious that African Ameri-
cans should have the highest level of American identity. Furthermore, controlling
for demographic differences did not substantially change this picture. The analyses
showed that while the resident-aliens had significantly lower American identity
scores than the dominant European Americans (t = –3.74, p < .001), African Ameri-
cans had unambiguously greater American identity scores than the dominant Euro-
pean Americans (t = 3.50, p < .001). Rerunning these regression analyses and recod-
ing ethnicity using blacks as the contrast group revealed that, everything else being
equal, blacks also had significantly greater American identity than all groups except
Latinos.
3. Gender had a significant relationship to patriotism in all the data sets; age was
significant in every data set except the UCLA data (where age differences were at-
tenuated). Patriotism was higher among males and older people. Income, education,
and similar class measures were usually of trivial importance.
4. The ideological asymmetry hypothesis posits that while the correlations be-
tween ingroup bias and certain system-justifying ideologies will be positive among
dominants, these correlations will be negative among subordinates.
126
The Contribution of Ethnic and National Identities to Political Conflict


References
Aiken, L. S., & West, S. G. (1993). Multiple regression: Testing and interpreting interac-
tions. Newbury Park, CA: Sage Publications. 
Alba, R. D. (1990). Ethnic identity: The transformation of white America. New Haven:
Yale University Press.
Allen, J. P., & Turner, E. (1997). The ethnic quilt. Population diversity in Southern 
California. The Center for Geographical Studies, Department of Geography,
California State University, Northridge.
Bar-tal, D., & Staub, E. (1997). Patriotism in the lives of individuals and nations.
Chicago: Nelson-Hall.
Blumer, H. (1960). Race prejudice as a sense of group position. Pacific Sociological 
Review,1, 3-5.
Dahl, R. A. (1961). Who governs? New Haven: Yale University Press.
de la Garza, R. O., Falcon, A., & Garcia, F. C. (1996). Will the real Americans
please stand up: Anglo and Mexican-American support of core American politi-
cal values. American Journal of Political Science, 40, 335–351.
Fang, C. Y., Sidanius, J., & Pratto, F. (1998). Romance across the social status con-
tinuum: Interracial marriage and the ideological asymmetry effect. Journal of
Cross Cultural Psychology, 29, 290–305.
Fredrickson, G. M. (1981). White supremacy: A comparative study in American and
South African history. Oxford: Oxford University Press.
Freedman, S. G. (1996). The inheritance: How three families and America moved from
Roosevelt to Reagan and beyond. New York: Simon and Schuster.
Gordon, M. (1964). Assimilation in American life: The role of race, religion, and national
origins. New York: Oxford University Press.
Greeley, A. M. (1971). Why can’t they be like us. New York: Dutton. 
Greeley, A. M. (1976). Ethnicity: A preliminary reconnaissance. New York: Wiley.
Gurr, T. R., & Harff, B. (1994). Ethnic conflict in world politics. Boulder, CO: 
Westview.
Hofstetter, C. R., Feierabend, I. K., & Klicperova-Baker, M. (1999). Nationalism
and ethnicity: A community study. Paper delivered at the Annual Meeting of the
International Society for Political Psychology, Amsterdam, The Netherlands,
July 17–21, 1999.
Horowitz, D. L. (1985). Ethnic groups in conflict. Berkeley: University of California
Press.
Huckfeldt, R. & Kohfeld, C. W. (1989). Race and the decline of class in American poli-
tics. Urbana: University of Illinois Press.
Jackman, M. (1994). The velvet glove: Paternalism and conflict in gender, class and race
relations. Berkeley: University of California Press.
Kelley, R. (1979). The cultural pattern in American politics: The first century. New York:
Knopf.
Kleppner, P. J. (1970). The cross of culture. New York: Free Press.
Kosterman, R., & Feshbach, S. (1989). T
oward a measure of patriotic and national-
istic attitudes. Political Psychology, 10, 257–274. 
Levin, S., & Sidanius, J. (1999). Social dominance and social identity in the United
States and Israel: Ingroup favoritism or outgroup derogation? Political Psychology,
20, 99–126.
Communal and National Identity in a Multiethnic State
127


Levin, S., Sidanius, J., Rabinowitz, J. L., & Federico, C. (1998). Ethnic identity, 
legitimizing ideologies and social status: A matter of ideological asymmetry. 
Political Psychology, 19, 373–404.
Lijphart, A. (1977). Democracy in plural societies: A comparative exploration. New
Haven: Yale University Press.
Lijphart, A. (1979). Religious vs. linguistic vs. class voting: the “crucial experiment”
In comparing Belgium, Canada, South Africa, and Switzerland. American Political
Science Review, 73 (June), 442–458.
Lipset, S. M., & Rokkan, S. (1967). Cleavage structures, party systems, and voter
alignments: An introduction. In S. M. Lipset & S. Rokkan, (Eds.), Party systems
and voter alignments (pp. 1–64) New York: Free Press.
Novak, M. (1971). The rise of the unmeltable ethnics. New York: Macmillan.
Parenti, M. (1967). Ethnic politics and the persistence of ethnic voting. American
Political Science Review, 61, 717–725.
Petrocik, J. R. (1981). Party coalitions: Realignments and the decline of the New Deal
party system. Chicago: University of Chicago Press.
Petrocik, J. R. (1998). Reformulating the party coalitions: The “Christian Democra-
tic Republicans.” Paper delivered at the 1998 annual meeting of the American
Political Science Association. Sep 3-6. Boston Marriot Copley Place Hotel. 
Polsby, N. W. (1963). Community power and political theory. New Haven, Yale Uni-
versity Press.
Powell, C. L. (1995). My American journey. New York: Random House.
Powell, G. B. Jr. (1982). Contemporary democracies: Participation, stability, and violence.
Cambridge: Harvard University Press.
Pratto, F., Sidanius, J., Stallworth, L. M., and Malle, B. F. (1994). Social dominance
orientation: A personality variable predicting social and political attitudes. Jour-
nal of Personality and Social Psychology, 67, 741–763.
Salins, P. D. (1997). Assimilation, American style. New York: Basic Books.
Schlesinger, A. M., Jr. (1992). The disuniting of America. New York: Norton.
Sidanius, J. (1993). The psychology of group conflict and the dynamics of oppres-
sion: A social dominance perspective. In S. Iyengar & W. McGuire (Eds.), Explo-
rations in political psychology (pp. 183–219). Durham, NC: Duke University Press. 
Sidanius, J., Feshbach, S., Levin, S., & Pratto, F. (1997). The interface between eth-
nic and national attachment: Ethnic pluralism or ethnic dominance? Public Opin-
ion Quarterly, 61, 103–133.
Sidanius, J., & Pratto, F. (1993). The inevitability of oppression and the dynamics of
social dominance. In P. Sniderman & P. T
etlock (Eds.), Prejudice, Politics, and the
American dilemma. (pp. 173–211). Stanford University Press. 
Sidanius, J., & Pratto, F. (1999). Social dominance: An intergroup theory of social hierar-
chy and oppression. New York: Cambridge University Press.
Sidanius, J., Pratto, F., & Rabinowitz, J. (1994). Gender, ethnic status, ingroup at-
tachment and social dominance orientation. Journal of Cross-Cultural Psychology,
25, 194–216.
Skerry, P. (1993). Mexican-Americans: The ambivalent minority. New York: Free Press.
Smith, T. W. (1991). What do Americans think about Jews? Working papers on contempo-
rary anti-Semitism. New York: The American Jewish Committee, Institute of
Human Relations.
T
ajfel, H., & Turner, J. C. (1986). The social identity theory of intergroup behavior.
128
The Contribution of Ethnic and National Identities to Political Conflict


In S. Worchel & W. G. Austin (Eds.), Psychology of intergroup relations (pp. 7–24).
Chicago: Nelson-Hall.
Walch, T. (1994). Immigrant America: European ethnicity in the United States. New
York: Garland Publishing.
Wolfinger, R. E. (1965). The development and persistence of ethnic voting. Ameri-
can Political Science Review, 59: 896–908.
Communal and National Identity in a Multiethnic State
129


This page intentionally left blank 


PART III
THE CONTRIBUTION OF 
SOCIAL IDENTITY TO 
VIOLENT INTERGROUP 
CONFLICT


This page intentionally left blank 


Robert W. White
6
Social and Role Identities 
and Political Violence
Identity as a Window on Violence in Northern Ireland
Much of the violence in the contemporary world appears, on the surface, to
involve two groups of people engaged in “them versus us” behavior. From
this view, group members develop a social (group) identity in comparison
with members of some other group. This comparative self-definition leads
to heightened differences across some social category, for example, across
tribes or ethnic groups, and, ultimately, violence. Examples of intergroup
violence based on social identity might include Tutsi versus Hutu, Arab ver-
sus Israeli, and Serb versus Croat. I suspect, however, that much of this vio-
lence does not reduce to a straightforward “them versus us” situation.
It has been argued that violence in Northern Ireland, as an example, in-
volves individuals with competing social identities based on religion—of
Irish Protestants (the majority community) in conflict with Irish Catholics
(the minority community); for example, Whyte, 1990, p. 97; Cairns, 1982.
This view is subject to question. The Irish Republican Army (IRA) has
been the most violent actor in Northern Ireland since 1969 (as measured
by deaths caused). Typically, the IRA is viewed as a Catholic defense orga-
nization. Yet, the origins of the IRA lie in the Irish Protestant community,
and there have been some Protestant members of the organization (see
McKittrick, Kelters, Feeney, & Thornton, 1999, pp. 462–463; White,
1997a). 
The complexity of the political violence in Northern Ireland may be
demonstrated by a relatively straightforward examination of who has killed
whom since 1969.1 The examination shows that there are three broadly de-
fined violent actors: Irish Republican paramilitaries, Protestant (or “Loyal-
ist”—loyal to the British crown) paramilitaries, and the security forces.2
133


T
able 6.1 presents the number of persons killed by each actor for each 
“status” of victim in the Northern Ireland conflict (Sutton, 1994). This is
one of the world’s most studied conflicts, and these data are consistent with
other attempts to enumerate the dead (see White, 1997a; McKittrick, et al.,
1999, pp. 15–19, 1473–1493; Whyte 1990, p. viii).
The primary target of Protestant paramilitaries is Catholic civilians. Six
hundred and twelve (67 percent) of their 911 victims were Catholic civil-
ians. Only 26 (of 911—3 percent) of their victims were Irish Republicans.
Irish Republicans killed the most people. And, Irish Republicans killed a
large number of Protestant civilians (20 percent, or 377/1,926). However,
the single largest group of Irish Republican victims was members of the
British army. Adding in members of the Royal Ulster Constabulary (RUC),
134
The Contribution of Social Identity to Violent Intergroup Conflict
table 6.1 Agency Causing Deaths and the Status of the Victims, Northern 
Ireland, 1969–1993
Status of Agent
Security Forces
Status of
Irish Republican
Protestant
British
RUC-
Victim
Paramilitaries
Paramilitaries
Army 
UDR/RIR1
Other
T
otal
British Army
494
1
3
2
3
503
RUC
280
6
2
1
3
292
UDR/RIR
196
3
2
0
1
202
Republican
154
26
123
18
5
326
Paramilitaries
Protestant
27
61
11
2
1
102
Paramilitaries
Catholic
181
612
134
30
30
987
Civilians
Protestant
377
111
17
6
27
538
Civilians
Political
25
31
1
2
1
60
Activists
Prison
23
2
0
0
0
25
Officers
Former
62
0
0
0
0
62
Security
Officers
Other
107
58
3
1
7
176
T
otals
1926
911
296
62
78
3273
Source: Sutton, 1994.
1. The Royal Ulster Constabulary and the Ulster Defence Regiment/Royal Irish Regiment. The RUC is
the local police force in Northern Ireland. The UDR, a locally recruited regiment of the British Army, was
replaced by the Royal Irish Regiment.


the police force in Northern Ireland, and the Ulster Defence Regiment
(UDR; a locally recruited regiment of the British Army, later replaced by
the Royal Irish Regiment; RIR) shows that about half (970 of 1926) of the
victims of Republican paramilitaries were members of the security forces.
Irish Republicans do not exclusively, or even primarily, target the Protestant
community in Northern Ireland (see also White, 1997a, 1997b; Bruce,
1997; Drake, 1998).
State agents have killed a relatively large number of people in Northern
Ireland. The British Army killed 296 people. The Royal Ulster Constabu-
lary and the Ulster Defence Regiment/Royal Irish Regiment killed 62 peo-
ple. The two communities have not suffered equally at the hands of the
state. It is not surprising that state agents have killed a large number of Irish
Republican paramilitaries, given that Republicans shoot and bomb them.
However, state agents killed 164 Catholic civilians compared to 23 Protes-
tant civilians (and 131 Republican paramilitaries). In Northern Ireland the
state is not a neutral actor (see also White & White, 1995; White, 1999;
Thatcher, 1993, pp. 384–385). T
able 6.1 suggests that between 1969 and
1993, Irish Republicans and the security forces confronted each other with
violence, while Protestant paramilitaries directed their violence against
Catholic civilians (see also note 3).
My objective is to draw on the concepts of social and role identity to pro-
vide insight into the complexity of violence in places like Northern Ireland.
This violence does not reduce to a “them versus us” model, but rather in-
volves a complex mix of identities. I suspect that several conflicts in the
world today, upon in-depth examination, will be shown to be as complex as
the Northern Ireland conflict. With respect to Northern Ireland, I argue
that the violence of Protestant paramilitaries may be understood in “them
versus us,” or in Irish Protestant versus Irish Catholic, terms. Protestant
paramilitaries target their opponents because they are Catholic or Irish
Catholic. Irish Republican paramilitary activity is best understood as a com-
bination of behavior based on a social identity and behavior based on a role
identity. However, for Republicans the social identity dimension of the con-
flict is not Irish Catholic versus Irish Protestant in Northern Ireland, but
instead Irish Nationalist versus the British government and its agents, in
Britain and Ireland. Also, an Irish Republican role identity (described be-
low) complements the Irish social identity. Finally, the British view the con-
flict as an Irish problem—of Irish Protestants versus Irish Catholics (them
vs. them). This has important implications for the behavior of British elites,
British soldiers, and Protestant and Republican paramilitaries, as demon-
strated below.
Like many other conflicts in the world, the violence in Northern Ireland
is centuries in the making. Historical background is therefore presented
first. An in-depth examination of the various identities at work in Northern
Ireland follows.
Social and Role Identities and Political Violence
135


The Historical Setting: Background to the Complexity
Norman/English soldiers invaded Ireland in the 12th century. From then
until the mid-sixteenth century, political violence in Ireland was primarily
motivated by what today is referred to as national identity. The English
found a people different from themselves in language, customs, and outlook
on life. For the next four hundred and so years there were numerous violent
conflicts between native Irish and invading English (Cronin, 1980).
In the 16th century, with the Reformation, religion became important.
The Irish and Old English in Ireland remained Catholic. Most people in
Great Britain embraced Protestantism. In 1601, Irish Chiefs rebelled and
lost. A few years later they fled to the European continent. Their lands—in
Ulster in the North of Ireland—were confiscated and given to English and
Scottish settlers. In the 1640s, the native Irish in Ulster rebelled. They were
supported by the Old English in Ireland. Both groups were Catholic. From
this point religion and national identity became intertwined in Anglo-Irish
politics. When Oliver Cromwell invaded in 1649 he did so with particular,
Protestant religious zeal. His zeal was also a function of the fact that it was
Irish Catholics who were being pacified; Cromwell never treated Catholics
in England as harshly as he did Catholics in Ireland (Fraser, 1973/1997;
Cronin, 1980, pp. 488–489). There was a further confiscation of Irish land
and property, which were placed in trustworthy, Protestant hands. In the
late 1680s, William of Orange seized the throne in England. James II, a
Catholic, fled to Ireland. He was supported by his fellow Catholics in Ire-
land, who sought relief from Protestant domination. William’s army de-
feated James’s army at the Battle of the Boyne, and for the third time in a
century the Irish suffered the results of a lost war (e.g., O’Leary & Mc-
Garry, 1993). 
These 17th century wars established a Protestant ascendancy in Ireland.
By the end of the century roughly 10 percent of Irish land was in Catholic
hands, though perhaps 90 percent of the population was Catholic (i.e., “na-
tive Irish”; O’Leary & McGarry, 1993, pp. 66–69). The wars were about 
religion and political and social control of Ireland. The native Irish were
virtually excluded from Irish social and political affairs. In the process, a
siege mentality was generated among the Protestant elite. Placed in control
of a conquered people were a Protestant Ascendancy that was vastly out-
numbered and constantly fearful.
In the late 18th century the political philosophy of Republicanism was
introduced into the mix. The United Irishmen, the first Irish Republicans,
were founded by Presbyterian merchants and manufacturers in Belfast and
Dublin. They were influenced by the French and American Revolutions,
and by a desire for economic and political freedom. They found allies in
Irish Catholics, who sought political and social relief. Republican ideology
is a political philosophy, not a religious or ethnic one. In the 1790s, it was a
radical philosophy. The United Irishmen sought such reforms as universal
136
The Contribution of Social Identity to Violent Intergroup Conflict


(male) suffrage and no property qualification (see Cronin, 1980). The
United Irishmen rebelled in 1798. Unlike the republicans in the United
States and France, they failed.
In response to the violence in Ireland, the Act of Union (effective Janu-
ary 1, 1801) disbanded an independent Irish parliament and created the
United Kingdom of Great Britain and Ireland. This did not end political 
violence in Ireland. Irish Nationalists rejected the Union and aspired for an
independent Irish nation. Irish Republicans sought to achieve this nation
through force of arms. There were Republican rebellions in 1803, 1848,
and 1867. The rebellions failed, but there remained in Ireland persons
committed to an independent Irish Republic. Throughout the 1870s and
1880s there was Republican-inspired violence. Throughout the 19th cen-
tury, the general Protestant community in Ireland—“Unionists”—sup-
ported the Union. This was especially so in northeast Ireland, where the
descendants of the 17th-century settlers feared that “Home Rule” would
lead to “Rome Rule.” In Belfast in the 1880s, for example, there was wide-
spread rioting in opposition to a Home Rule Bill for Ireland (see Cronin,
1980; Lyons, 1973).
Northern Ireland dates from events in 1916–1922. In 1916, there was a
rebellion the day after Easter. The rebels lost, but they reorganized as the
Irish Republican Army (IRA) and the political party Sinn Féin (founded
1905). The two organizations worked to break the connection with the
United Kingdom. They established a revolutionary Irish government in
Dublin and carried out military activities against the British presence in Ire-
land. The IRA and Sinn Féin were successful throughout most of Ireland.
In the northeast, the IRA and Sinn Féin were opposed by both the state’s
security agents—for example, the Royal Irish Constabulary—and by a
Protestant paramilitary organization, the Ulster Volunteer Force. In re-
sponse to the political unrest, the British passed the Government of Ireland
Act (1920), which partitioned Ireland. The Act created Northern Ireland,
which comprised six of the nine counties of the province of Ulster, and the
Irish Free State. The parliamentary headquarters for Northern Ireland was
established in Belfast. Dublin became the seat of power for the Free State.
At its creation, Northern Ireland was approximately two-thirds Protestant
and one-third Catholic. The Irish Free State (declared a republic in 1949),
was (and is) approximately 95 percent Catholic.
In 1921, negotiations between Republicans and British representatives
led to a treaty. The IRA and Sinn Féin split over the Anglo-Irish Treaty
(1921), and a civil war ensued. Those Republicans opposed to the treaty
maintained the IRA and Sinn Féin, and fought against former comrades
who accepted the treaty and the new Irish Free State. Pro-treaty forces won
the war, and, among other things, this cemented partition. Those who re-
mained in Sinn Féin and the IRA refused to recognize partition and the
governments in Dublin and Belfast. They also maintained their organiza-
tions. From 1939 to 1945 and from 1956 to 1962, the IRA engaged in un-
Social and Role Identities and Political Violence
137


successful paramilitary campaigns in England and Northern Ireland, in an
attempt to create a united Ireland (see Bell, 1979).
In Northern Ireland, the Protestant community embraced the new
province. The Catholic community, both in Northern Ireland and in the
Irish Free State, largely rejected it. Thus, partition created a double minority
situation. In Northern Ireland Protestants dominated a minority Catholic
community who believed they had been gerrymandered into a Protestant
state. Within Ireland as a whole the Protestant community was in the minor-
ity. The government of the Free State/Republic rejected Northern Ireland
and claimed the entire island for itself.
In the 1960s, there was a civil rights campaign in Northern Ireland. The
campaign sought to redress grievances that Nationalists/Catholics had
faced since the founding of the province. Civil rights protests led to coun-
terdemonstrations by Protestant extremists, which led to widespread riot-
ing in Northern Ireland in August 1969. The British government sent in
troops to restore order. It is out of events in the mid-to late 1960s that the
three actors of table 6.1 are derived: Irish Republican paramilitaries, Protes-
tant paramilitaries, and the security forces (Bell, 1979, 1993).
Organizationally, Irish Republicans can legitimately trace themselves to
the rebellion in 1848. Veterans of 1848 were involved in rebellious activity
in the 1860s, and some of those involved in the 1860s (and 1870s and 1880s)
took part in the 1916 Easter Rising (see Bell, 1979, 1987, p. 136). In 1969
and early 1970, the IRA and its political wing, Sinn Féin, split again and the
Provisional IRA (hereafter, IRA) and Provisional Sinn Féin (hereafter, Sinn
Féin) were created. The founders of the Provisionals were Republican vet-
erans, some with activist histories that dated to 1916. Contemporary
Protestant paramilitary organizations date from the mid-1960s and the cre-
ation of the Ulster Volunteer Force (named for the UVF of the 1910s). In
1971 the Ulster Defence Association (UDA) was started as a coordinating
body for vigilante groups. The Ulster Freedom Fighters (UFF) became the
paramilitary wing of the UDA (see Flackes & Elliott, 1989, pp. 272–277,
278).
Political violence in Northern Ireland is hundreds of years in the mak-
ing. It is not simply about religion, but instead involves a complex mix of
national identity, religion, and political aspirations. Social identities and
role identities, as described below, provide an insight into this complexity.
Identities: Social (Group) and Role
According to T
ajfel (1981, p. 255), social identity is that “part of an indi-
vidual’s self-concept that derives from his knowledge of his membership of
a social group.” T
ajfel typically refers to groups based on race, ethnicity, or
religious identification. For T
ajfel, the need to view oneself positively may
heighten differences between groups and lead to conflict. While T
ajfel fo-
138
The Contribution of Social Identity to Violent Intergroup Conflict


cuses on intergroup relations, Turner, Hogg, Oakes, Reicher, & Blackwell
(1987) extend the social identity perspective to emphasize intragroup be-
havior. They define social identity as self-categorizations that “define the
individual in terms of his or her shared similarities with members of certain
social categories in contrast to other social categories” (Turner, Oakes,
Haslam, & McGarty, 1994, p. 454). Thoits and Virshup (1997) summarize
these two perspectives by noting that they focus on “we’s,” in the sense that
social identities are “collective identities.” The focus is on how the indi-
vidual and the group mesh together. While the social groups around which
a collective “weness” develops are often described in terms of race, gender,
and ethnicity, a social identity may also develop around a political group-
ing—for example, the Ulster Volunteer Force or the Irish Republican
Army. My use of the term “social identity” is consistent with Thoits and
Virshup’s (1997) usage, in which they refer to a collective “weness” that ex-
ists in comparison to a collective “other.”
Role identities focus on “me’s” (Thoits & Virshup, 1997). Role identities
act as “internalized positional designations” (Stryker, 1980, pp. 60–61; see
also McCall and Simmons, 1978) that locate and define individuals within a
group. Role identities refer to individual-level self-conceptions of the re-
quirements attached to social positions. Individuals internalize these re-
quirements, which become identified as a part of the self. Role identities
have “salience” for individuals (Stryker, 1980, 1968). In certain situations
one role identity may be more salient for an individual than another role
identity. In addition, there is the potential for significant role conflict and
role strain—when the expectations associated with roles conflict or compete
(Stryker, 1980). My use of the term “role identity” follows the definition of
Stryker and refers to self-conceptions of the requirements attached to a so-
cial position that develop from social interaction—for example, by learning
to be a Volunteer in the Irish Republican Army from other IRA Volunteers.
The social or group (collective) identity is conceived of as a group-level
phenomenon, where the individual conceives of the self as one of “us”—
often in contrast to “them.” The focus is on cognitive processes, comparing
one’s group to some other group. As Thoits and Virshup (1997, p. 115)
state, “Group- or category-based identities are collective-level self-concep-
tions; they are identifications of the self with a collectivity.” In contrast, role
identities focus on who the individual is. Instead of comparing “us” vs.
“them,” the individual considers “Who am I?” by drawing on social interac-
tion and incorporating the expectations of other people into their self-
conceptualization. Both kinds of identity are social, but in different ways
(see also Stryker, 2000; Klandermans and de Weerd, 2000). For those who
employ the concept of social identity, the self is a product of comparison
and categorization processes in which “the perceiver appraises the self in re-
lation to others.” For those who employ the concept of role identity, the self
is a product of incorporating others’ expectations and appraisals, of apprais-
ing the self from the perspective of others—for example, in role-taking 
Social and Role Identities and Political Violence
139


exercises (see Thoits & Virshup, 1997, pp. 120–122). In terms of violent 
political activists in Northern Ireland, persons may kill on behalf of their
group, and/or they may kill because they have internalized such behavior as
part of the role of being a particular political activist. Indeed, activists may
kill because they are motivated simultaneously by both a social identity and
a role identity.
In the following, I argue that the violence of Protestant paramilitaries is
best understood from a social identity perspective. Protestant paramilitaries
act in defense of their social group by killing members of the other (Catholic)
social group. In contrast, I argue that Irish Republican paramilitaries are mo-
tivated by both social and role identities. Irish Republicans target the security
forces in Ireland because these forces represent an oppressor of their “Irish-
ness” (the British government) and because they have internalized the role
requirements of being an Irish Republican. Finally, I argue that the British
perspective may be understood in social identity terms. However, instead of
taking a “them versus us” perspective—for example, British forces versus
Irish Republican forces—the British take a “them versus them” perspective
and view the violence as an Irish problem, involving Irish Protestants versus
Irish Catholics.3
Data Sources: Accounts from Activists
The accounts below are from very different sources. None of the data were
collected with this particular research topic in mind. The accounts from
Irish Republicans are from my ongoing research on the causes and conse-
quences of small-group political violence. These accounts are taken from
interviews that centered on why the respondents became involved in the
Irish Republican movement. The Republican accounts represent my inter-
pretation of how Irish Republicans may be understood in terms of social
and role identities. This interpretation is based on interaction with Irish
Republicans over a 17-year period. I have formally interviewed perhaps 100
Irish Republicans, and have informally met with an unknown number of
others. Many of the respondents were low-level Sinn Féin workers, such as
advice center staffers. Included among the respondents are several promi-
nent Republicans and several veterans of the IRA. Many of the respondents
were born and raised in families where a parent or other significant adult
was active in Republican politics or had strong Republican sympathies (see
White, 1993; White & Fraser, 2000).
The accounts from Protestant paramilitaries, British elites, and British
soldiers are taken from secondary sources. These sources include interviews
with journalists, memoirs, and published accounts of speeches. I did not
collect these data. My interpretation of the accounts of Protestant paramili-
taries, British elites, and British soldiers, therefore, is qualified. With re-
spect to Protestant paramilitaries, I present the accounts to illustrate the
140
The Contribution of Social Identity to Violent Intergroup Conflict


conventional wisdom with respect to these people—that is, that the motive
of these paramilitaries is sectarian and is based on a strong identification
with being Protestant. This conventional wisdom is based on solid social
science research (e.g., Bruce, 1992). The statements from British elites and
British soldiers are consistent with survey data, noted below, which shows
that Northern Ireland is not a high-priority item for the British public.
This is one indicator that British people may interpret the conflict as one of
“them versus them.”
Northern Ireland: The Actors and Their Identities
In Northern Ireland, political actors represent two communities that reside
there and one external interest group: the Protestant community, the Irish
Nationalist community, and British representatives. Within the Protestant
community, there are Unionists and Loyalists. Unionists want Northern
Ireland to remain part of the United Kingdom. Loyalists are Unionists who
take an extreme position on the Union and are willing to defend it with
paramilitary activity. Irish Nationalists, who are predominantly Catholic,
seek an end to the Union and a united Ireland. Among Irish Nationalists are
Irish Republicans, who are willing to use paramilitary activity to bring
about a united Ireland. British interests are represented by political ap-
pointees, for example, the secretary of state for Northern Ireland, and per-
sons involved with the security apparatus, such as the British army.
The Protestant Community: Unionists.
In many ways, the Unionist mentality is a siege mentality. In spite of
their political power in Ireland, Unionists have always been in the minority,
and they have, therefore, always perceived their position to be threatened.
When it was created, Northern Ireland offered Unionists a gerrymandered
majority situation, but this was not guaranteed to last forever. Historically,
the Protestant ascendancy in Ireland maintained their position only with
the help of the English and British state. The British parliament created
Northern Ireland as a haven for pro-British Protestants; the Protestant
community in Northern Ireland did not do this on its own.
T
oday, most Protestants in Northern Ireland are Unionists (there are
some Catholic Unionists). In addition, many Protestants in Northern Ire-
land identify not with Ireland but with Britain. When asked in 1968, on the
eve of the current violence, which term best describes themselves (Irish,
British, Ulster, Sometimes British-Sometimes Irish, Anglo-Irish), only 39
percent of Protestants chose “British.” An additional 32 percent chose “Ul-
ster.” Perhaps surprising, the first choice of 20 percent of Protestants to this
question was “Irish” (see Rose, 1971, p. 485). In 1978, in the midst of a vio-
lent conflict, the question was repeated on a survey by Edward Moxon-
Social and Role Identities and Political Violence
141


Browne. Moxon-Browne found that the number of Protestants whose first
identity was “British” increased significantly, to 67 percent. The number
who chose “Ulster” fell to 20 percent, and only 8 percent chose “Irish”
(Moxon-Browne 1983, pp. 5–7).
The Protestant Community: Loyalists
Loyalists (loyal to the British crown) tend to be working-class Unionists
who view violent opposition to Irish nationalism as both justified and ne-
cessary; they support, or are, Protestant paramilitaries (see Bruce, 1992, 
pp. 14–15; Nelson, 1984). There is general agreement that Protestant para-
militaries are motivated by sectarianism, by deep-seated mistrust, and by
hatred of Catholics. Steve Bruce, an authority on Protestant paramilitaries,
stated in 1994:
Although leaders of the UDA [Ulster Defence Association] and UVF [Ulster
Volunteer Force] have periodically tried to confine the violence of their mem-
bers to some notion of “legitimate targets” (that is, republicans) or have
claimed to be trying so to restrict the deaths, very many of their victims were
chosen simply because they were Catholics. Even in these more publicity-
conscious days there are loyalist paramilitary leaders who say that they like to
kill republicans but an ordinary Catholic will do. (1994, p. 125)
The data in table 6.1 support this interpretation. The majority of victims
of Protestant paramilitaries are Catholic civilians. Because Catholics tend to
be Irish Nationalists, they are by (Protestant paramilitary) definition trai-
tors and subject to violence. Protestant paramilitaries punish the general
Catholic community when they are threatened or want to hit back. Because
Irish Republicans are hard to target, the objective is to create enough fear in
the Catholic community that they, in turn, will force the Republican para-
militaries to quit. At its most basic level, Loyalists view themselves as fight-
ing for tribal survival (Bruce, 1992, p. 55). 
In identity terms, Protestant paramilitaries view the world from a
Protestant social identity, in opposition to Catholics. Consider the follow-
ing from a former Protestant paramilitary, John White. White was a lead-
ing member of the Ulster Freedom Fighters (UFF). He describes his up-
bringing:
I remember my mother, an uneducated working-class woman, saying to me,
Son, we don’t go into a united Ireland, and me saying, why not, Mammy, and
I remember her saying, because you’ll not get a job, and you’ll probably not
get a house, and they’ll murder us, you know. And my Da, he just had a con-
firmed hatred for Roman Catholicism. I remember one time my brother
brought a Roman Catholic to the house, a girlfriend, and my Da told her to
get out of the house and not come back again. So, you know, I thought there
must be some genuine fears there, when even my Da was saying, no bringing
T
aigs home to this house. (Stevenson, 1996, p. 62)4
142
The Contribution of Social Identity to Violent Intergroup Conflict


In 1976, White confessed to the 1973 murders of Paddy Wilson and
Irene Andrews. Wilson was a member of the Social, Democratic, and
Labour Party (SDLP), a moderate, middle-class political party associated
with the Catholic community in Northern Ireland. The SDLP then, and
now, advocated a united Ireland, but through peaceful means. Now out of
prison, John White commented on why Wilson, who was publicly known
to be unarmed, was a target:
My analysis then was that he [Wilson] supported the concept of a united Ire-
land; the IRA were trying to pursue a united Ireland; therefore, by his
rhetoric, he was supporting the IRA. It may have been warped at that time,
but that was the way I viewed it when I was twenty two years of age. (Steven-
son, 1996, pp. 59–61)
John White’s account is not unique among former Protestant paramili-
taries. The following is from Alex Calderwood, a veteran of the Ulster De-
fence Association (UDA). He states, “I grew up hating Catholics.” 
I admit that I was a sectarian bigot. . . . It was my background and the soci-
ety that programmed me in a sense, but I was responsible for what I done. I
was a bigot and I hated Catholics. (Stevenson, 1996, p. 65)
A central theme in the accounts of Protestant paramilitaries is a sense of
“us versus them,” Protestant versus Catholic. The account of another for-
mer Protestant paramilitary, William Smith, is consistent with this:
In the seventies these communities were close-knit communities. . . . You
still had the extended family situation, and word-of-mouth and rumor was far
more valid than printed matter or television. But there’s no one thing, it was
the whole situation at the that time. You had segregated schools, you had seg-
regated areas, you had a whole history dealt down to you by word-of-mouth
about troubles, the IRA, and all that—that was all part of your dogma the
whole way through your life. So whenever the bogeyman appeared in 1970,
this was the bogeyman we were told about. The bogeyman appeared and the
rest became part of life—there was thousands and thousands, I mean, the vast
majority of people were involved in it. So when your community felt threat-
ened, you felt threatened. (Stevenson, 1996, p. 67)
Based on these accounts, it appears that persons who became Protestant
paramilitaries were motivated by an “us versus them” perspective. These
accounts suggest that the paramilitaries learned this perspective because
they were encouraged, by family, friends, and the Protestant community
around them, to view things comparatively—Protestant versus Catholic.
Thus, Catholics are people to be hated because “they’ll murder us.” 
Beginning in the early 1970s, large numbers of people were recruited
into Protestant paramilitary organizations like the Ulster Volunteer Force,
the Ulster Defence Association, and the Ulster Freedom Fighters. It can be
expected that members of these groups developed social identities that were
particular to these groups—that is in addition to their Protestant social
Social and Role Identities and Political Violence
143


identity they probably also developed social/collective identities based on
their membership in a particular paramilitary group. As members of these
groups they may also have developed role identities (see below). However,
the key to understanding these groups and their members remains their
Protestant social identity. This social identity is common to most if not all
Protestant paramilitaries, and is probably the most important motivator of
their behavior.
The Irish Community: Nationalists
An Irish Nationalist is a person who wants all of Ireland to be ruled by Irish
people (Irish people being defined as people living in Ireland, even if they
identify themselves as British). Irish Nationalism predates the Reformation
and the Act of Union (effective January 1, 1801; Cronin, 1980). T
oday, most
Irish Nationalists are Catholic. With respect to Northern Irish politics, the
terms “Catholic” and “Nationalist” are often used interchangeably, but they
are not synonymous. Catholics in Northern Ireland do not exclusively iden-
tify themselves as Irish. When Richard Rose (1971) surveyed the Northern
Irish public in 1968, he found that when asked which term “best describes
the way you think of yourself” (British, Irish, Ulster, sometimes British-
sometimes Irish, Anglo-Irish), 76 percent of Northern Irish Catholics chose
the “Irish” option. Only 15 percent of Catholics identified themselves as
“British,” and even fewer (5 percent) identified themselves as “Ulster.” T
en
years later, in the midst of violence, when Edward Moxon-Browne included
the question on his survey, the percentage had fallen; 69 percent of
Catholics chose “Irish” as their first option. Again, only 15 percent of
Catholics identified themselves as “British,” and few (6 percent in 1978)
identified themselves as “Ulster.” Also in 1978, when asked “which of these
would you least like to call yourself? British, Irish, Ulster, sometimes
British-sometimes Irish, Anglo-Irish, Other?”, less than 2 percent (1.7 per-
cent) of Catholics chose “Irish”—that is, very few Catholics rejected the
Irish national identity (see Moxon-Browne, 1983, pp. 6–9). In contrast to
the Protestant community, the national identity of Catholics/Nationalists in
Northern Ireland appears to be more stable (see also Moxon-Browne,
1983). 
The Irish Community: Irish Republicans
At its most basic level, an Irish Republican is someone who “seeks to sepa-
rate Ireland from England by force” (Cronin, 1980, p. 23). At their core, all
Republicans agree on two things: (1) the British government is the cause of
political strife in Ireland, and (2) a British withdrawal from Ireland can be
gained only through force of arms (see White & Fraser, 2000). These be-
liefs are evident in accounts from Republicans and may be interpreted in
both social and role identity terms. 
144
The Contribution of Social Identity to Violent Intergroup Conflict


Consider the following from Respondent 1. He describes events that led
him into the Irish Republican Army. Recognizing that he is Irish and that
the British were killing “our people” was essential for his recruitment into
the IRA. In the early 1970s he went to school in England:
RESPONDENT 1 [Hereafter, R]: I went to England and as I say everything
then was at a distance, it wasn’t as immediate. . . . But then Bloody Sunday
[in which British troops shot dead 14 Nationalist civilian protesters, in 1972]
happened and again this incredible sense of outrage. And then, just through-
out the year, watching the British, some of the British people—their attitudes
to us, the Irish. And it was just—fuck sake. The attitudes were just that—
you’re “Paddy.” They automatically call you Paddy when you go over there.
I’d say, “My name’s Ruairí.” “Sure, you’re Paddy. You’re all called Paddy and
Mick.” I said, “No, my name’s Ruairí.” I took it from my friends there because
it was usually a joke, there was no malevolence in it. But, there were people
who used it as a derogatory term. And just throughout the year, watching
things going at home. And following along Bloody Sunday I decided that
what I was doing, which I wasn’t happy with anyway, that I was wanting to go
home more and more. And then when I left England to travel for a couple of
months, hitched around, and I came home. And as soon as I came home I got
into Sinn Féin and it was a natural progression from there. Because my beliefs
were fairly well formulated at the time. The British were killing our people, they
were locking them up and they were nothing more than Stormont [the
Northern Ireland parliament]. (emphasis added) 
Similar comments were offered by a Protestant Republican (see also Hynd-
man, 1996). He was born and raised in East Belfast. By his background he
was an unlikely candidate for involvement with Irish Republicans. When he
moved to England, he discovered that he was Irish.
INTERVIEWER [Hereafter, I]: You mentioned our people, you said they
were killing our people or shooting our people. Do you consider yourself
Irish? 
R: Oh yes, I am Irish you see. 
I: But I mean, the typical—
R: (Laughs) I know it’s uh—I understand what you mean. Given my back-
ground, right? For a long time I thought I was British. I didn’t really know
what I was, actually. When I was younger I went to England to work due to
the high unemployment at home. And, then I was made very aware of my iden-
tity, and it was Irish. I was left in no doubt that I was a Paddy. And uh, you
know the old stories about people going to try to find flats, you know, some-
where to live in London, where the signs are in the window, “No Irish,” “No
Coloureds.” It was a good thing to experience because it did give me a per-
ception, it did give me an education which I might not necessarily have gotten
if I had stayed in the North [of Ireland]. But certainly that made me very
aware of my identity. Then moving from my Loyalist area, through other lo-
cations, finally to end up in a Republican area was a great experience. Living
there which had always—had been a place which I had visited quite a lot, but
Social and Role Identities and Political Violence
145


living in it is a lot different from visiting. T
o experience the continual daily
harassment, the constant harassment. It’s a great education as well. T
o see the
poverty as well. [RESPONDENT 2, emphasis added]
These two respondents acquired an Irish social identity by personally expe-
riencing discrimination and devaluation of their Irishness. The develop-
ment of this social identity helped to make them potential recruits to Irish
Republicanism. Their recruitment was also influenced by their acquiring a
Republican role identity.5
Respondents 1 and 2 became Republicans when they were adults. Many
Republicans, however, are born and raised in Republican families and 
they are influenced by Republicanism at an early age. One result of this in-
fluence is an Irish social identity. Respondent 3 was raised in a Republican
family: 
R: the only solution to our problems, and it has been my opinion all down
through the years, is a united Ireland. It’s the only way forward for us. There
is no other way. The other way is a creeping, cringing, crawling spirit of the
Irish nation. We desire freedom, we seek freedom, and it’s not too much to
demand. All we’re seeking is our own country. . . . I think it only makes
good common sense that if you are the owner of a house, I always make this
analogy, that if you’ve got a four-roomed house, and there’s one person occu-
pying the house. And there’s quite a large family of you. And you make a de-
mocratic decision time after time for the person that’s occupying your house,
undemocratically, for to remove . . . to get out of the house. That democra-
tic decision is taken. But the man won’t move. What’s your alternative? Get
him by the scruff of the neck and throw him out. It’s the only way. Our people
has taken the democratic decision in this country here several times to ask Britain to
leave this country. So as we can apply our own national government. But she
hasn’t adhered to that democratic decision, and that’s why this struggle is still
going on. (Respondent 3, emphasis added)
Respondent 4 was also raised in a Republican family. He stated the following:
R: [When I joined, my goal for the Movement] was to see the complete free-
dom of this country. In other words, what the Republican goal always was, to
break the connection with England and to establish an Irish Republic. An Irish So-
cialist Republic. And that was my reason for joining the Movement and that is
why I got involved. (Respondent 4, emphasis added)
Like Protestant paramilitaries, Republicans—both from Republican fami-
lies and not—develop a social identity that places them in an “us versus them”
situation. The “them,” however, is the British government, not the Protes-
tant community in Northern Ireland. 
There is evidence that Irish Republicans develop a role identity that
complements their Irish social identity, and that this role identity develops
through social interaction with family members and/or peers who are Re-
publicans. Consider the following from the three respondents raised in 
Republican families:
146
The Contribution of Social Identity to Violent Intergroup Conflict


I: Why in general did you get involved in the Movement?
R: Well, Bob, I was involved with the Republican Movement at a very early
age because of my father being a member of the Republican Movement as
well. From 1916 ‘til the Civil War my home was a place for anybody who
came on the run, or to seek help. It was always given to them. So I was sort of
brought up, you know what I mean, within Republican circles, my father be-
ing a Republican. And uh, I suppose it rubbed off on me. I always had Repub-
lican ideals, you know? (Respondent 3)
R: Well, the reason why I decided to get involved in the Republican Move-
ment was that I was born into the Republican tradition. Both on my father’s
and my mother’s side. And that is one of the reasons. And reading Irish his-
tory I realized and came to the conclusion that the Republican Movement
were the true inheritance of the Republican faith. That is the reason why I 
became [involved]. 
. . .
R: [Reading the question] Did any particular person have a strong influence—
eh, to get involved? Well, I wouldn’t just say to get involved, but, there was
one woman, she was my grandmother. Incidentally my four grandparents
were teachers. But this particular—my mother’s mother, she was an outstand-
ing republican sort of thing, you know? When she died in 1969—She never
wavered. . . . She was strong in her Republican [beliefs] to the day she died.
And when she—her house, as a matter of fact, was the headquarters of the
Irish Republican Army [in her area]. . . . she was a very highly respected
woman sort of thing there. . . . And she had a big influence on my life there
because as you may know you’re—well, in this country anyway your grand-
children go on holidays with their grandparents. . . . She never wavered, as
I said. And she had a big influence. And even up to this day, if any questions
come up [I would refer to her Republicanism]. (Respondent 4)
R: Republicanism in our house was a—sort of what would be termed now as
pure Republicanism, in that my father wouldn’t of been very sympathetic to-
ward a socialist line of thinking. It would of been more of a Republican prin-
cipled directly. And, uh, seen the great danger in socialist tendencies within it.
And, to some extent that would of influenced my direction for a number of
years. And even yet to some extent that you would fear a socialist—too big a
socialist involvement within it. And uh, he was very much just a straight line
Republican. 
I: Is that why you would have gone with Provisional rather than Official
[when the Republican Movement split in 1969/70]?
R: That’s right. That would of been a big influence actually. And uh, not only
from hearing it in the past but actually hearing him talking about it at that
particular time, eh, that the split come. My direction like would of been influ-
enced a big part by his sort of talking to me. (Respondent 5)
It appears that these children of Republicans were raised such that they ac-
quired Republican principles and learned the role of being a Republican—
for example, to help other Republicans “on the run” and to never “waver”
in your commitment. This acquisition and learning occurred through inter-
Social and Role Identities and Political Violence
147


action with other Republicans. Accounts from Republicans not born and
raised in Republican families suggest that they also acquire a role identity
through interaction with both their family members (even though they are
not Republican) and their peers who are Republicans. The following is
from Respondent 1. He is not from a Republican family. He joined the IRA
in the early 1970s. He was strongly influenced by his personal experiences
with peers who were involved in the Irish Republican Movement. His in-
volvement stems from a family background that stressed being a responsible
person, and his social interaction with people who he believed were being
responsible—that is, persons who were willing to risk their lives in a just
struggle against a much stronger opponent (and his development of an Irish
social identity, as noted earlier).6
R: The main reason that I got involved with the Republican Movement is that
I was brought up to believe that you should treat people as equals and that no-
body had the right to walk over you, nor did you, more importantly, did you
have the right to walk over anybody else. And the Republican Movement, it
may have historical background, yes, but essentially it got off the ground be-
cause it came out of the community and it was a community that accepted it
because it was its own people. And therefore I suppose you could say it was
my contemporaries that, if anybody influenced, would have influenced me to
join. But it was just a natural thing because it was the time of ferment, of po-
litical ferment that was in it. I felt that I couldn’t close my eyes to it and that I
had a responsibility to do something. I’m not sure, mind you, that would be
everybody’s reply—it sounds terribly altruistic [laughs]. But the point is that
that’s just the way I felt about it. I just felt that I had had a pretty good life
compared to most of the lads that would have been in the movement in Derry.
I come from a middle class background and I never really had to worry about
anything and I just felt there was a responsibility for me to do something. But
it wasn’t as clear cut as that. 
. . .
R: I didn’t join because my friends were in the movement. That’s not the rea-
son I joined. I was motivated to join the Republican Movement because I
knew that the people who were in it were sincere. I mean, a 19-year-old kid or
an 18-year-old kid who makes a conscious decision to go out and fight a vastly
superior army on the streets, to take on a government and all its forms, and to
change society—He may not think in terms of vast political thought about
how society’s going to be changed by him. But if he makes that conscious de-
cision instead of happily waiting until he get his welfare check, and then goes
off and plays pool or drinks beer and goes to dances—I mean they’re de-
nounced as being mindless thugs but that is just not true. It could never for
me be a description of them. (Respondent 1)
As it was for Protestant paramilitary organizations, in the early 1970s
Irish Republican organizations received large numbers of recruits. As with
Protestant paramilitaries, the new Republican paramilitary recruits were in-
fluenced by a social identity. Unlike Protestant paramilitaries, the Republi-
148
The Contribution of Social Identity to Violent Intergroup Conflict


can recruits did not interpret their activity as Catholic versus Protestant,
but rather as Irish versus British. In addition, and also unlike Protestant
paramilitaries, the Republican paramilitary recruits joined a movement with
a lengthy history. Along with this history came established roles for mem-
bers of the Republican Movement—“she never wavered.” Out of interac-
tion with persons who were already Republican (be they family members 
or peers) the Republicans acquired a role identity that complemented their
social identity.
The British Perspective:“Them vs.Them”
The British interpretation of the violence is that it is an Irish problem in
which members of two Irish communities, Protestant and Catholic, cannot
get along. This is implied in the accounts of Respondent 1 and Respondent
2, one Catholic and the other Protestant, but both “Paddy” in Britain. In ef-
fect, the British deny the social identity of Protestant paramilitaries, who
view themselves as British, and they deny the political aspirations of Irish
Republican paramilitaries, who do not want to be citizens of the United
Kingdom. 
This is not a new phenomenon. In 1921, on a visit to Belfast, King
George V said:
I speak from a full heart when I pray that my coming to Ireland today may
prove to be the first step towards an end of strife amongst her peoples, what-
ever their race or creed. In that hope I appeal to all Irishmen to pause, to
stretch out the hand of forgiveness and conciliation, to forgive and forget and
to join in making for the land which they love a new era of peace, content-
ment and goodwill. (see Lyons 1973, p. 426)
Fifty years later, Reginald Maudling was the British Home Secretary re-
sponsible for Northern Ireland affairs at Westminster (from 1970-1972). 
In his memoirs, he described how the “Irish” problem perplexed him, an
Englishman.
It is very hard for an Englishman to understand the feelings of those who live
in Northern Ireland. The history of their struggles is a long one, and they
tend to cherish every moment of hatred in it. The deep divide between
Catholic and Protestant, which, incidentally, is more racial than religious, had
been handed down faithfully from generation to generation. (Maudling 1978,
p. 180)
Traditionally, Ireland is not that important for British elites. This makes
the violence that much more an “Irish problem.” After helping execute
Charles I, Oliver Cromwell spent nine and one-half months in Ireland,
pacified the country, and then returned to the real business at hand—
running England (Fraser, 1973/1997). In the early part of this century,
there were a series of debates and proposals associated with the Home Rule
[for Ireland] Bill in 1914. William Manchester, in his biography of Winston
Social and Role Identities and Political Violence
149


Churchill, describes the Irish situation as being, “foremost in everybody’s
mind, so certain to burst into flames at any moment” (1983, p. 463). But, as
a meeting concluded Sir Edward Grey read out a document from the For-
eign Office. It was an ultimatum from Austria to Serbia, part of the begin-
ning of what became World War I. The war developed, and the Irish issue
was “deferred.” Churchill later recalled, 
The parishes of Fermanagh and Tyrone faded back into the mists and squalls
of Ireland, and a strange light began immediately, but by perceptible grada-
tions, to fall and grow upon the map of Europe. (Manchester, 1983, p. 463)
The “Irish question” also has not had an impact on the voting British
public. Northern Ireland constitutes less than 3 percent of the population
of the United Kingdom. Since 1983, the province has sent only 17 (of 635)
MPs to Westminster. In response to Gallup polls on important issues in the
United Kingdom, Northern Ireland traditionally has “fluctuated between
nil mention and coming at the bottom of a list of ten or a dozen issues” (see
O’Malley, 1990, p. 226). 
Because Northern Ireland has not been a high priority for the British
public, British soldiers, recruited from that public, appear to be relatively
unaware of the politics of the situation (at least in the early years of the re-
cent conflict). Consider the following from two soldiers sent to Northern
Ireland in 1969. 
BRITISH SOLDIER 1: The boys knew where they were supposed to go in
terms of sealing off streets, so it was then a question of getting men down to
the corners and road junctions and giving them cover in case there was shoot-
ing. We hadn’t a clue what was going on really, but the feeling I got within half
an hour was one of relief, on both sides. The Irish are terribly curious and they
came out of all sorts of corners, and I think both Protestants and Catholics
were relieved that there might be some stability. . . . (emphasis added;
Arthur, 1987, p. 7)
BRITISH SOLDIER 2: For the first week I slept underneath an awning by a
baker’s shop. I’d chosen the baker’s shop because I knew we’d get fresh bread
and buns each day. This was the period of complete daze, immediately after
the initial rioting. No one knew what was going on, so we had to start talking to the
people to find out what might happen. We felt we were terribly welcome, particu-
larly by the Protestants, but they kept on saying: ‘Why have you got your ri-
fles pointing at us, and your backs to the Catholics, when they’re the ones
who started it?’ So immediately we came up against the sectarian problem:
which way do you face? Who do you defend? Which side do you protect? (emphasis
added; Arthur, 1987, pp. 9–10)
The perspective on Ireland of British elites, citizens, and soldiers may
best be described as viewing the conflict as “them versus them.” The degree
to which Britishers identify with Britain and adopt potential role identi-
ties—“As a British soldier my job here is to keep the peace”—may be less
important than the fact that they view what happens in Northern Ireland as
150
The Contribution of Social Identity to Violent Intergroup Conflict


external to themselves, or at least not of their own making, that it is a con-
flict caused by “others” and it is not a reflection of British policies or 
actions. 
By denying the identity of those Protestants in Ulster who identify
themselves as British, British elites contribute to the fears of these people.
By refusing to recognize the political claims of the Republicans, and by ac-
tively denying these claims, British elites leave Republicans with violence as
a means of expressing their political desires. Perhaps most important, view-
ing the conflict as external, or involving “them versus them,” contributes to
policy mistakes on the part of the British government, as described below. 
Implications
In Ireland there are three violent political actors: Protestant paramilitaries,
Irish Republican paramilitaries, and state agents. Each perceives the vio-
lence in Northern Ireland from a different perspective. These perceptions
have real consequences. This is demonstrated in the British policies of
criminalization and Ulsterization, which were implemented in the mid-
1970s. Their impact is still felt today.
In 1972, IRA prisoners led by Billy McKee, an IRA veteran of the 1930s
and 1940s, went on a hunger strike in order to gain political status. In the
midst of negotiations that led to a truce between the British government
and the IRA in late June, political status was granted to both Republican
and Protestant paramilitary prisoners. Criminalization revoked this status.
After March 1, 1976, persons convicted of paramilitary crimes were treated
as ordinary criminals. The British government denied the political nature
of the IRA’s activity and equated IRA violence with the sectarian violence of
Protestant paramilitaries. Ulsterization dates from 1977 and is designed to
make the Royal Ulster Constabulary the primary policing agent in North-
ern Ireland (e.g., Flackes & Elliott, 1989, p. 392 ; see also Bell, 1993, 
p. 517). As a result, the violence became that much more an Irish problem
(of “them vs. them”).
The implications of criminalization and Ulsterization were profound.
Irish Republican prisoners rejected criminalization. They refused to wear
prison clothes and wrapped themselves only in blankets. The protest esca-
lated, ultimately resulting in the hunger-strike of 1981. In an attempt to re-
gain political status, 10 Irish Republican prisoners fasted to death (see Bell,
1993). The behavior of the Republican prisoners on the strike may be un-
derstood in terms of a role identity and role behavior. Hunger-striking has a
lengthy history among Republicans; the hunger-strike weapon is part of the
Republican repertoire. The first Republican to die on a hunger strike in this
century was Thomas Ashe, in 1917. There were deaths by Republicans on
hunger strike in the 1910s, 1920s, 1940s, and the 1970s. People like Ashe, it
can be argued, were role models. One dimension of the role identity of be-
Social and Role Identities and Political Violence
151


ing an Irish Republican is to confront the state while in prison with the
weapon of the hunger strike. Billy McKee, the senior Republican hunger
striker in Belfast in 1972, was active in the IRA in the 1940s when then 
senior Republicans were on hunger strike. Some of those on the hunger
strike in 1981 were under the command of people like McKee (if not Mc-
Kee himself) when they were new recruits to the Republican Movement in
the early 1970s. 
The hunger strike in 1981 led to a mass mobilization of the general 
Nationalist community in Ireland. One of the strikers, Bobby Sands, was
elected as an abstentionist MP to Westminster. The chief beneficiary of the
hunger strike was the Republican movement. Both Sinn Féin and the IRA
attracted large numbers of new recruits (see Bell, 1993; White, 1993). For
the British government, who held firm against the hunger strike, it was a
disaster. Ulsterization probably saved the lives of British soldiers, but it, like
criminalization, did not stop Irish Republicans, who claim that they ‘target
the uniform’ rather than the religion of members of the security forces (see
Flackes & Elliott, 1989, p. 414, table 4). While killing a British soldier may
have more of an impact, Republicans were quite willing to kill members of
the RUC (see White, 1997a, 1997b; Bruce, 1997)). 
Criminalization and Ulsterization had very different results for Protes-
tant paramilitaries. Criminalization did not lead to deaths from hunger
strikes by Protestant paramilitary prisoners. Although there have been
Protestant paramilitary prisoners on hunger strike, none have maintained a
fast to the death. There are a number of reasons for this and two are impor-
tant here. First, it must be acknowledged that the conditions for Protestant
paramilitary prisoners have never been as severe as they have been for Irish
Republican paramilitaries. Pro-British Protestant paramilitary prisoners are
less of a threat to the state than are anti-state Republican prisoners (see
White, 1999). Second, and very important for present purposes, there is no
history of hunger striking to the death by Protestant paramilitaries. The
historic and contemporary “role models” (martyrs) that Irish Republicans
have—Thomas Ashe in 1917, Sean McCaughey in 1946, Michael Gaughan
in 1974, and Bobby Sands in 1981, for example—are not available to
Protestant paramilitaries. The role identity of a Protestant paramilitary
prisoner confronting the British government by dying on hunger strike
does not exist.
Indeed, at the beginning of the violence a Protestant paramilitary role
identity could not exist because such people, and such organizations, had
not yet formed. When the Ulster Volunteer Force and the Ulster Freedom
Fighters were created in the mid-1960s and early 1970s, a previous genera-
tion of Protestant paramilitaries was not there to serve as paramilitary lead-
ers (role models). The original UVF was largely recruited into the security
forces of Northern Ireland in the 1920s. That organization ceased to exist,
and its former members became police agents of Northern Ireland (see Far-
rell, 1980, pp. 35, 44–45). While some of the parents of the Protestant para-
152
The Contribution of Social Identity to Violent Intergroup Conflict


military recruits of the 1960s and 1970s were members of the RUC and the
Ulster Defence Regiment (UDR; e.g., T
aylor, 1999, p. 82), this is qualita-
tively different from being raised in a family with a tradition of involvement
in illegal paramilitary activity. A leader of a vigilante group described their
activity at the beginning of the current violence:
We never planned to go on the kill. There was no time that we sat down and said
‘That’s it. Stiff a T
aig.’ Mind, we planned doing something to the Ardoyne af-
ter the three Scots boys [three British soldiers] were killed [in 1971] but for
some reason it never came off. No, it was ground up. One or two volunteers
just started doing it. (taken from Bruce, 1992, p. 54; emphasis added)
Contrast this with the description of Billy McKee’s activities in the early
1970s. The interviewer, the journalist Peter T
aylor, asked a Republican who
was recruited back into the IRA by McKee to comment on his activities in
the early 1970s:
What did he [McKee] tell you about the policy towards British soldiers?
He didn’t indicate that there was going to be an immediate offensive
against the British army. He said, “These things take time. People have to be
trained. People have to be motivated. People have to be equipped. All this won’t 
happen overnight.” But the intention was there and it sounded good to me.
(T
aylor, 1998, p. 74; emphasis added)
For Protestant paramilitaries, criminalization reinforced their interpre-
tation that Republicans are out to destroy their community. While Republi-
cans state that they only target uniforms, not the ethnic/religious belief of
the occupant, Protestants (paramilitaries and in general) view attacks on the
RUC as sectarian because the force is 90 to 95 percent Protestant. In re-
sponse to the killing of members of the RUC, Protestant paramilitaries
killed Catholic civilians (see White, 1997a, 1997b; Bruce, 1997).
The accounts from Protestant paramilitaries suggested that they do not
interpret their behavior in terms of a role identity. However, these accounts
are from secondary sources and social and role identities were not the focus
of those who collected the accounts. Also, the accounts are from activists
who were not born and raised in Protestant paramilitary families. The vio-
lence in Northern Ireland has spanned a generation. Many of the Protes-
tant (and Republican) paramilitaries of the 1970s and early 1980s now have
adult children. It is very possible that with the continuity of the Protestant
struggle there has developed among Protestant paramilitaries a role identity
that has been passed to another generation in the same manner as Irish Re-
publicans pass on their role identity. T
oday, the leader of the Ulster Demo-
cratic Party, the political wing of the Ulster Defence Association, is Gary
McMichael. Gary McMichael is the son of John McMichael, a leading
Protestant paramilitary who was killed by the IRA in 1987 (Cusack & Mc-
Donald 1997, p. 267; Bruce, 1992, pp. 232–233). If a Protestant paramili-
tary role identity has developed, this identity should be consistent with the
Social and Role Identities and Political Violence
153


Protestant social identity, and the basic social identity differences among
Protestant paramilitaries, Irish Republican paramilitaries, and the British
should remain. 
Conclusion
Much of the violence in the world today appears to be caused by heightened
social identities, by groups that define themselves as “us versus them.” On
the surface, violence in Northern Ireland fits this pattern. The above case
study reveals, however, that there are three significant violent actors in
Northern Ireland, not two. These actors are Protestant paramilitaries, Irish
Republican paramilitaries, and the state’s security forces. Each of the actors
interprets the violence in different social identity terms. Protestant para-
militaries see themselves as British and they define themselves as Protes-
tants in contrast to Irish Catholics. Further, they perceive that Irish Repub-
licans act in a manner consistent with an Irish Catholic identity. Irish
Republicans, in contrast, act in a manner consistent with an Irish social
identity. Their conflict is not with Irish Protestants but with British parlia-
ments that lay claim to Irish territory. They view themselves as being in a
war of national liberation against the British government. Protestant para-
militaries interpret Republican activity differently. For Protestant paramili-
taries, targeting the British in Northern Ireland means targeting Protes-
tants in Northern Ireland—they view themselves as the British in Northern
Ireland. In the midst of these different interpretations, British people have
traditionally viewed the violence in Ireland from an outsider’s perspective.
The violence is between Protestant Irish and Catholic Irish, and the British,
as the lawful government, are caught in the middle. This adds to the com-
plexity of the situation, especially because British government actions are
subject to different interpretations by Protestant paramilitaries and Irish
Republican paramilitaries.
In addition to their Irish social identity, the data suggest that there is an
Irish Republican role identity and that this influences Republicans’ behav-
ior. This role identity develops through social interaction with other Re-
publicans. Although the data did not necessarily indicate the development
of a Protestant paramilitary role identity, there is no reason to believe that
such an identity has not developed over time. Similarly, British elites and
British soldiers probably have role identities. As with Republicans, these
role identities probably develop through social interaction. 
At the moment, there is a peace process underway in Northern Ireland.
It has been a long time coming, with origins that date from events in the
mid- to late-1980s. As of yet, peace has not been fully confirmed. Involved
in this peace process are the three general actors that are the subject of this
paper—Protestant paramilitaries, Irish Republican paramilitaries, and the
British government (also involved are several other actors, including non-
154
The Contribution of Social Identity to Violent Intergroup Conflict


violent political activists and the Irish Government in Dublin). The differ-
ent perceptions that the actors have of themselves and their counterparts
contribute to the difficulty of achieving a lasting peace in Ireland.
The Northern Ireland conflict is hundreds of years in the making. The
conflict is not unique. Many of the conflicts that attract our attention today
are decades, if not centuries, old. Serb versus Croat and Hutu versus Tutsi
are centuries old conflicts. There are more than two violent actors in
Northern Ireland. Other violent settings also involve more than two actors,
if not currently then historically. The British were key actors in the carving
up of Palestine. T
oday, the United States plays a key role in that conflict, as
Arab states and Palestinian paramilitaries are aware. Similarly, for the vic-
tims of the bombing of Kosovo (whether or not it was justified), the United
States and its allies were significant actors. An in-depth investigation of
many conflicts that appear to involve two actors engaged in “them versus
us” behavior may show them to be as complex as the Northern Ireland 
conflict.
Acknowledgment
This research was supported by grants from the Harry Frank Guggenheim Founda-
tion, the National Science Foundation (grant SES-8318161), and the IUPUI Office
of Faculty Development. I thank Stephen Heathorn, Shel Stryker, Peggy Thoits,
Eric Wright, the participants in the Symposium, and the reviewers for their com-
ments.
Notes
1. My use of the term “political violence” follows Tilly’s conception of the term
“any observable interaction in the course of which persons or objects are seized or
physically damaged in spite of resistance” (Tilly, 1978, p. 176). Political violence in
Ireland is not confined to Northern Ireland. The Irish Republican Army recruits
and carries out operations on an all-Ireland basis. Protestant paramilitaries have
killed people in the Irish Republic, and the IRA has been active throughout Ireland,
England, and on the European continent. Included in the Sutton (1994) data in
table 6.1 are persons who were killed outside of Northern Ireland, in events related
to the conflict there. By far the vast majority of persons killed because of the North-
ern Ireland conflict were killed in Northern Ireland (see White, 1997a). 
2. These actors are “broadly” defined because there are in fact several different
Republican, Protestant, and security force organizations active in Northern Ireland.
These include the Irish Republican Army, the Irish National Liberation Army, the
Ulster Volunteer Force, the Ulster Freedom Fighters, the British army, and the
Royal Ulster Constabulary. 
3. This is too simple a description of a complex situation. While Irish people are
in general “thems” to Britishers (as described below), there are probably degrees of
“themness.” In particular, one group of “them”—Ulster Protestants—identifies with
Britain, is devoted to the British crown, proudly flies the Union Jack, etc., while an-
Social and Role Identities and Political Violence
155


other group of “them”—Irish Nationalists—identifies with Ireland, rejects the
British crown, flies the Irish flag, etc. While Irish people may be “them,” some of
“them” are likely to be more tolerable than others. This helps to explain differences
in state repression, whereby Republican paramilitaries experience more repression
at the hands of the British than do Protestant paramilitaries (e.g., White, 1999). It
also helps explain the much larger number of Catholic civilians versus the relatively
small number of Protestant civilians killed by the security forces. Catholic civilians
come from the dis-loyal “them” community. They are subject to much more state
repression (see White & White, 1995).
4. “T
aig” is a slang term for a Catholic person. It is typically used by Protestants
in a derogatory fashion. It is a corruption of the Irish first-name, “T
adhg.”
5. In order to demonstrate the development of an Irish identity I present infor-
mation from two respondents who lived in England prior to their joining the IRA.
The life stories of these two respondents are such that their understanding of their
Irishness was practically forced on them, which makes for a good presentation. The
other Republican respondents (Respondents 3, 4, and 5), developed an Irish social
identity without first living in Britain.
6. Determining when a person acquires a role identity is difficult. Some people
may acquire role identities only after joining an organization and formally moving
into the role. Others may acquire role identities from interaction with persons in
roles prior to their actually joining a given organization. In a sense, they may find
themselves in the role, and then join the organization. Imagine the potential recruit
who goes out on an operation or two with his friends and quickly finds herself a 
regular member of an IRA unit. Similarly, consider the child of a Republican who, at
the age of 18, is interned under the assumption that he is a member of the IRA, but
is not. Interred without trial with other Republican prisoners, having been brought
up in Republican surroundings, this individual probably knows the role—or learns it
very quickly—of being an IRA prisoner, adapts quickly to prison, and joins the IRA.
References
Arthur, M. (1987). Northern Ireland: Soldiers talking. London: Sidgwick & Jackson.
Bell, J. B. (1979). The secret army: The IRA 1916–. Dublin: Irish Academy Press.
Bell, J. B. (1987). The gun in politics: An analysis of Irish political conflict, 1916–86. New
Brunswick, NJ: Transaction.
Bell, J. B. (1993). The Irish troubles: A generation of violence. New York: St. Martin’s.
Bruce, S. (1992). The red hand: Protestant paramilitaries in Northern Ireland. Oxford:
Oxford University Press.
Bruce, S. (1994). The edge of the Union: The Ulster Loyalist political vision. Oxford: 
Oxford University Press.
Bruce, S. (1997). Victim selection in ethnic conflict: Motives and attitudes in Irish
Republicanism. Terrorism and Political Violence, 9, 56–71.
Cairnes, E. (1982). Intergroup conflict in Northern Ireland. In H. T
ajfel (Ed.), Social
identity and intergroup relations (pp. 277–297) Cambridge: Cambridge University
Press.
Cronin, S. (1980). Irish nationalism: Its roots and ideology. Dublin: Irish Academy
Press.
Cusack, J. & McDonald, H. (1997). UVF. Dublin: Poolbeg Press.
156
The Contribution of Social Identity to Violent Intergroup Conflict


Drake, C. J. M. (1998). The role of ideology in terrorists’ target selection. Terrorism
and Political Violence, 10 (Summer), 53–85.
Farrell, M. (1980). Northern Ireland: The orange state. London: Pluto Press.
Flackes, W. D., & Elliott, S. (1989). Northern Ireland: A political directory. Belfast:
Blackstaff Press.
Fraser, A. (1973/1997). Cromwell: Our chief of men. London: Weidenfeld and Nicolson.
Hyndman, M. (1996). Further afield: Journeys from a Protestant past. Belfast: Beyond
the Pale Publications.
Klandermans, B., & de Weerd, M. (2000). Group identification and political
protest. In S. Stryker, T. J. Owens, & R. W. White, Self, identity, and social move-
ments. (pp. 68–90) Minneapolis: University of Minnesota Press.
Lyons, F. S. L. (1973). Ireland since the famine. London: Fontana.
Manchester, W. (1983). The last lion: Winston Churchill, 1874-1932, Visions of glory.
New York: Little, Brown and Co.
Maudling, R. (1978). Memoirs. London: Sidgwick & Jackson.
McCall, G., & J. L. Simmons. (1978). Identities and interactions. New York: Free
Press.
McGarry, J., & O’Leary, B. (1995). Explaining Northern Ireland. Oxford: Blackwell
Publishers.
McKittrick, D., Kelters, S. Feeney, B., & C. Thorton. (1999). Lost lives: The stories 
of the men, women, and children who died as a result of the Northern Ireland troubles.
Edinburgh: Mainstream Publishing.
Moxon-Browne, E. (1983). Nation, class and creed in Northern Ireland. Aldershot:
Gower Publishing.
Nelson, S. (1984). Ulster’s uncertain defenders: Protestant political, paramilitary, and
community groups and the Northern Ireland conflict. Belfast: Appletree Press; 
Syracuse, NY: Syracuse University Press.
O’Leary, B., & McGarry, J. (1993). The politics of antagonism: Understanding Northern
Ireland. London: Athlone Press.
O’Malley, P. (1990). The uncivil wars: Ireland today. Boston: Beacon.
Rose, R. (1971). Governing without consensus: An Irish perspective. Boston: Beacon
Press.
Stevenson, J. (1996). “We wrecked the place”: Contemplating an end to the Northern Irish
troubles. London: The Free Press.
Stryker, S. (1968). Identity salience and role performance: The relevance of 
symbolic interaction theory for family research. Journal of Marriage and the
Family, 30, 558–564.
Stryker, S. (1980). Symbolic interactionism: A social structural version. Menlo Park, CA:
Benjamin/Cummings.
Stryker, S. (2000). Identity competition: Key to differential social movement partic-
ipation? In S. Stryker, T. J. Owens, & R. W. White, Self, identity, and social move-
ments (pp. 21–40). Minneapolis: University of Minnesota Press.
Sutton, M. (1994). An index of deaths from the conflict in Northern Ireland, 1969–1993.
Belfast: Beyond the Pale Publications.
T
ajfel, H. (1981). Human groups and social categories: Studies in social psychology. Cam-
bridge: Cambridge University Press.
T
aylor, P. (1999). Loyalists. London: Bloomsbury.
Thatcher, M. (1993). The Downing Street years, 1979–1990. New York: Harper-
Collins
Social and Role Identities and Political Violence
157


Thoits, P. A., and Virshup, L. K. (1997). Me’s and we’s: Forms and functions of 
social identities. In R. Ashmore & L. Jussim (Eds.), Self and identity: Fundamental
issues. New York: Oxford University Press.
Tilly, C. (1978). From mobilization to revolution. Reading, MA: Addison-Wesley.
Turner, J. C., Oaks, P. J. Haslam, S. A. & McGarty. C. (1994). Self and collective:
Cognition and social context. Personality and Social Psychology, Bulletin 20,
454–463.
Turner, J. C., with Hogg, M. A., Oakes, P. J., Reicher, S. D., & Blackwell, M. S.
(1987). Rediscovering the social group: A self-categorization theory. Oxford: Basil
Blackwell.
White, R. W. (1993). Provisional Irish Republicans: An oral and interpretive history.
Westport, CT: Greenwood Press
White, R. W. (1997a). The Irish Republican Army: An assessment of sectarianism.
Terrorism and Political Violence, 9 (Spring), 20–55.
White. R. W. (1997b). The Irish Republican Army and sectarianism: Moving 
beyond the anecdote. Terrorism and Political Violence, 9, 120–131.
White, R. W. (1999). Comparing state repression of pro-state vigilantes and anti-
state insurgents: Northern Ireland, 1972-75. Mobilization, 4, 189–202.
White, R. W., & Fraser, M. R. (2000). Identities and long term movement 
activism: Republican Sinn Féin. In S. Stryker, T. J. Owens, & R. W. White 
(Ed.), Self, identity, and social movements. (pp. 324–346). Minneapolis: University
of Minnesota Press.
White, R. W. & White, T. F. (1995). Repression and the liberal state: Northern 
Ireland, 1969-72. Journal of Conflict Resolution, 39, 330–352.
Whyte, J. (1990). Interpreting Northern Ireland. Oxford: Clarendon Press.
158
The Contribution of Social Identity to Violent Intergroup Conflict


Ervin Staub
7
Individual and Group Identities in
Genocide and Mass Killing
Mass killing and genocide, in which a group turns against another with sys-
tematic and widespread violence, has been a tragic aspect of life in the 20th
century. How does such violence, especially its most extreme form, geno-
cide, come about? What is the motivation for such violence and how does it
evolve? What are the social conditions, characteristics of cultures, and psy-
chological processes of individuals that make it more likely? Most impor-
tant for the present volume, what role do individual and group identities
play in genocide and mass killing?
The U.N. Genocide Convention has defined genocide as “acts commit-
ted with intent to destroy, in whole or in part, a national, ethnic, racial or
religious group.” I see the words “in part” in the definition above as prob-
lematic. When some people in a group are killed, for example, to intimidate
the whole group, or when members of the population are killed in the belief
that this will eliminate support for guerrillas, as in Guatemala, this may best
be considered mass killing, not genocide. However, genocide and mass
killing have similar origins and prevention requires similar actions. Because
some nations objected to the inclusion of political groups, they have not
been included in the Genocide Convention. Some authors have called vio-
lence that aims to eliminate a political group, “politicide” (Harff & Gurr,
1990) while others, including myself, prefer the term “genocide”. Influ-
ences leading to genocide against political and other kinds of groups and
the preventive efforts required seem to be the same (Staub, 1989b).
I define “mass killing” as killing many members of a group without the
intention to eliminate the group (Staub, 1989b). Such killings can be part of
an evolution toward genocide. It is important for both understanding and
159


prediction, and therefore prevention, that similar influences lead to mass
killing and genocide. While the likelihood of group or collective violence is
predictable, its exact nature, whether it will be mass killing or genocide,
may not be.
In this chapter I first outline a conception of the origins of genocide and
other collective violence that I have developed over the last 20 years and
have applied to the analysis of a number of specific instances of genocides
and mass killing (see Staub, 1989b, 1996a, 1999b). I next examine elements
of this conception in detail as I extend it by exploring the role of identity in
the generation of such violence. I will look at different types of and aspects
of identities: individual identity and varied forms of it; social identity, the
extent to which a person’s identity is based on membership in the group;
group self-concept, or how the members see and define their group. I will
also discuss the psychological and social processes through which people
strengthen their identities in difficult times, often in destructive ways.
One of the important reasons for understanding the origins of genocide
and mass killing is to find ways to prevent them. Thus, in the last section 
of the chapter I will consider prevention and the role of identities in 
prevention.
The Origins of Genocide and Mass Killing
Over the years I have developed a model of the origins of genocide and
mass killing (Staub 1989b, 1996b, 2000). This is depicted schematically in
figure 7.1. Genocide or mass killing usually begins with some form of “in-
stigation.” Instigation may have little to do with the group that ends up as
the victim. Difficult life conditions in a society are such instigators: severe
economic problems, such as inflation, depression, and unemployment; po-
litical disorganization or political chaos; and rapid social changes.
Difficult life conditions usually frustrate fundamental human needs—
what I have called “basic needs.” All of the following psychological needs
may be frustrated by difficult life conditions: the need for security, for a
positive identity, for effectiveness and control, for connection to other peo-
ple, and for a meaningful comprehension of reality and of one’s own place
in the scheme of things. These needs are viewed here as universal, even
though culture may shape their strength, manifestation, and available av-
enues for fulfillment. Especially when these needs are reasonably fulfilled,
another important need also enters—the higher need for spirituality or
transcendence, the need to go beyond a focus on one’s own self, needs, and
concerns (Staub, 1989b, 1996a, 1998a, 1999a, 1999b, 1999c).
Conflict between groups is another instigator. One type of conflict has to
do with vital needs, such as territory needed for living space. Conflicts can
be intractable, often for psychological reasons, such as a group’s identity be-
160
The Contribution of Social Identity to Violent Intergroup Conflict


ing tied to living in a certain territory. Another type of conflict occurs when
a powerful group limits the opportunities and resources of a subordinate,
powerless group.
Group conflict also frustrates basic needs. What needs it affects will de-
pend on specific conditions, but it is likely to affect security and identity.
Conflict between superordinate and subordinate groups, as well as the un-
even effects of difficult life conditions on different groups, may give rise to a
sense of injustice. While justice may not be a fundamental need in the same
way as the needs described to this point, it is nonetheless a powerful human
motive.
The instigation and the resulting frustration of basic needs leads to psy-
chological and social processes that turn a group against others. These in-
clude individuals turning to some group for identity; people elevating the
group by devaluing other groups and taking action against their members;
the ingroup scapegoating outgroups; and ideologies or visions of social
arrangements that promise a better life but also identify enemies.
One example is Rwanda. All the instigators were present there. There
were increasing economic problems, starting in the late 1980s. The inva-
sion of a Tutsi-led group of rebels in 1990, consisting of refugees from vio-
lence against Tutsis in previous decades or the children of refugees, began a
civil war. Political demands by varied groups added to political chaos and
disorganization. The Tutsis were scapegoated for all the problems. An in-
tense ideology of “Hutu power” was developed and propagated, with the
Individual and Group Identities in Genocide and Mass Killing
161
figure 7.1 Influences and Processes Contributing to Genocide and Mass Killing


Tutsis defined as enemies to be destroyed (des Forges, 1999; Prunier, 1995;
Staub, 1999b).
These psychological and social processes often lead to action against the
devalued, scapegoated group, or a group that is seen as an ideological en-
emy, and thus is the starting point for an evolution toward possible mass
killing or genocide. Individuals and groups change as they harm others.
They justify their actions by devaluing their victims more and more. They
become desensitized to the suffering of their victims. The standards of 
acceptable social behavior change, allowing and even encouraging violence.
Institutions change or new ones are established that serve discrimina-
tion and violence. In the end, there may be a reversal of morality: killing 
the scapegoat or ideological enemy may become a “higher good” (Staub,
1989b).
Unless halted by bystanders, the violence evolves in intensity. Unfortu-
nately, both internal bystanders (members of the involved groups) and ex-
ternal bystanders (other nations and the international community) tend to
remain passive. This affirms and encourages the perpetrators. At times 
outsiders—groups and nations—support the perpetrators because of past
ties or a desire to promote what they define as their own national interests.
When the kind of psychological and social processes here described do not
lead to genocide, usually bystanders are involved in a positive way—they in-
tervene in some way to alter the processes leading to collective violence
(Staub, 1989b, 1999b, 2000).
Certain characteristics of cultures make the genocide or mass killing
process more likely. A history of devaluation of a group of people is a pri-
mary characteristic. Very strong respect for authority and a monolithic soci-
ety with a limited range of values, in contrast to a pluralistic society, are also
important. Unhealed wounds in a group owing to past victimization, which
lead members of the group to feel easily threatened and respond with what
they see as defensive aggression, make it more likely that such a genocidal
process unfolds in response to instigation. The nature of individual identi-
ties that the society has fostered and the nature of group self-concept are
also important. 
The influences here described do not act separately and independently.
They interrelate and join in meaningful ways. I will examine here some as-
pects of this interrelation; others can be seen and deduced from the subse-
quent discussion (see figure 7.1 and Staub, 1989b). It is the frustration of
basic needs by instigating conditions that leads group members, whose indi-
vidual identity is shaken, to turn to the group for identity, to focus more on
their social identity, or to “give themselves over” to an identity group. This
frustration also leads to scapegoating and the creation of destructive ideolo-
gies (which identify enemies), that turns the group against another group.
The kind of ideology that is adopted, and the scapegoat and ideological en-
emy that is selected, very much depend on characteristics of the culture. 
Whether the group turns against another depends in part on whether 
162
The Contribution of Social Identity to Violent Intergroup Conflict


the devaluation of some group is part of the culture, the nature of that de-
valuation, and how intense it is (Staub, 1989b, 1999b). It depends on past
victimization of the group and unhealed wounds, which make it more vul-
nerable to threat. It depends on the strength of “authority orientation,”
how deeply and unconditionally people are accustomed to be led, to respect
and obey authorities and the extent they are accustomed to gain support
from their group. The stronger this orientation, the more they will be af-
fected by difficult life conditions and other instigators, and the more diffi-
cult it will be to stand on their own. Their need to turn to new authorities
who might be able to guide and support them will be greater.
A history of devaluation of a group, which all members of a society tend
to absorb to some extent, a strong authority orientation, and a monolithic
rather than pluralistic-democratic culture, as well as other cultural charac-
teristics (Staub, 1989b), all make it less likely that people will oppose initial
actions against a victim group, or initial “steps along the continuum of de-
struction.” The passivity of bystanders has the effect of allowing and even
encouraging the evolution of violence. 
The psychological processes that initiate and maintain the increasing 
violence are fostered both by intense instigators and the presence of cul-
tural characteristics. The social conditions that constitute the instigators af-
fect both perpetrators and internal bystanders. As the violence evolves, it
constitutes culture change; further evolution happens in a changing and
changed cultural context. 
Identity and the Origins of Mass Killing and Genocide
I now describe how individual and group identities are involved in the paths
to genocide or mass killing depicted in figure 7.1. T
o begin, severe eco-
nomic problems deeply impact people. Unemployment, or the inability to
provide for oneself and one’s family owing to severe inflation, makes people
feel insecure and ineffective. Such conditions call into question people’s
personal identity and self-worth: Who am I if I have no work and cannot
take care of the basic material needs of myself and my family? At the same
time that identity is questioned, ties to one’s group can also weaken because
the problems originate in society and social conditions.
Intense political conflict and disorganization also affect people’s experi-
ence of their group and perception of the world. Rapid social change, which
often occurs at times of economic and political crises, adds to social chaos
and disorganization. Since substantial social change requires adjustment,
even positive change places demands on people. 
The different elements of difficult life conditions, which may occur sepa-
rately but frequently are present in some combination, frustrate many of
the basic needs. In difficult times, focusing on their own needs, people be-
come disconnected from each other. At times when they need it most, peo-
Individual and Group Identities in Genocide and Mass Killing
163


ple lack connection and support. The social upheaval makes their old
worldview inadequate in providing a comprehension of reality: an under-
standing of how people, society, and the world operate, and of their own
role or place in the scheme of things. 
As a result, when a society experiences persistent, intense life problems,
its members desperately need a renewed basis for identity, security, a sense
of effectiveness, connection to other people, and comprehension of reality.
They turn for identity and the fulfillment of other basic needs to a new
group, or to leaders who strengthen the group through scapegoating out-
siders or offer it new identity, meaning, and hope through ideologies that
promise a better life and offer the fulfillment of basic needs.
Identity, Scapegoating, and Ideology
Three types of “identities” are discussed in this chapter: first, personal identi-
ties, or the ways in which individuals answer the question, who am I?; sec-
ond, social identities, or the extent to which individual identity is rooted in,
or connected to, the group; third, group self-concepts, or the socially shared
way members perceive and experience their group. Relevant to all of them
is a group’s culture and the social and political institutions that both express
and shape that culture. The nature of individual identity is shaped by cul-
ture; in turn, the identities that characterize individuals maintain the cul-
ture. Social identity theory has focused on how people categorize them-
selves (and others) as members of groups, how they identify themselves
with certain groups, and the tendency they have to enhance themselves by
comparing their group favorably with other groups (T
ajfel, 1982; Turner,
1987; Myers, 1999). It emphasizes the psychological experience of rooted-
ness in and connection to the group. 
But social identities also have content, as defined by the culture. “We”
are intelligent and hardworking, or easygoing and ready to enjoy life, or
brave and ready to fight for our rights or for maintaining important values.
The content of social identity is what I refer to as group self-concept.
While social identity theory stresses the tendency to favorably evaluate
one’s group, group self-concept can be positive or negative (superior or
weak and vulnerable). As in the case of individual identity, consciously held
views and evaluations of the group can exist side by side with feelings about
and valuations of the group that may not be consciously held, such as a
sense of vulnerability or weakness.
Individual identity is, in part, a social identity rooted in groups to which
one belongs ranging from family to nation (T
ajfel, 1978, 1982; Turner,
1987; Bar-T
al & Staub, 1997). It is not surprising, therefore, that in difficult
times people try to strengthen their identity through personal identification
with some group. A good social identity—an identity based in a group that
is seen in a positive light, for example, as effective and powerful—can also
164
The Contribution of Social Identity to Violent Intergroup Conflict


be a good source of security, feelings of effectiveness and control, and con-
nection. The values and worldview of the group also provide a comprehen-
sion of reality. 
Scapegoating some other group can strengthen both individual and
group identity and also help fulfill other needs. If another group is viewed
as responsible for life problems, then neither I nor my group are to blame.
Scapegoating at such times is usually a social process in that it is done to-
gether with like-minded others. This not only builds identity but strength-
ens security and connection. 
Ideologies are almost always a part of the genocidal process. I see ideolo-
gies as visions of social arrangements that tell people how to live life. They
are blueprints for the organization of societies and relations among indi-
viduals. The ideologies created or adopted in difficult times offer the satis-
faction of basic needs. They promise people a better life. Sometimes their
stated purpose is to improve life for all human beings—what I have called
“better world ideologies.” Communism has been such an ideology. Nation-
alistic ideologies promise better life to people in a particular group by en-
hancing the group’s wealth, power, or purity (Staub, 1989b).
Often, ideologies have both better life and nationalistic components.
This was true of the Nazi ideology. The Khmer Rouge ideology was osten-
sibly a better world ideology, focusing on total social equality. But it was
group based, looking to the group’s history as a simple agricultural society
for ways to create this social equality. Without explicit elaboration it also
included a strongly nationalistic agenda. This expressed itself in the espe-
cially intense violence against the Vietnamese in Cambodia, in hostility to
Vietnam, an ancient enemy, and in genocidal violence against other minori-
ties in Cambodia. 
Legitimizing ideologies (Levin, Sidanius, Rabinotitz, & Frederico, 1998;
Staub, 1989b, 1999b) offer justifications for the power and privilege of
dominant groups. Violent reactions by powerful groups to demands by sub-
ordinate groups for rights and privilege have been one of the most frequent
sources of mass killing since World War II (Fein, 1993). In my view, this 
response of dominant groups is not simply to a threat to their power and
privilege. They also respond to a threat to their identity and worldview
(Staub, 1989b). They have come to see themselves as rightfully occupying
positions of power and influence, owing to their background, history, dili-
gence, intelligence, bravery, or other aspects of their presumed nature. The
world is arranged the “right way” with them in power (see Sidanius &
Petrocik, this volume).
People usually hold or turn to and adopt ideologies as part of a group or
a movement. At times some people create or put together ideas that form
the ideology, as did Hitler (1925) or the leaders of the Khmer Rouge in
Cambodia (Staub, 1989b). However, it is when shared among a group of
people that an ideology comes to life. The ideology defines the group and
gives the group and its individual members a positive social identity. It also
Individual and Group Identities in Genocide and Mass Killing
165


provides a basis for connection. Participating in the movement can make
people feel effective and in control. The goals of the movement become the
motivating force for members. The ideology and the movement that arise
from it often become central sources of people’s identity and comprehen-
sion of reality.
While human beings need positive visions in difficult times, the ideolo-
gies that emerge usually identify another group as an enemy. This group is
seen either as the cause of problems, or as standing in the way of the ide-
ology’s fulfillment, or both. They, therefore, become the objects of hostility.
Harmful actions begin to be directed at them. 
The Role of Individual Identities in Genocide
Certain individual identities that a culture promotes make the genocidal
process more likely. Research and theory have identified differences in the
type of self-concept or individual identity that is shaped by individualist and
collectivist cultures (Triandis, 1994). Differences in women’s and men’s
identities in Western societies, including the United States, have also been
noted (Sampson, 1988). It has been proposed, for example, that male iden-
tity is more autonomous, more like the identity shaped by individualistic
cultures, while female identity is more relational, like identities shaped by
collectivist cultures.
Considering conditions that give rise to genocide, both relational/collec-
tivist and autonomous/individualist identities may have vulnerabilities. Dif-
ficult life conditions may especially frustrate autonomous persons. They
have learned to stand on their own, but now they are unable to effectively
fend for themselves or their families. American soldiers who became pris-
oners of war in the Korean War had more difficulty in resisting brainwash-
ing than, for example, Turkish soldiers, seemingly because the former were
trying to face it alone while the latter supported each other (Kincaid, 1959).
Relational identities may also create difficulties. People with such identi-
ties may feel interconnected with other people in a way that makes it espe-
cially difficult to speak out against and oppose the direction the group is be-
ginning to take. As their group or society begins to scapegoat, identify an
enemy, discriminate, or take other harmful actions against another group,
they are more likely to remain silent and go along. This makes more proba-
ble the evolution of increasing discrimination and violence. 
I have suggested that relational/collectivist identities may be differenti-
ated into embedded and connected (Staub, 1993a). The embedded identity,
which I just described, is characterized by a strong connection to other peo-
ple that also embodies dependence on the group and an inability to separate
from it. In contrast, people who have developed connection to others as in-
herent to their identities, but who also have learned to stand on their own,
to be separate, have connected selves or identities. 
The nature of personal identity has implications for peoples’ relationship
166
The Contribution of Social Identity to Violent Intergroup Conflict


to their group. People with connected selves may be less likely to give
themselves up to the group in difficult times. They may be less likely to re-
linquish their own identity and give themselves over to a social or group
identity than people with embedded selves. Seemingly paradoxically, be-
cause they can use others for support more effectively, those with connected
self-concepts may be even less likely to do this than people with autono-
mous/individualist identities. As a result, people with connected selves may
be less likely than those with either embedded or autonomous identities to
remain passive as the evolution of harm-doing begins. They may be more
likely to speak out and oppose actions contrary to their own values, or to
the values they believe the group ascribes to. This is extremely difficult 
to do but is essential in halting the evolution of increasing violence. 
I am suggesting that identity-related differences in individuals’ relation-
ship to the group have implications for group violence. Related to this, I
have proposed that some individuals are “constructively” patriotic while
others are “blindly” patriotic (Staub, 1997). In research exploring this we
find that constructive patriots express love for their group, but also a will-
ingness to question and criticize the policies and practices of the group that
are contrary to caring values and concern for human welfare, or contrary to
what they see as important group values (Schatz & Staub, 1997; Schatz,
Staub, & Lavine, 1999). Although this has not been assessed in the research,
I expect that constructive patriots would have connected identities. In con-
trast, blind patriots, who I expect to have either individualist or embedded
self-concepts, love their group but are unwilling to criticize it or to question
its policies or practices. Different identity types and associated relationships
to the group have clear implications for supporting or opposing destructive
views and group practices. 
Other classifications of individual identities that are relevant to genocide
are possible. Some people may have secure identities (Staub, 1993a). Pre-
sumably their basic need for security has been fulfilled in the course of their
lives, especially early lives. Perhaps they also have had experiences with tak-
ing effective action to protect themselves. As a result they feel reasonably
secure in the world. They would be less affected by instigating conditions,
such as difficult life conditions, and less moved to the psychological and so-
cial processes that were described as a consequence of such conditions.
Others, perhaps due to the frustration of the need for security early in life,
including victimization or other traumas, may develop insecure identities,
making it more likely that these psychological processes leading to genocide
unfold in response to instigating conditions.
Authority Orientation and Identity
Observation and research indicate that in many cases of genocide or mass
killing the society has been characterized by strong respect for authority
(Gourevich, 1998; Kressel, 1996; Smith, 1998; Staub, 1989b,1999b). “Au-
Individual and Group Identities in Genocide and Mass Killing
167


thority orientation” in the culture is a term I prefer to “authoritarian per-
sonality” because it focuses purely on respect for authority and hierarchy,
without implications of repression of sexual feelings and other characteris-
tics that have been described as aspects of the authoritarian personality.
Authority orientation affects the nature of identity in a number of ways.
Individuals who have embedded selves are connected to others through net-
works of dependent relationships. People in an authority-oriented culture
also seem deeply embedded in the group, but this is based on their orienta-
tion to authority. The experiences they have as this orientation is developed
in them, normally from early childhood on, diminishes the evolution of a
strong independent self. 
In the face of difficult life conditions or group conflict, people who have
relied on leaders for guidance and protection will find it more difficult to
bear the threat, anxiety, and frustration of basic needs. When policies and
practices are instituted in a group that harm others, they will be less willing
to speak out and to oppose the authorities and the rest of the group. Strong
respect for authority also makes obedience to immoral orders by authorities
more likely.
The Nature of Group Self-concept
Weakness, Superiority, and a Positive Self-Concept. The nature of group self-
concept (by which I mean the way members consensually see and experi-
ence their group) is also important. How do members of a group, people
who identify with a group, see their group? What is their conception and
experience of the group? And, how do differences in group self-concept
predispose toward mass killing or genocide? I have hypothesized (Staub,
1989b) that both a sense of weakness and vulnerability, and feelings and be-
liefs of superiority, may make it more likely that groups will turn to and en-
gage in systematic violence against others. When a group feels weak and
vulnerable, life problems in the society or conflict with other groups will in-
tensify vulnerability and the need to defend oneself. Elevating the group,
scapegoating another group, creating “positive visions” that identify ene-
mies, and self-defense through aggression become more likely. 
A sense of superiority, when frustrated, may also become a powerful
source of psychological processes and actions that contribute to genocide.
Feeling superior makes individuals and groups vulnerable to disconfirma-
tion of their superiority and, as a result, likely to experience narcissistic
wounds (Baumeister, 1997). However, a self-concept of superiority is differ-
ent from a high self-esteem. Feelings of superiority may be the result of a
history of success by a society in many realms combined with the usual eth-
nocentrism of groups that leads to interpretation of this success as evidence
of the group’s superiority. But more often, a belief in one’s superiority is ex-
aggerated and may represent a compensatory identity. In the lives of most
individuals and groups there are enough small and large experiences of inef-
168
The Contribution of Social Identity to Violent Intergroup Conflict


fectiveness, failure, and loss that, under normal conditions, a more moder-
ate sense of self would develop.
A positive identity in individuals is likely to arise from experiences of
nurturance and affirmation and from success in overcoming obstacles.
Comparable experiences on the group level can give rise to positive group
self-concepts. The absence of these may lead to low self-concept. Their ab-
sence, or negative experiences, overlaid with experiences of success but
without incorporating the difficult, painful experiences into one’s identity,
can lead to a feeling of superiority, which is what I regard as a compensatory
identity. In children a defensive self-esteem or defensive egotism, which was
found to be different from high self-esteem, was associated with more bul-
lying of others. Defensive self-esteem was ascribed to children who de-
scribed themselves and were described by others as having positive self-
esteem, but were also described by others as “always wants to be the center
of attention; thinks too much of himself or herself; can’t take criticism”
(Salmivalli, Kaukiasnen, Kaistaniemi, & Lagerspetz, 1999, p. 1271).
Germany may be a good example of having a superior group self-
concept. The group self-concept of Germans early in this century was ex-
tremely positive, of a kind that may be described as a superior compen-
satory identity. This was due to recent military successes, economic
strength, a history of cultural achievement, effective civic institutions, and
an internal climate that propagated self-satisfaction with German culture
and society beyond what is “normal” ethnocentrism (Staub, 1989b). 
These positive experiences might account for what seems was an early
20th-century German sense of superiority. But Germans also had a history
of problems and suffering. Many Germans died in wars in previous cen-
turies. During the Thirty Years’ War (1618–1648), a third of Germany’s
population died (Craig, 1982). After struggles among many smaller units,
Germany became a single country only late in the 19th century. History
leaves its mark, and perhaps underlying the outward sense of group superi-
ority there was also a deep-seated national vulnerability. Germany was then
frustrated by a series of events beginning with the loss of World War I.
These included a revolution that overthrew the monarchy, hyperinflation,
economic depression, political chaos, and actions by other countries that
Germans felt were humiliating to them (e. g., the French occupying the
Ruhr district to ensure the receipt of compensation Germany was forced to
pay after World War I). Whether it was defensive egoism, or normal ethno-
centrism, or even a “genuine” sense of superiority, the nature of their group
identity was likely important in Germans’ reactions to the intense life prob-
lems following World War I. 
A decade ago I wrote: “in groups and in individuals very high self-evalu-
ation often masks self-doubt. Persistent life difficulties may contradict high
self-evaluation and bring self-doubt to the surface. Even when there is no
underlying self-doubt, a very high self-evaluation may be associated with
limited concern for others. Among individuals, a moderately positive self
Individual and Group Identities in Genocide and Mass Killing
169


concept is most strongly associated with sensitivity and responsiveness to
other people” (from Staub, 1989, p. 55; see also Jarymowitz, 1977; Reese,
1961). This still seems a reasonable conclusion. Both individuals and groups
have to value themselves in order to value other people, but not value them-
selves so strongly that others do not matter, or so highly and unrealistically
that setbacks are experienced as devastating blows to collective and personal
self-worth. 
Cultural Devaluation and Ideologies of Antagonism 
A form of group identity that I have called “ideology of antagonism” is es-
pecially important in genocide and collective violence. Inherent in group
identity is a differentiation between “us” and “them.” In my view a devalua-
tion of “them” is not inherent in, but easily arises out of, this differentiation
(see Brewer, this volume). There are many factors that develop a persistent
devaluation of one group by another, which then becomes part of the
group’s culture. They include justification of past mistreatment of the other
group, fear of the other’s difference, and the need to create identity when
one group has split off the other. (The latter may be an important origin of
Christian anti-semitism [Staub, 1989b]). Devaluation embedded in a cul-
ture is a crucial cultural source of a group turning against another.
At times, two groups have had a long history of conflict with mutual vio-
lence. The evolution of increasing violence has already occurred, and as a
result new violence easily flares up. In the course of this past history, not
only the view of the other but the identity of each group as well has come to
be defined around this enmity: the other is my enemy, and I am the enemy
of the other. In short, group self-concept comes to center on violent con-
flict with the other group (see Kelman, this volume).
The Hutus and Tutsis in Rwanda seem to be an example of ideologies of
antagonism. Up to the 20th century, the Tutsis were dominant, but group
membership was defined by a complex combination of ethnic origin and
class, with rich Hutus sometimes acquiring a classification as Tutsis. Under
colonial rule by the Belgians, Tutsi dominance was intentionally enhanced,
with Hutus receiving increasingly bad treatment. They became an under-
class with limited rights and privilege. In 1959, in a violent revolt, they
killed about 50,000 Tutsis. The Hutus ruled the country after it gained in-
dependence in 1962. There were significant massacres of Tutsis in the
1960s and 1970s, with continued discrimination and more limited, sporadic
violence against them after that.
In the 1990s, a Tutsi-led group, consisting mainly of descendants of
refugees over the past few decades, invaded Rwanda. In the course of the
following few years, a complex political and military situation developed in
the country. Dominant Hutus increasingly advocated “Hutu power,” with
media propaganda explicitly or implicitly promoting the destruction of 
Tutsis. The genocide of 1994 followed. It was stopped by the victory of the
170
The Contribution of Social Identity to Violent Intergroup Conflict


Tutsi-led army over the government (see des Forges, 1999; Prunier, 1995;
Staub, 1999b).
Cultural devaluation is extremely persistent, and ideology of antagonism
is even more so. Neither tend to go away on their own, unattended, simply
with the passage of time. They remain part of the culture, expressed in lit-
erature, in everyday relations, and in the life of institutions. They remain
part of the identity of individuals and groups. Active social processes are re-
quired to change them.
Victimization, Healing, and Identity
When a group has been victimized in the past, without significant healing,
it has an increased likelihood of becoming a perpetrator of mass violence
(see Eriksen, this volume). People who experience trauma are deeply af-
fected (Herman, 1992; Janoff-Bulman, 1992; McCann & Pearlman, 1990).
Group victimization diminishes the different selves—the individual self, 
social identity, and group self-concept. Who am I and who are we, if we are
so treated? Since individual identity is partly based on group identity, even
those members of the group who are not present and have not been indi-
vidually in danger are deeply affected.
Victimized groups are more likely to respond to threat with violence. In
case of conflict, their understandable self-focus makes it difficult for them
to consider the needs of others. Feeling insecure, their self vulnerable, and
perceiving the world as a dangerous place, they experience threat as more
intense than it is. They are more likely to engage, therefore, in aggression
which they see as defensive (Staub, 1998b). 
Healing following victimization makes it less likely that a group turns
against another and perpetrates violence. Acknowledgment of the group’s
suffering by others and the expression of caring and empathy contribute to
healing. The group itself can also promote healing by directly engaging
with its past experience, including certain kinds of memorials and rituals of
mourning and remembrance (Staub, 1998b). Healing from both individual
trauma and group victimization involve repair of identity, as well as connec-
tions to other people (Pennebaker & Keough, 1999). The two are inher-
ently related. The reestablishment of trust in people and the capacity to ex-
perience connection to them help repair identity.
Perpetrators also need to heal. At times, their woundedness has con-
tributed to their actions. Their own violence further wounds them. It seems
impossible to inflict tremendous harm on others, to kill many people and
not become wounded or damaged. When perpetrators and victims continue
to live together, as in Bosnia and Rwanda, for reconciliation to take place,
some healing by both groups has to precede it. Reconciliation is required,
in turn, to stop a continuing cycle of violence.
In the case of the Armenians who survived the genocide by the Turks in
1915–1916, healing was interfered with by the denial of Turkey that a geno-
Individual and Group Identities in Genocide and Mass Killing
171


cide had taken place and by the complicity in this denial of many other
countries. The relationships of the recently created Armenia to other na-
tions, especially Turkey and Azerbaijan populated by ethnic Turks, have
been affected by this history and the identity it has helped create. That
identity has components of vulnerability, including a view of the world as
dangerous and, perhaps, an ideology of antagonism in relation to Turks. 
It is difficult to identify victim and perpetrator groups that have healed
and thus see the consequences of such healing. While there has been aware-
ness of the need for individuals to heal following the mass killing in Bosnia
and the genocide in Rwanda, there has been little prior consciousness of the
need for whole groups to heal. There have certainly been natural processes
that have moved groups toward healing, but there have probably been no
intentional efforts specifically aimed at this by victimized groups themselves
or by the community of nations or by other bystanders. 
A consciousness about the need for healing seemed to have emerged, at
least in contemporary times, with respect to individual survivors of the
Holocaust, and later Vietnam veterans, as well as victims of individual vio-
lence such as sexual or physical abuse (Herman, 1992; McCann & Pearl-
man, 1990). The aim of such healing has been to improve individual lives,
not reducing the likelihood of later group violence.
Germany might serve as an example of healing, even if far from com-
plete. Germany was wounded by the sequence of events that contributed to
the Holocaust and by its own horrific actions during World War II and the
Holocaust. Afterwards, the Germans were led to hesitantly engage with
their past through the Nuremberg trials, other trials, and through the rest
of the world reminding Germans of their conduct. At the same time, in-
stead of acting in a generally punitive and harsh way, the world offered sup-
port, help, and connection to Germany. This ranged from the Marshall
plan, to the willingness by French leaders to make connections with Ger-
many, to Germany’s becoming a member of the European Community, and
to other demonstrations of the world accepting Germany into the interna-
tional community. In different ways and for different reasons, this was true
of both West and East Germany. Reconnecting with others is an avenue to
healing for both victims and perpetrators. 
Changes in Identities as Harm Doing Evolves
As noted above and depicted in figure 7.1, genocide and mass killing are
usually the outcome of an evolution of increasing harm doing by perpetra-
tors. There may be one victim group, although in the course of this evolu-
tion the boundaries expand and violence toward other groups also becomes
probable. How do individual and group identities change along the way?
The perpetrators come to devalue members of the victim group more
and more. They justify their actions, in part, by the nature of the victim,
who is not only increasingly devalued but also increasingly excluded from
172
The Contribution of Social Identity to Violent Intergroup Conflict


the moral realm, the realm of people to whom moral values and moral con-
siderations apply (Opotow, 1990; Staub, 1989b). Perpetrators also justify
their actions by the “higher” values of their ideology. Along the way the
perpetrators may come to see themselves as people who are willing to make
the sacrifices required, including the killing of the victims, to fulfill impor-
tant goals. Thus, they may see themselves, as Himmler encouraged the SS
to see themselves, as bearing noble sacrifice for their higher cause (Hilberg,
1961). 
In addition to their identity changing in relation to the victims, perpetra-
tors are likely to become desensitized to human suffering in general. Deval-
uation tends to extend from the original victims to new groups, which 
prepares the way for the extension of violence (Staub, 1989a, 1989b). In
Germany, for example, after the euthanasia program, the killing of physi-
cally handicapped, mentally ill, and other “inferior” Germans, after the
genocide of the Jews and gypsies, and after violence against people in occu-
pied countries, the authorities began to consider the killing of physically
ugly German prison inmates (Lifton, 1986). 
As boundaries weaken and new groups are devalued and become poten-
tial or actual victims, the perpetrators’ identity as people who value human
beings and human welfare is likely to change. As this process unfolds perpe-
trators are likely to distance themselves from all human beings not a part of
their group, however that is defined. At the extreme, might it be the case
that even the welfare of one’s ingroup and of friends and family become of
less concern?
The identity of bystanders also changes. Internal bystanders—members
of the perpetrator group who are not themselves perpetrators—usually re-
main passive. They are also affected by instigating conditions. They have
grown up in a culture that imbues members with respect for authority and
devalues the victim group. Thus, they share a “cultural tilt” with perpetra-
tors. This cultural background, often in combination with fear of a repres-
sive system, reduces the motivation to oppose the group as it moves along a
continuum of increasing violence.
Passivity changes bystanders. It is difficult to see others suffer and do
nothing. T
o justify their passivity, reduce their guilt, and reduce empathy
with victims that would make them suffer, passive bystanders tend to dis-
tance themselves from victims, in part by increasingly devaluing them. This
makes it even more likely that they will remain passive. It also leads some
bystanders to join perpetrators.
For example, a group of psychoanalysts in Berlin remained passive as
their fellow Jewish analysts were the object of harassment and persecution.
They began to use racially appropriate Nazi terms to refer to Jewish col-
leagues, like “not pure Germans.” These served both as euphemisms and as
means to distance and objectify their fellow analysts. They accepted a rela-
tive of Herman Goring, untrained in psychoanalysis, as the head of their In-
stitute. They began to rewrite psychoanalytic theory to fit Nazi ideology
Individual and Group Identities in Genocide and Mass Killing
173


and doctrine. Some of them became participants in the euthanasia program,
and in the end some got involved in the extermination of Jews (Staub,
1989a).
The process in this specific instance may be an example of how by-
standers distance themselves, even from familiar others and from close asso-
ciates. Adopting at first an “objective,” disinterested stance, and introducing
ideological thinking change both the view of and emotional relationship to
the other. Through this process bystanders, as well as perpetrators, change
the nature of their own self. Those changes in identity are intertwined with
changes in society and culture. As violence evolves, the norms and standards
of society change, allowing and often promoting actions that were previ-
ously inconceivable. For instance, in Rwanda, neighbors killed neighbors
and people gave up members of their families to the killers (des Forges,
1999; Gourevich, 1998; Staub & Pearlman, in press).
I have suggested that these changes in perpetrators and bystanders repre-
sent changes in identity. Although identity is a broad concept, we may see it
as including a person’s self-concept and orientation to and evaluation of the
self, as well as values relating to and orientation toward and evaluation of
other people. Thus, I am proposing here that a person’s view of the world is
also part of personal identity (Epstein, 1980). Some or all of these elements
of identity will change in the course of a societal evolution toward genocide.
The changes in a person’s self-concept and in the group’s collective sense of
self mutually reinforce each other.
However, the negative changes in the view of the other, and the changes
in the self, as individuals and a society move along a “continuum of destruc-
tion” (Staub, 1989a, 1989b), are not the whole story. Even in very simple
societies there is variation in personality and, presumably, self-concept.
Only a small percentage of people have the “modal identity type” that may
be seen as characteristic of the society (Konner, 1978). In genocidal soci-
eties, in addition to perpetrators and passive bystanders, there are usually
some people who attempt to help victims. The rescuers of Jews in Nazi Eu-
rope have been the primary group of such helpers who have been exten-
sively studied (Fogelman, 1994; Oliner & Oliner, 1988; T
ec, 1986). But in
other societies as well, there have been heroes who have endangered them-
selves to save lives. 
I have heard several such stories in Rwanda, both in conversation and in
formal interviews I conducted with people whose lives were saved. In one
case, for example, a Hutu man who had worked for a woman sent another
man to help her after her husband was killed. The man who came, a
stranger carrying a Bible, repeatedly endangered himself as he faced off
people who came to the house wanting to take her away. The woman telling
this story was unclear about his motivation to help: she suggested his reli-
gion, his being a good man. Having some difficulty accepting such self-
sacrificial behavior, she wondered about possible selfish motives but could
not identify any.
174
The Contribution of Social Identity to Violent Intergroup Conflict


Once people begin to help in some limited way, they are further shaped
by their own actions. Like violence, so helping evolves. People who have
agreed to help for a short time, or to a limited extent, change as a result of
their actions. They become more committed to the welfare of those they
have helped and often to the welfare of human beings in general. They
come to see themselves as people who are willing to endanger themselves to
help others. Their values and identity change (Staub, 1989b, 1993b). 
The Role of Identity in the Prevention of Genocide and Mass Killing
The Role of Leaders 
How can genocide and mass killing be prevented and what is the role of dif-
ferent kinds of identities in prevention? How can people gain positive iden-
tities in difficult times without turning to destructive groups or making
their groups destructive? One way is for leaders to offer positive visions that
are constructive, visions that offer hope but are realistically aimed at solving
life problems without identifying enemies, visions that connect rather than
separate people. The New Deal under Roosevelt was one example of this.
This can be difficult and at times even dangerous for leaders when there is
already strong enmity against a potential victim group.
Healing by Victimized Groups 
An important contributor to prevention is healing by victimized groups.
Healing can help repair damaged identities and worldviews (as well as fulfill
other frustrated needs). It can recreate some sense of security and trust in
the world. As a result, it makes it less likely that a victim group becomes a
perpetrator. Acknowledgement of the group’s suffering, expressions of em-
pathy and sorrow, and other forms of affirmation by the rest of the world
can contribute to healing. Punishment of perpetrators can also help victims.
It can communicate to victims that what was done to them is not regarded
as normal and acceptable by the world. That message, in turn, makes it
seem less likely that it will happen again (Staub, 1998b, 1999c). 
Engaging with their painful experiences can also help survivors of group
violence heal. Trauma victims frequently avoid such engagement even
though they are often affected by what has happened to them in many ways:
through intrusive imagery, hyper arousal, negative feelings about the self,
mistrust and difficulty in relating to people (Herman, 1992; McCann &
Pearlman, 1990). Engagement can lead to integration of the self and recon-
nection with people (Pennebaker & Keough, 1999). It can lead to finding
meaning, perhaps in helping other people, working for the prevention of
similar violence against people, or living life in a different way. 
Healing makes reconciliation possible. In places like Rwanda and Bosnia,
Individual and Group Identities in Genocide and Mass Killing
175


where victims and perpetrators live together, or where mutual violence has
occurred, reconciliation is essential to prevent continued cycles of violence.
The self is to a significant extent defined by relationships to others. Ideo-
logies of antagonism, intense devaluation and hostility toward another, 
or strong fear of a neighbor are central aspects of identity. Reconciliation,
which when genuine requires forgiveness and overcoming devaluation and
antagonism, changes identity in central ways. As a result of healing and rec-
onciliation, feelings of diminished self-worth and vulnerability can lessen.
Mistrust in people and in the world can also be reduced. Both the desire for
revenge as a form of justice and the readiness to use force to defend oneself
and one’s group can decrease. Trust in people, trust in the world as a livable
place, and hope for the future can reemerge.
How can healing, reconciliation, and positive connection to others in
general be promoted in previously victimized groups? We have so far only a
limited repertoire of skills and techniques for this. T
o begin with, victimized
groups have to start with at least some desire for change, which, in case of
intense antagonism, may be lacking. When there is some openness, con-
texts for healing can be created by bringing members of different groups to-
gether. People can write about and talk to each other about what has hap-
pened to them (Pennebaker & Keough, 1999). They can talk about their
pain, sorrow, loss, and anger. As they listen to each other and respond with
empathy, pain, which has been a source of disconnection, can become a
source of connection (Staub & Pearlman, 1996).
This is one aspect of a project that we have been conducting in Rwanda
on “healing, forgiveness, and reconciliation.” In addition to such experien-
tial aspects of promoting healing and reconciliation, the training we have
used also had psycho-educational aspects. Participants learned about and
discussed the impact of trauma on people in the hope that this would help
them understand their own experience better. They learned about avenues
to healing. They heard about the fulfillment and frustration of basic needs
and they received information and engaged in discussion about the origins
of genocide (Staub, 2000; Staub & Pearlman, in press).
This last element of the training has had seemingly dramatic effects.
Learning about a number of other genocides and about how genocide
comes about seemed to reaffirm for participants their own humanity. If
these terrible things have been done to others as well, and if there are un-
derstandable human processes that give rise to them, then the victims in
Rwanda remain part of the human realm in spite of what happened to them.
In addition, understanding the forces that led to genocide seemed to hu-
manize the perpetrators, both in the survivors’ eyes and, perhaps, in their
own, thereby contributing to reconciliation. The type of training described,
which we conducted in Rwanda, can help not only victims but also mem-
bers of the perpetrator group. Genocide and mass killing, inflicting violence
on others, deeply wounds the perpetrators. It also wounds members of the
perpetrator group who have remained passive bystanders. They also are in
176
The Contribution of Social Identity to Violent Intergroup Conflict


need of healing. Healing, in turn, can open the way to reconciliation. As
reconciliation enhances a feeling of security, it furthers the possibility of
healing.
Some of the participants in our project used the approach provided in
the training, integrated with their own traditional approach, with groups 
in the community. Formal, experimental assessment showed that trauma
symptoms decreased among community members who participated in
meetings led by people we trained, both over time, and in comparison to
groups led by different trainers who used their organizations’ traditional
approach, as well as in comparison to no treatment control groups. Orien-
tation to members of the other ethnic group became more positive among
community members who participated in the meetings that used this 
integrated approach, both over time and in comparison to people in the
other groups (Staub, Pearlman, Gubin and Hagengimana, unpublished 
research).
Healing by leaders is also important. Like the other members of a vic-
timized group, so the people who emerge as leaders have been affected by
past victimization. In Bosnia, a number of the Serb leaders have had mem-
bers of their families killed by the Croat regime that killed many Serbs and
Jews during World War II. Outsiders who engage with such leaders need to
be aware of and attempt to address their wounds to individual and group
identity, if they are to effectively influence policies and practices (Staub,
1999c).
Hostile groups can also benefit from joint efforts in behalf of shared
goals. Depending on the nature of hostility and past injuries, this may re-
quire prior dialogue, problem solving in small groups (see Kelman, in this
book), or healing. But significant engagement by members of groups with
each other in the course of working together for shared goals can bring
benefits in overcoming devaluation and antagonism (Allport, 1954; Petti-
grew, 1997; Staub, 1989b, 1999c).
Culture Change
Another group-level issue is culture change. Cultural devaluation and au-
thority orientation shape identities. Monolithic cultures, with a limited set
of values and limited access of at least some groups of people to the public
domain, also make genocide more likely, while pluralistic cultures make
genocide less likely. When respect for authorities is moderate and there are
varied values and points of view that can be expressed in the public domain,
people are more able to tolerate uncertainty and therefore social upheaval
(Soeters, 1996).
Democracies, which are pluralistic, are unlikely to engage in genocide
(Rummel, 1994). This, however, seems especially true of “mature” democ-
racies. (Staub, 1996b) These are democracies in which laws are enforced, in
which all groups can participate in public life, and which have a “civil soci-
Individual and Group Identities in Genocide and Mass Killing
177


ety “ (i. e., well-established civic institutions). Germany was not this kind of
democracy at the time of the Weimar Republic, nor was Argentina at the
time of “the disappearances” (Staub, 1989b), when elected governments
were regularly replaced by military rule. Working to create democracy and
a civil society is one way to bring about culture change that makes genocide
less likely. Such a culture is likely to form individuals whose relationship to
the group is more akin to constructive, rather than blind, patriotism and
who are less likely to be embedded in their group.
Multiple group identities seem important in preventing genocide. When
people can gain identity and fulfill other basic needs through connection to
a variety of groups, they are less likely to give themselves over to and lose
themselves in a group that turns against others. Their ability to oppose de-
structive ideologies and practices will increase. 
However, these groups need to be reasonably independent of each other.
In many cases of genocide, subgroups of society, which could have been the
source of at least somewhat independent identities, have subordinated
themselves to or have become embedded in the state or in an ideological
movement. This was recently the case with the churches in Rwanda. Many
church leaders, both in the dominant Catholic Church and in the minority
Protestant churches, were long connected to the ruling elements of the
country. They became accomplices during the genocide, facilitating the
killing of Tutsis (des Forges, 1999; Gourevich, 1998).
Inclusive Caring, Moral Courage, and the Behavior of Bystanders
What type of socialization will lead children to speak out against harmful
practices within their groups? What is required for people to exert influ-
ence within their group so that it does not remain a passive bystander in the
face of violence against people outside the group? Warmth, affection, effec-
tive but not punitive guidance, allowing children reasonable autonomy,
helping them develop a sense of significance and a “voice,” and guiding
children to help and not harm others make it likely that they will develop
inclusive caring, connected identities and constructive patriotism, as well as
moral courage (Staub, 1996a). Such characteristics would make genocide
less likely.
Constructive patriotism has a value basis in caring about human welfare.
Holding such a value motivates criticism and opposition to harmful policies
and practices of the group. Caring about human welfare is a value that has
to develop in children through experiential learning. Without love and af-
fection and positive guidance in their own lives, without feeling valued,
children are unlikely to learn to value others.
T
o publicly oppose harmful practices and advocate benevolent ones of-
ten requires moral courage. This is not necessarily physical courage, but the
ability and willingness to tolerate conflict. The kind of individual identities
people develop and the nature of social identities, what the group requires
178
The Contribution of Social Identity to Violent Intergroup Conflict


of its members, are central to this. Even when there is no physical danger,
people often won’t speak out against harmful and potentially violent prac-
tices out of concern for ostracism or exclusion from the group. Is it possible
to create cultures in which opposition for the sake of human welfare is 
valued? 
It is relevant in this context that, according to one researcher, about one-
third of rescuers of Jews in Nazi Europe were marginal to their group (T
ec,
1986). They had a different religion, or one foreign parent, or moved to
their community from some other place, or may have been simply person-
ally different and therefore less taken in by or less invested in their commu-
nity. This marginality may have made them less identified with their group,
perhaps fearing less the possibility of ostracism and enabling them to have a
separate, independent perspective.
Inclusive caring means the development of caring for people beyond the
group. It has to do with the relationship of “us” and “them,” with the devel-
opment of an identity that makes individuals and groups not draw a line at
the boundary of their group that diminishes concern for the welfare of
those beyond the group. (Staub, 1996a). Many of the rescuers grew up in
families where this line between their group and minorities was not as
sharply drawn as in the rest of the group. The parents and the family en-
gaged with people outside the “ingroup,” often including Jews (Oliner &
Oliner, 1988).
Individual identities based on inclusive caring would lead people to want
their countries to be active bystanders, who would respond when there is 
violence against people outside their borders. Self-concepts that embody
inclusive caring would be likely to move individuals to exert influence on
their governments not to remain passive. This is essential for the preven-
tion of mass killing and genocide. While this has rarely happened, the 
limited examples that exist strongly suggest that if the community of na-
tions responds early, as harm doing just begins, the evolution of increasing
violence can usually be stopped. 
Even groups of individuals can exert significant influence. After the revo-
lution in Iran the Bahai, for a long time a devalued and persecuted group,
were immediately intensely persecuted. Over 200 Bahai were executed. The
international Bahai community immediately made representations to the
various governments where Bahai lived and to the international community.
The United Nations, as well as individual nations, passed resolutions con-
demning the killing and persecution. While the persecution did not stop,
the killings did (Bigelow, 1993), with some limited resumption, unfortu-
nately, in the 1990s. 
In places where violence could have been much greater, as in Northern
Ireland, or between Israelis and Palestinians, usually external bystanders
had some role in limiting violence. Internal bystanders, members of differ-
ent groups reaching out to each other, perhaps enabled by the behavior of
external bystanders, also seems to have had a role (Staub, 1999b). In North-
Individual and Group Identities in Genocide and Mass Killing
179


ern Ireland, the British military and political presence provided support to
political and social leaders who have worked for opportunities in jobs and
more social equality for Catholics, as well as to community organizations
that brought members of the two groups together for dialogue and other
forms of engagement with each other (Cairns & Darby, 1998; Staub, 1999b;
see, however, White, this volume).
With early, committed, and resolute actions by the community of na-
tions, the evolution of violence can probably be inhibited much of the time
without the use of military force (Staub, 1999b). But for this to happen re-
quires important changes in identities: in individuals’ views of themselves
and their nations and in nations’ views of themselves and of the interna-
tional community. It requires committed bystanders and the creation of an
international system that leads to action when action is required. 
Summary and Conclusions
Identities of individuals, the nature of the relationship to the group in
which they invest their identity, and group-self concept are all important in
the generation of genocide, mass killing, and other harmful behavior by
groups toward other groups. They are also important in prevention. How-
ever, identities must be considered in the context of other aspects of indi-
vidual and group psychology. Identity is a concept that refers to certain
structures in individual personality with related processes. In reality, of
course, human beings (and cultures and societies), are unitary wholes with
interrelated structures and processes.
Important concepts related to identity that have relevance to genocide
and mass killing include values, the way the self and others are divided (“us”
and “them”), orientation to authority, the individual’s relationship to the
group, and basic human needs (figure 7.1). One may conceptualize all these
as part of identity, or separate them and see them as concepts with their
own structures and processes. I find it useful to separate these and then con-
sider the relationships among them. As I suggested in this chapter, the na-
ture of a person’s identity and group self-concept will certainly affect orien-
tation to authority, divisions between “us” and “them,” and so on. 
I have proposed a series of steps that can take members of a group from
instigating factors (left side of figure 7.1) to full-fledged genocide or mass
murder (right side of figure 7.1). Instigating factors include intergroup con-
flict and difficult life circumstances (e.g., poverty) that threaten basic needs
(e.g., need for security, sense of self-worth, connections to others). Behav-
iors that elevate the ingroup and derogate outgroups (e.g., scapegoating)
provide means of coping with these threats to the self. Derogation of the
outgroup can encourage initial acts of violence against the outgroup. Initial
violence, in turn, sets in motion a cycle of further violence and polarization
180
The Contribution of Social Identity to Violent Intergroup Conflict


that becomes tied to self and group identity. This cycle may eventually esca-
late to mass killing and genocide. The development of this chain of events is
moderated by both cultural variables (top of figure 7.1) and the behavior of
bystanders (bottom of figure 7.1). For example, ideologies that legitimize
violence and a group’s history of being a victim of violence are two cultural
factors that can increase the likelihood that initial instigators will lead to
mass killing or genocide. On the other hand, involvement of bystanders
(whether ingroup members or outsiders, such as the international commu-
nity) can mitigate the march from instigators to eventual genocide.
One of the important relationships suggested by this chapter is that be-
tween identity and basic needs (figure 7.1). Security, connection, and effec-
tiveness are important from birth on, and their fulfillment and frustration
begin at birth. When we consider motivation for action that may have to do
with identity, under many conditions a complex of motives and needs will
be involved. When people engage in destructive ways of fulfilling their need
for identity (for example, by joining an ideological movement that aims to
destroy “enemies”), they are also usually seeking to satisfy other basic needs
that I have proposed: security, connection, a sense of effectiveness and con-
trol, and the need for comprehension of reality. The same is true when peo-
ple act in positive ways. Reconciliation with others may arise from and has
effects on identity, but also fulfills varied basic needs and expresses and af-
firms certain values.
Individual identity is a storehouse of and is shaped by all the experiences
a person has, at least all significant experiences. For example, it is shaped by
a history of fulfillment and frustration of basic needs. Some may regard it,
therefore, as superordinate to other aspects of personality and their related
processes. In my view, however, the most useful theory will maintain and
develop conceptual distinctions. The important task for basic theory build-
ing is to identify the interrelations of concepts and related processes.
How are these concepts, underlying structures, and processes interre-
lated in the tragic domain of human life this chapter addresses? Is identity a
more basic need than others? When a basic need such as security or con-
nection is fulfilled, does that automatically strengthen identity? How can
basic needs be fulfilled constructively for whole groups of people? Will 
answers to these and other questions help in halting and preventing group
violence? How can we apply gains we make in theoretical understanding to
help us in the practical realm, in preventing violence?
These questions point to an obligation of psychologists. The obligation
is even deeper than applying knowledge to the great practical issues of hu-
man life. It is to be purposeful in our efforts to gain knowledge, whether in
the laboratory or through experience in real-life settings. The obligation is
to be guided by a vision of reducing human suffering and enhancing human
welfare, especially when the knowledge we seek relates to such important
matters as the origins and prevention of genocide and mass killing.
Individual and Group Identities in Genocide and Mass Killing
181


References
Allport, G. W. (1954). The nature of predjudice. Reading, MA: Addison-Wesley.
Bar-T
al, D., & Staub, E. (1997). Introduction: The nature and forms of patriotism.
In D. Bar-T
al & E. Staub (Eds.), Patriotism in the lives of individuals and groups.
Chicago: Nelson-Hall.
Baumeister, R. F. (1997). Evil: Inside human violence and cruelty. New York: W.H.
Freeman.
Bigelow, K. R. (1993). A campaign to deter genocide: The Baha’i experience. In H.
Fein, (Ed.), Genocide watch. New Haven: Yale University Press.
Cairns, E., & Darby, J. (1998). The conflict in Northern Ireland. American Psycholo-
gist, 53, 754–760.
Craig, G. A. (1982). The Germans. New York: New American Library.
des Forges, A. (1999). Leave none to tell the story: Genocide in Rwanda. New York:
Human Rights Watch.
Epstein, S. (1980). The self-concept: A review and the proposal of an integrated
theory of personality. In E. Staub (Ed.), Personality: Basic aspects and current re-
search. Englewood Cliffs, NJ: Prentice-Hall.
Fein, H. (1993). Accounting for Genocide after 1945: Theories and some findings.
International Journal of Group Rights, 1, 79–106.
Fogelman, E. (1994). Conscience and courage: Rescuers of Jews during the Holocaust.
New York: Anchor Books.
Harff, B., & Gurr, T. R. (1990). Victims of the state genocides, politicides and group
repression since 1945. International Review of Victimology, 1, 1–19.
Gourevich, P. (1998). We wish to inform you that tomorrow we will be killed with our
families. New York: Farrar, Straus and Giroux.
Herman, J. (1992). Trauma and recovery. New York: Basic Books.
Hilberg, R. (1961). The Destruction of the European Jews. New York: Harper & Row.
Hitler, A. (1925). Mein Kampf. Translated by Ralph Manheim. Boston, MA:
Houghton Mifflin Company.
Janoff-Bulman, R. (1992). Shattered assumptions. New York: The Tree Press.
Jarymowitz, M. (1977). Modification of self-worth and increment of prosocial sensi-
tivity. Polish Psychological Bulletin, 8, 45–53.
Kincaid, E. (1959). In every war but one. New York: Norton.
Konner, M. (1978). Social and personality development: An anthropological 
perspective. In M. Lamb (Ed.), Social and personality development. New York:
Holt.
Kressel, N. J. (1996). Mass hate: The global rise of genocide and terror. New York:
Plenum Press.
Levin, S., Sidanius, J., Rebinowitz, J. L., & Frederico, C. (1998). Ethnic identity, 
legitimizing ideologies and social status: A matter of ideological asymmetry.
Political Psychology, 19, 373–404.
Lifton, R. J. (1986). The Nazi doctors: Medical killing and the psychology of genocide.
New York: Basic Books.
McCann, L. I., & Pearlman, L. A. (1990). Psychological trauma and the adult survivor:
Theory, therapy, and transformation. New York: Bruner/Mazel.
Miller, A. (1982). For your own good: Hidden cruelty in child-rearing and the roots of 
violence. New York: Farrar, Straus and Giroux.
Myers, D. (1999). Social psychology. New York: McGraw-Hill.
182
The Contribution of Social Identity to Violent Intergroup Conflict


Oliner, S. B., & Oliner, D. (1988). The altruistic personality: Rescuers of Jews in Nazi 
Europe. New York: Free Press.
Opotow, S. (Ed.). (1990). Moral exclusion and injustice. Journal of Social Issues, 46 (1).
Pennebaker, J. W., & Keough, K. A. (1999). Revealing, organizing and reorganizing
the self in response to stress and emotion. In R. J. Contrada, & R. D. Ashmore
(Eds.), Self, social identity, and physical health. (pp. 101-121) London: Oxford. 
Pettigrew, T. F. (1997). Generalized intergroup contact effects on prejudice. Person-
ality and Social Psychology Bulletin, 23, (2), 173–185.
Prunier, G. (1995). The Rwanda crisis: History of a genocide. New York: Columbia
University Press.
Reese, H. (1961). Relationships between self-acceptance and sociometric choices.
Journal of Abnormal and Social Psychology, 62.
Rummel, R. J (1994). Death by government. New Brunswick, NJ: Transaction 
Publishers.
Salmivalli, C., Kaukiainen, A., Kaistaniemi, L., & Lagerspetz, K. M. J. (1999). Self-
evaluated self-esteem, peer-evaluated self-esteem, and defensive egotism as pre-
dictors of adolescents’ participation in bullying situations. Personality and Social
Psychology Bulletin, 25, 1268–1278.
Sampson, E. E. (1988). The debate on individualism. American Psychologist, 47,
15–22.
Schatz, R., & Staub, E. (1997). Manifestations of blind and constructive patriotism.
In D. Bar-T
al & E. Staub (Eds.), Patriotism in the lives of individuals and groups.
Chicago: Nelson-Hall.
Schatz, R., Staub, E. & Lavine, H. (1999). On the varieties of national attachment:
Blind versus constructive patriotism. Political Psychology, 20. 151–175.
Simpson, C. (1993). The splendid blond beast. New York: Grove Press.
Smith, N. S. (1998). The psychocultural roots of genocide. American Psychologist, 53,
743–753.
Soetters, J. L. (1996). Culture and conflict: An application of Hofstede’s theory to
the conflict in the former Yugoslavia. Peace and Conflict: Journal of Peace Psychol-
ogy, 2 (3), 233–244.
Staub, E. (1989a). Steps along the continuum of destruction: The evolution of by-
standers, German psychoanalysts and lessons for today. Political Psychology, 10,
39–53.
Staub, E. (1989b). The roots of evil: The origins of genocide and other group violence. New
York: Cambridge University Press.
Staub, E. (1993a). Individual and group selves, motivation and morality. In T. Wren,
and G. Noam, (Eds.), Morality and the self. (pp. 337–359) Cambridge: MIT Press.
Staub, E. (1993b). The psychology of bystanders, perpetrators and heroic helpers.
International Journal of Intercultural Relations, 17, 315–341.
Staub, E. (1996a). Altruism and aggression in children and youth: Origins and cures.
In R. Feldman, (Ed.), The psychology of adversity. Amherst: University Massachu-
setts Press.
Staub, E. (1996b). Cultural-societal roots of violence: The examples of genocidal 
violence and of contemporary youth violence in the United States. American 
Psychologist, 51, 117–132.
Staub, E. (1997). Blind versus constructive patriotism: Moving from embeddedness
in the group to critical loyalty and action. In E. Staub & D. Bar-T
al (Eds.), Patri-
otism. Chicago: Nelson-Hall.
Individual and Group Identities in Genocide and Mass Killing
183


Staub, E. (1998a). Basic human needs and their role in altruism and aggression. Un-
published manuscript. Department of Psychology. University of Massachusetts
at Amherst.
Staub, E. (1998b). Breaking the cycle of genocidal violence: Healing and reconcilia-
tion. In J. Harvey (Ed.), Perspectives on loss: A sourcebook. Philadelphia: T
aylor and
Francis.
Staub, E. (1999a). A brighter future: The evolution of caring and nonviolent chil-
dren. Book in preparation.
Staub, E. (1999b). The origins and prevention of genocide, mass killing, and other
collective violence. Peace and Conflict: Journal of Peace Psychology, 5 (4), 303–336.
Staub, E. (1999c). The roots of evil: Social conditions, culture, personality and basic
human needs. Personality and Social Psychology Review, 3, 179–192.
Staub, E. (2000a). Genocide and mass killing: Origins, prevention, healing and 
reconcilation. Political Psychology, 21, 367–382.
Staub, E. (2000b). Mass murder: Origins, prevention and U.S. involvement. Encyclo-
pedia of violence in the United States. New York: Charles Scribner’s Sons, Reference
Books.
Staub, E., & Pearlman, L. (1996). Trauma and the fulfillment of the human poten-
tial. Workshop presented of the meetings of the International Society for Trau-
matic Stress Studies. November, San Francisco.
Staub, E., & Pearlman, L. A. (in press). Healing, reconciliation and forgiving after
genocide and other collective violence. In Forgiveness and Reconciliation, Radnor,
PA. T
empleton Foundation Press. 
Staub, E., Pearlman, L. A., Gubin, A., and Hagengimana, A. Healing, forgiveness
and reconciliation in Rwanda: Intervention and experimental evaluation. Manu-
script in preperation. Department of Psychology, University of Massachusetts at
Amherst.
T
ajfel, H. (1978). Social categorization, social identity and social comparison. In H.
T
ajfel (Ed.), Differentiation between social groups (pp. 61–76). London: Academic
Press.
T
ajfel, H. (1982). Social psychology of intergroup relations. Annual Review of Psy-
chology, 33, 1–39.
T
ec, N. (1986). When light pierced the darkness: Christian rescue of Jews in Nazi-occupied
Poland. New York: Oxford University Press.
Triandis, H. C. (1994). Culture and social behavior. New York: McGraw-Hill.
Turner, J. C. (1987). Rediscovering the social group: A self-categorization theory. New
York: Basic Blackwell.
184
The Contribution of Social Identity to Violent Intergroup Conflict


PART IV
THE ROLE OF SOCIAL 
IDENTITY IN REDUCING 
INTERGROUP CONFLICT


This page intentionally left blank 


Herbert C. Kelman
8
The Role of National Identity in
Conflict Resolution
Experiences from Israeli-Palestinian 
Problem-Solving Workshops
My colleagues’ and my work as scholar-practitioners has focused on analysis
and resolution of protracted, seemingly intractable conflicts between na-
tional, ethnic, or other kinds of identity groups, best exemplified by inter-
communal conflicts, such as those in Cyprus, Northern Ireland, Sri Lanka,
Bosnia, and apartheid South Africa. My own most intensive and extensive
experience, over the past quarter-century, has been with the Israeli-Pales-
tinian conflict, and my analysis draws primarily on that experience.1
Using the Israeli-Palestinian conflict as a case in point, this chapter ex-
amines the way in which issues of national identity can exacerbate an inter-
national or intercommunal conflict and the way in which such issues can be
addressed in conflict-resolution efforts. The chapter starts out with a brief
history of the Israeli-Palestinian conflict, setting the stage for the identity
issues at the heart of the conflict. It then proceeds to describe the struggle
over national identity between the two peoples, which has led them to per-
ceive their conflict in zero-sum terms, with respect to not only territory and
resources but also national identity and national existence. Next, it argues
that long-term resolution of this and similar deep-rooted conflicts requires
changes in the groups’ national identities, such that affirmation of one
group’s identity is no longer predicated on negation of the other’s identity.
Such identity changes are possible as long as they leave the core of each
group’s national identity intact. Furthermore, the chapter proceeds to ar-
gue, such changes need to be and can be “negotiated” between the two
groups. One venue for negotiating identity, described in the next section, is
provided by the problem-solving workshops between Israeli and Palestinian
elites that my colleagues and I have convened for many years. Finally, the
187


paper concludes with an illustration of the possibilities and limits of the ne-
gotiation of identity, based on a joint Israeli-Palestinian exploration of the
problem of Palestinian refugees.
The Israeli-Palestinian Conflict
The Israeli-Palestinian conflict is now more than a century old (see T
essler,
1994, for a comprehensive account of the history of the conflict, or
Mendelsohn, 1989, and Gerner, 1991, for shorter accounts). Its origins go
back to the birth of political Zionism at the end of the 19th century (see
Halpern, 1969, and Hertzberg, 1973). The early decades of the 20th cen-
tury brought to Palestine waves of Jewish immigrants who purchased land,
built settlements and social institutions, and clearly signaled their intention
to establish a Jewish homeland and ultimately a Jewish state in Palestine.
The growing Jewish presence was soon perceived as a threat by the Arab
population of the land, which was itself influenced by the development of
Arab nationalism and the construction of a specifically Palestinian identity
(see Muslih, 1988, and Khalidi, 1997). Violence first erupted in the 1920s
and has continued to mark the relationship between the two peoples ever
since.
During the period of the British mandate, which was established after
World War I, various formulas for the political future of Palestine were ex-
plored, including partition and establishment of a federal state, but none
was found to be acceptable to both the Arab and the Jewish populations (or
indeed to either one of them). In November 1947—in the wake of World
War II and the decimation of European Jewry—the United Nations Gen-
eral Assembly voted to end the British mandate over Palestine (on May 15,
1948) and to partition the land into a Jewish and an Arab state. The Zionist
leadership accepted the partition plan, with reservations. The Arab leader-
ship, both within Palestine and in the neighboring states, rejected it. Fight-
ing between the two sides broke out immediately after adoption of the UN
resolution and turned into all-out war after May 15, 1948, when the British
forces withdrew, the Jewish leadership in Palestine declared the indepen-
dent state of Israel, and regular armies from the neighboring Arab states
joined the fray. Fighting continued until early 1949.
In July of 1949, Israel and the Arab states signed armistice agreements
(though the state of war continued). The armistice lines became the official
borders of the State of Israel. These borders included a larger portion of
Palestine than the UN partition plan had allotted to the Jewish state. The
Arab state envisioned by the partition plan did not come into being. T
wo
parts of mandatory Palestine remained under Arab control: the West Bank,
which was eventually annexed by Jordan, and the Gaza Strip, which came
under Egyptian administration. The establishment of Israel and the war of
1948–49 also created a massive refugee problem, with the flight or expul-
188
The Role of Social Identity in Reducing Intergroup Conflict


sion of hundreds of thousands of Palestinian Arabs from their homes in the
part of Palestine that became the State of Israel. I shall return to the refugee
problem later in this chapter.
The map changed radically as a result of the Arab-Israeli war of June
1967—and, along with it, the political atmosphere in the Middle East. By
the end of the Six-Day War, as Israelis called it, Israel occupied the West
Bank and the Gaza Strip, thus extending its control over the entire territory
of mandatory Palestine. It also occupied the Sinai Peninsula and the Golan
Heights—Egyptian and Syrian territories, respectively. The new geopoliti-
cal and strategic situation created by the 1967 War led to the Palestinianiza-
tion of the Arab-Israeli conflict, bringing it back to its origin as a conflict
between two peoples over—and increasingly within—the land they both
claimed (Kelman, 1988).
The Palestinianization (or re-Palestinianization) of the conflict has mani-
fested itself in the actions of the Arab states, of the Palestinian community it-
self, and of Israel. Israel’s neighboring Arab states gradually withdrew from
the military struggle against Israel—though not before another major war in
1973—leaving it, essentially, to the Palestinians themselves. The disengage-
ment of the Arab states became dramatically clear with the 1977 visit to
Jerusalem of President Anwar Sadat of Egypt, the largest and most powerful
Arab state, an initiative that led to the Camp David accord of 1978 and the
Egyptian-Israeli peace treaty of 1979. The Palestinians took repossession of
their struggle, which in the years between 1949 and 1967 had been mostly in
the hands of the Arab states. Fatah, under the leadership of Yasser Arafat, and
other Palestinian guerrilla organizations grew in strength and eventually
took over the Palestine Liberation Organization (PLO), which was originally
a creature of the Arab League. Between the 1960s and the 1980s, the Pales-
tinian movement gradually shifted its emphasis from the liberation of all 
of Palestine through armed struggle against Israel to the establishment of an
independent Palestinian state in the West Bank and Gaza through largely 
political means. The end of the occupation became the immediate goal of the
movement and, with the onset of the intifada—the uprising in the West Bank
and Gaza—in December 1987, the occupied territories became the focal
point of its struggle.
On the Israeli side, the intifada further underlined the Palestinianization
of the conflict in the wake of the 1967 War. What had been largely an inter-
state conflict between 1948 and 1967 had now been internalized by Israel—
that is, transformed into a continuous confrontation with a resentful Pales-
tinian population, living under occupation within Israel’s post-1967
borders. Many Israelis were persuaded by the intifada that continuing occu-
pation was not tenable and that the Palestinians were indeed a people,
whose national movement had to find some political expression if there was
to be a peaceful accommodation between the two sides (Kelman, 1997c).
By the end of the 1980s, there was a strong interest on all sides in finding
a peaceful accommodation and an increasing recognition that some version
The Role of National Identity in Conflict Resolution
189


of a two-state solution would provide the best formula for a broadly accept-
able historic compromise. The political obstacles to such a solution, how-
ever, remained severe. A number of strategic and micropolitical considera-
tions—traceable, in particular, to the end of the Cold War and the
aftermath of the Gulf War—eventually brought the leaderships on both
sides to the negotiating table at the Madrid Conference in 1991 and the
subsequent talks in Washington. These talks, however, never developed
momentum. It was only after Prime Minister Yitzhak Rabin came into
power in Israel in 1992, at the head of a government led by the Labor Party,
and gradually (and reluctantly) concluded that Israel would have to deal di-
rectly with the PLO leadership in order to make progress in the negotia-
tions, that a breakthrough was finally achieved. This breakthrough oc-
curred in the secret Oslo talks, which culminated in the exchange of letters
of mutual recognition between the PLO and the State of Israel (which, in
my view, was the most significant achievement of the Oslo process) and the
Declaration of Principles signed in Washington in September 1993 (see
Kelman, 1997c, for further details).
The Oslo breakthrough occurred because a number of long-term and
short-term interests of the parties—to which I alluded briefly—persuaded
them of the necessity of reaching an agreement. “But a significant factor
contributing to the breakthrough was the conclusion, on both sides, that
negotiations were not only necessary but also possible—that they could
yield an acceptable agreement without jeopardizing their national existence.
This sense of possibility evolved out of interactions between the two sides
that produced the individuals, the ideas, and the political atmosphere re-
quired for negotiations” (Kelman, 1997d, p. 213). A variety of unofficial
contacts between the two sides, including the workshops to be described in
this chapter, contributed to creating this sense of possibility and the climate
conducive to negotiations (see Kelman, 1995).
Though the Oslo agreement was a genuine breakthrough in the effort to
resolve the Israeli-Palestinian conflict, the process of negotiating an interim
agreement and initiating final-status negotiations has faced numerous ob-
stacles and experienced dangerous setbacks, all reminding us of the in-
tractability of conflicts between identity groups, such as the Israeli-Palestin-
ian conflict. Significant changes in the conflict have taken place with the
establishment of a Palestinian Authority (PA) in the West Bank and Gaza
Strip, which constitutes the scaffolding for a Palestinian state. But the path
to a negotiated agreement has been obstructed by such events as the He-
bron massacre of Palestinians at prayer, the assassination of Yitzhak Rabin,
the suicide bombings against Israeli civilians, the continuation of the Israeli
settlement process in the West Bank, the deteriorating economy and quality
of life in the West Bank and Gaza, the corrupt and autocratic leadership of
the PA, and the three-year rule of Prime Minister Netanyahu in Israel.
The election of a new Israeli government in May of 1999 brought new
hope that serious negotiations of the final-status issues will be pursued. But
190
The Role of Social Identity in Reducing Intergroup Conflict


the negotiations failed and, in late September 2000, a new intifada broke
out, producing a violent Israeli response and a serious deterioration of the
peace process. Nevertheless I feel fairly confident that the negotiations will
eventually succeed in establishing a Palestinian state. I am less confident
that this new state will be truly independent, secure, and viable, and that its
establishment will form the basis of a long-term peace and a cooperative,
mutually enhancing relationship with Israel, conducive to ultimate recon-
ciliation between the two peoples. How the identity issues that divide the
two peoples are addressed in the negotiations and the public debates that
surround them will play a major role in determining the quality and dura-
bility of the solution achieved at the negotiating table.
The Struggle over National Identity
Although national identity is carried by the individual members of a na-
tional group and can thus be studied as a property of individuals, the pre-
sent chapter refers to it as a collective phenomenon—as a property of the
group. “Insofar as a group of people have come to see themselves as consti-
tuting a unique, identifiable entity, with a claim to continuity over time, to
unity across geographical distance, and to the right to various forms of self-
expression, we can say that they have acquired a sense of national identity.
National identity is the group’s definition of itself as a group—its concep-
tion of its enduring characteristics and basic values; its strengths and weak-
nesses; its hopes and fears; its reputation and conditions of existence; its in-
stitutions and traditions; and its past history, current purposes, and future
prospects” (Kelman, 1997a, p. 171; see also Kelman, 1998b).
In the Israeli-Palestinian and other such conflicts, the threat to collective
identity is a core issue in the conflict, which is integrally related to the
struggle over territory and resources. Both peoples and their national
movements claim the same territory, and each seeks ownership of that terri-
tory and control over its resources as the basis of an independent state that
gives political expression to its national identity. The integrity of this col-
lective identity is critical to each group for several reasons. First, the in-
tegrity of the national identity is an end in itself, in that the identity serves
as a source of distinctiveness, unity, and continuity for the group and of a
sense of belongingness for its members. Second, the national identity con-
stitutes the ultimate justification of the group’s claim to ownership of the
land and control of its resources. And third, the national identity provides a
focus for developing and maintaining the group’s distinctive culture, reli-
gion, and way of life. The collective identity of each group is bolstered by a
national narrative—an account of the group’s origins, its history, and its re-
lationship to the land—that explains and supports its sense of distinctive-
ness, its positive self-image, and the justice of its claims and grievances.
In conflicts such as that between Israelis and Palestinians, in which the
The Role of National Identity in Conflict Resolution
191


two sides live in the same space and claim ownership of the same territory, it
is not only the actions of the other but also the identity and the very exis-
tence of the other that are a threat to the group’s own identity. The other’s
identity and its associated narrative challenge the group’s claims to owner-
ship—at least to exclusive ownership—of the land and its resources. The
other’s presence in the same space, particularly if it is accompanied by 
demands for a share of the power and for recognition of the other culture,
religion, and/or language, is perceived as a threat to the integrity and cohe-
siveness of the group’s society and its way of life.
These dynamics lead to a view of the conflict as a zero-sum struggle, not
only around territory but also around identity (Kelman, 1987). Acknowl-
edging the other’s identity becomes tantamount to jeopardizing the iden-
tity—and indeed the national existence—of one’s own group. Each side
“holds the view that only one can be a nation: Either we are or they are.
They can acquire national identity and rights only at the expense of our
identity and rights” (Kelman, 1987, p. 354). Thus, over the course of the 
Israeli-Palestinian conflict, there has been a systematic tendency on each
side to deny the other’s identity as a people, the authenticity of the other’s
links to the land, the legitimacy of the other’s claims to national rights, and
the very existence of the other as a national group (Kelman, 1978, 1982).
Negation of the other’s identity and of the narrative in which it is embedded
becomes so important to the conflict that it is incorporated in the identity
that each group constructs for itself and in the narrative that it presents to
the world (Kelman, 1999).
The contrasting Israeli and Palestinian narratives about the creation of
the State of Israel in 1948 both rely on the negation of the other to bolster
the justice of their own cause. For Israelis, the creation of Israel represented
a rightful return of the Jewish people to its ancestral homeland. Establish-
ment of a Jewish state in Palestine did not, in their eyes, constitute an injus-
tice to the Arabs who resided there, because Palestinian Arabs were not a
people, distinct from the Arab inhabitants of surrounding countries, and
never exercised sovereignty in Palestine. Moreover, in the Israeli narrative,
the responsibility for the Palestinian refugee problem and the suffering of
the Palestinian Arab population rests with their own aggressive and incom-
petent leadership, which rejected all compromise and initiated violent at-
tacks in the effort to block the establishment of Israel. For Palestinians, by
contrast, the creation of Israel represented an act of usurpation by Euro-
pean settlers, who forcefully displaced the indigenous population and de-
stroyed their society, their property, and their way of life. In the Palestinian
narrative, Jews are a religious group, not a nation entitled to its own state,
and Zionism is a form of settler colonialism that imposed itself on a region
in which it has no roots. Each identity gains some of its strength and legiti-
macy from negating and delegitimizing the other.
Identities that rest in part on negation of the other inevitably take on 
an exclusivist and monolithic character (Kelman, 1997b). In the Israeli-
192
The Role of Social Identity in Reducing Intergroup Conflict


Palestinian conflict, such exclusivist and monolithic definitions of identity
have begun to give way in recent years. For significant segments of the two
population, however—and in some respects even for large majorities—such
definitions still prevail. 
In the Israeli-Palestinian case, a defining element of each group’s identity
is its relationship to the land and its history. Insofar as this relationship is
exclusive—that is, insofar as the group’s identity rests on the view that the
land and its history belong to it alone and that the other’s claims on them as
part of its own identity are illegitimate and inauthentic—there is little room
for conflict resolution. Conflict resolution becomes an option when the
parties accept the possibility that certain elements of identity may be shared
with the other, acknowledging that the other also has a profound attach-
ment to the land, anchored in authentic historical ties to it. Israelis and
Palestinians have been gradually moving toward acceptance of shared ele-
ments of identity as they have been searching for a political formula for
sharing the land. It has proved more difficult for the two sides, so far, to ac-
cept Jerusalem as a shared element of the two identities and to develop a
political formula to reflect that view.
Identities that rest on negation of the other also take on a monolithic
character; that is, all dimensions of the group’s identity—such as ethnicity,
religion, and language—tend to be viewed as highly correlated. The ide-
ology calls for complete correspondence between ethnic boundaries, politi-
cal boundaries, boundaries of emotional attachment, and boundaries of in-
tensive interaction. Self and other are, in principle, completely separated
along all of these lines. The Israeli-Palestinian conflict, as well as other pro-
tracted conflicts, particularly between identity groups living in close prox-
imity within a small space (such as Northern Ireland or Cyprus), might be
more amenable to resolution if there were some degree of disaggregation of
the monolithic identity, based on distinctions between different types of
boundaries. Such distinctions would allow for the development of a tran-
scendent identity—not in place of the particular ethnonational identities, but
alongside of them. In the Israeli-Palestinian case, a transcendent identity
could be fostered by separating the concept of the state, as a sovereign po-
litical entity, from that of the country, as a geographical entity. This distinc-
tion would allow the two communities to treat the entire country (Eretz
Yisrael or Palestine) as an object of common sentimental attachment and as
the framework for common instrumental pursuits (in such areas as develop-
ment and use of water resources, environmental protection, public health,
and tourism), while living in and identifying with separate political states
within that country.
The zero-sum view of identity and the mutual denial of the other’s iden-
tity that I have described create serious obstacles to conflict resolution. All
issues tend to become existential—matters of life and death for each side.
Compromise solutions that involve sharing of the land or agreeing on dif-
ferent boundaries for different purposes are likely to threaten exclusivist
The Role of National Identity in Conflict Resolution
193


and monolithic identities. The demonized other is not trusted to negotiate
in good faith and respect agreements. In short, when acceptance of the
other’s national rights and recognition of the other’s national identity are
seen as relinquishing the group’s own rights and jeopardizing its own iden-
tity, distributive solutions based on compromise are hard to achieve. Even if
the parties agree to make certain compromises in response to reality de-
mands and external pressures, these compromises are unlikely to lead to
durable changes in the relationship between the conflicting groups, con-
ducive to stable peace, mutually enhancing interaction, and ultimate recon-
ciliation. Lasting change requires mutual adjustments in collective identity.
Identity Changes
The stubborn resistance to change in collective identities is widely recog-
nized and taken for granted. Yet identities have to change, at least tacitly, if
protracted identity conflicts are to be settled and, certainly, if they are to be
resolved in a way that transforms the relationship and opens the way to rec-
onciliation. South Africa provides perhaps the best illustration of an arena
of intense, protracted conflict in which fundamental identity changes paved
the way to resolution and reconciliation, although it also illustrates the dif-
ficulties in changing the worldviews and the structural realities that became
entrenched in the apartheid era.
Despite their undeniable rigidities, identities are potentially changeable
(and indeed negotiable) for two reasons: First, unlike territory and resources,
they are not inherently zero-sum; though they are perceived and debated as
such in intense conflicts, it is in fact not the case that A’s identity can be rec-
ognized and expressed only if B’s identity is denied and suppressed. 
If the two identities are to become compatible, however, they have to be
redefined. And this points to the second reason for the potential change-
ability of group identities: they can be redefined because they are to a large
extent constructed. T
o view national identity as a social construction does
not imply that it is manufactured out of nothing. There may be cases in
which one can properly speak of an imagined past, invented to buttress a
newly formed identity (cf. Anderson, 1983). Generally, however, the social
construction of an identity draws on a variety of authentic elements held in
common within a group: a common history, language, or religion; common
customs, cultural expressions, experiences, values, grievances, aspirations
(Kelman 1997b). Typically, the social construction of an identity involves a
dual process of discovery (or rediscovery) and creation of such common ele-
ments (Kelman, 1997a). The social construction of the identity implies a
degree of arbitrariness and flexibility in the way the identity is composed
(which elements are admitted into it and which omitted from it), and in
what its boundaries are (who is included and who is excluded). These
choices depend on the opportunities and necessities perceived by the elites
194
The Role of Social Identity in Reducing Intergroup Conflict


that are engaged in mobilizing ethnonational consciousness for their politi-
cal, economic, or religious purposes (Kelman, 1997b). Serbs and Croats, for
example, share a common language and culture, but differ in religion and
historical experiences. Political leaders have at times focused on the simi-
larities in the effort to shape them into a single nation; at other times they
have magnified the differences to define them as separate—and mutually
antagonistic—nations.
Thus, although national identities are generally constructed out of real
experiences, these experiences can be ordered in different ways, resulting in
different boundaries and priorities. As a consequence, they can be—and
typically are—deconstructed and reconstructed. “In fact, the reconstruction
of identity is a regular, ongoing process in the life of any national group.
Identities are commonly reconstructed, sometimes gradually and some-
times radically, as historical circumstances change, crises emerge, opportu-
nities present themselves, or new elites come to the fore” (Kelman, 1997b,
p. 338). Clearly, therefore, there is room for maneuver in a group’s self-
definition, particularly with respect to the definition of group boundaries
and the priorities among different elements of the group’s identity.
Changes in identity over the course of a protracted conflict come about
through a combination of changed perceptions of the necessity and the possi-
bility of resolving a conflict that has become increasingly costly to the par-
ties. The mounting costs and dwindling prospects of governing Algeria for
the French, of maintaining apartheid for white South Africans, of the occu-
pation of Palestinian territories for Israelis, and of the armed struggle for
the Palestinians created the necessity for changes in identity: Algeria as an in-
tegral part of France, South Africa under exclusive white control, Israel
within the borders of Greater Israel, and Palestinian repossession of the en-
tire homeland were assigned lower priority in the national identities of
these groups as it became clear to a majority that these aspirations could not
be realized at an acceptable cost. 
What made it possible to change these priorities was often the discovery
that accommodation of the other’s identity need not destroy the core of the
group’s own identity, and that a compromise solution to the conflict was
therefore negotiable. This kind of learning can take place in the course of
official or unofficial interactions between the groups or their members, in-
cluding the problem-solving workshops that my colleagues and I have con-
ducted: In the course of Israeli-Palestinian workshops, for example, partici-
pants have learned to differentiate their image of the enemy by discovering
that there are potential negotiating partners on the other side, that there is
a distinction between the other’s ideological dreams and operational pro-
grams, and that the other has positive goals beyond destruction of their
group (Kelman, 1987). They were enabled to enter into the enemy’s per-
spective, thus discovering the historical sources of the other’s claims and
grievances, the depth of the other’s fears, and the authenticity of the other’s
sense of peoplehood. They began to visualize a different future, discovering
The Role of National Identity in Conflict Resolution
195


possibilities for mutually beneficial coexistence and cooperation. As such
experiences multiply, and as the learnings produced by them are infused
into the two political cultures, each group may gradually change its identity
by eliminating the negation of the other’s identity as an element of its own
identity and perhaps even admitting the possibility of a partnership as a new
element of its own identity.
Negotiating Identity
The changes I have described are often the result of an explicit or implicit
process of negotiating identity. At its core, national identity is clearly non-
negotiable; indeed, the very idea of negotiating identity sounds like an oxy-
moron. National identity is a collective psychological conception, which
cannot be dictated or prescribed by outsiders. A group of people who define
themselves as a nation cannot be told that they have no right to do so be-
cause their self-definition does not conform to some set of theoretical, 
juridical, or historical criteria for doing so, or because their nationhood is
inconvenient to others. Nor does it make sense to tell them how to draw the
boundaries of the group: whom to include and whom to exclude. People are
a nation if they perceive themselves as such and are prepared to invest 
energy and make sacrifices in terms of that perception (Kelman, 1978).
Neither Palestinians nor Israelis will give up the core of their identity: their
sense of peoplehood, their attachment to the land, their conviction about
the historical authenticity of their links to that land, their commitment to
their national culture, language, and way of life. Nor will they give up the
national narrative that substantiates the justice of their cause. 
But there are many elements that can be added to or subtracted from an
identity without jeopardizing its core. In fact, changes in less central ele-
ments of the identity are often advocated precisely in order to protect the
core of the identity. It was on that basis that the majority of Israelis and
Palestinians came to accept territorial compromise—that is, a shrinking of
the territorial dimension of their identity—as the best available option for
maintaining their national identity. The Peace Now movement in Israel, for
example, advocated withdrawal from the occupied territories largely on the
grounds that this was the only way in which Israel could maintain its char-
acter as both a Jewish state and a democratic state. Yehoshafat Harkabi
(1986), a former chief of Israeli military intelligence and a prophetic voice
in the debate about Israeli-Palestinian peace, explicitly advocated a “smaller
Israel”—a “Zionism of quality” rather than a “Zionism of acreage.” He 
argued that Israel had to choose between withdrawing from the West 
Bank and making way for a Palestinian state there, or annexing the West
Bank with the consequence that Israel would eventually become a Palestinian
state. On the Palestinian side, the territorial dimension of the Palestinian
identity has gradually changed as the movement reflected on its realistic 
196
The Role of Social Identity in Reducing Intergroup Conflict


options. The thinking of the PLO evolved from advocacy of a Palestinian
Arab state in the whole of Palestine, to a secular democratic state, and even-
tually to a Palestinian state alongside of Israel comprising the West Bank
and Gaza (cf. Muslih, 1990). Significant segments of both societies still re-
ject territorial compromise on religious or ideological grounds and link
their national identity to possession of the land in its entirety. But the Pales-
tinian and Israeli mainstreams have by now come to terms with a national
identity that finds its political expression in only part of the land, as evi-
denced by the opinion polls that are now conducted on a regular basis in
both societies.
Such changes in elements of identity are a legitimate subject for “nego-
tiation” between groups whose identities clash, because the identity that
one group chooses for itself has significant implications for the rights, in-
terests, and identity of the other. Whenever one group translates the self-
definition of its nationhood into action—“by making territorial claims, by
demanding an independent state, by seeking to redraw borders, by declar-
ing who is included in the national identity and who is excluded from it, or
even by selecting a name for itself” (Kelman, 1997b, p. 337)—the other is
inevitably affected. Each group, therefore, has a legitimate concern about
the way the other defines itself, the way it formulates its national identity. It
is not surprising, then, that identity issues play an important role in the for-
mal and informal processes of pre-negotiation and negotiation—that is, in
the efforts to open a way to the negotiating table and to reach agreement
around the table.
T
o some extent, identity issues are part of the subject matter of the offi-
cial negotiations. I have already referred to the territorial dimension of
identity: insofar as Israelis and Palestinians are negotiating on the basis of a
“land for peace” formula, they are accepting territorial limits to their na-
tional identities, which have, after all, been historically linked to the whole
of the land. Similarly, the mutual recognition between Israel and the PLO,
as expressed in the exchange of letters between Arafat and Rabin—which I
have described above as the most important breakthrough of the Oslo
agreement —can be viewed as a product of the negotiation of identity: an
act of acceptance and legitimization of the other who in the past had been
defined as the antithesis to the self.
Although redefined identities are thus promulgated around the official
negotiating table, the “negotiation” of identity is primarily an informal, un-
official process in which members of the conflicting parties explore and 
invent ways of accommodating their group identities to one another. The
purpose of negotiation in this looser sense of the term is not to produce 
political agreements, but to develop joint understandings and formulations
that can help pave the road to political agreements at the official level. Im-
plicitly and explicitly, this kind of “negotiation” has been a central focus for
problem-solving workshops between Israelis and Palestinians that my col-
leagues and I have conducted over the past quarter-century.
The Role of National Identity in Conflict Resolution
197


Problem-Solving Workshops
Problem-solving workshops are the central instrument of interactive problem
solving, an unofficial, third-party approach to the resolution of international
and intercommunal conflicts, derived from the pioneering work of John
Burton (1969, 1979, 1984) and anchored in social-psychological principles
(see Kelman, 1972, 1979, 1992, 1998a). A workshop is a specially con-
structed, private space in which politically involved and often politically in-
fluential (but generally unofficial) members of conflicting communities can
interact in a nonbinding, confidential way. The microprocess of the work-
shop provides them the opportunity to penetrate each other’s perspective;
to explore both sides’ needs, fears, priorities, and constraints; and to engage
in joint thinking about solutions to the conflict that would be responsive to
the fundamental concerns of both sides.
Let me describe a typical “one-time” workshop for Israeli and Palestin-
ian participants. By “one-time” I mean that this particular group of Israelis
and Palestinians convenes only for this one occasion. Some of the indi-
viduals may have participated in more than one such workshop, and the 
one-time workshops that we have held over the years have had a cumulative
effect within the two societies. But, until 1990, we made no effort to recon-
vene the same group of participants for another occasion. The workshops
take place under academic auspices and are facilitated by a panel of social
scientists knowledgeable about international conflict, group process, and
the Middle East. The workshops usually begin with two pre-workshop 
sessions, about four hours in length, during which the third party meets
separately with each of the two parties. The workshop itself typically lasts
about two and a half days, often scheduled over an extended weekend. The
participants include three to six Israelis and an equal number of Palestin-
ians, plus a third party of three or more members.
The Israeli and Palestinian participants have included parliamentarians,
leaders and activists of political parties or political movements, journalists,
editors, directors of think tanks, and politically involved academicians—that
is, scholars who not only publish academic papers but who also write for
newspapers and appear in the media, who serve as advisors to political lead-
ers, and some of whom move back and forth between government and
academia. Some of our participants have been former diplomats, officials,
or military officers, and many were later to become negotiators, ambas-
sadors, cabinet ministers, parliamentarians, and leading figures in the media
and research organizations. We look for people who are within the main-
stream of their societies and close to the center of the political spectrum. At
the same time, they have to be people who are at least willing to explore the
possibility of a negotiated solution and to sit down as equals with members
of the other party.
The central ground rule of problem-solving workshops is the principle
of privacy and confidentiality. In the early days of our work, confidentiality
198
The Role of Social Identity in Reducing Intergroup Conflict


was particularly important for the protection of our participants, because
the mere fact that they were meeting with the enemy was controversial and
exposed them to political and even physical risks. Confidentiality is equally
important, however, for the protection of the process that we are trying to
promote in workshops. We are trying to encourage the participants to talk
and listen to each other, rather than focus on their constituencies, on third
parties, or on the record. We want them to think out loud, to experiment
with ideas, to explore different options, without having to worry about how
others would react if their words in the group were quoted outside. We
want them to engage in a type of interaction that is generally not feasible
among parties engaged in a bitter conflict—a type of interaction that, in-
deed, deviates from the conflict norms that usually govern their behavior:
an interaction that is analytic rather than polemical, one in which the par-
ties seek to explore each other’s perspective and gain insight into the causes
and dynamics of the conflict; an interaction that is problem-solving rather
than adversarial, one in which the parties sidestep the usual attempt to allo-
cate blame and, instead, take the conflict as a shared problem that requires
joint effort to find a mutually satisfactory solution.
The agenda of a problem-solving workshop is designed to allow this kind
of interaction to unfold. The core agenda of a one-time workshop has four
components. First, each side is asked to discuss its central concerns in the
conflict—the fundamental needs that would have to be addressed and the
existential fears that would have to be allayed if a solution is to be satisfac-
tory to it. The parties are asked not to debate the issues raised, although
they may ask for clarification of what the other says. The purpose is for
each side to gain an adequate understanding of the other’s needs, fears, and
concerns, from the perspective of the other. Once they have demonstrated
that they understand the other’s needs to a significant degree, we move to
the second phase of the agenda: joint thinking about possible solutions.
What participants are asked to do in this phase is to develop, through an in-
teractive process, ideas about the overall shape of a solution for the conflict
as a whole, or perhaps, a particular issue in the conflict, that would address
the needs and fears of both sides. They are given the difficult assignment of
thinking of solutions that would meet not only their own side’s needs, but
the needs of both sides.
Once the participants have developed some common ground in this
process of joint thinking, we turn to the third phase of the workshop: dis-
cussion of the political and psychological constraints within the two soci-
eties that would create barriers to carrying out the ideas for solution that
have been developed in the group. We deliberately leave the discussion of
constraints to the third phase, so that it does not hamper the creative
process of jointly generating new ideas. Finally, depending on how much
progress has been made and how much time is left, we ask the parties to en-
gage in another round of joint thinking—this time about ways of overcom-
ing the constraints that have been presented. The participants are asked to
The Role of National Identity in Conflict Resolution
199


come up with ideas about what their governments, their societies, and they
themselves might do—separately or jointly—that would help to overcome
the barriers to negotiating mutually satisfactory solutions to the conflict.
The third party in our model enacts a strictly facilitative role. It does not
propose solutions, nor does it participate in the substantive discussions. Its
task is to create the conditions that allow ideas for resolving the conflict to
emerge out of the interaction between the parties themselves. The facilita-
tion of the third party, however, is an important part of the process. The
third party sets the ground rules and monitors adherence to them; it helps
to keep the discussion moving in constructive directions, tries to stimulate
movement, and intervenes as relevant with questions, observations, and
even challenges. It also serves as a repository of trust for parties who, by de-
finition, do not trust each other: They feel safe to come to the workshop
because they trust the third party to maintain confidentiality and to protect
their interests.
Workshops have a dual purpose: to produce changes, in the form of new
insights and ideas, in the individual participants; and to transfer these
changes into the political process at the levels of both public opinion and
decision making. These two purposes may at times create contradictory re-
quirements—most notably in the selection of participants (Kelman, 1992).
T
o balance these contradictory requirements, we look for participants who
are not officials, but politically influential. They are thus more free to en-
gage in the process but, at the same time, because of their positions within
their societies, any new ideas that they develop in the course of a workshop
can have an impact on the thinking of decision makers and the society at
large.
The Israeli-Palestinian workshops we carried out until 1990 were all
one-time events designed to create a climate conducive to movement to the
negotiating table. In 1990, we organized our first continuing workshop, in
which a group of high-level Israelis and Palestinians met periodically over a
three-year period (Rouhana & Kelman, 1994). In 1994, Nadim Rouhana
and I convened a Joint Working Group on Israeli-Palestinian Relations,
which is still in progress.2 For the first time in our program, this group has
been engaged in producing jointly authored concept papers on some of the
final-status issues in the Israeli-Palestinian negotiations and on the future
relationship between the two societies. One of three papers that the group
has published so far—a paper on the problem of Palestinian refugees and
the right of return (Alpher & Shikaki, et al., 1999)—provides the starting
point for my detailed illustration of the negotiation of identity, to be pre-
sented in the next section.
Much of the discussion in our workshops, from the beginning and until
this day, has focused, in effect, on a process of negotiating collective identi-
ties. In our experience, such a process can be productive only if it is based
on mutual respect for the core of the other’s identity and on the principle of
reciprocity. Each side must know that the other does not seek to undermine
200
The Role of Social Identity in Reducing Intergroup Conflict


its group identity, and each must take care not to undermine the other’s
identity. And each must know that the risks it takes in acknowledging the
other’s claims, rights, and authenticity will be reciprocated by the other’s ac-
knowledgment of its claims, rights, and authenticity.
Starting from the understanding that neither side is prepared to negoti-
ate the core of its identity—its peoplehood, its relationship to the land, the
basic justice of its cause—or the general lines of its national narrative, there
remain various elements of each group’s identity that can be “negotiated” in
the interest of mutual accommodation. Let me illustrate some of the possi-
ble changes in identity that can be and have been discussed in problem-
solving workshops and similar encounters and that have, over time, begun
to penetrate the Israeli and Palestinian political cultures.
1. Many members of both communities have become able to remove the
negation of the other’s identity as an integral part of their own identity.
Though the other may still be seen as an obstacle to achieving one’s own
national goals, the other is not as often seen as the antithesis of one’s own
identity whose demise is a condition for one’s own survival. Thus, many 
Israelis have come to accept the reality of Palestinian peoplehood, particu-
larly after observing Palestinians’ readiness to make sacrifices for their na-
tional cause during the intifada and Palestinians’ celebration of the signing
of the Oslo agreement in September 1993. Interestingly, Israelis saw paral-
lels between these events and their own struggle for statehood and celebra-
tion at attaining it—a significant degree of identification with the other
whose existence had previously been denied. Many Palestinians, on their
part, now recognize the right of Israelis to their state, on the grounds that
the state has existed for over half a century and that its dismantlement
would create a new injustice to the generations that were born into it. Very
few Palestinians, on the other hand, are prepared to acknowledge the his-
torical links of Jews to the land, which might be seen as justification for the
establishment of the Jewish state in the first place.
2. We have seen signs of softening of the exclusiveness of group identity,
which allows for the recognition that—despite the validity of one’s own
claims—the other too has valid claims. The recognition of elements of
identity shared with the other, such as a common identification with the
land, opens the way to political solutions based on sharing territory and re-
sources. In a recent workshop, for example, mainstream Israelis and Pales-
tinians were able to agree—much to everyone’s surprise, including their
own—on a formula for sharing Jerusalem: a united city containing the capi-
tals of both states. Public opinion polls on both sides also suggest greater
willingness to share Jerusalem than had been widely assumed (Segal, 1999).
3. Workshop participants have experimented with disaggregating the
monolithic nature of their identities, recognizing that there are different
boundaries of group identity (such as ethnic boundaries, political bound-
aries, boundaries of sentimental attachment) that do not necessarily coin-
cide. This recognition opens the way to the development of transcendent
The Role of National Identity in Conflict Resolution
201


identities, which might allow the two peoples to maintain a common at-
tachment to the country while “owning” only part of that country as their
political state. The concept of a “united country with divided sovereignties”
was discussed in one of our workshops in the early 1980s. In a more recent
workshop, the idea of establishing different kinds of boundaries was ex-
plored in the attempt to find solutions to the problem of Israeli settlements
in the areas in which Palestinians hope to establish their state.
4. Workshop discussions can help to identify outdated elements of
group identity, which refer to maximalist goals and dreams of glory or self-
aggrandizing images that have no current political relevance but poison the
climate for conflict resolution. Examples here might be Palestinian refer-
ences to the armed struggle as the way to eliminate the Zionist entity, or 
Israeli references to the Zionist project of making the desert bloom. Work-
shops have often sensitized participants to words and images that humiliate
and frighten the other and could be discarded with minimal cost to group
identity.
5. In the course of workshop discussions, participants may decide to re-
order the priorities within their national identities, such that certain ele-
ments (e.g., territorial ambitions), which may not have been given up but
have become too costly to pursue, are relegated to low priority and thus be-
come available for negotiated compromise. Thus, over time, Palestinians
(in our workshops and in the larger society) decided to give priority to end-
ing occupation and establishing a Palestinian state over recovering the lost
land in its entirety. Israelis gave priority to maintaining the Jewish character
of Israel over controlling the whole of the land.
6. Finally, workshop participants may “negotiate” changes in national
narratives that accommodate the other’s view of history as much as possible,
such as accepting a share of the responsibility for the course of the conflict.
The above-mentioned concept paper of our current Joint Working Group
on Israeli-Palestinian Relations on “The Palestinian refugee problem and
the right of return” (Alpher & Shikaki, et al., 1999) provides a good illustra-
tion of such an effort to negotiate identity. I shall conclude this chapter with
a more detailed discussion of this concept paper, trying to show what made
it possible to achieve common ground in this effort, what still remains to be
done, and possible directions for achieving further progress.
The Problem of Palestinian Refugees
From the point of view of the struggle over national identity, the problem
of Palestinian refugees is probably the most important and the most diffi-
cult issue to resolve in the Israeli-Palestinian conflict. It may ultimately be
more important and more difficult than the issue of Jerusalem, because it
bears directly on the contrasting views of the justice or injustice of the es-
tablishment of Israel presented by the two national narratives.
202
The Role of Social Identity in Reducing Intergroup Conflict


The problem of Palestinian refugees goes back to the 1948 war that 
followed the declaration of an independent State of Israel. In the course 
of that war, somewhere between 600,000 and 760,000 Palestinian Arabs—
about half of the Arab population of mandatory Palestine—were displaced
from their homes in what became the State of Israel (Morris, 1987, 
pp. 297–298). They were turned into refugees and most of them and/or
their descendants remain refugees to this day. The precise number of Pales-
tinian refugees today is in dispute. The figure given by UNRWA—the
United Nations Relief and Works Agency, which is charged with the wel-
fare of Palestinian refugees—is almost three and a half million (as of Janu-
ary 1998). This number includes Palestinians who registered as refugees
with UNRWA after losing their homes in 1948, as well as their descen-
dants. The majority of this population lives in refugee camps in Gaza, the
West Bank, Jordan, Lebanon, and Syria. Palestinian refugees in Lebanon
(some 350,000) and Syria (some 300,000) by and large remain stateless, as
do many Palestinians living elsewhere in the Arab world. 
Palestinians and Israelis have differed sharply in their account of the
Palestinian exodus in 1948. According to the Palestinian narrative, “the
Arab refugees were forcibly expelled by Jewish forces or left in a panic flight
to escape massacre and . . . they were helped on their way by occasional
massacres, committed by Jewish forces to keep them running” (Alpher &
Shikaki, et al., 1999, p. 173). According to some Israeli accounts, the Arab
population was urged by its own leaders to leave their homes in anticipation
of an early return after the defeat of the Jewish forces by the Arab armies.
The more recent writings of Israeli “revisionist historians” (e.g., Morris,
1987) support the claim that there were indeed some systematic efforts by
Jewish forces to expel the Arab population at various times and places. No
doubt, a combination of circumstances created refugees in a war that was
fought in the midst of the country’s towns and villages.
How the refugee problem is ultimately resolved raises major practical
and symbolic issues for both sides that reach to the core of their respective
identities. By practical issues, I refer to those relating to the implementation
of any agreement on the refugees—that is, to the concrete steps that will be
taken to deal with the plight of the current refugee population. By symbolic
issues, I refer to those relating to the principle of the right of return of the
refugees to their original homes.
At the practical level, Palestinians are concerned with normalizing the
status of the refugees to the satisfaction of the individual Palestinians in-
volved and of the entire Palestinian community. Normalization of the status
of individual refugees could take any one or combination of the following
forms: return to Israel (perhaps their original homes, where that is still fea-
sible), return to the Palestinian state, financial compensation, resettlement
in a country outside of Palestine, integration with citizenship in the state in
which they are now living, or Palestinian citizenship and permanent status
in the state in which they are now living. The options offered to the
The Role of National Identity in Conflict Resolution
203


refugees must be such as to satisfy the entire Palestinian community that
the final peace agreement was not achieved at the expense of any part of
that community.
At the symbolic level, Palestinians are asking for an Israeli acknowledg-
ment of the right of return as a matter of principle, even if they are prepared
to compromise on the implementation of that right. From the Palestinian
perspective, Israeli acknowledgment of the right of return presupposes ad-
mission of direct moral responsibility for the plight of the refugees. Such an
acknowledgment and admission would confirm the Palestinian narrative
about the consequences for the Palestinian people of the Zionist enterprise,
of Jewish settlement in Palestine, and of the establishment of Israel. In 
confirming the Palestinian narrative, such an acknowledgment would also
recognize the Palestinian national rights that are anchored in this narrative.
There seem to be two interrelated, but separable, reasons for Palestinians’
emphasis on this symbolic issue. Psychologically, they need an Israeli ac-
knowledgment that Palestinians have a right to return to the homes from
which they were wrongfully displaced, in order to satisfy their sense of justice
and be able to let go of the conflict. Politically, they need this acknowledg-
ment to support the validity of their claims (such as claims for compensation)
in the bargaining that lies ahead.
For Israelis, too, major practical and symbolic issues are at stake in the
resolution of the refugee problem. At the practical level, Israelis are pri-
marily concerned with the impact of a return of Palestinians in large num-
bers to Israel proper (i.e., to Israel within its pre-1967 borders). A large-
scale return of refugees (beyond a symbolic number of up to 100,000)
would be disruptive to the society because it would raise many divisive
questions about the ownership of homes, villages, and other properties that
have changed hands or been destroyed after the 1948 events. T
o what pre-
cise destination would refugees return whose former houses are now in-
habited by Jewish families or whose homes or villages no longer exist? Fur-
thermore, and most centrally, Israelis are concerned about the impact of a
large-scale return of Palestinians to Israel on the demographic balance and
the Jewish character of the state. In short, they see a large-scale return of
refugees to Israel proper as creating an existential threat to Israel by under-
mining the Jewish majority and the viability and stability of the state. There
is also concern among Israelis about a mass return of refugees to the future
Palestinian state because of the impact on the stability and economy of that
state, but this concern is shared by the Palestinians themselves, who are at
least as mindful as Israelis about the absorptive capacity of the Palestinian
state.
At the symbolic level, Israelis have serious concerns about acknowledg-
ing the principle of a Palestinian right of return, even if Palestinians agree
that this right will not be implemented through a large-scale return of
refugees to Israel proper. Acknowledgment of that principle is particularly
troubling, from the Israeli perspective, if it entails admission of moral re-
204
The Role of Social Identity in Reducing Intergroup Conflict


sponsibility for the plight of the refugees. Such an acknowledgment and ad-
mission would undermine the Israeli narrative about the rightful return of
the Jewish people to its ancestral homeland, the establishment of the State
of Israel, the Arab rejection of partition and attack on the new state, and the
resulting refugee problem. Again, there seem to be two reasons for Israelis’
stance on this symbolic issue and, in particular, for their reluctance to ac-
cept direct moral responsibility for the plight of the refugees. Psychologi-
cally, they are not willing to accept a national self-image that negates the
righteousness of their cause and implies that Israel was “born in sin.” Po-
litically, they are concerned that accepting the principle of the right of 
return and moral responsibility for the refugee problem would keep the 
issue open, even after a peace agreement has been signed, and make Israel
vulnerable to future claims. 
Our Joint Working Group on Israeli-Palestinian Relations discussed, in
considerable detail, the divergent narratives of the two communities; the
fundamental needs, fears, and concerns of each community that would have
to be addressed if a solution of the problem of Palestinian refugees is to be
acceptable to each; and the general principles that must underlie such a so-
lution. The joint concept paper that finally emerged from the group’s dis-
cussion (Alpher & Shikaki, et al., 1999) starts with a review of the two nar-
ratives, of the basic needs of each side, and of the principles that must
govern a solution. It then proceeds to present four possible solutions to the
refugee/right of return issue. Solutions 1 and 2 reflect the traditional posi-
tions—conceivably, the opening negotiating positions—of the Palestinian
and Israeli sides, respectively. Solutions 3 and 4 are compromise solutions,
from a Palestinian and an Israeli standpoint, respectively. These two solu-
tions represent serious efforts, on the part of each side—based on intensive
discussions within the group—to come up with a compromise that would
accommodate the concerns of the other side without jeopardizing its own
vital interests. The paper discusses the advantages and disadvantages of
each solution, highlights the areas of agreement between solutions 3 and 4
as well as the remaining gaps between the two sides, and points to possible
ways in which these gaps might be bridged.
Solutions 3 and 4 represent a considerable narrowing of the gap between
the two sides when compared to the opening positions of solutions 1 and 2.
Still, there remains a significant gap between the two positions, particularly
with respect to the symbolic issues that divide the two sides. Solution 3 calls
for an Israeli acknowledgment of the moral right of Palestinians to return
to their homes based on Israel’s acceptance of responsibility for creating the
refugee problem (although it recognizes that the actual exercise of the right
of return would be limited). Solution 4, the Israeli compromise proposal,
acknowledges that Israel shares practical (but not moral) responsibility for
the plight of the refugees, and accepts the right of return to the Palestinian
state, but not to Israel proper (although it agrees to the return of a limited
number of refugees to Israel). 
The Role of National Identity in Conflict Resolution
205


The third party tried very hard to encourage the participants to bridge
the remaining gaps between solutions 3 and 4 and try to come up with a
joint solution 5—a compromise of the compromises, as it were. We did not
succeed in this effort, however. In part, further movement toward a joint
position on the symbolic issues was blocked by the great difficulty of the as-
signment. We were dealing here with issues that reached to the core of each
group’s identity and associated narrative. For Israelis to acknowledge any-
thing more than shared practical responsibility for the refugee problem and
for Palestinians to accept anything less than an Israeli acknowledgment of
moral responsibility would undermine their respective narratives, and most
members of the group were not prepared to go that far. Furthermore, at
least some of the participants may have preferred to leave the final bridging
of the gap between the two sides to the official negotiators. Since that final
bridging would inevitably entail large concessions on one or both sides,
they may have felt that it could best be done in the context of an overall
agreement, where concessions on the refugee issue could be made with ap-
propriate trade-offs on other issues.
Even though the Joint Working Group did not come up with a joint so-
lution to the refugee problem, the concept paper that it produced repre-
sents a significant contribution to the final negotiation of this issue. The
publication of the paper is in itself significant, in that it is the joint product
of a group of politically mainstream and influential Israelis and Palestinians,
who worked on it over the course of two years and agreed to make it public.
It is also significant that the Israeli and Palestinian authors of the paper ex-
plicitly premise their discussion on the assumption of a two-state solution,
advocating the establishment of a viable, independent Palestinian state. In
the paper, each group recognizes the other’s narrative, without abandoning
its own. The different options for resolving the refugee problem are evalu-
ated in terms of their success in addressing the fundamental needs and fears
of each side. What emerged from this effort is a considerable narrowing of
the gap between the two sides, even though the paper does not entirely
bridge that gap. Solutions 3 and 4 represent substantial movement from the
initial positions reflected in solutions 1 and 2—movement based on each
side’s sincere attempt to accommodate the basic concerns of the other.
With respect to the practical issues—that is, the actual implementation
of the right of return—the gap between the two sides was substantially nar-
rowed in the Working Group’s paper. “Solutions 3 and 4 concur that, at the
practical level, the core of the solution involves four components: a ‘return,’
however defined, of a limited number of refugees to Israel proper; return of
a larger number to the Palestinian state; permanent absorption in host
countries (notably Jordan); and compensation. Both sides agree that a suc-
cessfully negotiated solution should close the file on the refugee issue”
(Alpher & Shikaki, et al., 1999, p. 184). The two solutions disagree about
some important details. For example, the Israelis frame the limited return
of refugees to Israel proper as part of a program of family reunification,
206
The Role of Social Identity in Reducing Intergroup Conflict


whereas the Palestinians frame it as part of the exercise of the right of re-
turn. With regard to compensation, solution 3 calls for both collective and
individual compensation for the refugees’ losses, whereas solution 4 calls
for collective compensation only. Furthermore, solution 4 stipulates a simi-
lar mechanism of collective compensation for Jewish refugees from Arab
countries (albeit without an operational link to the compensation of Pales-
tinian refugees), whereas solution 3 eschews any such linkage. These differ-
ences clearly require further negotiation. But what is important is that the
two solutions agree on the crucial practical issues of implementing the right
of return: Palestinians concede that there will be no large-scale return of
refugees to Israel proper; and Israelis accept the refugees’ right of return 
to the Palestinian state, agreeing to leave it up to Palestinian authorities to
control the inflow of refugees in accord with the state’s absorptive capacity.
With respect to the symbolic issues—those relating to the principle 
of the right of return—the concept paper represents significant movement
toward a mutually acceptable formulation, although there is still a wide gap
between the two positions. Solution 4, the Israeli compromise statement,
acknowledges that Israel “shares practical (but not moral) responsibility, to-
gether with the other parties to the process that culminated in the 1948 war,
for the plight and suffering of the refugees, and that rectification of that
plight by all parties is a central goal of the Arab-Israeli peace process”
(Alpher & Shikaki, et al., 1999, p. 181). This statement comes closer than
past statements by mainstream Israelis to acknowledging the suffering of
the Palestinian refugees and accepting a share of the responsibility for it,
and thus beginning to address the Palestinians’ sense of justice. The move-
ment reflected in this statement was generated by an intensive process of
negotiating identity within the Working Group. Though it represents sig-
nificant movement, this acknowledgment is not sufficient to meet the Pales-
tinians’ call for full Israeli acknowledgment of the principle of the right of
return, based on acceptance of direct moral responsibility for the plight of
the refugees. Solutions 3 and 4, thus, still remain far apart at the symbolic
level. This remaining gap is understandable in that the two solutions,
though they represent serious efforts at compromise, have their point of de-
parture in two sharply clashing national narratives. The question is whether
it might be possible to bridge that remaining gap through a further process
of negotiating identity.
Closing the gap at the symbolic level is clearly difficult because of the
sharp clash in national narratives that gives rise to it. Neither party can be
expected to give up the core of its narrative in the course of the current ne-
gotiations; it is likely to take several generations before such changes can
occur. But it is possible, as part of the process of negotiation and reconcilia-
tion, for the parties to make changes in how they formulate their narratives,
where they place their emphasis, how they present the other within their
own narratives, and what they require by way of acknowledgment or valida-
tion of their narratives. Such changes at the edges of the narratives can po-
The Role of National Identity in Conflict Resolution
207


tentially be “negotiated,” in the sense of being explored and jointly formu-
lated by the parties in a spirit of mutual accommodation.
Thus, in the case of the right of return, we can assume that Palestinians
will not abandon their conviction that they have a moral right to return to
the homes from which they feel they were wrongfully displaced by the 
Israelis, and that Israelis will not admit guilt for the plight of the refugees.
The question is whether it is possible to jointly formulate an Israeli ac-
knowledgment that would be substantially responsive to what Palestinians
need to hear about the right of return without requiring either side to aban-
don these core elements of their narratives. In the terms of my summary,
toward the beginning of this section, of the symbolic issues that the refugee
problem raises for the two sides, the question would be: Can the parties 
formulate an Israeli acknowledgment that would provide significant
(though probably not total) satisfaction of Palestinians’ sense of justice and
validation of their claims for restitution, without requiring Israelis to negate
the righteousness of their cause and making them vulnerable to future
claims?
Let me offer three possible approaches to answering this question, which
can be seen as three steps in the negotiation of identity that might further
narrow the gap between the two sides manifested in solutions 3 and 4.
First, it may be possible to redefine the conception of responsibility in a way
that would not imply guilt and thus be acceptable to the Israeli public, and
yet be viewed by Palestinians as confirming their narrative, validating their
claims, and satisfying their sense of justice. One possibility here is for Israel
to accept a degree of causal responsibility without necessarily accepting
guilt. An Israeli acknowledgment might, for example, describe the refugee
problem as a tragic and unjust outcome of a historical process that arose
with the establishment of the State of Israel. It might go on to say that,
while there are disagreements about the precise historical events and the
distribution of blame, Israel acknowledges that it was an active participant
in this historical process and, indeed, a beneficiary of it, and thus bears part
of the responsibility for these events. The statement might conclude with
an offer of compensation as an expression of that responsibility. A statement
of this kind clearly implies causal responsibility without explicitly acknowl-
edging guilt. It thus has the potential, if properly fine tuned, of an acknowl-
edgment consistent with both the Palestinian and the Israeli narratives.
Second, it may be possible to find a mutually acceptable moral basis for the
right of return, which satisfies Palestinians’ need for an Israeli acknowledg-
ment of the right of return as a moral principle without linking that moral
right to the refugees’ original homes in Israel. One approach here would be
to anchor the moral basis for the right of return in return to Palestine: 
Israelis might acknowledge the Palestinians’ moral right to return to Pales-
tine, based on the proposition that both peoples have legitimate claims to
the land and that sharing the land (by establishing two states within it) 
is, therefore, a just solution to the conflict. Although the two sides differ in
208
The Role of Social Identity in Reducing Intergroup Conflict


their views of the justice and injustice of the events of 1948, the two-state
solution allows them to develop a shared view of justice for the future. 
Israeli acknowledgment of the refugees’ moral right to return to Palestine,
while not fully addressing Palestinians’ sense of justice, would do so to a sig-
nificant degree. It would also strengthen the validity of Palestinian claims
by confirming (in a context of reciprocity) Palestinian national identity and
rootedness in the land. For Israelis, anchoring the moral basis for the
refugees’ right of return in return to Palestine, rather than to their original
homes, would greatly reduce their concerns about future claims, even
though it may not entirely eliminate them.
Third, in conjunction with the two approaches just discussed and possi-
ble other approaches to narrowing the gap between the two sides at the
symbolic level, it may be possible for each side to find ways to affirm the
identity of the other. Such affirmations, perhaps in the form of acknowledg-
ments that the other has integral links to the land and national rights, could
compensate each side for the concessions they make to resolve the refugee
issue: the Israelis for the risks they take in accepting causal responsibility for
the plight of the refugees, and the Palestinians for accepting something less
than the right of the refugees to return to their original homes. Affirmation
of the other’s identity can take place only in the context of a process of ne-
gotiating identity, based on the principle of reciprocity, in which each side
learns to uncouple negation of the other from affirmation of the self. Each
can affirm the other’s identity to the extent that it feels assured that it is not
thereby jeopardizing its own identity and its own claims.
The three approaches to narrowing the remaining gap between the Is-
raeli and Palestinian positions in our Working Group’s paper on the prob-
lem of refugees all represent efforts to negotiate identity. As such, they
search for ways whereby each group can accommodate the narrative of the
other without undermining its own narrative. In such a process, each group
is encouraged to affirm or protect the identity of the other if, in turn, its
own identity is protected or affirmed. Each may be willing to grant some-
thing more to the other than it was originally prepared to offer, or receive
something less from the other than it originally demanded, in return for a
substantial concession by the other at the symbolic level.
Conclusion
In the Israeli-Palestinian conflict, as in other protracted ethnic conflicts, the
ever-present disputes over territory, resources, and political control are ex-
acerbated by perceived threats to national identity and national existence
that underlie the actions and reactions of the opposing communities.
Threats to identity and existence create obstacles to the settlement of con-
flicts, even when both parties have concluded that a compromise agreement
is in their best interest. Moreover, even when specific issues in conflict are
The Role of National Identity in Conflict Resolution
209


settled and political agreements signed—often with the mediation of pow-
erful third parties—these agreements may not lead to stable peace, fruitful
cooperation, and ultimate reconciliation between the two parties, unless
they have formed a new relationship based on mutual respect for their na-
tional identities.
The experience of my colleagues and myself in problem-solving work-
shops on the conflicts in Israel/Palestine, Cyprus, Sri Lanka, and Northern
Ireland, and experiences with similar efforts in other parts of the world,
support the central argument of this chapter about the role of national
identity in intercommunal conflict and conflict resolution. A central feature
of such deep-rooted, protracted conflicts is the struggle over national iden-
tity between the two parties, which leads them to perceive their conflict in
zero-sum terms, not only with respect to territory and resources, but also
with respect to national identity and national existence. Long-term resolu-
tion of such conflicts and reconciliation between the former enemies re-
quires changes in the groups’ national identities; in particular, they require
a redefinition of each group’s identity so that affirmation of its own identity
is no longer predicated on negation of the identity of the other. Such iden-
tity changes are possible, provided they leave the core of each group’s iden-
tity and national narrative—its sense of peoplehood, its attachment to the
land, its commitment to the national language, culture, and way of life—in-
tact. Furthermore, such changes must and can be “negotiated” between the
two groups. It is possible for the groups to accommodate their national
identities and the surrounding narratives to one another, as long as each 
respects and acknowledges the other’s peoplehood and thus reassures the
other that the core of its identity will not be jeopardized by changes in 
the periphery. Although some negotiation of identity takes place around the
negotiating table, it is primarily a process that must engage the entire body
politic on each side, at all levels and through different media. Problem-
solving workshops provide a specially constructed arena for engaging in this
process and jointly developing new insights and formulations that can be
injected into the public debate and can penetrate the political culture on
both sides.
A central lesson from our experience is that national identity, though
very much part of the problem in ethnic conflicts, can also become part of
the solution. The way we talk about our identity affects the way we think
about it and ultimately the way we act on it. In groups that are caught up in
protracted conflict, identity depends on the conflict and is shaped by the
conflict: Many elements of identity are constructed as vehicles for pursuing
the conflict. It should be possible, within limits, to reconstruct these ele-
ments as vehicles for peace and reconciliation. What is needed is an invest-
ment of identity in conflict resolution and in a new relationship with the
former enemy. Development of such a new, transcendent identity confronts
many obstacles. It cannot bypass the political process of negotiating a mu-
tually acceptable agreement, nor can it be allowed to threaten or undermine
210
The Role of Social Identity in Reducing Intergroup Conflict


the particularistic identity of each group. But within these constraints, the
potential for reconstructing the national identities of former enemies in the
service of peace and reconciliation exists and needs to be nurtured.
Notes
1. The work reported here is carried out under the auspices of the Program on
International Conflict Analysis and Resolution (PICAR), which I direct at the
Weatherhead Center for International Affairs, Harvard University. I am grateful to
the Center for providing a home for PICAR, to the William and Flora Hewlett
Foundation for providing financial support for PICAR over a number of years; and
to Donna Hicks, Nadim Rouhana, and my other PICAR colleagues.
2. I am grateful to the organizations that have provided financial support for the
Joint Working Group over the years: the Nathan Cummings Foundation, the
Carnegie Corporation, the Ford Foundation, the Charles R. Bronfman Foundation,
the U.S. Information Agency, and the William and Flora Hewlett Foundation, as
well as the Renner Institut in Vienna and the Weatherhead Center for International
Affairs.
References
Alpher, J., & Shikaki, K., with the participation of the additional members of the
Joint Working Group on Israeli-Palestinian Relations (1999). Concept paper:
The Palestinian refugee problem and the right of return. Middle East Policy, 6 (3),
167-189. (Originally published as Weatherhead Center for International Affairs
Working Paper No. 98-7. Cambridge, MA: Harvard University, 1998.)
Anderson, B. (1983). Imagined communities: Reflections on the origins and spread of na-
tionalism. London: Verso. 
Burton, J. W. (1969). Conflict and communication: The use of controlled communication in
international relations. London: Macmillan.
Burton, J. W. (1979). Deviance, terrorism and war: The process of solving unsolved social
and political problems. New York: St. Martin’s Press.
Burton, J. W. (1984). Global conflict: The domestic sources of international crisis.
Brighton, England: Wheatsheaf.
Gerner, D. J. (1991). One land, two peoples: The conflict over Palestine. Boulder, CO:
Westview Press.
Halpern, B. (1969). The idea of a Jewish state (2nd ed.). Cambridge, MA: Harvard
University Press.
Harkabi, Y. (1986). Israel’s fateful hour. New York: Harper & Row.
Hertzberg, A. (1973). The Zionist idea. New York: Atheneum.
Kelman, H. C. (1972). The problem-solving workshop in conflict resolution. In
R.L. Merritt (Ed.), Communication in international politics (pp. 168–204). Urbana:
University of Illinois Press.
Kelman, H. C. (1978). Israelis and Palestinians: Psychological prerequisites for mu-
tual acceptance. International Security, 3, 162–186. 
Kelman, H.C. (1979). An interactional approach to conflict resolution and its appli-
cation to Israeli-Palestinian relations. International Interactions, 6, 99–122.
The Role of National Identity in Conflict Resolution
211


Kelman, H. C. (1982). Creating the conditions for Israeli-Palestinian negotiations.
Journal of Conflict Resolution, 26, 39–75.
Kelman, H. C. (1987). The political psychology of the Israeli-Palestinian conflict:
How can we overcome the barriers to a negotiated solution? Political Psychology,
8, 347–363.
Kelman, H. C. (1988, Spring). The Palestinianization of the Arab-Israeli conflict.
The Jerusalem Quarterly, 46, 3–15.
Kelman, H.C. (1992). Informal mediation by the scholar/practitioner. In J. Bercov-
itch & J. Rubin (Eds.), Mediation in international relations: Multiple approaches to
conflict management (pp. 64–96). New York: St. Martin’s Press.
Kelman, H. C. (1995). Contributions of an unofficial conflict resolution effort to the
Israeli-Palestinian breakthrough. Negotiation Journal, 11, 19–27.
Kelman, H. C. (1997a). Nationalism, patriotism, and national identity: Social-
psychological dimensions. In D. Bar-T
al & E. Staub (Eds.), Patriotism in the lives
of individuals and nations (pp. 165–189). Chicago: Nelson-Hall. 
Kelman, H. C. (1997b). Negotiating national identity and self-determination in 
ethnic conflicts: The choice between pluralism and ethnic cleansing. Negotiation
Journal, 13, 327–340.
Kelman, H. C. (1997c). Some determinants of the Oslo breakthrough. International
Negotiation, 2, 183–194.
Kelman, H. C. (1997d). Group processes in the resolution of international conflicts:
Experiences from the Israeli-Palestinian case. American Psychologist, 52, 212–220.
Kelman, H. C. (1998a). Social-psychological contributions to peacemaking and
peacebuilding in the Middle East. Applied Psychology: An International Review,
47(1), 5–28.
Kelman, H. C. (1998b). The place of ethnic identity in the development of personal
identity: A challenge for the Jewish family. In P. Y. Medding (Ed.), Coping with
life and death: Jewish families in the twentieth century (Studies in Contemporary Jewry
XIV) (pp. 3–26). Oxford: Oxford University Press.
Kelman, H. C. (1999). The interdependence of Israeli and Palestinian identities:
The role of the other in existential conflicts. Journal of Social Issues, 55(3),
581–600.
Khalidi, R. (1997). Palestinian identity: The construction of a modern national conscious-
ness. New York: Columbia University Press.
Mendelsohn, E. (1989). A compassionate peace: A future for Palestine, Israel, and the
Middle East (rev. ed.). New York: Farrar, Straus & Giroux.
Morris, B. (1987). The birth of the Palestinian refugee problem, 1947-1949. Cambridge:
Cambridge University Press.
Muslih, M. (1988). The origins of Palestinian nationalism. New York: Columbia Uni-
versity Press.
Muslih, M. (1990). T
owards coexistence: An analysis of the resolutions of the Pales-
tine National Council. Journal of Palestine Studies, 19 (4), 3–29. 
Rouhana, N. N., & Kelman, H. C. (1994). Promoting joint thinking in international
conflicts: An Israeli-Palestinian continuing workshop. Journal of Social Issues, 50
(1), 157–178.
Segal, J. M. (1999). Defining Jerusalem. Middle East Insight, 14 (1), 27–28, 51–54.
T
essler, M. (1994). A history of the Israeli-Palestinian conflict. Bloomington and 
Indianapolis: Indiana University Press.
212
The Role of Social Identity in Reducing Intergroup Conflict


Richard D. Ashmore
9
Lee Jussim
David Wilder
Jessica Heppen
Conclusion
Toward a Social Identity Framework 
for Intergroup Conflict
We began this volume with two goals: (1) to review and organize social sci-
entific knowledge, across several disciplines, about social identity and inter-
group conflict; (2) to set the stage for future work by highlighting the theo-
ries, concepts, and methods with the most promise. These goals provide the
basic structure for this concluding chapter. First, we take a look back: we
summarize the chapters, distill the major points, and identify areas of agree-
ment and disagreement. Second, we take a look forward: again drawing
from the preceding chapters, we present a framework that highlights the
ideas that we feel should be the highest priority as we and other social and
behavioral scientists attempt to construct a more comprehensive and coher-
ent social identity perspective on intergroup conflict.
A Look Back 
As noted in the introductory chapter, the authors were given considerable
freedom in how they carried out their task, and this shows in the types of
contributions they have produced. Both chapters in Part I are conceptual
analyses; however, Brewer presents a general theory, whereas Eriksen uses
conflicts in three societies to illustrate his key points. In Part II, the chap-
ters by Citrin, Wong, and Duff and Sidanius and Petrocik report empirical
investigations, both based on survey data. In Part III, White presents a de-
tailed analysis of the group-based violence in Northern Ireland, while Staub
describes the place of self and identity within a general framework for un-
derstanding genocide. In the single chapter in Part IV, Kelman delineates
213


the ways in which national identity is involved in problem-solving work-
shops designed to help ameliorate the Israeli-Palestinian conflict. We turn
now to a more in depth review of these chapters.
Part I: The Contribution of Individuals’ Identities and the 
Collective Identities of Social Groups to Intergroup Conflict
Although both Brewer and Eriksen present conceptual analyses, there are
clear differences between these contributions. Brewer details a very tight
theory addressing a very specific question, When does ingroup love become
outgroup hate? This theory is stated in the most general and abstract form
and is similar to the grand hypothetico-deductive theories of learning in the
1940s (e.g., Hull, 1943) and to social psychological models of the 1950s
(e.g., Festinger’s theories of communication [1950], social comparison
[1954], and cognitive dissonance [1957]). And, as is true of this latter group
of conceptual frameworks, it is individualistic in orientation, focusing on
the thoughts and feelings of individuals, though admittedly in a complex 
social environment. Further, the model seeks support and testing primarily
in laboratory experiments. 
Brewer begins with Sumner’s notion of ethnocentrism: social groups ex-
ist in a state of war with one another, and, as a consequence, ingroup love
and outgroup hate are functionally interconnected sentiments necessary for
group survival. Brewer, however, suggests that Sumner’s theory can be de-
composed into four separate propositions: (1) social categorization—indi-
viduals and groups partition the social world into categories, and this cate-
gorization involves ingroup assimilation and outgroup contrast, which
make social categories more distinct in individual and collective representa-
tions than in reality; (2) ingroup positivity—given that self is a member of
some groups and that self is positively valued, there is a tendency to view
one’s group (the ingroup) as good; (3) intergroup comparison—individuals
compare their group to other groups, which can lead to judgments, not just
that we are good but also that we are better than other groups; (4) outgroup
hostility—the preceding steps, together with certain additional conditions, lead
people to hate an outgroup and actively work to harm it. Brewer suggests
that these four propositions are not highly correlated, as Sumner argued,
but instead constitute a progression of possible relations between ingroup
orientations and feelings about, and behavior toward, the outgroup.
Brewer further proposes that social categorization and ingroup positivity
are likely universal. She also draws an important implication from the latter
principle. Much discrimination is driven by ingroup positivity and is, thus,
favoritism for one’s own group, rather than discrimination against an out-
group. She presents laboratory experimental evidence in favor of this pro-
posal. She also notes that an important task for future research will be to
develop ways of disentangling pro-ingroup from anti-outgroup discrimina-
tion in real-world settings.
214
Conclusion


This raises the question that Brewer began with, When does ingroup
love (and “discrimination for us”) turn to outgroup hate (and “discrimina-
tion against them”)? Brewer argues that it is not “realistic, objective con-
flict” that provides the significant impetus. Rather, she suggests that the an-
swer lies within her optimal distinctiveness theory of social identity, which
she links to evolutionary theory. Her argument is that humans evolved, not
as isolated individuals or even in families, but instead as members of larger
social groups. T
o allow the individual to learn from and cooperate with oth-
ers, humans had to trust other people. However, to trust all others would be
dangerous. Thus, human beings developed the propensity to trust all others
within their group (“depersonalized trust”). When groups encounter one an-
other, each operating according to this same orientation gives rise to uni-
versal ingroup-outgroup stereotypes: “we” are close knit and cooperative;
“they” are clannish and treacherous. 
The final ingredient in the progression to outgroup hate is provided by
emotions. When members of different groups interact, they feel uncom-
fortable with one another and misattribute this negative emotional state to
“They are___________.” Under conditions of limited contact and little 
social mixing, the dominant emotions are likely to be contempt (“They are
inferior and repugnant”), and this should motivate distancing (Karen 
Horney’s “moving away from” [Ashmore, 1970]). However, under condi-
tions of close contact and possible integration, people will likely feel 
invaded, and this will give rise to anger and fear (“They are violent and dan-
gerous”). According to Brewer, it is these social conditions and associated
feelings and attributions that turn ingroup love into outgroup hate and 
produce group-based harm-doing (Horney’s “moving against”). 
In contrast to Brewer’s development of a general theory of group identi-
fication and intergroup conflict with most evidence provided by laboratory
research, Eriksen is concerned with identity politics, especially how social,
especially ethnic, identities are manipulated by political elites, and experi-
enced and enacted by everyday citizens, in real-world ethnic conflicts. Ac-
cordingly, his data are not from controlled experiments, but rather are case
studies of three actual “ethnic conflicts,” though it is clear from his analysis
that he would not accept this label. Eriksen builds his case, not by a step-by-
step elaboration of a small set of ordered propositions as Brewer does, but
instead by first describing the current position in cultural anthropology
concerning ethnicity and its implications for intergroup conflict. This posi-
tion essentially argues that ethnicity is not primordial (a social category de-
fined by shared biological descent or ancestry and thus a grouping that is
fundamental and enduring for both the group itself and individual group
members) and not centrally about what a group of people share. Instead,
ethnicity is best seen as a set of socially constructed group and interpersonal
performances that communicate to self and others the boundaries between
groups, which are not fixed entities but rather collectives in flux. T
o capture
this view, many anthropologists have turned from studying ethnicity per se
T
oward a Social Identity Framework for Intergroup Conflict
215


to seeking to understand social identities. One implication of this position is
that group identity is conceptualized as both conscious and strategic—indi-
viduals are aware of their multiple social identities and deploy them situa-
tionally in order to achieve personal and group goals. 
Although Eriksen generally accepts this basic position, he notes that it
leaves the person and his or her everyday experiences out of the ethnicity-
identity-intergroup conflict equation. He proposes that, in addition to con-
ceiving ethnicity as about social construction, boundaries, and strategic en-
actment, it is necessary to consider the personal identities of individuals and
how these can become fused with social identities. More specifically with
regard to intergroup conflict, he suggests that, when the lived experience of
individual group members is consonant with the key elements of a social
identity currently proffered by authorities, it sets the stage for destructive
intergroup conflict fought in the name of that group identity. Thus, for ex-
ample, although Bosnians and Serbs in Bosnia shared many cultural ele-
ments and in multiple ways were quite similar to one another, they had
lived separate lives in the private sphere, with little intermarriage and few
cross-group close friendships. This situation of separate private spheres al-
lowed group myths and pejorative outgroup beliefs to be part of the every-
day experience of members of each group. Thus, when leaders espoused the
separate and conflicting religiously based social identities of Bosnian and
Serb, the citizenry had personal selves and everyday experiences that made
these appeals resonate. 
Although they take two very different approaches and seem to offer in-
commensurate, if not incompatible, frameworks for understanding inter-
group conflict, the Part I contributions are actually much closer than might
appear at first glance. Brewer starts with the individual and does a nice job
of identifying and ordering the cognitive and affective factors that set the
stage for ingroup attraction and, in turn, outgroup hostility. She concludes
by describing factors that moderate the model she has constructed. These
moderators are all societal variables (e.g., social structural complexity), 
and it is precisely such factors that Eriksen elaborates in his contribution.
Eriksen draws the circle complete by his insistence that only when socially
propagated group identities fit with the lived experience of individuals and
their personal sense of self will the stage be set for ethnically based violence.
In complementary ways, then, both Brewer and Eriksen note the impor-
tance of individual and social factors in the ways in which social identity is
implicated in intergroup hostility.
The foregoing should not be construed as meaning that Brewer and
Eriksen are in total agreement about the personal and societal underpin-
nings of group-based harm doing. The clearest point of friction concerns
the role of objective conflict between groups as a precipitator of hostile re-
lations. Although he never uses the term, Eriksen at several points presents
ideas consistent with the realistic group conflict (RGC) view of intergroup
conflict (Ashmore & Del Boca, 1976; Boulding, 1962). According to this
216
Conclusion


perspective, groups come into conflict over limited material resources—
land, energy sources such as oil, and the like. Whereas Eriksen identifies
objective competition as crucial to conflict development and proliferation,
Brewer sees such “realistic” competition as much less important. In fact, she
might be read as suggesting that tangible conflicts come only after the stage
is set by the elements offered in her model.
The second major disagreement between Brewer and Eriksen is equally
instructive. Brewer’s presentation is all about the here and now, about indi-
viduals’ thoughts and feelings about me, us, them, in the present moment.
Eriksen, on the other hand, grounds his analysis in history. For each of his
three examples, he describes what happened in the past, not just as stage
setting but also as crucial to understanding why the groups are fighting
now. This difference reflects, in part, two contrasting disciplinary para-
digms. Brewer is a psychological social psychologist, and present-day prac-
titioners of this social science continue to be much influenced by Kurt
Lewin’s proscription against “historical” or “genetic” accounts: do not ex-
plain a person’s current behavior by recourse to what happened to that per-
son in her childhood; instead account for action on the basis of the contem-
poraneous field of forces acting on the person. Although Lewin’s ahistorical
approach referred to the life history of the individual, many of his students
who, in turn, shaped today’s social psychologists, took Lewin’s approach
more generally to mean that the important determinants of action are cur-
rent social factors. As a consequence, not only has the life history of the in-
dividual been left to developmental psychologists, but the past of social
groups has been left for historians. In contrast, social and cultural anthro-
pologists, viewing culture as, in part, something that one generation passes
to the next, have been much more inclined to see a group’s history as im-
portant to understanding its present organization and functioning—Thus,
Eriksen’s descriptions of current conflicts feature historical information. 
One important next step will be to bring these perspectives together.
One way, which will be elaborated below, is to extend the notion of life
story and personal narrative into the analysis of social identity and group-
based conflict. The basic notion is that, as individuals and groups, we make
sense of who we are now, in part, by looking back and creating stories—
about self, about our ethnic group, about our nation. And it is these stories,
as part of personal and group identities, that may be crucial in setting the
stage for violence and oppression done in the name of us vs. them.1
Part II: The Contribution of Ethnic and National Identities to 
Political Conflict in the United States
The two chapters in the second part follow very well from Eriksen’s contri-
bution. Methodologically, he suggests that it is useful to look closely at spe-
cific intergroup conflicts; substantively, ethnic conflict within nation-states
is a major issue around the globe and is not limited to Third World or de-
T
oward a Social Identity Framework for Intergroup Conflict
217


veloping countries. In Part II, the focus is on patterns of ethnic divergence
and convergence in just one nation, the United States, arguably the most
economically advanced country in the world. Further, only one type of con-
flict is considered: political contentions among ethnic groups.
Citrin, Wong, and Duff’s central concern is American national identity:
What is it? Are there ethnic group differences in beliefs about American
identity? How do individuals prioritize ethnic and national identity? How
do ethnic and national identity relate to policy preferences? Citrin and his
co-workers set the stage for their empirical research by carefully laying out
their notion of social identity, in this case national identity. Following T
ajfel
(1978) they conceive of national identity as involving both “identification
as,” or self-categorization (e.g., “I am an American”) and “identification
with,” or affective attachment to the nation. Regarding the latter, they fur-
ther distinguish patriotism, or love for one’s country (“Being American is
very important to me”) from chauvinism, or feelings of national superiority
and degrading of other countries (“If only other countries were like
America the world would be a much better place”). 
Citrin et al. suggest that they go beyond T
ajfel (a point we return to be-
low in considering the multiple dimensions of social identity) by adding a
third component to national identity—normative content. The use of the
word “normative” is deliberate and does not mean average. Instead, norma-
tive conceptions are assertions about how society should be organized, how
people should behave to make society work well, and how society should
treat groups and individuals in the nation. As political scientists, Citrin et al.
bring an important theme to the volume—normative as pre- and pro-scrip-
tive, not just descriptive. This reflects the subfield of political theory and
the issue of how political systems should be constructed. Given the empha-
sis on ethnicity and national identity, Citrin and his colleagues stress three
different normative views of how ethnic and national commitments should
be related in multiethnic states: liberal, in which national membership de-
pends on sharing a set of ideals (e.g., respect for the individual, freedom)
and not shared ethnicity (hence, it is often termed civic, not ethnic, nation-
alism); nativism, in which American identity is equated with one ethnic
group, white Anglo Saxons, and thus, to be an American means accepting
the values and other characteristics of this dominant ethnic group; multicul-
turalism, in which ethnicity is prized above national identity.
Citrin et al. also preface their data reporting by introducing symbolic
politics theory (Sears, 1993), which stresses that political orientations are
socialized in childhood and maintained by adult engagements (e.g., work).
According to the authors, national identity is widely socialized in the
United States, and, thus, there should be more ethnic consensus than con-
flict concerning national identity. 
Citrin et al. address their questions about national identity and predic-
tions from symbolic politics theory using survey data collected from both a
national sample and Los Angeles County respondents. Concerning their
218
Conclusion


first question, What is American identity?, they found considerable agree-
ment that, in regard to political issues, most Americans define self as “just
an American” (self-categorization) and feel considerable pride in and patri-
otism for America, but not high levels of chauvinism. Also, there was con-
siderable support for nativistic beliefs (though the pattern was uneven) and
even more support for assimilation notions. 
In answer to their second question, Are there ethnic group differences in
views on American identity?, Citrin et al. concluded that there are few dif-
ferences in defining American identity except on issues pertaining to ethnic
groups per se (e.g., how the nation treats minority groups). A most intrigu-
ing finding was that, on average, African Americans were more nativist than
whites. This was especially true for endorsing being a Christian as a crite-
rion for being “a true American.” As the authors note, this suggests yet an-
other important social identity, religion. 
Turning to the third question, How do individuals prioritize ethnic and
national identity?, Citrin and colleagues found that, when given the option,
about 50 percent of African Americans chose “both ______ [my ethnicity]
and American,” about one-third chose ethnicity as their primary identity,
and about one-sixth chose “just American.” Hispanics and Asians responded
similarly, but were somewhat less likely to select “just American” and some-
what more likely to choose “both _______ [my ethnicity] and American”
and ethnic identity. For whites, on the other hand, “just American” was far
and away the predominate response. 
The final question was, How do ethnic and national identity relate to
policy preferences? As predicted by symbolic politics theory, race (given
small sample sizes, only whites and blacks were compared) did not predict
views on multicultural policies or international affairs (military spending or
protectionism), whereas national identity did. On race-related policies, race
did predict policy preferences. However, when ethnic identification was
added to the predictive equation, felt closeness to one’s ethnic group did not
enhance support for race-based policies for blacks, but for whites felt close-
ness increased disapproval of programs that benefited minority groups.
That it was among whites that identification with one’s group related to po-
litical issues regarding race may seem paradoxical to some given that the
practice of identity politics is most often associated with minority groups.
The Citrin et al. contribution highlights three points. First, social identity
is multidimensional. Their chapter considers self-categorization, affective
attachment, and normative content. Below we identify additional possible
facets of social identification. Second, a particularly important aspect of ex-
periencing ethnicity, though often neglected, is the “shoulds” attached to
one’s group. These are related to the notion of normative content, but they
focus, not on broad issues of what it means, for example, to be a true Ameri-
can, but on the very specific group norms for how “one of us” should act—
the food we should eat, the way we should talk, and the like—and the pres-
sures an individual feels from members of her or his own group to abide by
T
oward a Social Identity Framework for Intergroup Conflict
219


these “rules.” T
o address this issue, Contrada et al. (in press) have recently
developed a self-report instrument to assess individual differences in own-
group pressure to behave in specific ways. Third, as Citrin and his colleagues
note, social identities are not necessarily political identities. This raises the
crucial question, How do people get from social definition to action? 
In the second chapter of Part II, Sidanius and Petrocik take up much the
same general topic as Citrin and his collaborators; the relationships be-
tween ethnicity and national attachment in the United States. And like 
Citrin et al., Sidanius and Petrocik identify three perspectives on this issue:
melting pot (each ethnic group contributes something to an emergent 
national American identity, and, in the process, members of these ethnic
groups lose their own sense of ethnicity and become identified with, and 
attached to, the nation); pluralism (individuals and groups retain their
uniqueness, but all agree to be tolerant and to also identify with the nation);
group dominance model (in all societies, groups are arranged in a hierarchy
of prestige and power, and this means that groups not only maintain their
distinctive identity, but also that, for those at the top, ethnocentrism and
patriotism are positively linked, but for those at the bottom this correlation
is negative). Again in parallel with the preceding chapter, Sidanius and
Petrocik assess with survey data the predictions they derive from each per-
spective. These data come from five different samples, from university stu-
dents, to Los Angeles residents, to a national sample. 
Sidanius and Petrocik address three specific issues: the salience of ethnic
identity; the relation between ethnicity and patriotism; the relationship be-
tween patriotism (attachment to the nation) and ethnocentrism (by which
they mean a negative attitude toward American ethnic groups relative to
one’s own group). In regard to the first issue, Sidanius and Petrocik con-
clude that ethnic identity is salient for Americans and that this is especially
true of minority-group Americans. With respect to the second issue, Sida-
nius and Petrocik find that white Americans are highest in patriotism and
African Americans lowest. Concerning the third issue, Sidanius and Petro-
cik report that, for white Americans, patriotism and ethnocentrism are of-
ten positively correlated, but the correlation for minority Americans is gen-
erally negative. 
Of the many excellent points in the Sidanius and Petrocik contribution,
we underline just one, their group dominance model. Individuals develop
their intergroup attitudes and ethnic groups come into conflict within soci-
ety, and all known societies contain group dominance structures: Some so-
cial categories are privileged and others are disadvantaged. Thus, the soci-
etal context for intergroup relations is never neutral or even-handed, and
this is true even in economically advanced nations with strong democratic
traditions. This is a particularly important reminder at this time in the
United States, where we surmise that a large portion of the dominant white
populace believes that the civil rights and women’s movements have some-
how leveled the playing field for all Americans and now it is time to treat
220
Conclusion


everyone as individuals and not as group members. Without denying the
social changes that have occurred over the past four decades, it is simply not
true that American society and culture treat all people and groups as equal. 
The chapters by Citrin et al. and Sidanius and Petrocik present a para-
dox: both address the same basic issue (the interplay between ethnic and na-
tional identity in the United States), they deal with two similar folk models
(though they use somewhat different labels—liberal [Citrin et al.] and plu-
ralism [Sidanius & Petrocik]; nativist [Citrin et al.] and Americanization
and melting pot [Sidanius & Petrocik]; see also Fredrickson [1999] for a
historical review of models of American ethnic group relations), they both
draw on similar previous studies (especially that by de la Garza, Falcon, &
Garcia, 1996), and they take the same basic methodological approach (sur-
vey data; drawing in one case from the same survey). But they reach quite
different conclusions: Citrin and co-workers see a small ethnic divide in the
United States and little tension between ethnic and national identity,
whereas Sidanius and Petrocik perceive a large ethnic divide and consider-
able conflict between ethnic and national attachment. 
There are three factors that make this paradox more apparent than real.
First, they use different items to assess ethnic/national self-categorization.
Second, the two chapters are not really directed at precisely the same set of
phenomena. Citrin et al. predict policy preferences from ethnicity and eth-
nic identity, whereas the primary outcome variable of interest to Sidanius
and Petrocik is the correlation between national attraction (especially pa-
triotism) and prejudice against U.S. ethnic groups other than one’s own
group. Third, and we believe most important, in reporting results, Citrin
and colleagues concentrate on absolute levels of attachment to the United
States (which are generally high) and play down group differences (which,
for some measures, are substantial), whereas Sidanius and Petrocik stress
group differences and deemphasize absolute levels of national attachment
by ethnic group members. 
The Citrin et al. and Sidanius and Petrocik chapters underline four im-
portant points. First, how one asks people to report their group self-catego-
rization is crucial. Some items implicitly pull for “just American,” whereas
others pull for ethnicity and nationality self-codings either instead of, or in
addition to, labeling self as “American.” Our recommendation is that sur-
veys concerned with ethnicity and national identity should use the 1995
LACSS item (described by Citrin et al.) that offers the respondent the fol-
lowing choices: “just an American,” “_______ [whichever national or ethnic
group the person sees self as a member of],” “both American and ______
[self-categorized ethnic group].” This seems to us to provide the most neu-
tral way of getting at how respondents define and categorize self in terms of
a superordinate national identity (“American”) and more particular ethnic
self-codings. Although the results reported by Citrin et al. are from a small
and local sample, they suggest that the modal white American sees self as
“just American,” while minority Americans define self in three very distinc-
T
oward a Social Identity Framework for Intergroup Conflict
221


tive ways: about half select a dual identity, about one in three categorizes
self in terms of ethnicity only, and about one in six regards self as just
American.
Second, the Citrin et al. and Sidanius and Petrocik chapters remind us
that it is easy to slip from ethnic identity, or identification, to ethnicity, or
simple ethnic group differences. Although both chapters report analyses
dealing with individual differences in group identification (especially self-
labeling), they also report analyses in terms of ethnic group differences.
However, simple ethnic group differences do not necessarily mean differ-
ences as a function of ethnic or national identity. In addition, such an ap-
proach obscures the wide variation within any social category in the content
and degree of identification with the group. For example, what other social
identity, intergroup attitude, and public policy preference differences are
there among minority group Americans who self-label as “just American,”
“American and ____,” and minority group label only?
Third, ethnic groups are not monolithic. Obviously there is variation by
age, socioeconomic status, and a variety of other variables that Citrin et al.
and Sidanius and Petrocik partial out in some analyses. At the same time,
some of these partialed-out variables constitute important social identifica-
tions. For example, there is considerable variation in the African-American
(and white) community in terms of religious commitment, especially defin-
ing self as a “fundamentalist Christian.” 
Fourth, just as ethnic and national identity are multifaceted, so too is 
national attachment. Both chapters note that there is a positive form of na-
tional attachment (patriotism) and a more virulent brand (chauvinism/
nationalism—basically “We are the best in the world . . . and others
should emulate us”), as well as issues of willingness to leave the country and
beliefs about what makes someone a true American. 
The foregoing summary suggests that the Citrin et al. and Sidanius and
Petrocik chapters have not resolved the question of how ethnicity and eth-
nic identity relate to national identity and attachment. They have, however,
provided some important clues. In general, there is fairly high attachment
to the United States in all groups, but there are also substantial differences
between groups, especially on issues pertinent to ethnicity. More impor-
tant, the two chapters begin to point us toward the necessary next genera-
tion of research. This will need to use multiple measures of ethnic or na-
tional self-categorization, especially items that do not tilt toward or away
from ethnic (or national) identification, and multiple measures of national
attachment. Thus, future work will be directed at two important questions:
How do the various facets of ethnic and national identity (self-categori-
zation, felt closeness, and the like) relate to the various components of 
national identity? How, in turn, do these sometimes competing and some-
times complementary forms of self-identification relate to variables perti-
nent to intergroup conflict (e.g., attitudes toward other groups, political
policies and programs)? 
222
Conclusion


Part III: The Contribution of Social Identity to Violent Intergroup Conflict
White continues the focus of Part II to look intensively at one conflict, but
the site of his analysis is Northern Ireland, not the United States, and the
type of conflict is deadly violence, not political strife. White begins his
chapter in a very dramatic fashion by presenting a table indicating who has
killed whom in the conflict. This table not only helps him make some very
important points about the conflict but also reminds the reader that the ul-
timate outcome of identity-based disputes is that people may kill one an-
other. Of the many excellent points made by White, we highlight three: (1)
the past importantly shapes the contemporary conflict; (2) the violence in
Northern Ireland is three- not two-sided; (3) identity is involved in the con-
flict in multiple ways. 
First, the current fighting is rooted in a long history, and that past not
only shapes the present conflict but also sets parameters for how the hos-
tility might be reduced. White describes how the violence in Northern Ire-
land began with the Norman/English invasion of Ireland over 700 years
ago, how religion was introduced into the conflict in the 16th century with
the Reformation, how a Protestant ascendancy was created during the 17th
century, how Northern Ireland was created in the early part of this century,
and how the current protracted violent confrontation began in the 1960s. 
This history is important to the current conflict in three major ways.
1. It underlines the importance of losing for group definition. On multi-
ple occasions over seven centuries, the Irish have lost battles and wars to the
English. These repeated defeats have been incorporated into the group nar-
rative of the Republicans of Northern Ireland. This parallels the history
and national narrative of Serbs as described by Eriksen (this volume; see
also Staub this volume). This 700-year record also suggests why the Irish
Republican Army may be so reluctant to “decommission,” which, as this
chapter is being written (spring 2000), may derail the recently agreed to
peace accord. The Irish have lost many times to the English; they can not
accept another “defeat,” which is how many perceive decommissioning,
even if the current agreement is widely supported by Protestants and
Catholics.
2. The Protestants in Ireland were imported by the British and took
over land and status from defeated Irish. They have always been in a mi-
nority in Ireland as a whole (though a majority in Northern Ireland), and in
White’s words, have a “siege mentality.” It is likely that being a numerical
minority with political and economic majority status (as whites in Southern
Africa and the Tutsi in Rwanda until recently) also has important implica-
tions for social identity. Groups that have superior power but are numeri-
cally in the minority (power majority, numerical minority) likely stereotype
the other group (power minority, numerical majority) as, at the same time,
both immature and childlike (to justify their inferior position) and poten-
tially violent in a primitive way (to make sense of past uprisings and fears of
T
oward a Social Identity Framework for Intergroup Conflict
223


what might happen if “they” won). This is much the same set of images that
characterized U.S. whites’ views of blacks during the slavery period (Boskin,
1970). In complementary fashion, the self-image of power majority yet nu-
merical minority groups features traits such as grown up and mature (re-
sponsible, productive, and the like) and civilized.
3. As White notes, a long history of fighting an external oppressor also
sets the stage for role identities in this struggle, a point to which we return
below. Here we simply note that often roles are hard to give up, thereby
making intergroup conflict difficult to reduce.
Second, what looks to an outsider as a simple two-sided religious con-
flict—Northern Ireland Catholics versus Protestants—is, upon more care-
ful analysis, really a multisided political conflict with a religious gloss.
White shows that there are three key contestants—Protestant paramili-
taries (who are just a portion of the Protestant community); Irish Republi-
cans (most IRA members are Catholic, but not all Catholics are affiliated
with the IRA, and some IRA members are Protestant); and the state, espe-
cially British authorities. At the end of his analysis, White notes that there is
actually a fourth important party, the people and government of the Repub-
lic of Ireland, and, with the current efforts at peace, the role of that party
will likely be increasing. The United States has also been involved in the re-
cent efforts to resolve the conflict, and thus there is a fifth “side.” The cru-
cial point for the present volume is that apparently “us against them” con-
flicts are seldom just two-sided; and the various parties are likely to have
different “identity shares” in the conflict, which leads to the next major
point we wish to underscore about White’s chapter.
Third, identity is implicated in multiple and diverse ways in the conflict.
T
o uncover how identity enters into the political violence, White focuses on
accounts given by the various participants—quite simply how do various
participants describe the conflict and their role in it? White begins his
analysis by using Thoits and Virshup’s (1997) distinction of “we” from “me”
social identities. The former are collective social identifications (member of
some social category [e.g., nation, ethnic group]), whereas the latter are role
identifications (occupant of a particular role [e.g., parental, occupational]).
White argues that for Republicans their involvement reflects both a collec-
tive social identity (“we”) and role identity (“me”). Importantly, the collec-
tive self-definition is not me as Irish Catholic versus Irish Protestants, but
me as Irish Nationalist versus the British. The role identity is a set of enact-
ments that the person has learned from either growing up in a Republican
home or as part of joining a Republican organization. White illustrates a
Republican role identity by describing the hunger strikes of the early 1980s.
He notes that hunger strikes have a long history in Republican efforts to get
the British out of Ireland, and this means not only that these are part of the
“Republican prisoner” role but also that there are specific men who have
gone on hunger strikes (and often died) and who now serve as role models
(see also Aretxaga, 1993). 
224
Conclusion


White’s suggestion that role identities can be important parts of involve-
ment in intergroup conflict is a significant addition to role identity models
(McCall & Simmons, 1966; Stryker,1987). Just as the normal everyday ac-
tivities of society require roles, so, too, do protracted conflicts. What might
these conflict roles be? Obviously, there are the roles of active fighter (sol-
dier) and prisoner. But in addition, “armies” require that other functions be
fulfilled. One important function is intelligence, or the collecting of infor-
mation to assist in waging the fight. White describes a British soldier who
slept outside a bakery; it is possible that the owner of the bakery could have
provided intelligence reports to either the Republican or Protestant para-
militaries and, if so, would have taken on a conflict-related role identity. 
In addition, when an army is an insurgent one, other roles are necessary,
among them, fund-raiser, politician, and civilian support. In one of the Re-
publican narratives reported by White, the respondent describes his grand-
mother, of whom he said, “She never wavered.” She may also have supplied
a safe house to paramilitaries and fulfilled additional functions in support of
others with active fighting roles, and, in so doing, her role of “supporter”
would have been active and not passive. 
According to White, the Protestant Loyalists and British actors in
Northern Ireland construe their involvement solely in “we” social identity
terms. For the Protestant paramilitaries, the conflict is “us against them,”
where “We are British Protestants” and “They are Irish Catholics.” For the
English, White argues that the conflict is “them versus them,” a religious
conflict among the Irish. 
White acknowledges possible role identities for Protestants and British,
but we believe that he underplays them. Although in the earliest years of the
current “troubles,” the Loyalists may have been relatively unorganized with
little sense of history, certainly as the conflict evolved it is probable that
Protestant paramilitary groups became relatively well organized and devel-
oped systems of roles. Role identities are even more probable for the British
or English first-line participants in Northern Ireland, the military person-
nel. It is likely that many English soldiers construed their involvement there
as “me as a soldier doing a soldier’s job,” and not as “we British.”
White’s primary data are verbal accounts, and gathering such narratives
in the midst of an armed conflict is very difficult. This difficulty raises some
questions about White’s conclusions. As he notes, the verbalizations he re-
ports are not factual histories or descriptions, but instead they are stories
that a person has told to someone else. As with any story, the content of the
account depends, in part, on who is talking to whom for what purpose. This
raises the most general question: Are the narratives given by the three
groups of protagonists (Republicans, Protestant paramilitaries, English sol-
diers and elites) comparable? This can be partitioned into two more specific
queries: Who gave the accounts? T
o whom were the accounts given and
how did the respondent think the accounts would be used? Regarding the
first or “who?” question, White acknowledges that the three groups were
T
oward a Social Identity Framework for Intergroup Conflict
225


not equated: many of the Republicans were Sinn Fein members; Protestant
paramilitary accounts were from former paramilitaries; the English voices
were from soldiers at the start of the current conflict (late 1960s) and politi-
cal leaders. Regarding the “to whom and how used?” question, the Republi-
cans were all interviewed by White himself, and it is likely that they were
assured confidentiality and told that the conversations were for research
purposes. The Protestant paramilitary stories and statements by English ac-
tors were taken from secondary sources, most from memoirs and remem-
brances or from official statements. The question of how the respondents
thought the accounts would be used is less certain, but at least some had
clear public relations and political implications. As a consequence, these
statements may have been tilted toward a “we” social identity perspective
and away from explaining the individual’s roles in the conflict. 
While White presents a detailed analysis of just one violent conflict,
Staub describes a general framework for understanding genocide that is
based on, and applied to, several different murderous group conflicts. Basi-
cally, this model indicates that mass killing and genocide are the end result
of a set of personal and social processes that are set in motion by one of two
general types of instigators: difficult life circumstances (e.g., hyperinflation)
and intergroup conflict. These instigators threaten basic needs, including
those related to identity. This sets the stage for a spiral toward large-scale
systematic violence as follows: the psychological elevation of the ingroup
(and generally a derogation of the outgroup), which facilitates initial harm-
doing; this, in turn, produces altered societal institutions and individual
identities. These lead to the final step in the model, mass killing and geno-
cide. T
o this causal sequence, Staub adds two moderators. The first, culture,
involves a variety of societal-level factors (e.g., history of devaluation of 
an outgroup); the second, bystanders, indicates individuals both within 
and outside the group who are not (at least initially) involved in the harm-
doing.
Among the many insights provided by Staub’s model, we highlight five.
First, both group and individual identity play important roles in intergroup
violence. Regarding the former, Staub suggests that ethnic groups and na-
tions vary in their “group self-concept,” and this shared notion of “us” can
lead some groups toward mass killing and genocide while directing others
away from such harm-doing. In particular, Staub highlights as a cultural
factor predisposing toward systematic intergroup violence, “unhealed vic-
timization,” the belief that “we” have been hurt coupled with no opportu-
nity to personally and socially come to terms with this wound. 
Second, and closely related to the first point, a personal and societal
sense of vulnerability (“They have killed our people before, and they could
attack us at any time”) can not only predispose a group toward violence but
also make preemptive violence more likely. As Eriksen noted regarding
Serbs in the former Yugoslavia and White with regard to the Irish Republi-
cans, these groups had histories of losing, which justified, in the partici-
226
Conclusion


pants’ minds, acts of “getting them before they get us.” This type of mind-
set helps sustain these conflicts 
Third, regarding the through-time set of factors leading to intergroup
violence, most important for the present volume is that these involve
changes in individual and group identities. Those in a society who engage
in initial violence are more receptive to self-definitions that make their
harm-doing not only acceptable but, in some cases, also morally right; those
who at the outset are bystanders can come to see killing or support for
killing as part of how “I ought to behave” and alter their self-conception 
accordingly. 
Fourth, in terms of culture moderators, it is possible to identify societal-
level identity types that are more likely to participate in mass killing, and
these are not neatly divided along Eastern versus Western modal self-
concept configurations. In fact, as Staub notes, both the autonomous indi-
vidual self more common in Western industrialized societies and the more
social and connected self-configuration that is more common in Eastern
and less industrialized nations have their own vulnerabilities to genocide.
The challenge for the next generation of researchers is to assess these types
and test Staub’s suggestions. (See Norenzayan et al. [1999] and Lyengar et
al. [1999] for additional ideas about different types of cultures and the im-
plications for intergroup strife.) 
Fifth, third parties play significant roles both in creating conditions ripe
for large scale group violence and as a major resource in limiting or stop-
ping genocide. Staub describes how Belgians built on existing Hutu-Tutsi
dominance systems and added the notion of “blood” to help make the soci-
ety more manageable for them. Years later this made the violence more
likely. 
Part IV: The Role of Social Identity in Reducing Intergroup Conflict
Staub’s emphasis on the importance of bystanders, especially external by-
standers, as crucial to preventing the escalation to mass killing and genocide
leads very nicely into the final part, which focuses on reducing conflict. The
chapter in this section, by Kelman, describes problem-solving workshops as
a conflict-reduction method that features third parties. Kelman begins his
chapter by describing the history of the Israeli-Palestinian conflict. Just as
Eriksen and White suggest that the past is important to understanding why
and how intergroup harm-doing develops, Kelman proposes that efforts to
reduce conflicts must take into consideration the historical foundations of
current strife. Next, Kelman lays out his notion of national identity. He ar-
gues that, “Although national identity is carried by the individual members
of a national group and can thus be studied as a property of individuals, the
present chapter refers to it as a collective phenomenon—as a property of
the group.” In the simplest terms, national identity is the shared and con-
sensual definition of us as a group and nation; this includes our claims to
T
oward a Social Identity Framework for Intergroup Conflict
227


land and continuity over time, our beliefs about who we are, and what
makes us distinctive. Kelman suggests that national identity is accompanied
by a “national narrative—an account of the group’s origins, its history, and
the relationship to the land.” Thus, for Kelman, social identity as a nation
involves both a self-definition (national identity) and group story (national
narrative). 
In seeking to understand how national identity might underlie conflict,
Kelman argues that in protracted intergroup fights, especially in which
both sides have claim to the same land, the groups’ national identities can
become exclusive (e.g., “This land belongs to us and can belong to no one
else”) and monolithic (all the multiple dimensions of nationhood [religion,
ethnicity, history, and the like] become seen as highly intercorrelated). Both
of these properties make conflict resolution more difficult, identity exclu-
sivity by making sharing impossible by definition, and identity unidimen-
sionality by making us all the same and reducing any possibility that in
some ways they are like us. One final aspect of protracted conflicts such 
as that between the Israelis and Palestinians is that the process of fighting
can become part of the national identities—“We are the people who fight
them.” This echoes Staub’s observation that engaging in intergroup vio-
lence changes individuals. Kelman proposes that intergroup conflict can
change whole nations and become part of the shared collective self-defini-
tion and the consensual national narrative.
Kelman then describes how identity change, though difficult, is still pos-
sible. Here he makes two very significant points. First, although people and
nations see their identities as fixed, they are, in fact, not. Identities, like all
belief systems, are not immutable essences, but rather are socially con-
structed and, as such, are modifiable. Second, unlike land and other physical
objects, identities are not zero-sum; the definition of “us” need not include
the defeat or devaluation of “them”.
How might these possible changes in national identity be achieved? Kel-
man describes the problem-solving workshops that he and his colleagues
have been conducting for many years in the Middle East. These feature
third parties who serve both to moderate the discussions and to provide an
unbiased resource for both sides. Another key aspect of the workshops is
that the participants are influential people in the Israeli and Palestinian
communities (e.g., business leaders) but not formal political leaders. Thus,
they have “clout” but do not need to represent a political position. 
The workshops attempt to find ways to bridge the gaps between the two
sides. Kelman highlights how this involves altering national identities. For
example, workshop participants have considered what aspects of their na-
tional identity might be outdated and thus might be psychologically dis-
carded. Kelman concludes his contribution with a detailed description of
how one workshop grappled with a crucial issue in the Israeli-Palestinian
conflict: what to do with, and for, the Palestinian refugees. A key feature of
his analysis is that this question has both material (what Kelman terms
228
Conclusion


“practical”) and symbolic facets. On the material or objective side, how can
Israel and Palestine accommodate the large number of refugees in other
countries? On the symbolic side, how can each side resolve the refugee is-
sue in such a way that it does not do violence to its national identity and
group narrative?
A Look Forward: A Social Identity Framework 
for Intergroup Conflict
As the preceding “look back” indicates, our authors have done an excellent
job of addressing the first goal of this volume: to survey what the social and
behavioral sciences have learned about the ways in which social identity and
intergroup harm-doing are interconnected. They have also succeeded in
addressing our second goal in putting together this book: to identify the
needed next steps in moving toward a more complete account of social
identity and intergroup conflict. In this section we pull together and orga-
nize the next-generation conceptual and methodological tools identified by
our contributors in the form of a social identity framework for intergroup
conflict that is schematically depicted in figure 9.1. It is important to em-
phasize that this is not a tight deductive model such as proposed by Brewer;
instead, we offer a heuristic scheme that identifies important constructs and
suggests how they might be interconnected. We present this framework as
follows. First, we provide a brief overview of the scheme. Second, the ad-
vantages of phrasing the framework in terms of a general threat or coping
paradigm are described. Third, the details and implications of the frame-
work are presented.
Overview of Framework 
The figure depicts three general sets of constructs—threat and appraisal;
destructive coping: intergroup harm-doing; and social identity—and identi-
fies these as mutually influencing one another. The arrow from “threat and
appraisal” to “destructive coping: intergroup harm-doing” marks this as a
general threat/coping framework as used in psychological analyses of stress
(Lazarus & Folkman, 1984). When members of a group perceive actual or
possible harm to themselves and other ingroup members (threat), they seek
to makes sense of this (appraisal), and if the other group is somehow judged
as responsible for the threat, they may cope with the appraised threat de-
structively by harming that group and its members. 
The return arrow from “destructive coping” to “threat” indicates that in-
tergroup harm-doing can escalate threat and appraisal: when one group in-
flicts harm on another, it raises the possibility of retaliation (i.e., threat).
The figure indicates that threat and appraisal and harm-doing are each con-
nected to social identity, and also that social identity moderates the link of
T
oward a Social Identity Framework for Intergroup Conflict
229


threat and appraisal to intergroup violence. Social identity consists of both
group (socially shared) and individual (personal) representations about own
and other group. Group identities that feature own group as victim and
other group as oppressor or aggressor heighten objective threat and ap-
praisal of danger; the reverse path indicates that being threatened and see-
ing another group as responsible is likely to influence individual and group
definition (e.g., living in a state of war, declared or not, with another group
will help create social and personal identities congruent with this state). 
Similarly, the “intergroup harm-doing” to “social identity” arrow indi-
cates that groups that engage in outgroup violence will have their personal
and group self-definitions altered, again to bring them into a psychologi-
cally consistent relation (e.g., “we are the people who are making the world
safe/pure/ . . . by subjugating/killing . . . them”). Finally, “social iden-
tity” moderates the link between “threat and appraisal” and destructive cop-
ing.” Groups that have a socially shared sense of “us” as victim and “them”
as dangerous are more likely to cope with threat via intergroup violence
than groups that do not have such a view of “us” and “them.”
The figure is obviously simplified in several respects. First, the frame-
work is from the perspective of only one group in relation to just one other
group. As White’s chapter in particular informs us, this is seldom the case in
real-world conflicts, and the framework, if it is to prove useful, will need to
be expanded to include additional groups beyond “us” and “them.” Second,
there are many other instigators to intergroup violence than threat. Often
one group aggresses against another for material gain because the group’s
leaders believe that they can get away with it. For example, the people of
West Africa in the 17th through 19th centuries posed no threat to Ameri-
cans; slaves were introduced into the United States simply because Euro-
230
Conclusion
figure 9.1 Social Identity Framework for Intergroup Conflict
&


peans and other slave traders had more power than Africans, and American
plantation owners wanted cheap labor. Third, threat, even if it is linked
clearly to some outgroup, does not always lead to violence. Fourth, there
are many inputs into intergroup conflict in addition to threat and appraisal
and social identity. At the societal level, type of economic system and politi-
cal structures are important, and, at the individual level, personality and so-
cial attitudes are likely significant contributors to group-based harm-doing.
Fifth, the figure suggests a static view of the links among the variables de-
picted. This is not the case. As Staub’s chapter makes very clear, there is a
through time path to mass killing and genocide. Thus, our figure stands for
many causal paths that unfold through time. For example, a group might
experience a large influx of immigrants to their country as an economic
threat (“They will take our jobs”); this may, in turn, lead to discriminatory
laws and outbreaks of violence against that group; if continued over time,
these actions will likely come to be prominent parts of views of us and them;
whether this would escalate to widespread and systematic harm-doing such
as genocide would depend on threat and appraisal and social identity ac-
cording to figure 9.1. 
Advantages of Threat/Coping Framework 
There are four major advantages to using a general threat/coping paradigm.
First, as has been demonstrated at the individual level, unpredictable and
uncontrollable threats are experienced as particularly stressful, and such
threats often lead people to seek the comfort and security of the ingroup
(Worchel, 1999). For most people, the world today is characterized by un-
predictable and uncontrollable forces and events—modernization in many
developing countries and globalization of the economy for all nations (Erik-
sen, this volume), rapid technological change, and often unplanned social
change. Thus, the model suggests that intergroup strife will be more, not
less, likely as we enter the 21st century, even though much of the globe is
experiencing considerable material progress. 
Second, as stress and health workers have demonstrated, it is fruitful to
distinguish problem- from emotion-focused coping. In the former, the per-
son, upon encountering a situation that signals current or future harm, at-
tempts to deal directly with the source of threat (e.g., upon learning that she
has high blood pressure, which it is explained to her increases the risk 
of health problems, the person changes her diet to reduce salt and fat).
Emotion-focused coping, on the other hand, involves attempts to reduce
the fear and anxiety that accompany threats. Thus, using the same example,
the newly diagnosed hypertensive individual could drink alcohol to avoid
thinking about the mysterious and scary condition about which she has just
learned. As applied to intergroup conflict, the 1929 Stock Market Crash
ushered in a worldwide depression. In the United States, one response was
the set of political measures termed the New Deal, which among other
T
oward a Social Identity Framework for Intergroup Conflict
231


things sought to provide assistance for the economically afflicted. This
problem-focused coping can be contrasted with the rise of Nazism in Ger-
many at about the same time. Hitler’s programs sought more to alleviate
feelings of despair and national impotence following the loss of World War
I and the economic and political turmoil of the Weimar years, and these led
to the “Final Solution” and the Holocaust.
Third, there are clear individual differences in styles of coping, and the
preceding chapters suggest that there are likely reliable societal and group
differences in styles of coping as well. Thus, for example, Staub notes that
societies differ in degree of authority orientation. It is likely that those high
in authority orientation will more likely accept violence-against-outgroup
“solutions” for threats offered by leaders even if they contradict societal
norms about harm-doing.
Fourth, and very important for the present analysis, there is evidence
that self and identity are implicated in the threat-coping process (Contrada
& Ashmore, 1999). Thus, there is reason to believe that an identity-inter-
group conflict model based on a threat-coping framework may be fruitful.
Framework Details and Implications
Destructive Coping: Intergroup Harm-Doing.
As depicted in figure 9.1, our
model seeks to account for intergroup harm-doing, by which we mean
harm by members of one group to those of another group based on group
membership. As suggested in the figure, we see intergroup harm-doing as
multidimensional. Intergroup violence can involve physical and/or psycho-
logical hurting. For example, during the slavery and Jim Crow eras in the
United States, African Americans were subjected to both physical violence
(e.g., beatings, lynchings, poor housing) and psychological violence (e.g.,
societal stereotypes of innate inferiority and racial etiquette that made in-
terracial contacts demeaning for blacks). T
oday the balance is much shifted
toward psychological harm-doing (persisting cultural stereotypes), though
physical violence against blacks has not been eliminated. 
Another important dimension is the degree to which harm-doing is sys-
tematized. At one extreme, members of one group engage in local and
spontaneous acts of violence against outgroup individuals. At the other ex-
treme, there are elaborate societal systems for hurting the other group. In
Nazi Germany, there was an organized set of institutions for carrying out
the “Final Solution.” It is not just open violence that can be highly system-
atized. The slavery system in the United States, the caste system in India,
and apartheid in South Africa all were undergirded by complex sets of 
social institutions (legal, educational, and the like) that oppressed certain
groups. 
There are two other important dimensions of intergroup harm-doing
not depicted in the figure: (1) extent—how widely across the other group is
the violence extended? and (2) duration—how long has the harm-doing
232
Conclusion


been going on? As both increase so too does the strength of the links of in-
tergroup harm-doing to social identity and threat. 
Threat and Appraisal. According to our model, a significant contributor to
intergroup harm-doing is threat and appraisal. Threat and appraisal are
highly related but separable concepts. Just as work on social problems has
profited from a distinction between actual conditions in the social environ-
ment and the public “seeing” and labeling of these as problematic (Thomp-
son & Fine, 1999), we too feel it is useful to distinguish the objective threat
and the individual and collective recognition and interpretation (or ap-
praisal) of this condition (see also Northrup, 1989). Thus, for example, dur-
ing the 1860s many Chinese immigrated to the United States to work on
the transcontinental railroad. The societally projected image of Chinese
during this period was relatively benign. However, when the railroad was
completed and many Chinese became competitors for jobs, unions and
other groups promoted public representations (in newspapers and the like)
of the Chinese as dangerous, and these came to be shared widely enough
that local discrimination occurred and national immigration restrictions
were implemented. Thus, it was not the sheer numbers of Chinese that
mattered or even the economic competition with whites, but the shared so-
cial representation of Chinese as dangerous. 
Threat, like harm-doing, is viewed as multidimensional. Figure 9.1 iden-
tifies three important dimensions: material, symbolic, link to other group.
Material and symbolic threat are often treated as opposite ends of a single
continuum. We believe, however, that it is best to consider these as poten-
tially independent dimensions. As noted above, material threat is the core of
realistic group conflict (RGC) approaches to intergroup conflict (Ashmore
& Del Boca, 1976). These are often contrasted with “symbolic politics” ap-
proaches (see Citrin et al., this volume; Sears et al., 1999). In symbolic poli-
tics (SP) models, groups come into “unrealistic” non-self-interested conflict
over symbolic issues. For example, Whose language should be spoken in
this country? Whose history should be taught in our schools? Our view is
that the conflict between the RGC and SP positions is more apparent than
real. Material things, especially land, often acquire symbolic meaning. And
symbols may have significant material implications. Continuing the same
examples, deciding “whose language” will, in turn, partially shape which
group’s members can make the better living economically. Whose history is
taught will legitimate and buttress the economic and political status of some
groups and undermine that of others. Thus, even symbolic conflicts have
important material implications.
In his contribution to this volume, Kelman stresses that the Israeli-Pales-
tinian conflict has both material, or “practical,” and symbolic facets, and
both of these need to be addressed in attempts to reduce the conflict. The
other chapters in this book tend to emphasize either material or symbolic
threat. As noted above, Brewer’s primary argument is in terms of symbolic
T
oward a Social Identity Framework for Intergroup Conflict
233


threat, whereas Eriksen stresses material conflicts. Citrin and his colleagues
were guided by symbolic politics theory; Sidanius and Petrocik’s group
dominance model is, at its core, a materialist framework—group dominance
in the service of material outcomes. 
The third dimension of threat depicted in the figure is link to the other
group. Some threats directly and clearly involve “them.” Some examples:
Chinese competing with whites in the United States after the completion of
the transcontinental railroad, Hutus and Tutsi attacking one another in
Rwanda (Staub, this volume), and laws in India designed to set aside jobs for
the lower castes (Eriksen, this volume). At the same time, intergroup harm-
doing can be triggered by threats that do not objectively arise from the
other group. In Germany in the 1920s, the hyperinflation and other eco-
nomic problems did not arise because of the actions of Jews. Hitler identi-
fied this group as a cause, however, and this appraisal was widely accepted
given the longstanding societal representations of Jews as sneaky and dan-
gerous (Goldhagen, 1996). Examples of types of threats that do not directly
involve the outgroup that yet can start a cycle of intergroup harm-doing are
economic instability and hardship (poverty, globalization that fractures 
traditional economic practices and procedures, inflation), political insta-
bility and corruption (changing governments, lack of societal agreement on
political processes, widespread graft by public officials), and rapid social
change (large-scale movements from rural to urban areas, changes in family
roles, and other social changes associated with “modernization”). 
In sum, our conceptualization of threat underlines two major points.
First, threats can be about tangible things (material threat) and about intan-
gibles as represented in valued symbols (symbolic threat). Further, a given
threat can have both a material and symbolic component. Second, conflict
between groups can begin with factors that are not objectively related to the
groups at all. Quite simply, some nongroup-related threats can initiate the
set of processes that lead to intergroup violence. In these cases, leaders di-
rect appraisal to “them” as the cause and away from the economic, political,
and social factors, leading individuals to see themselves and their groups as
in danger. In such cases, groups cope with the emotions aroused by the
threat rather than with the threat itself.
There are two other important dimensions of threat that are not de-
picted in the figure, and these parallel those for intergroup harm-doing: (1)
extent—how widely across the ingroup is the threat experienced? and (2)
duration—how long has the threat been in existence? As both increase, so
too do the links of threat to intergroup harm-doing. 
Appraisal is necessary for threat to lead to collective group-based vio-
lence. Although the figure suggests that this is a static phenomenon, it is
not. Instead, appraisal unfolds over time. T
o capture this idea we have iden-
tified both primary and secondary appraisal. Primary appraisal is an initial
making sense of the threat: What is the problem? What or who is responsi-
ble? Thus, for example, when a society with many foreign workers (as is
234
Conclusion


true of many nations) encounters an economic downturn, it might be ex-
plained relatively spontaneously as “them taking jobs from us,” and this
could lead, without much thought or organization, to isolated attacks on
the foreign workers.
However, for violence to be systematic and widespread requires mobiliz-
ing a larger segment of society and providing more organized paths to ac-
tion. This involves secondary appraisal guided by both group and individual
social identities as indicated in the figure. Regarding the former, group
identities must first be invoked to make sense of the threat. As Brewer,
Eriksen, and Staub note, leaders are crucial in this process (see also
Worchel, 1999). At least initially, these leaders are not likely to be national
political leaders but instead proponents of “fringe” groups, labor leaders, or
local religious figures. And, in an early coping stage, the efforts they direct,
though more organized than a spontaneous attack or a “race riot,” are still
relatively reflexive.
But what makes individual citizens ready to accept a group-based (“them
against us”) explanation for their troubles? Eriksen’s chapter points us in the
right direction when he notes that the shared group social identities prof-
fered by leaders must be consonant with the social and personal identities of
individuals and also fit with their daily experience and enactment of these
selves. This melding of a group identity with the identities of individuals
can result in mobilization of substantial numbers of individuals for action,
in this case a second stage of harm-doing.
As actions against the outgroup develop and widen, the ingroup engages
in secondary appraisal where the “guilt” of the outgroup is solidified in so-
cietally projected images, and harm-doing is explained as a necessary or
perhaps even a noble activity.
If mobilization is to be enduring, individuals will have to accept some
personal costs. This raises the crucial question: Under what circumstances
will individuals and groups be willing to endure personal hardship to either
inflict harm or to experience harm? Stated somewhat differently, what are
the factors that determine whether threat leads to mobilizing individuals
and the group to undertake action? As figure 9.1 suggests, we believe that
part of the answer lies in the content and structure of both group and indi-
vidual social identity. With regard to the group narrative, a national or eth-
nic story that features being a victim will increase the likelihood of mobi-
lization, and this will be true if there is also the sense of group vulnerability
(Eriksen, White, Staub, this volume). At the individual level, we suggest
that the relative lack of valued personal selves makes a society susceptible to
group-based mobilization. Thus, for example, in a nation in which men and
women are not able to successfully enact work and family roles because of
poverty, discrimination, and the like, they will be more likely to be suscepti-
ble to social identities that involve outgroup violence or support others 
doing harm.
An important implication of this line of reasoning is that protracted con-
T
oward a Social Identity Framework for Intergroup Conflict
235


flicts create social identities, at both the group and individual level, that
make the progression to widespread and systematic violence more likely.
Harming others based on group membership or standing by while others
do this changes people. Cognitive consistency models suggest that people
and groups would develop personal and shared reasons that justify this 
behavior. These have included belief systems specifying that they are not
human (and, thus, norms about hurting others do not apply), are inferior
humans (and, thus, need to be taken care of as slaves), or are powerful and
dangerous (and, thus, must be killed before they destroy us). These ratio-
nalizations become part of transformed group and individual identities 
as depicted by the arrow from intergroup harm-doing to social identity in
figure 9.1. Given the focus of this volume on social identity, we turn now to
a more detailed analysis of the group and individual identity portion of the
scheme.
Social Identity. As indicated at the top of figure 9.1, and suggested at several
places above, social identity is here viewed as a significant input to the
threat-intergroup harm-doing relationship. In the proposed framework, we
distinguish group from individual social identity. 
• Group Social Identity. Group social identities are group-level phe-
nomena. This means that both the referent and locus of the identity is 
the group. By referent we mean what the identity is about, the group—
America/Americans; Serbia/Serbians, and so on. The content of the iden-
tity is the set of things that are thought to “define,”—for example, America
and Americans, including geography (e.g., 50 states, mostly between
Canada and Mexico), politics (e.g., democracy, two-party system), econom-
ics (e. g., capitalism), morals and values (e.g., caring, freedom), history (e.g.,
revolution, civil war), and the like. By locus of the identity, we mean that
the group is also the level of representation of the identity. Such representa-
tion takes both psychological and socio-structural form. Regarding the for-
mer, labeled “Social Representations” in the figure, the group identity is a
set of beliefs that are widely accepted in the society, even though many indi-
viduals do not know or endorse all of the beliefs (Thompson & Fine, 1999).
The socio-structural form of group identity, “Societally Projected Images”
in figure 9.1, is manifested in two primary and intertwined ways: (1) it is
woven into the political, economic, and social structure of the society, often
in implicit ways (e.g., institutional racism in the United States [ Jones,
1997]); (2) it is conveyed by the societal mechanisms of socialization such as
schools and the media (e.g., cultural racism in the United States [ Jones,
1997, Worchel, 1999]). 
In addition to the rather static group social identity notion (“We are
_____”), we propose that there is a group-level analog to life story and per-
sonal narrative at the individual level (McAdams, 1997; Gergen & Gergen,
1988). That is, just as individuals develop stories that specify not only what
“I am like now” but also “how I got here” and “where I will likely go in the
236
Conclusion


future,” we propose that groups develop narratives that tell the story of
their group. This proposed construct is explicit in only one of the foregoing
chapters; Kelman makes national narrative a central part of his analysis, but
he draws no link to life story or self-narrative at the individual level. The
notion of group narrative, however, is implicit in several other contribu-
tions. For each of the three case studies of intergroup conflict described by
Eriksen, a group narrative was propagated by political leaders who identi-
fied “us” as noble and deserving and “them” as dangerous (e.g., in Bosnia
the Serbian national narrative featured past defeats at the hands of powerful
outsiders, especially the Turks, and the local Muslim community was de-
fined as representing external invaders even though this definition was inac-
curate). In describing the role identities of Republicans, White emphasized
the importance of a family history of involvement in the cause and the fact
that this history was often passed down from one generation to the next.
We suggest that, as one generation of Republicans socialized the next, they
not only told the story of “our family’s” involvement but also told the story
of “We Republicans,” the narrative of this group. 
There are four additional important points about group narrative:
1. Group stories can lie relatively dormant for periods of time. For
example, in Yugoslavia during the Tito period, the divergent group
narratives of the Serb, Croat, and Muslim communities were not in
the foreground of societally projected images or social representa-
tions of the citizenry.
2. At certain times and under certain conditions, there are forces
pushing for demonstrating and featuring a national narrative.
Threat, in the form of societal crises or outgroup actions, can
bring group stories to prominence and political leaders can use
such narratives (Brewer, Eriksen, this volume).
3. Group narratives can change. As Kelman noted, history is not
fixed, and, thus, groups can alter their story (Rothman, 1997).
4. Group narratives often contain folk theories of ethnicity, and these
generally do not match scientific understanding of this concept. As
Eriksen notes, most current-day anthropologists reject the notion
that ethnicity is primordial, and yet it is precisely this version of
ethnicity that activists and politicians use in mobilizing people to
action on the basis of ethnicity.
• Individual Social Identity. At the individual level, identity components
are generally partitioned into social identities and personal identities (Ash-
more & Jussim, 1997b), and we follow that convention here even though
drawing the line between personal and social is not always easy (Deaux,
1992). Social identities are those self definitions that identify self in terms
of, or in relation to, other people and groups (e.g., “I am a teacher” which
specifies a role that I and other teachers enact in relation to other people
who fill the role of student; “I am an American,” which specifies that I am 
T
oward a Social Identity Framework for Intergroup Conflict
237


a member of a social category along with others). Personal identities are
self-definitions that identify me as a unique individual (e.g., “I am obedi-
ent”). Personal identity also includes self-narrative (an organized story
about me, again as an individual). 
Given the thrust of this volume, we feature social identity here, though
we briefly describe below how individual-level personal identity might be
important to intergroup conflict. We follow Thoits and Virshup (1997) in
partitioning individual-level social identities into “we’s” (i.e., self as a mem-
ber of a collective group [e.g., “I am an African American”]) and “me’s” (i.e.,
self as fulfilling a role [e.g., “I am a mother”]). 
1. Individual social identity: “we selves.” In the application of social identity
to intergroup conflict, “we selves” are obviously important and have been
the most often studied. Thus, in this volume, all authors have been con-
cerned with how individuals define self in relation to an ethnic or national
group. And, they have done this in a variety of ways: self categorization in
terms of an ethnic label such as black or white, felt closeness to one’s group,
beliefs about one’s group and another group, and the like. This diversity 
of approaches raises an important question for the next generation of work
on social identity and conflict: what are the components of “we” social 
identities? 
Most researchers regard self-definition in terms of collective social
groups as multidimensional, but there is far from consensus on what these
dimensions are. T
o illustrate, table 9.1 lists the distinctions and components
that three research groups have proposed for ethnic identity, racial identity,
and cultural identity, perhaps the most often studied types of “we” social
identities. On the same line are listed dimensions that are the same or
highly similar. It can be seen that there are five areas of partial agreement
on what are the major dimensions of “we” social identity. T
wo of the three
articles included: (1) salience/centrality; (2) evaluation and feelings/private
regard; (3) values and attitudes/ideology; (4) interest and knowledge/famil-
iarity; (5) behaviors and practices/behavior and language.
At the same time, no dimension is mentioned by all three authors. And,
there are three components listed by just one author: (1) self-identification,
(2) commitment, (3) public regard. De la Garza et al. and Sellers and his as-
sociates may have omitted self-identification because they were proposing
scales to measure discrete dimensions of a specific social identification, and,
in so doing, they may have taken for granted that the first step in social
identification is placing self in a social category. At the same time, it is likely
that a comprehensive assessment of “we” social identity will require atten-
tion to self-identification, including such issues as preferred ethnic label and
the possibility of self-subtyping. Just as perceivers discern subtypes within
age (Brewer, Dull, & Lui, 1981), gender (Clifton, McGrath, & Wick, 1976),
and racial categories (Devine & Baker, 1991), it is likely that people may de-
fine self in terms of such subtypes (e.g., Eriksen noted that Yugoslavians liv-
ing in Belgrade and other urban areas had very different views of being a
238
Conclusion


Serb, Muslim, or Croat than those in rural areas, and this may involve sub-
typing). Regarding commitment, most would agree that something akin to
the Eriksen/Marcia notion of the developmental path that the person took
to her or his current sense of group identification is a potentially important
component of racial/ethnic/cultural identity. Cross’s (1995) work is a re-
lated but different approach to social identity as a developmental construct.
Finally, even though only Sellers et al. explicitly include a public regard di-
mension as part of social identity, it is likely that others would regard this as
a potentially important dimension.
T
able 9.1 illustrates one other complication in conceiving “we” social
identities: to what extent must the “other” group be included within the
concept of collective social self-definition? De la Garza and associates are
very clear that Latino cultural identity includes familiarity with and lan-
guage skills relevant to both the ingroup and the “other” group. Should as-
pects of identification in terms of “them” routinely be included in assess-
ments of social identities?
As is apparent, table 9.1 suggests a large number of possible components
of “we” social identities. As researchers grapple with which are necessary to
include, it is important to consider construct validity. In the simplest terms,
where does collective social identity end and related variables (causes, cor-
T
oward a Social Identity Framework for Intergroup Conflict
239
table 9.1 Dimensions of “We” Social Identity Proposed by Three Research
Groups
AUTHOR(S)
Phinney (1995)
de la Garza et al. (1995)
Sellers et al. (1998) 
CONSTRUCT
Ethnic identity
Latino(a) cultural identity
African-American 
racial identity
DIMENSIONS
Self-identification
—
—
Salience
—
Centrality
Evaluation of
—
Private regard
and
feelings about
group
—
—
Public regard
—
Values and attitudes
Ideology
Interest in
Familiarity with
—
and
Latino(a) culture
knowledge of
American culture
group
Ethnic behaviors 
Behavior 
—
and practices
and
language
Spanish proficiency
English proficiency
Commitment
—
—


relates, and consequences) begin? For example, is “interest in and knowl-
edge about” one’s ethnic group part of one’s social identity, or is it best seen
as an outcome predictable from identification? This question becomes even
more difficult regarding behavior: are ethnic behaviors part of social iden-
tity or an outcome that can be used to assess the construct validity of social
identity measures? 
One of the most important messages of the present volume is that even
conflicts that appear to be simple “us against them” on the basis of one so-
cial identity distinction (e.g., religion in Northern Ireland) also involve
multiple other social identities. These other social identifications are not
just regarding ethnic group and nation, which have been the focus of the
foregoing chapters. For example, in Part II of the volume addressing ethnic
and national identity in the United States, there was evidence that gender,
socioeconomic status, and religion were associated with ethnic and national
identifications and with the intergroup attitudes and policy positions under
investigation. Thus, as highly social creatures, people have multiple identi-
fications with collectives, and these interrelate with one another in complex
ways to shape orientations toward other groups and toward political issues.
2. Individual social identity: “me selves.” As White’s chapter reminds us,
how people relate to their own and other groups depends not just on collec-
tive self-definitions (“we’s”), but also on thinking of self in terms of social
roles (“me’s”). As noted above, role identities pertinent to group conflict are
especially likely to arise and be important in protracted conflicts where
many “supporting” roles are needed to carry on the fight. Also as noted
above, there is a second major way in which role social identities are impor-
tant to understanding intergroup conflict: to the extent that societal condi-
tions allow individuals to have and successfully fulfill valued role identities
(e.g., husband and father), it will be more difficult to recruit them for doing
violence against others. On the other hand, when poverty, oppression,
modernization, and other social forces make it difficult for large numbers of
people to fulfill social roles, it makes it easier for them to be drawn to social
identities that provide feelings of positive evaluation and potency even if
they also explicitly or implicitly involve doing harm to others. 
In addition, the relation between social role organization and group
boundaries is likely to have important implications for group-based vio-
lence. Eriksen’s observation that Bosnians and Serbs had separate private
spheres means that they did not have role identities (husband-wife, in law,
and the like) across social category lines. Social roles organized largely along
in-group lines may make it easier to recruit people to intergroup harm-do-
ing. Further, it is likely that friendships, too, were predominantly within
group. We suggest that cross-group friendship is likely to reduce inter-
group violence (Ashmore, 1970; Pettigrew, 1998). If I see myself as Richard
(Bosnian) friend of Mel (Serb) and we have extensive first-hand experience
with one another, it will be harder to mobilize either of us for “us against
them” identity conflict.
240
Conclusion


Personal Identity.
Although social identities are most pertinent to inter-
group conflict, we suggest that personal identifications (self-construals that
identify me as a unique individual) are also likely important. We offer two
illustrative examples. First, although Staub treated “authority orientation”
as a societal-level construct, it is likely to have an individual-level represen-
tation also, with substantial individual differences even within a society high
(or low) in this orientation. T
o the extent that I define myself as obedient to
authority, I will be a more likely recruit for group-based violence fomented
by political leaders. Second, as noted above, we believe that self narratives
are important to identity at both the group and individual level. White’s
chapter suggests that self stories are most likely to be conflict-related in so-
cieties characterized by protracted conflict. White notes that many current-
day Republicans come from families with a long commitment to the cause
of freeing Ireland from England. Further, in these families, that tradition is
a central part of the stories that parents and grandparents tell to children,
and, in turn, this history of one’s family’s involvement becomes part of the
individual’s personal story. Thus, just as Kelman noted that the group nar-
rative can feature “We are the people who fight them,” so, too, at the indi-
vidual level, a central part of an individual’s narrative can be that “My family
and I are people who fight them.”
Whose Social and Personal Identities? It is useful to distinguish three general
types of actors in intergroup conflict situations:
1. Elites—individuals who have political, economic, or information
power. As Brewer, Eriksen, Staub, and Kelman note, elites play a
pivotal role in social strife, often in propagating “us versus them”
social identities to mobilize citizens (Worchel, 1999).
2. Agents of violence—people who actively and directly carry out the
harm-doing (e.g., in White’s chapter, the Protestant and Republi-
can paramilitaries and English soldiers).
3. “Ordinary _____s” (as per Goldhagen’s [1996] “ordinary Ger-
mans”)—general citizens or average group members. Whatever
the merits of Goldhagen’s arguments, he has highlighted a crucial
point, How do ordinary citizens become participants in organized
violence between groups? Staub suggests that bystanders who “do
nothing” while organized violence is done will be touched, possi-
bly even taking on the identity of the hurting party.
Caveats
We reiterate that the proposed framework is not a tight deductive theory,
but instead a heuristic scheme intended to highlight the possible roles of
identity in intergroup conflict. Thus, it is necessary to note two significant
caveats.
T
oward a Social Identity Framework for Intergroup Conflict
241


One of the most important caveats about the model is what is left out.
Many factors that contribute to intergroup violence are not depicted. As
noted above, these omissions are made to underscore the importance of so-
cial identity. At the same time, as with any framework, our scheme hides as
well as highlights. Even some factors that we feel are quite important (e.g.,
the role of political, economic, and religious leaders in promoting identity-
based conflict) are only indirectly represented in the figure (e.g., as soci-
etally projected images under group social identity).
A second caveat concerns context. As depicted in figure 9.1, we have es-
sentially frozen in time a conflict between two groups and also wrenched
that conflict out of its regional and global context. As is clear in the chapters
by Eriksen, Sidanius and Petrocik, White, and Kelman, however, many eth-
nic and national conflicts unfold over a relatively long time period, and, as
noted above, early events profoundly influence the current situation
(Fisher, 1990; Northrup, 1989; Rothman, 1997). 
White’s chapter highlights the “space/place” context caveat for our
model: In the case of Northern Ireland, England/Britain is an “outside”
(not “them” or “us” if one focuses just on the two groups in Northern Ire-
land) yet potent factor (and if White is correct, the “them” party for Repub-
licans). Further, White notes but does not emphasize other important out-
siders—the Republic of Ireland and the United States. The differing
responses of the world community to the conflict in Yugoslavia and Rwanda
suggest the importance of regional context, together with associated strate-
gic interests and belief systems (e.g., racism). The fighting in Kosovo, on
the edge of Europe, led to a military campaign by NATO, whereas the
genocide in Rwanda resulted in action from the world community that was
“too little too late” (New York Times, December 16, 1999).
Identity and Conflict Reduction
Given that we anticipate a future volume of the Rutgers Series on Self and
Social Identity completely devoted to reducing intergroup conflict, here we
discuss just two issues important to defusing intergroup conflict: social psy-
chological mechanisms and types of interventions.
Regarding social psychological processes, several of the preceding chap-
ters noted that getting people to accept higher order, superordinate identi-
ties might be useful in ameliorating intergroup strife (Eriksen, Staub, Kel-
man). Brewer, too, raised this possibility, but also noted that according to
her optimal distinctiveness model and the present extension and elabora-
tion of it, overarching identity units could create conflict under some con-
ditions. The possible utility of higher order identities for conflict reduction
is the direct focus of Gaertner and his colleagues. In multiple experiments,
they have demonstrated that intergroup bias is significantly reduced when
242
Conclusion


members of different groups adopt a superordinate or common group iden-
tity (e.g., Gaertner, Dovidio, Nier, Ward, & Banker, 1998)
Perhaps the furthest extension of the concept of superordinate categories
is that of “global culture.” The goal of global culture is to unite all people
on the basis of characteristics shared as humans, thereby changing group
boundaries so members of currently different groups see themselves as be-
longing to the same group (Worchel, 1999). However, it is important to
note that global culture violates the basic need of identity uniqueness as
specified by Brewer’s optimal distinctiveness model (see Brewer, this vol-
ume). It also requires that individuals deny all previous group identities
(e.g., national, ethnic) in favor of the global identity. Thus, although the
notion of global culture builds on research on higher order identities, it
conflicts with evidence that the benefits of superordinate identities occur
even when subordinate identities are also strong. These findings suggest
that intergroup conflict reduction does not require that people give up their
ethnic identity or other relatively specific collective social identities (Gaert-
ner et al., 1998; Citrin et al., this volume).
Although the notion of higher order identities is appealing and Gaertner
and his co-workers have provided evidence for such an approach, we would
note that the power of transcendent identifications is not just inside the
head. Rather, we believe that it is action based on shared goals that is im-
portant. This was Sherif’s (1958) basic insight when he introduced the idea
of “superordinate goals”—goals that members of both groups desired but
that could not be achieved by either group on its own; as a consequence the
groups had to work together. Ashmore’s (1970) shared coping model took
Sherif’s logic a step further by proposing that shared goals induced coop-
erative action and that such behavior promoted cross-group positive affec-
tive ties, especially friendships, and these emotional ties promoted stereo-
type destruction, dissonance reduction, and the like that, in turn, produced
prejudice reduction (see also Pettigrew, 1998). The importance of cross-
group friendships as a mediator of intergroup contact fits with Eriksen’s
emphasis on lived experience. Just as the separate private spheres of Bosni-
ans and Serbs provided fertile ground for religiously based hostility, so, too,
could changes that promote cross-group friendship help in reducing con-
flict and making citizens of Bosnia more resistant to identity politics appeals
from leaders.
The second issue related to reducing intergroup strife to be addressed is
implementation of conflict reduction strategies. Here we discuss two ap-
proaches: educational programs designed to promote learning about other
groups, and third-party interventions. Many school-based programs have
been designed to foster intergroup learning (Aboud & Levy, 1999; Gurin,
1999). Bilingual and multicultural educational programs and materials have
been shown to have positive effects on intergroup relations among students
(Bigler, 1999; Genesee & Gandara, 1999; Stephan & Finlay, 1999). The
T
oward a Social Identity Framework for Intergroup Conflict
243


goal of these programs is to create an environment that allows groups to
maintain their cultural identity and, at the same time, accept other groups
(Fisher, 1990). Through these programs, individuals may see that they are
less different from members of other groups than they thought (Stephan &
Finlay, 1999). The danger, of course, is that individuals may find learning
about their group highly involving and satisfying, while ignoring or deni-
grating lessons about other groups. If so, such curricula might unintention-
ally magnify group boundaries and increase intergroup friction.
Another approach to conflict reduction, designed for groups engaged in
active, often violent conflict, is third-party intervention (e.g., Worchel,
1999; Rothman, 1997). Although the positive role of third-parties in reduc-
ing conflict is most obvious in Kelman’s chapter in the present volume, it is
also clear in the contribution by Staub. Fisher (1990) suggests that it is im-
portant to distinguish different types of third-party interventions, and pre-
sents a taxonomy of four types of third-party interventions that differ in
terms of whether or not the third-party focuses on the subjective (including
perceptions and evaluations of the conflict) and objective (including con-
flicts over both scarce resources [material] and over values [symbolic]) as-
pects of the conflict. Third-party consultation, the type performed by Kel-
man and his colleagues, is seen as the most desirable because it involves a
clear emphasis on both the objective and subjective aspects of the conflict
(Fisher, 1990). In contrast, there are numerous examples of conflict inter-
ventions carried out by other nations or aggregations of nations that focus
only on objective facets of the conflict. Worchel (1999) suggests that such
third-party interventions can sometimes save ethnic groups from persecu-
tion and destruction, but that they rarely lead to long-term peace. It is pos-
sible that these interventions are often unsuccessful, in part, because they
fail to adequately address the identity issues of the conflict. 
The chapters by Kelman and Staub raise an important question regard-
ing third-party intervention: When should third parties intervene? At the
roundtable discussion at the conference associated with this volume there
was clear agreement that early intervention was desirable, though Citrin
pointed out that there is little political incentive for leaders to intervene un-
til major conflict has emerged with clear evidence of harm-doing. Further,
Rothman (1997) notes that negotiating too early in a conflict may instead
escalate identity conflicts.
Summary and Suggestions
In this concluding chapter we both summarized the preceding chapters and
offered a conceptual framework that might guide future research.
Part I of the volume was, in part, concerned with the issue of the locus of
social identity—does social identification reside in the individual or in soci-
ety? Brewer and Eriksen did not resolve this starting question, and, in fact,
244
Conclusion


these authors answered that social identity is both “in” society and “in” the
individual. Thus, social identity, as with self and identity more generally, is
both personal and social (Ashmore & Jussim, 1997b). Brewer and Eriksen
identified the individual and societal building blocks of social identity, espe-
cially as they are linked to conflict between groups.
In Part II we asked the contributors to address one type of conflict (po-
litical strife) in one society (the United States). Not only did they share this
charge, the authors also used a similar method, survey research. Yet given
this shared topic and empirical approach, Citrin et al. and Sidanius and
Petrocik reached contrasting conclusions, the former seeing generally low
levels of ethnically based political conflict and the latter concluding that 
the ethnic divide is wide and deep. As noted above, the difference in con-
clusions is more apparent than real, reflecting, in part, methodological and
interpretational differences. The Part II authors leave us with many im-
portant ideas, including the multifaceted nature of ethnic and national 
identification.
In Part III, White and Staub addressed violent conflict, the first in an
analysis of Northern Ireland and the latter with a general model of the fac-
tors contributing to mass killing and genocide. These two chapters dovetail
well in highlighting the importance of history and action for understanding
the links between identity and conflict. White detailed the history of the 
violence in Northern Ireland and showed how the distant past was influen-
tial in shaping the most recent protracted fighting beginning in the late
1960s and is still much alive as the contestants seek to build on, and extend,
the 1998 peace accord. Staub underlined that identities are not just in 
the head but, instead, both influence action and are shaped by action (and
inaction). 
Kelman, in the single Part IV chapter, described the problem-solving
workshop approach that he and his colleagues have developed to ameliorate
protracted conflicts. In so doing he underlined the central role that national
identity plays in maintaining such conflict, and, at the same time, he offered
the very optimistic message that national identity as a social construction
can be altered.
The second part of this concluding chapter was devoted to a “look
ahead.” We presented a social identity framework for intergroup strife.
This scheme, depicted in figure 9.1, identifies three major constructs: threat
and appraisal, destructive coping: intergroup harm-doing, and social iden-
tity. It is proposed that these are mutually causally linked and further that
social identity moderates the threat and appraisal to intergroup harm-doing
link.
We offer three general suggestions for theorists and researchers under-
taking the next generation of work on social identity and intergroup con-
flict. First, we believe it will yield dividends if intergroup relations re-
searchers looked to the concepts and methods of the general field of “self
and social identity.” Except for White’s use of role identity models, all the
T
oward a Social Identity Framework for Intergroup Conflict
245


chapters in this book were guided by social identity perspectives specifically
intended for understanding intergroup relations (e.g., T
ajfel, 1981; Turner,
Hogg, Oakes, Reicher, & Blackwell, 1987). There is, however, a broad in-
terdisciplinary field concerned more generally with the issues of self and so-
cial identity (Ashmore & Jussim, 1997a). We urge scientists interested in in-
tergroup conflict to look more systematically at this field. We illustrated
what we have in mind above in linking “national narrative” as described by
Kelman to the individual-level concept “life story” (McAdams, 1997) and
personal narrative (Gergen & Gergen, 1988).
Second, we suggest that those interested in the interface between iden-
tity and group conflict consider a broader range of conflicts than here ad-
dressed. Because we anticipate a future volume directed at identity and gen-
der, we did not include female-male relations in this volume. Yet gender
and other forms of intergroup conflict are likely to share some features with
ethnic and national conflict but also to have some unique facets.
Third and finally, we hope that the general framework we have proposed
will generate both empirical work and conceptual debate. Our broader
hope is that the heuristic scheme, and the volume as a whole, will spur oth-
ers to develop a more comprehensive and coherent social identity paradigm
for intergroup conflict
Note
1. History is also important for knowing the “objective” ways in which present
day material realities and symbols are constructed. For example, the Rwandan geno-
cide was made more likely by the distant and recent past of Rwanda: Several cen-
turies ago, the indigenous Hutu were subdued by Tutsi invaders (Smith, 1998), who
were physically different (e.g., taller); the Tutsis implemented a feudal system that
subjugated the numerically superior Hutus and propagated a value system that ex-
alted Tutsi physical attributes. Early in the twentieth century, colonial Belgians used
these economic and value systems to help maintain order and their control over
both Hutus and Tutsis. In addition, they added the notion of “blood” to explain
group differences. These historical “facts” shape today’s material and symbolic 
realities that guide action.
References
Aboud, F. E., & Levy, S. R. (1999). Are we ready to translate research into pro-
grams? Journal of Social Issues, 55, 621–625.
Aretxaga, B. (1993). Striking with hunger: Cultural meanings of political violence in
Northern Ireland. In K. B. Warren (Ed.), The violence within: Cultural and political
opposition in divided nations (pp. 219–253). Boulder, CO: Westview Press.
Ashmore, R. D. (1970). Prejudice: Causes and cures. In B. E. Collins (Ed.), Social
psychology: Social influence, attitude change, group processes, and prejudice. Reading,
MA: Addison-Wesley.
246
Conclusion


Ashmore, R. D., & Del Boca, F. K. (1976). Psychological approaches to understand-
ing intergroup conflict. In P. A. Katz (Ed.), Towards the elimination of racism
(pp. 73–123). New York: Pergamon Press.
Ashmore, R. D., & Jussim, L. (Eds.) (1997a). Self and identity: Fundamental issues
(Vol. 1). New York: Oxford University Press.
Ashmore, R. D., & Jussim, L. (1997b). T
oward a second century of the scientific
analysis of self and identity. In R. D. Ashmore & L. Jussim (Eds.), Self and iden-
tity: Fundamental issues (Vol. 1, pp. 3-19). New York: Oxford University Press.
Bigler, R. S. (1999). The use of multicultural curricula and materials to counter
racism. Journal of Social Issues, 55, 687–705.
Boskin, J. S. (1970). Sambo. In G. Nash and R. Weiss (Eds.), The great fear: Race in
the mind of America. New York: Holt, Rinehart and Winston.
Boulding, K. E. (1962). Conflict and defense: A general theory. New York: Harper.
Brewer, M., Dull, V. & Lui, L. (1981). Perceptions of the elderly: Stereotypes as
prototypes. Journal of Personality and Social Psychology, 41, 656–670.
Clifton, A. K., McGrath, D., & Wick, B. (1976). Stereotypes of woman: A single
category? Sex Roles, 2, 135–148.
Contrada, R. J., & Ashmore, R. D. (Eds.) (1999). Self, social identity, and physical
health (Vol. 2, pp. 3–20). New York: Oxford University Press.
Contrada, R. J., Ashmore, R. D., Gary, M. L., Coups, E., Egeth, J. D., Sewell, A.,
Ewell, K., & Goyal, T. (in press). Measures of ethnicity-related stress: Psycho-
metric properties, ethnic group differences, and associations with well-being.
Journal of Applied Social Psychology.
Cross, W. E. (1995). The psychology of Nigresence: Revising the Cross Model. In 
J. G. Ponterotto, J. M. Casas, L. A. Suzuki, and C. M. Alexander (Eds.), 
Handbook of multicultural counseling (pp. 93–122). Thousand Oaks, CA: Sage 
Publications.
Deaux, K. (1992). Personalizing identity and socializing self. In G. Breakwell (Ed.),
Social psychology of identity and the self-concept (pp. 9–33). London: Academic Press.
de la Garza, M. F-O., Newcomb, M. D., & Myers, H. F. (1995). A multidimensional
measure of cultural identity for Latino and Latina adolescents. In A. M. Padilla
(Ed.), Hispanic psychology: Critical issues in theory and research (pp. 26–42). Thou-
sand Oaks, CA: Sage Publications.
de la Garza, R. O., Falcon, A., & Garcia, C. (1996). Will the real Americans please
stand up: Anglo and Mexican an-American support of core American political
values. American Journal of Political Science, 40, 335–351.
Devine, P. G., & Baker, S. M. (1991). Measurement of racial stereotype subtyping.
Personality and Social Psychology Bulletin, 17, 44–50
Festinger, L. (1950). Informal social communication. Psychological Review, 57, 271–
282.
Festinger, L. (1954). A theory of social comparison process. Human Relations, 7,
117–140.
Festinger, L. (1957). A theory of cognitive dissonance. Stanford: Stanford University
Press.
Fisher, R. J. (1990). The social psychology of intergroup and international conflict resolu-
tion. New York: Springer-Verlag.
Fredrickson, G. M. (1999). Models of American ethnic relations: A historical per-
spective. In D. A. Prentice & D. T. Miller (Eds.), Cultural divides: Understanding
and overcoming group conflict (pp. 23–34). New York: Russell Sage.
T
oward a Social Identity Framework for Intergroup Conflict
247


Gaertner, S. L., Dovidio, J. F., Nier, J. A., Ward, C. M., & Banker, B. S. (1999).
Across cultural divides: The value of superordinate identity. In D. A. Prentice &
D. T. Miller (Eds.), Cultural divides: Understanding and overcoming group conflict.
New York: Russell Sage.
Genesee, F., & Gandara, P. (1999). Bilingual education programs: A cross-national
perspective. Journal of Social Issues, 55, 665–685.
Gergen, K. J., & Gergen, M. M. (1988). Narrative and the self as relationship. In 
L. Berkowitz (Ed.), Advances in experimental social psychology (Vol. 21, pp. 17–56).
New York: Academic Press.
Goldhagen, D. (1996). Hitler’s willing executioners: Ordinary Germans and the holo-
caust. New York: Alfred A. Knopf.
Gurin, P., Peng, T., Lopez, G., & Nagda, B. A. (1999). Context, identity, and inter-
group relations. In D. A. Prentice & D. T. Miller (Eds.), Cultural divides: Under-
standing and overcoming group conflict (pp. 133–170). New York: Russell Sage.
Hull, C. L. (1943). Principles of behavior. New York: Appleton-Century-Crofts.
Iyengar, S. S., Lepper, M. R., & Ross, L. (1999). Independence from whom? Inter-
dependence with whom? Cultural perspectives on ingroups versus outgroups. In
D. A. Prentice & D. T. Miller (Eds.), Cultural divides: Understanding and overcom-
ing group conflict (pp. 273–301). New York: Russell Sage.
Jones, J. (1997). Prejudice and racism (second edition). New York: McGraw-Hill.
Lazarus, R. S., & Folkman, S. (1984). Stress, appraisal, and coping. New York:
Springer.
McAdams, R. C. (1997). The case for unity in the (post)modern self: A modest pro-
posal. In R. D. Ashmore & L. Jussim (Eds.), Self and identity: Fundamental issues
(Vol. 1, pp. 46–78). New York: Oxford University Press.
McCall, G. J., & Simmons, J. L. (1966). Identities and interactions. New York: Free
Press. 
New York Times, 16 December 1999.
Norenzayan, A., Choi, I., & Nisbett, R. E. (1999). Eastern and Western perceptions
of causality for social behavior: Lay theories about personalities and situations. In
D. A. Prentice & D. T. Miller (Eds.), Cultural divides: Understanding and overcom-
ing group conflict (pp. 239–272). New York: Russell Sage.
Northrup, T. A. (1989). The dynamic of identity in personal and social conflict. In
L. Kriesberg, T. A. Northrup, & S. J. Thorson (Eds.), Intractable conflicts and their
transformation (pp. 55–82). Syracuse, NY: Syracuse University Press.
Pettigrew, T. F. (1998). Intergroup contact theory. Annual Review of Psychology, 49,
65–85. 
Phinney, J. S. (1995). Ethnic identity and self-esteem: A review and integration. 
In A. M. Padilla (Ed.), Hispanic psychology: Critical issues in theory and research
(pp. 57–70). Thousand Oaks, CA: Sage Publications.
Rothman, J. (1997). Resolving identity-based conflict in nations, organizations, and com-
munities. San Francisco, CA: Jossey-Bass.
Rosenberg, S. (1997). Multiplicity of selves. In R. D. Ashmore & L. Jussim (Eds.),
Self and identity: Fundamental issues (Vol. 1, pp. 23–45). New York: Oxford Uni-
versity Press.
Sears, D. O. (1993). Symbolic politics. In S. Iyengar and W. J. McGuire (Eds.), 
Explorations in political psychology (pp. 113–149). Durham: Duke University Press.
Sears, D. O., Citrin, J., Cheleden, S., & Van Laar, C. (1999). Cultural diversity and
multicultural politics: Is ethnic Balkanization psychologically inevitable? In 
248
Conclusion


D. A. Prentice & D. T. Miller (Eds.), Cultural divides: Understanding and overcom-
ing group conflict (pp. 35–79). New York: Russell Sage.
Sellers, R. M., Smith, M. A., Shelton, J. N., Rowley, S. A. J., & Chavous, T. M.
(1998). Multidimensional model of racial identity: A reconceptualization of
African American racial identity. Personality and Social Psychology Review, 2, 18–39.
Sherif, M. (1958). Superordinate goals in the reduction of intergroup conflict.
American Journal of Sociology, 63, 349-363
Smith, D. N. (1998). The psychocultural roots of genocide: Legitimacy and crisis in
Rwanda. American Psychologist, 53, 743–753.
Stephan, W.G., & Finlay, K. (1999). The role of empathy in improving intergroup
relations. Journal of Social Issues, 55, 729–743.
Stryker, S. (1987). Identity theory: Developments and extensions. In K. Yardley &
T. Honess (Eds.), Self and identity: Psychosocial perspectives (pp. 89–103). New York:
Wiley. 
T
ajfel, H. (1978). Differentiation between groups: Studies in the social psychology of inter-
group relations. London: Academic Press.
T
ajfel, H. (1981). Human groups and social categories: Studies in social psychology. Cam-
bridge: Cambridge University Press 
Thoits, P. A., & Virshup, L. K. (1997). Me’s and we’s: Forms and functions of social
identities. In R. D. Ashmore & L. Jussim (Eds.), Self and identity: Fundamental 
issues (Vol. 1, pp. 106–136). New York: Oxford University Press.
Thompson, L., & Fine, G. A. (1999). Socially shared cognition, affect and behavior:
A review and integration. Personality and Social Psychology Review, 3, 278–302.
Turner, J. C., Hogg, M. A., Oakes, P. J., Reicher, S. D., & Blackwell, M. S. (1987).
Rediscovering the social group: A self-categorization theory. Oxford: Basil Blackwell.
Worchel, S. (1999). Written in blood: Ethnic identity and the struggle for human har-
mony. New York: Worth Publishers.
T
oward a Social Identity Framework for Intergroup Conflict
249


This page intentionally left blank 


Aborigines, 53
abstract communities, 60–61, 62
accentuation principle, 30
acculturation, 110–11
Act of Union (1801; Great Britain;
1801), 137, 144
affect, 72, 74–75, 81
affirmative action, 8, 93, 96
African Americans, 11, 76, 78, 107,
126n.2
disengagement by, 8
ethnic group identification, 86–87,
89, 96, 117, 219
immigration and, 104
oppression of, 103, 125
patriotism and, 11, 83, 84–85, 95,
110, 119–24, 220
policy preference issues, 93–94, 219
pride in America, 83, 84
racism toward, 23
religion and, 85, 219, 222
slavery and, 224, 230–31, 232
agents of violence, intergroup conflict
role, 241
aggression
defensive, 162, 171
intergroup, 26–28, 35
See also harm-doing; violence
Albanians (ethnic), in Yugoslavia, 49
Algeria, 195
Alien Act (1798; U.S.), 104
aliens, resident. See resident-aliens
altruism, 29
Americanization, of immigrants, 76,
105, 221
ancestry
American perspectives, 76
shared, 43, 215
anger, 33, 215
Anglo-Irish Treaty (1921), 137
antagonism
group-based, 65
ideologies of, 170–71, 172, 176
toward outgroups, 10, 17, 26–28, 32,
34, 36
anthropology, 7, 12
of identity, 43–47, 215
political, 62–66
anti-Semitism, 170, 234
apartheid, 44, 194, 195, 232
appraisal, 229–31, 233–36, 245
Arab-Israeli conflict, 188–189
Camp David accord (1978), 189
Palestinianization of, 189–91
251
Index of Subjects


Arab-Israeli conflict (continued)
See also Isreal-Palestinian conflict;
Palestinians
Arab-Israeli war (1967), 189
Arab League, 189
Arab nationalism, 188, 188
Arabs, 133, 155, 188–89, 192, 203. See
also Palestinians
Argentina, 178
Armenia, 172
Armenians, 8, 11, 171–72
Asian Americans, 11, 78, 86–87
Chinese nineteenth-century immi-
grants, 233, 234
ethnic group identification, 117, 219
multiculturalism and, 89
national loyalty and, 102–3
patriotism and, 97, 119–21, 123–24
pride in America, 83
assimilation
American perspectives, 72, 76–77,
82–85, 92, 93, 105, 219
intragroup, 22, 28, 29, 214
of self, 27
See also melting pot
atrocities, 8. See also ethnic cleansing;
genocide; mass killing
attachment
ingroup, 17, 18, 20, 28, 29, 34, 35
psychological, 74, 81, 219
authority orientation, 232
genocide, mass killing, and, 162, 163,
167–68, 173, 177, 180
personal identity and, 241
autonomous identities, 166, 167, 227
avoidance, 32–33
Azerbaijan, 172
Bahai, 179
Balkanization, 72, 76, 97
basic needs, 160–64, 168, 180, 181, 226
Belgium, 170, 227, 246n.1
better world ideologies, 165
Bharatiya Janata Party (BJP), 54, 55
bias, 7
ingroup, 17, 18, 20, 24–27, 34, 36
BJP
. See Bharatiya Janata Party
black Americans. See African Americans
blame, 33
blind patriotism, 75, 167, 178
Bosnia-Herzegovina, 52, 53, 56–59, 237
ethnic cleansing, 101
intergroup conflict in, 11, 49–50, 64,
187, 216
need for healing, 171, 172, 175–76,
177
Bosnians (people), 6, 8, 10, 49–50, 216,
240, 243
British mandate (Palestine), 188
Burundi, 59
bystanders, to violence, 162, 163,
173–74, 179–80, 181, 226, 227,
241
Cambodia, 11, 165
Camp David accord (1978), 189
Canada, 58–59
caring, inclusive, 178, 179
caste system, India, 55–56, 57, 61, 232,
234
categorization. See self-categorization;
social categorization
Catholics. See Roman Catholics
chauvinism
American views on, 81–85, 92, 93, 95,
219, 222
patriotism vs., 74–75, 218
children, values development in, 178
Chinese immigrants, American nine-
teenth-century hostility toward,
233, 234
Christians
African American, 85, 219, 222
anti-Semitism and, 170
in Fiji, 51
in India, 54
in Yugoslavia, 49–50
citizenship, 77
ethnic definitions of, 101
Civil Rights Act (1964; U.S.), 77
civil rights movements
in Northern Ireland, 138
in United States, 105, 220
civil society, 177–78
class, 48, 62
class-based social movements, 63
cognitive balance, 22
cognitive dissonance, 214
252
Index of Subjects


collective identities, 56–61, 66, 144
Israeli-Palestinian conflict and, 191,
194, 200–201
self and, 59, 139
Yugoslavia and, 50
See also group identity; social identity
collectivist cultures, 34, 166
communal differences, national identity
and, 101–25, 220–21
communication, theories of, 214
communism, 49, 165
communities
abstract, 60–61, 62
imagined, 45, 60, 65
comparison
intergroup, 19, 23–26, 27, 32, 214
modernization as trigger for, 57
social, 24, 34, 214
compensatory identities, 168–69
competition, 31, 65
for resources, 24, 28, 56, 60, 217
social, 24–26, 35
confidentiality, workshops and, 198–99,
200
conflict, 7, 8, 12–13. See also intergroup
conflict
conflict resolution
national identity and, 187–211
workshops on, 7, 12, 187, 190,
195–202, 210, 214, 227, 228, 245
conflicts of interest, 28, 34
connected identity, 166–67, 181, 227
conquest, 3, 8, 27
constructive patriotism, 167, 178
constructivism, 44, 47, 53, 62, 66, 215
consumerism, in India, 55
contempt, 18, 19, 32–33, 215
content of identity, 72, 74, 75–76, 164,
218, 219
cooperation, 23, 28–30, 33–34, 35, 36
coping
destructive, 229–30, 232–33, 245
problem– vs. emotion–focused,
231–32
shared, 243
styles of, 232
corruption, intergroup harm-doing and,
234
courage, moral, 178–79
criminalization (British policy), 151–
53
Croatia, 49
Croats, 48–50, 59, 133, 155, 177, 195,
237
Crusades, 11
cultural flows, 45–46
cultural identity, 238–39, 244
culture, 66
ethnicity and, 43–46, 215
Fiji, 52
global, 243
as intergroup conflict factor, 34, 46
Yugoslavian issues, 48–51
See also ethnic groups; genocide; mass
killings
Cyprus, 187, 193, 210
Dalits (Untouchables caste), 55, 57
Declaration of Principles (PLO and Is-
rael; 1993), 190
defensive aggression, 162, 171
democratic societies, genocide and,
177–78
destructive coping, 229–30, 232–33,
245
devaluation of outgroups, 161–63,
170–73, 176, 177, 226
differentiation, 20–22, 29–33, 35, 36,
170
discrimination, 7, 10, 84
apartheid policy, 44, 194, 195, 232
against outgroups, 18, 20, 23, 25,
26–27, 35, 36, 214–15
See also prejudice
disgust, 32
distinctiveness, 34, 36
ethnic, 72
of Israeli-Palestinian conflict issues,
191
optimal, 10, 21–22, 29–30, 32–33,
215, 242, 243
positive, 25, 26, 27
distrust, outgroup, 30–36
diversity
American identity and, 71, 76–77,
82–83, 103–6
as conflict source, 102, 106–7
immigration as source of, 92
Index of Subjects
253


dominance
group, 106–9, 119, 124, 125, 220–21,
234
social, 11, 103, 106–10, 122, 124, 125
Easter Rising (1916; Ireland), 137, 138
economic instability, intergroup harm-
doing and, 163, 234
education, 8, 61
about conflict reduction, 243–44
egalitarianism
in India, 55–56, 57
in United States, 95
Egypt, 188, 189
Egyptian-Israeli peace treaty (1979;
Camp David accord), 189
elites, intergroup conflict role of, 241
embedded identity, 166–67, 168, 178
emotion-focused coping, 231
emotions, 32–33, 215. See also specific
emotions
empathy, 23, 171, 175, 176
enemies
identification of, 161, 162, 170–71
image differentiation, 195–96
engagement of experiences, 175
England. See Great Britain
equality, social, 57
Eritrea, 59–60
essentialism, 44, 45, 46, 65
Ethiopia, 59–60
ethnic cleansing, 4, 11, 27, 33, 49, 65,
101
ethnic conflict, 59–60. See also inter-
group conflict
ethnic grievances, 114, 118
ethnic groups, 3, 7, 45
minority group identification with,
11, 86–87, 89, 95–96, 117, 219,
221–22
multiculturalism and, 77
reciprocal attitudes among, 18–19
in United States, 71, 72, 76, 77–78,
86, 95–96, 103–8, 218–22
See also specific ethnic groups
ethnic identity, 3, 5, 6, 8–11, 72
components of, 238–39
intergroup conflict and, 42–66, 215
multiculturalism and, 77
in United States, 5, 10–11, 78, 86–87,
89, 93–96, 108–9, 116–18, 124,
219–22
See also social identity
ethnicity, 42, 62, 75, 215–16
culture and, 43–46, 215
group narratives and, 237
Marxist view, 48
monolithic identity and, 193
symbolic politics theory and, 78, 219
U.S. perspectives on, 76, 77, 86, 109,
119–22, 218–22
ethnic pride, 72, 77
ethnocentrism, 11, 29, 30, 44
American perspectives on, 109, 110,
117, 122–24, 220
generalized, 112, 113, 117
group orientations, 19–28
hierarchical, 112–13, 122–23
Sumner’s theory of, 17–19, 214
superiority feelings and, 168, 169
European Community, 172
exclusionary patriotism, 107–8, 123–24,
125
exclusivist identities, 192–94, 201, 228
experience. See personal experience
exploitation, 3, 10
Fatah, 189
favoritism
ingroup, 17, 23–26, 34, 72, 75, 78,
214
outgroup, 25
fear, 32–33, 215
Fiji, 6, 10, 42–43, 51–53, 56–58, 60, 61,
63, 64
Fijians, 6, 10, 51–53, 56–58, 64
fission of groups, 64, 65
France, 169, 172, 195
friendship, cross-group, 240, 243
functional theory, 18
fusion of groups, 64, 65, 87
Gaza Strip, 188, 189, 190, 197, 203
GEI. See generalized ethnocentrism
gender, intergroup conflict and, 246
gender-based identity, 5, 166
gender-based social movements, 63
gender relations, in Fiji, 52
254
Index of Subjects


generalized ethnocentrism, 112, 113,
117
General Social Survey (1996), 79–86,
92, 115–16, 126n.2
genocide, 3, 6–8, 10, 11, 213, 226, 231,
245
definition of, 159
ethnic cleansing as, 4, 11, 27, 33, 49,
65, 101
Holocaust as, 11, 172, 173–74, 232
individual and group identities in, 12,
159–81
origins of, 160–63
prevention of, 175–80, 181
Genocide Convention (U.N.), 159
Germany, 8, 169, 234
Holocaust and, 11, 172, 173–74, 
232
Nazism and, 165, 173–74, 179, 232
Weimar Republic and, 178, 232
World War II and, 172, 173–74, 179
Yugoslavia and, 48
global culture, 243
globalization, 43, 66, 71, 231, 234
goals
common, 33–34, 243
superordinate, 9, 33, 243
Government of Ireland Act (1920;
Great Britain), 137
Great Britain
conflict in Northern Ireland and, 12,
134–38, 140–41, 144–45, 149–52,
154, 155–56nn.1–3, 180, 223–26,
242
Fiji and, 51, 52
mandate over Palestine, 155, 188
group-based hierarchies, 107–8, 220
group dominance model, 106–9, 119,
124, 125, 220–21, 234
group identity, 3–5, 22, 74, 216
changes to, 194, 201–2, 210–11, 227,
228
genocide, mass killing, and, 159–65,
170–71, 178, 180–81, 226
in social identity framework, 230,
235, 236–37
See also ingroups; outgroups; social
identity
group integration, 63, 65
group narratives, 210, 228, 235, 236–37,
241, 246
Israeli and Palestinian, 191–92, 202–9
See also myths, national
groups
integration into, 63, 65
ostracism from, 179
self-concepts of, 164, 168–70, 171,
174, 180, 226
GSS. See General Social Survey
Guatemala, 159
Gulf War, 190
Guyana, 53
Gypsies, 49, 173
harm, infliction of, 26–27, 214, 215, 
226
harm-doing, intergroup, 3, 10, 229–36,
245
hate, 17, 18, 25, 27, 32–34, 36, 37n.1,
214–15
healing, victimization and, 162, 163,
171–72, 175–77
HEI. See hierarchical ethnocentrism
hierarchical ethnocentrism, 112–13,
122–23
hierarchies, group-based, 107–8, 220
Hindus, 6
in Fiji, 51
in India, 10, 54–61, 65
hindutva movement, 10, 54–58, 61, 65
Hispanic Americans, 11, 78, 125
ethnic group identification, 86–87,
89, 117, 219
national loyalty and, 102–3
patriotism and, 83–84, 85, 97, 109,
111, 119–21, 123–24
pride in America, 83, 84
See also Mexican Americans
history, role of, 217, 223, 227, 245,
246n.1
Holocaust, 11, 172, 173–74, 232
Jewish Palestinian homeland and, 188
Home Rule, Ireland and, 137, 149
hostility, outgroup, 17–19, 27, 28,
33–36, 37n.1, 75, 78, 97, 214, 216
human welfare, caring about, 178–79
Hungarians, in Yugoslavia, 49
hunger strikes, 151–52, 224
Index of Subjects
255


Hutus, 59, 133, 155, 161–62, 170–71,
174, 234, 246n.1
identification, 238
ingroup, 17–36, 214–15
with vs. as, 72, 73–74, 218
identity, 12–13, 245
anthropology of, 43–47, 215
autonomous, 166, 167, 227
collective. See collective identity
compensatory, 168–69
connected, 166–67, 181, 227
definitions of, 5
embedded, 166–67, 168, 178
ethnic. See ethnic identity
exclusivist, 192–94, 201, 228
gender-based, 5, 166
genocide, mass killing, and, 160,
161–81, 213
group-based, 3, 4, 5, 139. See also
group identity
individual. See self
insecure, 167
as intergroup conflict factor, 3, 4, 7–9
monolithic, 192–94, 201, 228
national. See national identity
negation of, 192–93, 196, 201, 210
negotiation of, 196–97, 200–202,
207–10
racial, 11, 238–39
role, 135, 138–40, 144, 146–48,
151–54, 156n.6, 224–25, 237, 240,
245
secure, 167, 181
social. See social identity
social construction of, 194–95
superordinate, 242–43
in threat-coping process, 232
transcendent, 193, 201–2, 210–11,
243
identity politics, 42–47, 53, 58–66, 215,
243
in United States, 4, 94–95, 97, 219
ideological asymmetry, 124–25, 126n.4
ideologies
of antagonism, 170–71, 172, 176
of better world, 165
genocide, mass killing, and, 161–62,
165–66, 170–71, 173–74, 178, 181
monolithic identity and, 193
nationalistic, 165
imagined communities, 45, 60, 65
immigration, 231
to Palestine, 188
United States and, 8, 71, 76–77, 92,
96, 97, 104–5, 110, 233
Immigration and Nationality Act (1965;
U.S.), 77
inclusion, need for, 21–22, 29–30, 32,
33, 35, 36
inclusive caring, 178, 179
India, 6, 42–43, 60, 63
caste system, 55–56, 57, 61, 232, 234
hindutva movement, 10, 54–58, 61,
65
Muslim population, 4, 54–58, 60, 61
nationalist movement, 74
Indians (Fiji people), 6, 10, 51–53,
56–58, 64
individual-group discontinuity effect,
31
individual identity. See individual social
identity; self
individualism, U.S. perspectives, 109,
111, 126n.1
individualist cultures, 34, 166
individual social identity, 230, 235, 236,
237–40
ingroup positivity, 18–19, 20–23, 32, 36,
214
ingroups, 10, 17–36, 214–16
comparison and competition issues,
23–26, 34
ethnocentrism and, 17–19, 29, 214
favoritism toward, 17, 23–26, 34, 72,
75, 78, 214
formation of, 17, 19, 28–35
optimal distinctiveness theory, 21–22,
29–30, 32–33, 215
outgroup antagonism by, 26–28, 36
positivity toward, 18–19, 20–23, 32,
36, 214
scapegoating by, 161, 162, 164, 165,
180
social categorization and, 20, 214
injured social identities, 12
injustice, 57–58, 161
insecure identities, 167
256
Index of Subjects


instigators of violence, 160–63, 173,
180, 181, 226, 230
Institute for Social Research (University
of Michigan), 114
Institute for Social Science Research
(UCLA), 79
instrumentalism, 44, 45, 46–47, 53, 62
integration, group, 63, 65
interactive problem solving, 198
interdependence, 33, 34
cooperative, 35, 36
obligatory, 29
intergroup comparison, 19, 23–26, 27,
32, 214
intergroup conflict
definition of, 6–7
diversity as source of, 102
ethnic and national identity as issues
in, 42–66, 215, 222
gender and, 246
identity’s role in, 3, 4, 7–9
ingroup identification and, 17–36,
214–15
Israeli-Palestinian, 12, 187–211,
227–28
in multiethnic states, 101–2
multilevel analysis of, 7, 9
in Northern Ireland, 133–55, 223–26
realistic group conflict view of, 8,
216–17, 233
reduction of, 5, 7, 9, 12, 36, 242–44
self ’s role in, 3, 4
social identity and, 4, 5, 7, 8–12,
213–46
social identity framework for, 213,
229–42, 245
as violence instigator, 160–61, 180,
226
intergroup contrast, 22, 214
intergroup harm-doing, social identity
framework and, 229–36, 245
intergroup relations, 6, 11, 17, 18,
35–36
conflict-reduction strategies, 243–44
social categorization’s role in, 19–20
social identification and, 22–23
intermarriage, 57
in Fiji, 52
in India, 56
in United States, 105
in Yugoslavia, 50
international conflicts, 3, 187, 198. See
also war; specific conflicts and wars
interventions, as conflict-reduction
strategy, 243, 244
intifada, 189, 191, 201
intolerance, 32
intragroup assimilation, 22, 28, 29, 
214
invasion, feelings of, 33
IRA. See Irish Republican Army
Iran, 179
Ireland, 136–38, 149, 155–56nn.1, 3,
223, 224, 242
home rule, 137–38
See also Northern Ireland
Irish Free State, 137–38
Irish Nationalists, 135, 137, 138, 141,
142, 144, 152, 224
Irish National Liberation Army, 155n.2
Irish Republican Army, 12, 133,
137–38, 140, 145, 151, 152,
155nn.1, 2, 223, 224
Irish Republicans, 134–37, 140–42,
144–49, 151–53, 223–26, 237
paramilitaries, 11, 133–35, 137–38,
140, 148–49, 151, 154,
155–56nn.2, 3
self-narratives, 241
Islam. See Muslims
Israel, 133, 179, 187–211, 227–29, 233
Camp David accord (1978), 189
creation of, 188, 189
Declaration of Principles (1993), 
190
national identity concerns, 191–97,
200–201, 207–9, 214
overview of Palestinian conflict,
188–91
problem-solving workshop participa-
tion, 7, 12, 198–202, 210
Israeli-Palestinian conflict, 188–91
Arab states and, 188–89
Palestinian refugee issue, 188–89,
192, 202–9, 228–29
Palestinian uprising, 189, 191, 201
peace moves, 189–90
See also Arab-Israeli conflict
Index of Subjects
257


Jews, 8, 177, 179
anti-Semitism and, 170, 234
Holocaust and, 11, 172, 173–74, 232
in United States, 11, 104
Zionism and, 188, 192, 196, 202, 204
See also Israel
Joint Working Group on Israeli-Pales-
tinian Relations, 200, 202, 205–7
Jordan, 188, 203
justice, 161
Palestinian sense of, 204, 207, 208,
209
Khmer Rouge, 165
kinship
ethnicity vs., 57
identity politics and, 58, 60, 61, 64
Yugoslavian issues, 48, 50, 51
Korean War, 166
Kosovars, 8
Kosovo, 8–9, 48, 57, 58, 101, 155, 242
LACSS. See Los Angeles County Social
Surveys
language
conflict over, 8
in Fiji, 51, 52, 53
in India, 54
monolithic identity and, 193
in United States, 72, 82
in Yugoslavia, 48
Latino National Political Survey, 83
Latinos. See Hispanic Americans; Mexi-
can Americans
leaders, role of, 175, 177, 235, 237, 242
Lebanon, 203
liberalism, 76, 77, 78, 95, 218, 221
liberalization, in India, 55
life conditions, 160, 161, 163, 166, 168,
180, 226
literary studies, postcolonial movement
and, 44
Los Angeles County Social Surveys
(1994, 1995), 79, 83, 86–87, 221
Los Angeles County Social Surveys
(1997, 1998), 113–14
love
of country, 74–75
ingroup, 17, 18, 34, 214–15
Loyalists, Irish, 141, 142–44, 225
loyalty
identity politics and, 58, 64
ingroup, 28, 29–30, 33, 34
national, 71, 75, 101–3
pluralism and, 106
Macedonia, 49
Madrid Conference (1991), 190
Mandal commission (India), 56
Maoris, 53
marriage. See intermarriage
Marshall plan, 172
Marxism, 48
mass killing, 3, 6–8, 10, 11, 231
definition of, 159
individual and group identities in, 12,
159–81, 226, 245
origins of, 160–63
prevention of, 175–80, 181
See also genocide
material threat, 233–34
Mauritius, 53
media, 55, 61, 236
melting pot, America as, 11, 92, 103,
104–9, 116–19, 122, 124, 220, 221
Mexican Americans, 83–84, 109–11,
125
minimal group paradigm, 18
minority groups, 8
ethnic identifications by, 11, 86–87,
89, 95–96, 117, 219, 221–22
multiculturalism and, 77, 82
state-sanctioned violence against, 3,
11
subtle racism toward, 23
in United States, 71, 72, 77, 82–87,
89, 95, 96, 107, 219–22
See also specific minority groups
misattribution, 32, 215
modernity, 44–45, 47, 65, 66
modernization, 45, 234, 240
as conflict factor, 10, 56–57, 231
in Fiji, 53
monolithic identity, 192–94, 201, 228
monolithic societies, 162, 163, 177
Montenegro, 49
moral courage, 178–79
moral superiority, 10, 32–33, 34
258
Index of Subjects


Mughals, 55, 57
multiculturalism, 42, 44
United States and, 71, 76–78, 81–83,
89, 92–93, 96, 103, 218
multiethnic states, 75, 76, 78, 87, 218
communal and national identity in,
101–25
multiple voices, 44, 45–46
murder. See genocide; mass killing
Muslims
in Fiji, 51
in India, 4, 6, 54–58, 60, 61
in Yugoslavia, 49, 50, 52, 58, 59, 237
See also Arabs
myths, national, 6, 10
identity politics and, 58, 61
Yugoslavia and, 50, 216, 237
See also group narratives
narratives. See group narratives; self-
narratives
National Election Study (1992; 1996),
114–15
national identity, 3, 5, 6, 8–11, 245
American, 71–97, 101–25, 218–22
changes in, 187, 194–96, 201–2,
210–11, 228
definition of, 191
intergroup conflict and, 42–66
Israeli-Palestinian conflict issues and,
12, 187–211, 214, 227–28
negotiation of, 196–97, 200–202,
207–10
Northern Ireland hostilities and, 136
See also myths, national; social iden-
tity
nationalism, 10, 42, 44–46, 65, 115
Anderson’s (Benedict) view, 60
Arab, 188
civic vs. ethnic, 75
Indian hindutva movement and, 54,
55
as integrating ideology, 87, 89
patriotism vs., 74–75
Quebecois, 58–59
nationalistic ideologies, 165
nationalist movements, 8, 74
Nationalists, Irish, 135, 137, 138, 141,
142, 144, 152, 224
nationality, 48, 75
U.S. perspectives on, 76, 77, 86, 95,
105
National Opinion Research Center, 79
nations, 3, 7
definition of, 74
ethnic groups and, 44–45
loyalty to, 71, 75, 101–3
minority groups in. See minority
groups
See also specific countries
Native Americans, 76
nativism, 76–77, 82–85, 92, 218, 219,
221
NATO (North Atlantic Treaty Organi-
zation), 242
Nazism, 165, 173–74, 179, 232
Ndembu, 58
needs, basic, 160–64, 168, 180, 181, 226
negation of identity, 192–93, 196, 201,
210
negotiation of identity, 196–97,
200–202, 207–10
NES. See National Election Study
New Deal, 175, 231–32
Northern Ireland, 7, 133–55, 187, 193,
210, 213, 242, 245
activist accounts concerning, 140–51,
225–26
bystanders to violence in, 179–80
historical background of hostilities in,
136–38
hunger strikes in, 151–52, 224
political violence in, 133–55, 155n.1,
223–26
social and role identities in, 11–12,
135, 138–40, 143–44, 146–48,
151–54, 156nn.5, 6, 223–26, 237
Norway, 58
Nuer, 61, 63
Nuremberg trials, 172
obligatory interdependence, 29
one-time workshops, 198–200
oppression
of African Americans, 103, 125
institutionalized, 6
optimal distinctiveness, 10, 21–22,
29–30, 32–33, 215, 242, 243
Index of Subjects
259


“ordinary” people, intergroup conflict
role of, 241
Oslo agreement (1993), 190, 197, 201
ostracism, 179
outgroups, 10, 17–36, 214–16
American perspectives on, 73
antagonism toward, 10, 17, 26–28,
32, 34, 36
devaluation of, 161–63, 170–73, 176,
177, 226
distrust of, 30–36
ethnocentrism and, 18, 19, 214
favoritism toward, 25
hatred of, 17, 18, 25, 27, 32–33, 34,
36, 37n.1
hostility toward, 17–19, 27, 28,
33–36, 37n.1, 75, 78, 97, 214, 216
intergroup comparison and, 23–26,
27, 32
scapegoating of, 161, 162, 164, 165,
180
social categorization and, 20, 214
subtle racism toward, 23
PA. See Palestinian Authority
Pakistan, 55
Palestine, 155, 188–90, 192, 197, 203,
208–9
Palestine Liberation Organization, 189,
190, 197
Palestinian Authority, 190
Palestinians, 179, 187–211, 227–29, 233
Declaration of Principles (1993), 190
national identity concerns, 191–97,
200–201, 207–9, 214
Oslo agreement (1993), 190, 197, 201
overview of Israeli conflict with,
188–91
problem-solving workshop participa-
tion, 7, 12, 198–202, 210
refugee issue, 188–89, 192, 200,
202–9, 228–29
See also Arab-Israeli conflict; Israeli-
Palestinian conflict
passivity toward victims, 173–74, 179
patriotism, 115
blind, 75, 167, 178
chauvinism vs., 74–75, 218
constructive, 167, 178
exclusionary, 107–8, 123–24, 125
in United States, 10–11, 78, 81,
82–85, 92, 95, 97, 109–11, 119–24,
126nn.2–4, 219–22
Peace Now movement (Israel), 196
personal experience, 47
identity politics and, 58–61, 65, 66
personal identity. See self
PLO. See Palestine Liberation Organi-
zation
pluralism
in American society, 11, 103, 105–9,
118–19, 122, 124, 220, 221
genocide and, 177
pogroms, 27
policy issues, U.S., 73, 78, 87–96, 219,
221
immigration as, 76–77, 96, 104, 105
impact of racism on, 77
political anthropology, 62–66
political groups, 3, 72, 139, 159
political instability
genocide and mass killing origins
and, 163
intergroup harm-doing and, 234
political science, 7, 12
political violence, in Northern Ireland,
133–55, 155n.1, 223–26
politicide, 159
politicization, 34
politics
American ethnic group concerns
about, 117–18
of identity, 4, 42–47, 53, 58–66,
94–95, 97, 215, 219, 243
of power, 34
positive distinctiveness, 25, 26, 27
positive self-concept, 169
postcolonialism, 44
power politics, 34
power structure, 10
prejudice, 7, 8, 10, 221, 243
against outgroups, 17, 23, 25, 36,
37n.1
See also discrimination
pride
in America, 10, 78, 79, 81, 83, 84, 95,
219
ethnic, 72, 77
260
Index of Subjects


primary appraisal, 234–35
primordialism, 44, 45, 76, 215, 237
prisoners of war, 166
problem-focused coping, 231–32
problem-solving workshops, 7, 12, 187,
190, 195–202, 210, 214, 227, 228,
245
Protestants (in Northern Ireland),
133–38, 140, 141, 151, 153, 154,
155–56n.3, 223–26
Loyalists, 141, 142–44, 225
paramilitaries, 11, 12, 133–35, 137,
138, 140–44, 149, 151–54,
155–56nn.1–3, 224–26
Unionists, 137, 141–42
Protestants (in Rwanda), 178
Provisional Irish Republican Army. See
Irish Republican Army
Provisional Sinn Féin. See Sinn Féin
psychology, 12
punishment, 175
Quebecois nationalism, 58–59
race, policy preferences and, 93–95, 96,
219
racial identity, 11, 238–39
racism, 23, 77
Rashtriya Swayamsevak Sangh, 54
realistic group conflict, 8, 216–17, 233
reconciliation
identity change as factor in, 194,
210–11
need for, 171, 172, 175–77, 181
refugees, Palestinian, 188–89, 192, 200,
202–9, 228–29
reification, 45, 46
relational identities, 166
religious groups, 7
in Fiji, 51, 52
in India, 10, 54–56, 57, 60
monolithic identity and, 193
Northern Ireland hostilities and,
11–12, 133–38, 140–44, 149, 151,
153, 154, 155–56nn.3, 4, 223–25
in United States, 71, 72, 104, 219,
222
in Yugoslavia, 48, 49–50
See also specific religious groups
Republicanism, 136–37, 140. See also
Irish Republicans
resident-aliens, in United States, 11,
119–22
resources
competition over, 24, 28, 56, 60, 217
as Israeli-Palestinian conflict issues,
187, 191–92, 201
revenge, 57–58, 176
RGC. See realistic group conflict
rhetoric, political, 58
right of return, Palestinian refugees
and, 203–9
RIR. See Royal Irish Regiment
ritual, 58, 60
role identity, 135, 138–40, 144, 146–48,
151–54, 156n.6, 224–25, 237, 240,
245
Roman Catholics
in Northern Ireland, 133, 135–38,
140–44, 153, 154, 155–56nn.3, 4,
180, 223, 224
in Rwanda, 178
in United States, 104
Romans, 11
Royal Irish Regiment, 135
Royal Ulster Constabulary, 134–35,
137, 151–53, 155n.2
RSS. See Rashtriya Swayamsevak Sangh
RUC. See Royal Ulster Constabulary
Rwanda, 11, 59, 101, 174, 223, 234,
246n.1
church leaders’ actions in, 178
conflict instigators, 161–62
ideologies of antagonism and, 170–71
need for healing in, 171–72, 175–77
world community and, 242
See also Hutus; Tutsis
Sami, 58
scapegoating, 33, 161, 162, 164, 165,
180
SDLP
. See Social, Democratic, and
Labour Party
secondary appraisal, 234–35
secure identities, 167, 181
security forces (Northern Ireland),
133–35, 138, 140, 151, 152, 154,
155n.2
Index of Subjects
261


segmentary organization, 64, 65
self, 3, 4, 215–16, 245–46
collective identity and, 59, 139
definitions of, 5
genocide, mass killing, and, 12, 159,
160, 162–67, 169, 171, 174, 176,
179, 180–81, 213, 226
identity politics and, 47, 65
intergroup comparisons and, 23–24
monolithic identity and, 193
in optimal distinctiveness theory, 22,
30, 32
role identities and, 139–40
social identification and, 21, 87, 238
in social identity framework, 230,
235, 236, 237–41
in threat-coping process, 232
self-categorization, 3, 10, 20, 222
Americans and, 81, 86–87, 219, 221
social identity and, 72, 73–74, 139,
218, 238
self-concept, group, 164, 168–70, 171,
174, 180, 226
self-esteem, 21, 22, 25, 72, 169
self-evaluation, 24, 25
self-narratives, 238, 241, 246
self-worth, 21, 26, 176
Serbia, 49
Serbs, 6, 59, 61–62, 223, 226
Bosnia and, 10, 177, 216, 243
Croats and, 133, 155, 195
group narrative by, 58, 216, 237
Kosovo and, 8–9, 57, 58
role identities of, 240
Yugoslavia and, 48–50
shared coping, 243
Simmel’s Rule, 63, 65
Sinhalese-T
amil conflict, 58
Sinn Féin, 137, 138, 140, 152, 226
situational identification, 45–46
Six-Day War (1967), 189
slavery, 3, 8, 10
in Fiji, 51
in United States, 224, 230–31, 232
Slavs, 49–50
Slovenes, 48–49
Slovenia, 48, 49
social categorization, 10, 19–20, 29, 33,
35, 214
social change, intergroup harm-doing
and, 234
social class. See class
social comparison, 24, 34, 214
social competition, 24–26, 35
Social, Democratic, and Labour Party,
143
social dominance theory, 11, 103,
106–10, 122, 124, 125
social equality, 57
social identification, ingroup positivity
and, 18–19, 20–23, 32, 36, 214
social identity, 3–13, 47, 213–46, 228
Americans’ ethnicity and, 116–18,
124
class and, 48
communal differences and, 102,
108–9
definition of, 5–6, 21, 138–39
dimensions of, 73–74, 218, 219, 238
genocide, mass killing, and, 160, 162,
164–65, 171, 178–79
injured, 12
intergroup conflict framework and,
213, 229–42, 245
intergroup conflict reduction and, 5,
7, 9, 12, 36, 242–44
intergroup conflict role of, 3, 4, 7–9
multilevel analysis of, 7, 9
multiple aspects of, 178, 240
Northern Ireland hostilities and,
11–12, 138–40, 143–44, 146, 148,
154, 156n.5, 223–26
optimal distinctiveness and, 10,
21–22, 29–30, 32–33, 215, 242, 
243
social categorization’s role in, 19–
20
types of, 71
as violence factor, 5, 11–12, 133, 135,
154, 160, 162
See also group identity
social identity theory, 3, 10, 21, 27, 164
U.S. national identity and, 72–75,
77–78, 87, 97, 124
socialization, 31, 78, 84, 86, 218, 236
social movements
class-based, 63
gender-based, 63
262
Index of Subjects


social psychology, 7, 10, 17
conflict-reduction mechanisms,
242–43
social representations, 236, 237
social structure, 34–35
social welfare. See human welfare
societally projected images, 236, 237
sociology, 7, 12
Somalia, 59
South Africa, 187, 223
apartheid policy, 44, 194, 195, 232
Sri Lanka, 187, 210
status of groups, 24, 25, 35
stereotypes, 7, 8, 10, 223–24, 232, 243
categorization and, 20
outgroup, 30–31, 215
Sudan, 59
suffering
acknowledgment of, 171, 175
desensitization to, 173–74
images of past, 57–58, 61
superiority
beliefs of, 168–69
moral, 10, 32–33, 34
superordinate category identity, 36
superordinate goals, 9, 33, 243
superordinate identities, 242–43
symbolic politics theory, 73, 78–79, 84,
87, 95, 218, 219, 233, 234
symbolic threat, 233–34
symbolism, political, 58
Syria, 189, 203
T
aig (term), 142, 156n.4
T
amils, 58
territory, Israeli-Palestinian conflict
over, 187, 191–93, 196–97, 201,
202
third-party interventions, 243, 244. See
also problem-solving workshops
Thirty Years’ War (1618–1648), 169
threat, 28, 32–34, 35, 36
feelings of, 162, 163, 165, 171, 209,
237
material vs. symbolic, 233–34
in social identity framework, 229–36,
245
threat-coping frameworks, 229, 231–32
tolerance, 77, 95, 177
traditionalism, 55
transcendent identity, 193, 201–2,
210–11, 243
tribal societies, 64
Trinidad and T
obago, 53
trust, 31, 36, 171, 175, 176
ingroup, 10, 28–30, 34, 35, 215
outgroup distrust, 30–36
Turkey, 171–72
Turks, 8, 9, 11, 57, 172
Tutsis, 59, 133, 155, 161–62, 170–71,
178, 223, 234, 246n.1
Ulster Defence Association (UDA),
138, 142, 143
Ulster Defence Regiment (UDR), 135,
153
Ulster Democratic Party (UDP), 153
Ulster Freedom Fighters (UFF), 138,
142, 143, 152, 155n.2
Ulsterization (British policy), 151, 
152
Ulster Volunteer Force (UVF), 137,
138, 142, 143, 152, 155n.2
U.N. See United Nations
Unionists, Irish, 137, 141–42
United Irishmen, 136–37
United Kingdom. See Great Britain
United Nations, 179
Genocide Convention, 159
Palestine partition plan, 188
United Nations Relief and Works
Agency, 203
United States, 5–8, 10–11, 155, 217–22,
240, 245
American identity conceptions in,
75–78, 81, 84, 92, 218–19
Americanization of immigrants in,
76, 105, 221
chauvinistic views in, 81–85, 92, 93,
95, 219, 222
communal and national identity in,
101–25, 220–21
communal conflict in, 102, 106–8
historical perspectives on ethnic di-
versity in, 103–6, 221
identity politics in, 4, 94–95, 97, 219
immigration issues in, 8, 71, 76–77,
92, 96, 97, 104–5, 110, 233
Index of Subjects
263


United States (continued)
multiculturalism in, 71, 76–78,
81–83, 89, 92–93, 96, 103, 218
national identity’s meaning in, 71–97,
218–20
Northern Ireland conflict and, 224,
242
patriotism in, 10–11, 78, 81, 82–85,
92, 95, 97, 109–11, 119–24,
126nn.2–4, 219–22
policy issues in, 73, 77, 78, 87–96,
104, 105, 219, 221
pride in country in, 10, 78, 79, 81, 83,
84, 95, 219
slavery in, 224, 230–31, 232
unit relationships, 22
universal stereotypes, 30–31, 215
University of California, Los Angeles
(UCLA), Institute for Science 
Social Research, 79
University of Michigan, Institute for
Social Research, 114
UNRWA. See United Nations Relief
and Works Agency
Untouchables (Indian caste), 55, 57
UVF. See Ulster Volunteer Force
values, genocide, mass killing, and, 162,
177, 178–79, 180
victimization, 162, 163, 171–72,
175–77, 226
Vietnam, 165, 172
violence, 8, 216
agents of, 241
bystanders to, 162, 163, 173–74,
179–80, 181, 226, 227, 241
instigators of, 160–63, 173, 180, 181,
226, 230
in Northern Ireland, 133–55, 155n.1,
223–26, 245
as revenge for past, 57–58
social identity’s role in, 5, 11–12, 133,
135, 154, 160, 162
state-sanctioned, 3, 11
See also aggression; ethnic cleansing;
genocide; harm-doing; mass
killing; war
vulnerability, sense of, 163, 168, 172,
176, 226
war, 3, 6, 10, 27
in Ireland, 136, 137, 223
in Middle East, 188–89
in Rwanda, 161
Sumner’s theory of, 18
in Yugoslavia, 47–51
See also international conflicts; specific
wars
weakness, sense of, 168
Weimar Republic, 178, 232
West Bank (Middle East), 188, 189,
190, 196, 197, 203
women
identity and, 166
in United States, 76
women’s movement, 220
workshops, problem-solving, 7, 12, 187,
190, 195–202, 210, 214, 227, 228,
245
World War I, 169, 188
World War II, 10, 172, 173–74, 177,
179. See also Holocaust
Yugoslavia, 42–43, 47–52, 57, 61, 63,
77, 237
ethnic cleansing policy, 4, 49, 101
world community and, 242
See also Bosnia-Herzegovina; Kosovo;
Serbs
Zambia, 58
zero-sum perspective, 194, 210
discriminatory behavior and, 26
in Israeli-Palestinian conflict, 12,
187, 192, 228
social competition and, 24–25
Zionism, 188, 192, 196, 202, 204
264
Index of Subjects


Aboud, F. E., 243
Abrams, D., 21
Aiken, L. S., 123
Alba, R. D., 86, 105, 111
Alexander, king of Yugoslavia, 48
Allport, G. W., 7, 20, 177
Alpher, J., 200, 202, 203, 205, 206, 207
Anastasio, P. A., 9, 23
Anderson, Benedict, 45, 60, 62, 194
Andrews, Irene, 143
Appiah, K. A., 77
Arafat, Yasser, 189, 197
Aretxaga, B., 224
Arthur, M., 150
Ashe, Thomas, 151, 152
Ashmore, Richard D., 3–14, 71, 213–49
Bachman, B. A., 9, 23
Baker, S. M., 238
Banaji, M. R., 21
Banker, B. S., 97, 243
Barber, B., 71
Baron, R. A., 23
Bar-tal, D., 75, 115, 164
Barth, F., 43, 46, 62
Bauman, Zygmunt, 42, 66
Baumeister, R. F., 5, 168
Beike, D., 23–24
Bell, J. B., 138, 151, 152
Bettencourt, B. A., 26
Bigelow, K. R., 179
Bigler, R. S., 243
Billig, M., 18
Blackwell, M. S., 73, 139, 246
Blanz, M., 27
Blumer, H., 107
Boskin, J. S., 224
Boulding, K. E., 216
Bourhis, R., 25
Brewer, Marilynn B., 5, 7, 10, 17–41,
75, 87, 170, 213, 214–17, 229,
233–34, 235, 237, 238, 241, 242,
243, 244–45
Brimelow, P., 76
Bringa, T., 50
Brown, R., 17, 18, 25
Bruce, Steve, 135, 141, 142, 152, 153
Bundy, R., 18
Burton, John, 198
Cairns, E., 133, 180
Calderwood, Alex, 143
Campbell, D. T., 18, 25, 30, 31
Caporael, L. R., 29
265
Index of Names


Chapman, M., 62
Charles I, king of England, 149
Cheleden, S., 77
Chirot, D., 3, 71
Churchill, Winston, 149–50
Citrin, Jack, 5, 6–7, 10, 71–100, 109,
213, 218–22, 233, 234, 243, 244,
245
Clifton, A. K., 238
Cohen, Abner, 44, 62
Cohen, A. P., 47
Connor, W., 74, 75, 87
Contrada, R. J., 220, 232
Cornell, S., 50, 51
Craig, G. A., 169
Crocker, J., 21
Cromwell, Oliver, 136, 149
Cronin, S., 136, 137, 144
Cross, W. E., 239
Cusack, J., 153
Dahl, R. A., 106
Danziger, K., 5
Darby, J., 180
Davis, J., 79
Deaux, K., 237
de la Garza, R. O., 72, 83, 87, 103, 106,
109, 110, 111, 221, 238, 239
Del Boca, F. K., 8, 216, 233
Deschamps, J-C., 20
des Forges, A., 162, 171, 174, 178
Devine, P. G., 238
de Weerd, M., 139
Diehl, M., 18
Dietz, M., 74
Doise, W., 20
Dorr, N., 26
Dovidio, J. F., 9, 22, 23, 97, 243
Drake, C. J. M., 135
Duff, Brian, 5, 10, 71–100, 213, 218
Dull, V., 238
Eidheim, H., 58
Elliott, S., 138, 151, 152
Epstein, A. L., 58
Epstein, S., 174
Erickson, G., 21
Eriksen, Thomas Hylland, 5, 6, 7, 10,
42–68, 171, 213, 214–17, 223, 226,
227, 231, 234, 235, 237, 238–39,
240, 241, 242, 243, 244–45
Evans-Pritchard, E. E., 61, 63, 64
Falcon, A., 72, 83, 103, 221
Fang, C. Y., 125
Fanon, Frantz, 44
Farley, R., 71
Farrell, M., 152
Featherstone, M., 43
Federico, C., 125
Feeney, B., 133
Feierabend, I. K., 115
Fein, H., 165
Feshbach, S., 74, 75, 81, 103, 112, 115
Festinger, L., 24, 214
Fincham, F., 26–27
Fine, G. A., 233, 236
Finlay, K., 243, 244
Fisher, R. J., 242, 244
Fiske, S. T., 31
Flackes, W. D., 138, 151, 152
Flament, C., 18
Fogelman, E., 174
Folkman, S., 229
Fordham, S., 8
Ford, T. E., 23
Fortes, M., 64
Fraser, A., 136, 149
Fraser, M. R., 140, 144
Frederico, C. M., 78, 165
Fredrickson, G. M., 221
Freedman, S. G., 105
Frey, D., 23
Fried, A., 74
Friedman, J., 42
Froystad, K., 61
Funk, C. L., 28, 95
Furnivall, J. S., 50
Gaertner, S. L., 9, 23, 97, 242–43
Gandara, P., 243
Gandhi, Rajiv, 54
Garcia, C., 72, 83, 103, 221
Gaughan, Michael, 152
Geertz, Clifford, 44
Gellner, Ernest, 44–45, 46, 61, 62
Genesee, F., 243
George V, king of England, 149
266
Index of Names


Gergen, K. J., 236, 246
Gergen, M. M., 236, 246
Gerner, D. J., 188
Gerstle, G., 77
Giddens, A., 42, 47
Glazer, N., 77, 97
Gleason, P., 76
Gluckman, Max, 63
Goldberg, D. T., 77
Goldhagen, D., 3, 234, 241
Gordon, Milton, 77, 105
Goring, Herman, 173
Gourevich, P., 167, 174, 178
Greeley, A. M., 105
Green, D. P., 81
Greenfeld, L., 3, 71, 75
Greenwald, A. G., 21
Grey, Edward, 150
Gubin, A., 177
Gurin, G., 74
Gurin, P., 74, 243
Gurr, T. R., 3, 101, 159
Gurtman, M. B., 22
Haas, E. B., 72
Hagengimana, A., 177
Halpern, B., 188
Handelman, D., 58
Handler, R., 59
Hanna, B. A., 30
Hannerz, U., 43
Hanning, P., 27
Hansen, T. B., 56
Harff, B., 101, 159
Harkabi, Yehoshafat, 196
Hartmann, D., 50, 51
Hartz, L., 78
Haslam, S. A., 7, 139
Heider, F., 22
Heppen, Jessica, 213–49
Herman, J., 171, 172, 175
Hertzberg, A., 188
Hewstone, M., 21, 26–27
Higham, J., 76
Hilberg, R., 173
Himmler, Heinrich, 173
Hinkle, S., 18, 25
Hirt, E. R., 21
Hitler, Adolf, 55, 58, 165, 232, 234
Hobsbawm, E., 62
Hochschild, J. L., 84
Hofstetter, C. R., 115, 125
Hogg, M. A., 20, 21, 73, 139, 246
Holland, D., 4, 5, 46, 47
Hollinger, D. A., 71, 77, 89
Horney, Karen, 215
Horowitz, D. L., 56, 78, 97, 101
Hull, C. L., 214
Huntington, S., 48, 53
Huntington, S. P., 76
Hyndman, M., 145
Ignatieff, Michael, 48, 65
Insko, C. A., 31
Jackman, M., 107
James II, king of England, 136
Janoff-Bulman, R., 171
Jarymowitz, M., 170
Jaspars, J., 26–27
Jenkins, R., 47
Johnson, G., 23
Jones, J., 3, 236
Jones, James, 96
Jussim, Lee, 3–14, 71, 213–49
Kaistaniemi, L., 169
Kapferer, B., 58
Kateb, G., 77
Kaukiasnen, A., 169
Keegan, J., 3
Kelley, H. H., 32
Kelley, R., 104
Kelly, J. D., 51
Kelman, Herbert C., 3, 5, 6, 7, 12, 75,
170, 177, 187–212, 213–14,
227–29, 233, 237, 241, 242, 244,
245, 246
Kelters, S., 133
Kennedy, C., 21
Keough, K. A., 171, 175, 176
Khalidi, R., 188
Kincaid, E., 166
Kinder, D. R., 83, 84
Klandermans, B., 139
Kleppner, P. J., 104
Klicperova-Baker, M., 115
Konner, M., 174
Index of Names
267


Kosterman, R., 74, 75, 81, 112, 115
Kramer, R. M., 30
Kreisberg, L., 3
Kressel, N. J., 167
Laclau, E., 65
Lagerspetz, K. M. J., 169
Lash, S., 43
Lau, R., 78
Lavine, H., 167
Lazarus, R. S., 229
LeVine, R. A., 30
Levin, S., 26, 78, 103, 125, 165
Levy, S. R., 243
Lewin, Kurt, 217
Lifton, R. J., 173
Lijphart, A., 101, 102, 103
Lind, E. A., 89
Lind, M., 76
Lipset, S. M., 78, 103
Lui, L., 238
Lyengar, S. S., 227
Lyons, F. S. L., 137, 149
Major, B., 21
Malanchuk, O., 74
Malle, B. F., 107
Manchester, William, 149–50
Manzi, J., 25, 26
Marcia, 239
Martin, P., 82
Maudling, Reginald, 149
McAdams, R. C., 236, 246
McCall, G. J., 139, 225
McCann, L. I., 171, 172, 175
McCaughey, Sean, 152
McClintock, C., 26
McDonald, H., 153
McDonald, M., 62
McGarry, J., 136
McGarty, C., 139
McGrath, D., 238
McKee, Billy, 151, 152, 153
McKittrick, D., 133, 134
McMichael, Gary, 153
McMichael, John, 153
Meertens, R. W., 23
Mendelsohn, E., 188
Merelman, R. M., 82
Meyer, G., 20
Miller, A. H., 74
Miller, D., 72, 74, 77, 89
Miller, D. T., 87
Morris, B., 203
Moxon-Browne, Edward, 141–42, 144
Moynihan, D. P., 77, 97
Mullen, B., 25
Mummendey, A., 25–26, 27
Muslih, M., 188, 197
Muste, C., 72, 77
Myers, D. G., 6, 164
Myrdal, G., 96
Neidenthal, P. M., 23–24
Nelson, S., 142
Netanyahu, Benjamin, 190
Nier, J. A., 97, 243
Norenzayan, A., 227
Northrup, T. A., 3, 233, 242
Novak, M., 105
Oakes, P. J., 7, 20, 73, 139, 246
O’Connell, A., 27
Ogbu, J. U., 8
Ohbuchi, K., 89
Okin, S. M., 77
O’Leary, B., 136
Oliner, D., 174, 179
Oliner, S. B., 174, 179
O’Malley, P., 150
Opotow, S., 173
Orwell, George, 75
Otten, S., 26, 27
Parenti, M., 105
Pearlman, L. A., 171, 172, 174, 175,
176, 177
Pennebaker, J. W., 171, 175, 176
Perdue, C. W., 22
Petrocik, John R., 5, 6–7, 11, 101–29,
165, 213, 220–22, 234, 242, 245
Pettigrew, T. F., 23, 177, 240, 243
Platow, M., 27
Polsby, N. W., 106
Portes, A., 82
Powell, C. L., 106
Powell, G. B., Jr., 102
Pratto, F., 103, 107, 125
268
Index of Names


Premdas, R., 51
Prentice, D. A., 87
Prentice-Dunn, S., 23
Prunier, G., 162, 171
Rabin, Yitzhak, 190, 197
Rabinowitz, J. L., 78, 125, 165
Ranger, T., 62
Raz, J., 72, 77
Reese, H., 170
Reich, R. B., 71
Reicher, S., 20, 73, 139, 246
Reingold, B., 72, 81
Rex, J., 44
Roediger, D. R., 86
Rogers, R. W., 23
Rokkan, S., 103
Roosens, E. E., 45
Roosevelt, Franklin, 175
Rose, Richard, 141, 144
Rothman, J., 237, 242, 244
Rouhana, Nadim, 200
Rubin, M., 21
Rummel, R. J., 177
Ruscher, J., 31
Rust, M. C., 9, 23
Sachdev, I., 25
Sadat, Anwar, 189
Said, Edward, 44
Salins, P. D., 105
Salmivalli, C., 169
Sampson, E. E., 166
Sanders, L. M., 83
Sands, Bobby, 152
Savarkar, Veer, 55
Schaar, J. H., 74, 75
Schatz, R., 167
Schlesinger, A. M., Jr., 44, 72, 77, 106
Schöpflin, G., 48
Schopler, J., 31
Schreiber, H., 25–26
Schwartz, S. H., 26
Sears, D. O., 28, 73, 77, 78, 79, 83, 84,
87, 95, 218, 233
Segal, J. M., 201
Sellers, R. M., 238, 239
Shave, R., 27
Shaw, J. S., 25, 26
Sherif, M., 7, 9, 33, 243
Shikaki, K., 200, 202, 203, 205, 206,
207
Shore, C., 62
Sidanius, Jim, 5, 6–7, 11, 25, 26, 78, 85,
101–29, 165, 213, 220–22, 234,
242, 245
Simmel, Georg, 63
Simmons, J. L., 139, 225
Simon, B., 25
Skerry, Peter, 125
Smith, Anthony D., 44, 45, 62
Smith, C., 25
Smith, E. R., 32
Smith, N. S., 167
Smith, P., 21
Smith, Rogers, 76
Smith, T., 79
Smith, William, 143
Soeters, J. L., 177
Stahelski, A. J., 32
Stallworth, L. M., 107
Stangor, C., 23
Staub, Ervin, 3, 5, 7, 12, 74, 75, 81, 115,
159–84, 213, 223, 226, 227, 228,
231, 232, 234, 235, 241, 242, 244,
245
Stephan, C., 32
Stephan, W. [G.], 32, 243, 244
Stevenson, J., 142, 143
Strathern, M., 43
Streich, G., 82
Struch, N., 26
Stryker, S., 139, 225
Sugawara, I., 89
Sullivan, L. A., 23
Sullivan, J., 74
Sumner, W. G., 17–19, 214
Sutton, M., 134, 155n.1
T
ajfel, H., 3, 4, 5, 7, 10, 17, 18, 20, 21,
26, 31, 72, 74, 87, 117, 124, 164,
218, 246
T
akagi, E., 29
T
ambiah, S. J., 47
T
amir, 87
T
aylor, C., 74
T
aylor, Peter, 153
T
ec, N., 174, 179
Index of Names
269


T
esser, A., 25
T
essler, M., 188
Thatcher, M., 135
Thoits, P. A., 3, 4, 72, 73, 139, 140, 224,
238
Thompson, L., 233, 236
Thornton, C., 133
Thorson, S. J., 3
Tilly, C., 155n.1
Tinker, H., 51
Tito, Josip Broz, 48, 237
T
onkin, E., 62
Triandis, H. C., 34, 166
Turner, J. C., 3, 7, 10, 17, 18, 20, 21, 24,
72, 73, 117, 124, 139, 164, 246
Turner, P
., 21
Turner, V., 58, 60
Turton, D., 47
T
yler, R. B., 22
T
yler, T., 89
Urry, J., 43
van der Veer, P., 55
van Laar, C., 77
Virshup, L. K., 3, 4, 72, 73, 139, 140,
224, 238
Walch, T., 105
Ward, C. M., 97, 243
Warner, R. S., 71
Waters, M., 86
Weber, J. G., 21, 23
Weber, Max, 75
West, S. G., 123
Wetherell, M., 20
White, John, 142–43
White, Robert W., 5, 7, 11–12, 133–
58, 180, 213, 223, 224–27, 
230, 235, 237, 240, 241, 242, 
245
White, T. F., 135, 155–56n.3
Whyte, J., 133, 134
Wick, B., 238
Wilder, David, 3–14, 213–46
Wilkes, 20
William of Orange, 136
Williksen-Bakker, S., 52
Wilson, Paddy, 143
Wolfe, A., 72
Wolfinger, R. E., 105
Wong, Cara, 5, 10, 71–100, 213, 218
Wood, J. V., 24
Worchel, S., 231, 235, 236, 241, 243,
244
Young, I. M., 72, 77
Zaller, J., 79, 87
Zillman, D., 21
270
Index of Names