Professional Documents
Culture Documents
In a recent paper, Taylor & Brown have argued that, although research in social psychology needs to take
into account the social context of social behaviour, the theories should aim at the explanation of individual
behaviour. The present paper argues against this view as it applies to some important issues in social
psychology. First, it is contended that the ‘individualistic’bias of research in social psychology derives from
the nature of the theories dominating the discipline. Second, a theory of inter-group behaviour is briefly
outlined in order to show that its structure and aims are different in some important ways from the
individualistic theories. The bias of these theories which is due to their assumption that social behaviour
takes place inside homogeneous and unstructured social systems is illustrated using as examples the theory
of belief similarity in prejudice and equity theory.
I find much to agree with in the paper by Taylor & Brown (1979) (which will be referred to as
T & B from now on). Their argument goes in three stages. First, they agree with the criticisms
some of us have made of the near-monopoly of the ‘individual’ and ‘inter-individual’ research in
social psychology. This led to the neglect of a broader perspective in which social behaviour
would be seen as a ‘dynamic interaction ’ between psychological processes and the nature of the
social systems inside which this behaviour takes place. T & B provide a summary of some of
these earlier criticisms: (i) too much research on individual and inter-individual behaviour rather
than groups; (ii) ‘naive extrapolations from individual to the groups’; (iii) divorce of the studies
of interpersonal phenomena from their ‘wider societal context ’.
The second stage of the T & B argument is that these criticisms are justified with respect to
social psychological research, but that the theories will have to continue dealing with individuals
in order to remain within the domain of the fundamental interests and preoccupations of
psychology. The third stage of the argument consists of attempting to show, using as examples
two theories formulated by the ‘critics’ (Moscovici, 1976; Tajfel, 1974, 1978; Tajfel & Turner,
1979): (i) that these theories remain essentially ‘individual’; (ii) that their value lies precisely in
the fact that they have remained ‘individual’ while managing to contextualize social behaviour in
its interaction with group phenomena; and (iii) that the sooner this is openly acknowledged, the
easier it will be for these theoretical developments, which T & B consider important, to
influence beneficially the ‘mainstream ’ of social psychology.
In other words, T & B state that social behaviour originates from, and pertains to, individuals.
Whatever ‘non-individual ’ variables may affect it (such as ‘groups ’, ‘social context ’, etc.), useful
social psychological theories must remain at this individual level. The two recent theories
discussed by T & B have done precisely this, despite their protestations to the contrary. Like
M. Jourdain we were speaking prose, but we didn’t know it.
My agreement with T & B stops at the end of the first sentence of the above summary
paragraph. The statement about social behaviour originating from, and being performed by,
individuals is entirely unexceptionable and trivially true. But immediately afterwards the
difficulties begin. The bourgeois-gentilhomme may have been speaking prose without knowing it,
but Molikre never informed us whether the result was good prose or bad prose. To say that in
the last analysis we are concerned with individual social behaviour is as true as it is meaningless
until and unless some useful and interesting statements are made about the characteristics of this
* This paper started offas a brief rejoinder to the article by Taylor & Brown. As the work proceeded, it
became clear that the issues raised by Taylor & Brown were too important to be dismissed or discussed in a
few summary paragraphs. The ‘rejoinder’ became longer.
OOO7-1293/79/0602-0183 $02.00/0 @ 1979 The British Psychological Society
184 Henri Tajfel
behaviour and the kind of theoretical approach which will be needed to understand these
characteristics.
inter-group behaviour which was discussed by T & B in their paper, although it would be idle to
pretend that they were not much hazier in the early stages than they are now in hindsight.
The first of the above three questions requires an answer in two stages. We need, first of all, a
definition of a ‘group’ which refers to the way the notion is constructed by those inside the
system. This definition must enable us to make the transition from what we assume is
‘constructed ’ by the individuals involved to data showing whether our assumptions were correct
or incorrect. The notion of social categorization discussed by T & B is undoubtedly and crucially
implied in this process of definition. If we did not know about the human capacity to categorize
the environment in certain ways, we could not even begin to make the assumption that
collections or heaps of individuals in a social system can be cognitively organized into a complex
matrix of overlapping categories. But this by itself tells us nothing about the nature of the
categorizations and their uses or effects in social behaviour.
The second stage of the answer to the question about the ways groups are ‘constructed’
concerns the fact that a theory of inter-group behaviour has for its aim the explanation (or a
better understanding) of certain uniformities of inter-group behaviour. It is therefore necessary
to state what are the basic conditions for ‘groups’ to be constructed in such a way that the
consequent behaviour of members of one group towards another shows uniformities rather than
a random variation from individual to individual. The theory makes two testable and
interdependent propositions about these conditions. They consist of stating that certain
uniformities of inter-group behaviour will appear if (i) members of a group believe that there is
no possibility for them (or at least that there are considerable difficulties) of moving from one
group to another; and that (ii), consequently or in interaction with the former condition, the
boundaries between the groups are fairly sharply drawn. One of the results of the conjunction of
these two conditions is that it may powerfully determine the course of interpersonal interaction
independently of the individual characteristics of the interactants or of the detail of their past or
present personal relations (cf. Tajfel, 1978, ch. 2, for a detailed discussion both of the definition
of ‘groups ’ and of the differences between inter-individual and inter-group behaviour).
The second question above concerned the psychological effects of the above conditions. This
is the aspect of the theory briefly summarized by T & B and presumably considered by them to
be more fundamental than the others. It is the sequence from social categorization to social
identity and social comparison (see Tajfel, 1974; Turner, 1975; Tajfel & Turner, 1979; Tajfel,
1978, ch. 3). There is no doubt that T & B are quite correct in stating that the focus of these
psychological effects is in individuals. It could hardly be anywhere else. What is more, as I
wrote earlier, there are some important similarities in these processes as they appear in
individual, inter-individual or inter-group behaviour. There are also some important differences
which have been discussed elsewhere.
The third question formulated earlier concerned the relationships between, on the one hand,
social inter-group constructions and their effects, as just discussed, and, on the other hand,
various forms of social reality. Providing a set of testable answers to it is quite fundamental if
the theory is to have any predictive value at all. There are several reasons for this requirement
being fundamental. All the theoretical considerations summarized so far would not be capable of
making clear-cut and testable distinctions between a multigroup social system which retains its
stability and one which is in the process of change, attempted change or resisted change in the
relations between the groups of which it is composed. The fact that individuals in a social
system perceive the system as having sharply drawn and fairly impassable group boundaries
does not, by itself, lead to the conclusion that they will engage in common action either to
change it or to preserve it. The fact that they are capable of making social comparisons which
are relevant to their self-image does not, by itself, mean that these comparisons will result from
the perceived position of their group as relating to other groups rather than remaining at an
inter-individual level (cf. Turner, 1975, 1978, for a more detailed discussion). The construction of
Individuals and groups in social psychology 187
a social system in terms of sharply drawn social categories and the capacity to categorize and
compare oneself with others in certain ways and for certain purposes are the necessary
conditions for the appearance of certain forms of inter-group behaviour; they are not suficient
conditions.
The translation from potentiality to actual social behaviour must be sought elsewhere. ‘Social
reality ’ can be described or analysed in terms of socio-economic, historical or political
structures. Such descriptions or analyses are not within the competence of the social
psychologist. But he can ascertain that, for whatever reasons, the system of the relations
between social groups is perceived by the individuals located in the various parts of the system
as being capable or incapable of change, as being based on legitimate or illegitimate principles
of social organization. He can also ascertain whether, as a result, group actions are being
considered or undertaken in common by those who feel that their location (and ‘social identity’
as defined by T & B and elsewhere) in the system is capable or incapable of change, ‘secure’ or
‘insecure ’. A combination of these shared interpretations of social reality with the location of
social groups within the system as perceived by their members provides the possibility of
formulating a number of hypotheses. These hypotheses and some of the research deriving from
them have been described elsewhere (cf. ch. 4 and most of the following chapters in Tajfel,
1978).
It will now be clear, I hope, why I find it difficult to agree with T & B who, after having
summarized in their paper the sequence of social categorization - social identity - social
comparison, and described it as containing the ‘basic postulates’ of the theory, relegate all the
rest to the single following sentence: ‘It is from these basic postulates that a variety of important
concepts emerge including social mobility, social change, cognitive alternatives, legitimacy and
stability’. It is the use of this truncated version of a more complex structure which enables them
to write immediately below that their aim was ‘to point out that the focus of the theory is
ultimately the individual ’, although ‘group processes are of course fundamental to it ’. The focus
of the theory is not the individual but the explanation of uniformities of inter-group behaviour.
No one would deny that ‘ultimately’ we are concerned with ‘individuals ’ who behave in one
way or another. But a clear distinction must be made between theories which are ‘individualistic’
and one which is concerned with socially shared patterns of individual behaviour. An
‘individualistic’ theory contains the (most often) unstated assumption that individuals live and
behave in a homogeneous social medium. This medium consists of a collection of undifferentiated
individual particles which are assumed to relate to each other inter-individually following the
laws of ‘basic’ psychological processes. There is no room in this vision of randomly floating
particles for the cognitive and socially shared organization of the system within which the
particles float. Or if it is admitted that the lines along which the system is structured, both
‘objectively’ and ‘subjectively’, have a great deal to do with the social behaviour of its
individual elements, this is considered to be no more than a set of ‘variables ’ superimposed on
something more ‘fundamental’. It has been the contention of this paper so far that this kind of
approach can get us nowhere very far in understanding those crucial uniformities of social
behaviour which pertain to the psychological aspects of the social systems in which we live, i.e.
it will get us no nearer to an adequate social psychology of social conflict, social stability, social
change, social movements or social unrest. T & B wrote that ‘the individual should and will
remain the ultimate target of understanding’. My view is that if this were really to be the case,
then social psychology ‘should and will remain’ as incapable as it has been until now of
providing any new psychological insights into some of the most important aspects of our
functioning in society. In the next section of this paper, I hope to support this statement using as
examples two theories which were also briefly mentioned by T & B in their paper.
188 Henri Tajjel
processes. And even if many of us who wish to ignore them are entirely free to do so, we do not
have the right to imply through the conclusions we draw from our work that our cosy and
equitable inter-individuality can reach beyond the blinkered vision of social reality which we
have selected for our special consideration. As T & B wrote, the<ipdividual is the target of
understanding for social psychology; but he cannot remain the ody target.
References
BERKOWITZ, L. & WALSTER,E. (eds) (1976). Equity Monographs in Social Psychology, vol. 2. London:
Theory: Toward a General Theory of Social Academic Press.
Interaction. Advances in Experimental Social TAJFEL,H. (1972b). La catCgorisation sociale. In S.
Psychology, vol. 9. New York: Academic Press. Moscovici (ed.), Introduction 6 la psychologie
BILLIG,M. & TAJFEL,H. (1973). Social sociale, vol. 1 . Paris: Larousse.
categorization and similarity in intergroup TAJFEL,H. (1974). Social identity and intergroup
behaviour. European Journal of Social behaviour. Social Science Information, 13, 65-93.
Psychology, 3, 27-51. TAIFEL, H. (ed.) (1978). Differentiation Between
CADDICK, B. (1977). The sources of perceived Social Groups: Studies in the Social Psychology of
illegitimacy in intergroup behaviour. University of Intergroup Relations. European Monographs in
Bristol, Mimeo. Social Psychology, vol. 14. London: Academic
DOISE,W. & SINCLAIR, A. (1973). The Press.
categorization process in intergroup relations. TAJFEL,H. & TURNER, J. C. (1979). An
European Journal of Social Psychology, 3, integrative theory of social conflict. In W. Austin
145-1 57. & S. Worchel (eds), The Social Psychology of
EISER,J. R. & STROEBE,W. (1972). Categorization Intergroup Relations. California: Brooks/Cole.
and Social Judgment. European Monographs in TAYLOR,D. M. & BROWN,R. J. (1979). Towards a
Social Psychology, vol. 3, London: Academic more social social psychology? British Journal of
Press. Social and Clinical Psychology, 18, 173-180.
FESTINGER, L. (1954). A theory of social TAYLOR,D. M. & GUIMOND, S. (1978). The belief
comparison processes. Human Relations, 7, theory of prejudice in an intergroup context.
117-140. Journal of Social Psychology, 105, 11-25.
MOSCOVICI,S. (1976). Social Influence and Social TURNER, J . C. (1975). Social comparison and social
Change. European Monographs in Social identity: Some prospects for intergroup behaviour.
Psychology, vol. 10. London: Academic Press. European Journal of Social Psychology, 5, 5-34.
ROKEACH, M. (1968). Beliefs, Attitudes and Values: TURNER, J. C. (1978). Social categorization and
A Theory of Organization and Change. San social discrimination in the minimal group
Francisco: Jossey-Bass. paradigm. In H. Tajfel (ed.), Differentiation
TAJFEL,H. (1959). Quantitative judgment in social Between Social Groups: Studies in the Social
perception. British Journal of Psychology, 50, Psychology of Intergroup Relations. London:
16-29. Academic Press.
TAJFEL, H. (1969). Cognitive aspects of prejudice. WALSTER,E., BERSCHEID, E. & WALSTER,G . W.
Journal of Social Issues, 25 (4), 79-98. (1976). New directions in equity research. In L.
TAJFEL,H. (1972~).Experiments in a vacuum., In Berkowitz & E. Walster (eds), Equity Theory:
J. Israel & H. Tajfel (eds), The Context of Social Toward a General Theory of Social Interaction.
Psychology: A Critical Assessment. European New York: Academic Press.
Requests for reprints should be addressed to Henri Tajfel, Department of Psychology, University of Bristol,
8-10 Berkeley Square, Bristol BS8 IHH, England.