Philosophical Reflections on the Identity Discourse of Social Psychology

by Thomas Wren

(Forthcoming in Personal and Moral Identity, edited by B. Musschunga, Amsterdam: Kluwer, 2001 or 2002)

A recent discussion of identity in the context of multicultural education in the United States opens with the following sentence: "To speak of education in the 1990s is inevitably to speak of identity" (Hoffman, 1998, p. 324). This reading of the current literature is, I think, quite accurate, since what was once a purely philosophical issue and then a clinical one has now become a (if not the) major social and educational issue in North America and I think Europe, as well as anywhere else in the world where people are self-consciously trying to clarify who they are. To be sure, philosophers continue to examine the concept of identity, and the present volume is only one of many attempts to develop what might be called a logical geography of identity. But it seems reasonable to ask whether philosophical analysis has anything to say about such seemingly non-philosophical questions as those posed by two multicultural theorists who recently asked:

What is the effect of the corporate commodification of black culture? How does such commodification shape and reshape black students' identities? White students' identities? How do such corporate and media depictions reconfigure the struggle for social justice? What are the openings they create for positive resistance to oppression? How do teachers incorporate these issues into the curriculum and lived world of the classroom for positive effect?... (Kincheloe and Steinbert, 1997, p. 105)

I do not intend to address these questions directly in this essay, since my concern is the second-order one of how those who ask (and try to answer) them are using the word "identity" as a category of human experience. As the passage just quoted illustrates, there is a consensus among those who now write on the subject that social identity, which following current American practice I will usually call "cultural identity," is forged in a social and political context. This context includes -- but is considerably broader than — classical anthropological notions of culture as a worldview or shared way of life. (1) However, identity is also a matter of profound importance for the psychological development and well-being of individual agents. Identity theorists sometimes emphasize the social/cultural side of identity, sometimes the psychological side, but seldom distinguish between them. And when they do, the distinction is often flawed, as I will try to show by discussing certain developments in social psychology.

The concept of identity has a long and rather untidy philosophical history that antedates its role in current discussions by multicultural educators, anthropologists, psychologists, and scholars of cultural studies. For this reason, I will begin with some general philosophical remarks about how we talk about personal identity, followed by two (very!) rapid reviews of ways that this concept has been constructed, first in the western philosophical tradition and then in the literature of contemporary social psychology. In both cases, I believe there is a two-way street between scholarly theory and social practice, but I will not pursue that Gramscian theme in these pages. Suffice it to say that most of us reflect on our personal life by using "folk theories" of identity that have been shaped in important respects by these intellectual traditions.

Identity Discourse

According to the anti-essentialist view of personal identity — which is the prevailing view among social theorists as well as the perspective from which I have written this essay — there is no determinate, de facto "hidden self"that, like a jack-in-the-box, suddenly reveals itself when conditions are right. Identities are narratives, not things. Like other narratives, identities are fashioned in discourse, and hence "in specific historical and institutional sites within specific discursive formations and practices, by specific enunciative strategies" (Hall, 1996, p. 4). However, there do seem to be general features of current identity discourse that constitute a rudimentary kind of natural logic without invoking essences that are "always already there" as fixtures of the universe. Philosophers and others interested in conceptual clarity can, I think, analyze this logic without resorting to the postmodern jargon deployed by those who, like Stuart Hall in the passage just cited, want to critique and destabilize this discourse. For instance, when we say of a person that he or she has a certain identity, we are sometimes trying to give a psychological description, but more often are picking that person out of the general population and assigning him or her to some sub-population such as the population of male teenagers or unemployed harpsichord players. And when we say the person has a certain cultural identity, we often mean to define the sub-population in cultural terms, using that adjective narrowly and anthropologically (Inuits, Maori, Swedes, etc.) or broadly and sociologically (ethnic Americans, gays, the deaf, etc.). (2) But sometimes we are trying to do something quite different. Instead of assigning the person to a certain population, we might be trying to say how he or she feels about the group in question. In simple terms, this sort of identification is an attempt to spell out the subjective relationship the person has with the group. What is important here is not membership per se, but endorsement, solidarity, affection, or some other mode of consciousness. The former sort of identification is a matter of demographics, the latter a matter of internalized affiliation.

Finally, just as groups can be identified without referring to their members, so also individuals can be identified without referring to the groups they happen to belong to. At the simplest extreme, I can just point at you while saying something like "That person over there" or, in direct address, "You there!" This is what some philosophers call an indexical reference, for the same reason that we call the digit used for pointing our "index finger": such statements single things out by indicating their relationship to the speaker. They are not based on generic categories that would have the same meaning no matter who uttered them. The three-word sentence "Socrates is dead" is true no matter who says it (unless it's Socrates, of course). But the truth value of those other three little words "I love you" depends utterly on who is talking (and to whom).

Indexical identifications have their place in the general discourse about individual identity, but not a central place. Much more common are identifications that purport to describe how individuals are related to themselves, using terms taken from either folk psychology or more systematic theories of personality. In either case, these psychological terms are themselves a mix of ordinary language and special theoretical categories. Accordingly, when we try to clarify their use we often find that etymological considerations have heuristic value. For instance, the word "identity" is derived from the Latin idem or "same." Thus the second definition of "identity" in the Oxford English Dictionary is: "the condition or fact of a person or thing being that specified unique person or thing." Like other dictionary definitions, this one goes on to emphasize the static, almost Parmenidean character of identity, declaring that the aforesaid condition or fact is "a continuous unchanging property." For this reason psychologists who speak of a person's identity in developmental terms must be seen as stretching the classical meaning of the term, though of course that is just the way living languages work. However, not all psychologists who write about personality are developmentalists, just as not all are Freudians or social learning theorists. For many personal identity theorists, the unit of analysis is a stable personality trait such as the disposition to use a single schema (self-concept) to think of oneself. In their theories neither the trait nor the schema is conceptualized as fluid or dynamic even though these theorists allow that people do change their self-concepts and other self-ascribed properties, especially when they seem markedly discrepant with the views other people have of them.

Telling the Story

With these considerations we move from logic to the story of how the concept of personal identity has been constructed in the western intellectual tradition. (3) The plot synopsis is as follows. For centuries, personal reality was understood — i.e., socially constructed — in terms of relatively static traits, which moralists called virtues. It was the task of children and young adults to acquire virtues, typically by acting as though they had already acquired them, so that eventually these traits became what Aristotle called one's "second nature." The classic conceptualizations of a person's or thing's nature as something stable, unchanging, and so on were reminiscent of the etymology of the word "identity," though in those times the word itself was seldom used in that connection. The phenomenon of personal growth was not denied — no one doubted that children needed to be educated — but its teleology was understood organically, in the sense that the outcome of a growth process is determined by the fixed essence or internal structure of the organism. This view of development eschewed novelty: for classical thinkers there was a right way for an acorn to grow, and also a right way for a person to grow. If as sometimes happened the growth process went wrong, the outcome was unnatural. In an oxymoron tolerated by every classical philosopher and educator, a vicious nature was considered an "unnatural nature," never a false or literally inauthentic one.

The classical construction of human nature and personal identity continued into the middle ages and early renaissance, when it was supplemented by the belief that everyone had his or her divinely appointed place in the feudal or post-feudal social order. Identity was a self-evident matter of nature-plus-position, and authenticity consisted in being true to one's life station. The model of identity as what we now call personal authenticity began to emerge at some point in the history of modern European thought — some would say in the 16th century with Martin Luther, whereas others would say the 17th century with G. Vico, the 18th century with J.-J. Rousseau, the 19th century with G.W.F. Hegel, or even with John Dewey at the beginning of the 20th century. Regardless of who gets the credit for opening the discussion, what happened was that the old notion of person as social niche-holder loosened up, so that one's person or selfhood was distinguished from one's position in society. More precisely: hierarchical roles became less and less definitive of personal identity as political movements challenged their legitimacy, as Rousseau did when he proposed that aristocratic titles be replaced with the non-preferential title citoyen — which wasn't really a title at all, strictly speaking, since every citizen possessed it (see Taylor, 1994). The climax of this political challenge to the preeminence of social roles was, of course, the French Revolution, but other liberation movements before and after it are important parts of the story too. During this time the concept of personal authenticity became increasingly individuated. First it was a kind of shortcut for discovering the same set of common truths and life goals that had been represented in the texts of classical philosophers and, within the Christian tradition, in scripture and the works of the church fathers, theologians, councils, and so on. For instance, the Protestant notion of individual conscience was presented as a better, surer way of discovering the Divine Plan, not a way of inventing one's own alternative to that Plan. However, once Romanticism became established, personal authenticity was understood as a mode of constructing meaning rather than of discovering it. And that is where we seem to be today.

Over the last half century or more, the search for personal identity has been understood as "finding one's way," "discovering who I am," or — the formula I find least metaphorical — "deciding what is most important" (important for oneself, that is). Implicit in these and similar expressions is the idea that identity is something to be won and, correlatively, that the attempt is not always successful. Since Freud, the child's struggle has usually been understood as gaining control over impersonal forces constituting the Id, whereas the adult's struggle was originally understood as the emergent Ego's attempt to overcome personal but nonetheless alien forces such as the Superego, False Consciousness, Internalized Stereotypes, and so on. Psychologists have sometimes discussed identity in terms of the first sort of struggle, sometimes in terms of the second. Thus we read in Freudian and neo-Freudian literature that the Ego (the seat of personal identity) has its own developmental pattern and even its own drive-like "executive functions," such as mastery strivings or various tendencies toward self-regulation. Similar models were produced by non-Freudians such as Jean Piaget, Neal Miller, and Walter Mischel (see Wren, 1990). Eventually the discourse of executive functions gave way to the discourse of self-interpretation, as represented in the work of Erik Erikson, for whom identity formation was much more closely tied to social milestones (e.g., entering the work force) than it was for the earlier Freudians. Subsequent work on identity formation by non-Freudians and post-Freudians such as object relations theorist Nancy Chodorow (1978) has moved the focus from social milestones to social structures of domination, with the result that the psychological processes of identity formation are now regularly featured in discussions of feminism, gay studies, racial consciousness, and other multicultural issues. However, this new focus has called into question the essentialist language that seemed so natural in earlier theories that regarded identity as a stable, and stabilizing, property of the self.

The Unsubstantial Self (Person): The Philosophical Story

Regardless of what one might think about the complex philosophical debates over methodological individualism (e.g., whether concepts of collectives are anything more than the sum of features found in the concepts of their individual members), it seems difficult if not impossible to discuss the concept of identity without eventually invoking the terms "self' or "person." (4) However, identity and self/person are separate albeit related concepts, and have separate albeit intertwined histories, the first of which we just considered. Let us now turn to the second.

Once again etymologies have heuristic value. As noted above, the term "identity" is derived from the Latin word for same. In contrast, "person" is derived from the Latin word for something that is eminently changeable, namely a theatrical mask (persona). Etymologists speculate that the latter term was coined by the Romans because they knew that in the large amphitheaters of ancient Greece the actors' lines were sounded through — per + sona -- little megaphones that were part of the outsized masks they wore as part of the dramatic spectacle. Whether or not this explanation is historically accurate, the fact remains that a mask is a mask, something to be put on and taken off, an impermanent thing that derives its intelligibility from its role as a representation of something besides itself. Indeed, a mask need not even be a thing in the sense of a solid substance, as can be seen in sentences such as "His courtesy was a mask for his villainous intentions." In this usage, the word "mask" is like "dance" or "song," all of which refer to actions that can be verbalized as nouns but are not by that simple linguistic maneuver converted into real things, essences, or pieces of furniture in the universe. Unfortunately for the present discussion, we have no such expression as "He has been personing all day," though we do use "impersonating" as a verb.

The theme of non-substantiality is also found in the etymology of the word "self," though here as before we must remember that such considerations are only heuristic devices, not philosophical arguments or historical evidence for drawing conclusions about the way the world is. The origin of "self' and its various European cognates seems to be unknown, but their earliest known uses have to do with strength and weakness, depending on the inflection. Thus the Old English word self meant "strong," and selfa meant "weak." It was used to emphasize a pronoun as in "She did it herself," and a bit later to suggest the idea of something being one's "own" or, more oddly still, the idea of something being "peculiar." However, in Middle English it took on the reflexive use it still has today, which led to the modern substantival use of "self" as an inner reality or essence, roughly equivalent to the Platonic-Christian notion of an immaterial soul.

What do these etymological considerations suggest, beyond the bare-bones proposition that the self is something other than a substance or entity? If we go back to the image of the actor's mask, we notice that the point of wearing a mask is to communicate. We also notice that masks are used to perform, and that the various meanings which an actor's mask takes on in the course of a dramatic performance are all fundamentally relational. In a word, it would make no sense to wear a mask unless one were in a social situation.

In western philosophy (5) the substantial self does not appear as a distinct theme until the beginning of the so-called modern era, when René Descartes (1641/1986) uttered his famous Cogito ergo sum. For him, to say "I think" was to say "I am" right on the spot, since he thought that each person "is in the strict sense only a thing that thinks" (ibid., p. 18). Here Descartes is doing much more than simply emphasizing the cognitive dimension of human life: he is claiming that our life simply is our thought. To say we are thinking things is to define our essence; to speak in what he called "the strict sense." Furthermore, the thinking activity Descartes had in mind is utterly solitary. The Cartesian self — i.e., the "thing that thinks" — is a substance that has no access to reality other than through its own mental representations (ideas), which it nonetheless hopes mirror that reality. The hope is warranted, Descartes claimed, by a line of reasoning that begins with the Cogito and involves his famous thought experiments of an Evil Genius, our innate Idea of God, and so on. Whatever one might think of the logical quality of his much-criticized argument, it was an astonishingly bold attempt to demonstrate the existence (or better, the conceivability of such an existence) of an inherently asocial entity at the core of human consciousness.

Descartes' idea of the self as an immaterial substance set the agenda for the next three centuries of philosophical debate about how individuals come to have knowledge about the world and, correlatively, about themselves. Throughout this long and still-continuing debate, the image of an isolated knower has remained intact, in spite of spirited dissent from Descartes' original view of the self's knowability (Locke), immateriality (Berkeley, Hume), or logical status (Kant). As the 21st century begins, few thinkers are willing to be characterized as having a Cartesian conception of the self, but most if not all find it necessary to recall it as a way of laying out their own positions. Some differentiate their views by stressing the embodied character of the self, but most of today's writers about the self stress its social character. Virtually every contributor to the contemporary literature on social or cultural identity subscribes to the view that personal identity is socially constructed. But what does this mean?

The idea that our selves are socially constructed appears in at least three general modalities of theoretical discourse, which I will describe very briefly here. The first modality is discourse about how our identities are shaped. Nearly everyone, scholar or person in the street, recognizes that he or she would have been a much different person if the circumstances of birth, childhood, economics, and other social conditions, etc., had been different. The assumption here is that our psychological states and structures are formed in the course of social interaction, though it is usually accompanied by another assumption to the effect that we were not totally passive in the formative process — i.e., that it was indeed an inter-action.

This commonsense view is compatible with a Cartesian conception of self, since it is logically possible that social influences on our thought and behavior are completely external to the self, as when long exposure to the sun's light changes the color of an old painting or even that of our own skin. But there is no such compatibility when the discourse moves to the second modality, which concerns not how we think and behave but how our very capacities to think and behave have been shaped by our social world. Examples of such discourse are the arguments of many feminists that our sexual capacities are themselves socially constructed, as well as the gender-specific stereotypes which have influenced the way we exercise those capacities. The philosophers David Bakhurst and Christine Sypnowich have made the same point, concluding that "in all these arguments, the social serves, not to mold an antecedently present self or identity, but to bring that self into existence" (1995, p. 6).

The third modality of discourse about the self is still further removed from the Cartesian view. Here the issue under discussion is not how the "capacities" of the self are shaped by social interactions but rather how the very conception of selfhood emerges from such interaction. Perhaps "the conception of selfhood" is a bit misleading, since this rather high-sounding phrase suggests the sort of discourse that takes place at philosophy conferences or in the faculty lounge of a university psychology department, whereas I have more everyday contexts in mind. However, in many (most?) cultures there is a continuity between technical and lay discourse about the self, and usually a mutuality between them, as ethnolinguists have shown (see Shweder, 1991, Shweder & LeVine, 1984). For instance, a western man need not have read Freud to think of his sexuality in terms of drives, energy, pressures, and so on. And a woman from India need not have an academic degree in order to think of her sexuality in the way charted by the Hindu scriptures and transmitted through her culture. I speak here of "culture" rather than "social interaction" (Mead, 1934), in order to avoid the impression that basic concepts such as selfhood emerge from one-on-one exchanges such as a conversation between friends or even from the more intimate relationships between children and parents. Such exchanges can be important, to be sure, but they are themselves conditioned by the larger cultural reality. Furthermore, it is hard not to think of the social interaction episodically, by which I mean as a sequence of discrete events.

Before we leave the topic of selfhood and its social construction, two more philosophical distinctions are in order. The first is simply that social constructions are not arbitrary decisions. There are many constraints, some imposed by history and economics, some by psychology and biology. As A. Musschenga (private communication) has reminded me, you cannot build a skyscraper from wood. The second qualification is equally general. Philosophers and social theorists who prefer the social constructivist perspective over essentialism must be careful not to oversimplify the opposition between these two philosophical alternatives, if only to avoid making the rather embarrassing mistake of essentializing social constructivism. (To see how easy it is to make this mistake, see Fuss, 1989.) Of the many thinkers who insist on the social character of the self, some are much closer to the essentialist view than others are. If essentialism can be described as the general belief that there is a certain way that things "are," then philosophers like Alasdair MacIntyre or psychologists like Lev Vygotsky would have to be counted as essentialists in spite of their sensitivity to the social, historically conditioned, and generally elusive quality of the self. Their accounts of the social formation of the self are offered in an attempt to tell their readers "how things are," which is to say to give readers an insight into what must eventually be named, pace Aristotle, as human nature. In contrast, other theorists call into question the very reality of the self, and hence can hardly claim to have uncovered a human or any other sort of nature. Instead they dismiss all essentialist views, including those about the self, as the aftermath of the "modern" age (so called to distinguish it from the more tradition-oriented middle ages) that began with Descartes, which is why these writers are usually called postmodernists. A large part of their discussion of the self and cultural identity consists in attempts to "deconstruct" these notions by showing that the discourse which uses these terms is historically conditioned in every respect, and that the modernists' homespun, seemingly natural metaphysical ambition (i.e., to discover how things are) is itself a delusion. This is the central tenet of the postmodernist theories of identity proposed by the social philosopher Michel Foucault, the feminist thinker Judith Butler, and others who have had such an impact on the politics of cultural identity. Backhurst and Sypnowich summarize the postmodernists' position quite nicely. For such writers, "the self is a fiction of modernity, constructed by discourses such as metaphysical philosophy and psychotherapy and many of the practices of everyday life. As such, selves have no nature or essence, they are not parts of the furniture of the world; they are discursive effects, products of our forms of representation" (p. 8).

"Cultural Identity": The Social Psychology Story

Keeping in mind this philosophical context, let us now turn to the standard social psychological conception of identity, found in various theories of personality that treat (i.e., construct) identity as an independent variable. But first, a few general remarks concerning the umbrella term "cultural identity." (6) The umbrella is not only wide but elastic. In current English usage, the expression "cultural identity" refers equally well to a property of an individual person (my cultural identity) or a property of a group (the identity of our culture). It can be used in a positive sense to indicate ethnic pride, or negatively, to denote the crudest sort of ethnocentrism. And like the term "multiculturalism," it is associated with a wide variety of often-conflicting social and theoretical positions. In political contexts, it suggests identity politics and even anti-liberalism or at least what political theorists now call communitarianism. In clinical and other psychological contexts, cultural identity is thought of either as a relatively high level of self-actualization or, at the other extreme, as a rather regrettable stage of conformism, just a little better than indoctrination and brainwashing. Outside the relatively narrow sub-discipline of experimental social psychology, psychologists divide on whether cultural identity enhances or competes with personal agency, whether it is something one has or something one does — in short, whether it is inside, outside, or the same as one's self. Sociologists and philosophers find still other issues connected to the concept of cultural identity, such as whether groups determine individual consciousness or vice versa, and whether selves and/or groups are stable entities or whether they are like Heraclites' river, so fluid that one can only name them but never truly know or possess them.

However, in spite of this diversity there are commonalities, and even orthodoxies, in the literature of modern social psychology. Cultural identity is nearly always discussed oppositionally: it is a figure that takes its shape when contrasted with the conceptual background and social horizon of cultural assimilation. More specifically, it is those who would resist the rhetoric and social practice of assimilation who wind up talking about cultural identity; when assimilationists speak of cultural identity they only do so in order to characterize their opponents' position. This tendency is also found in discourse that on the surface seems purely factual rather than persuasive, such as classical ethnography and the heavily quantitative discourse of mainstream social science in North America and Britain.

In modern social psychology, especially that branch called personality theory, the orthodox discourse treats personal identity as an epiphenomenon of social identity and an amalgam of self-concept and self-esteem. This social scientific construction grew out of a more general discussion in 20th century social theory concerning how individuals are related to groups. (7) The latter theme can be traced back to the grand speculations of Emile Durkheim and Max Weber and, through them, to Rousseau and Kant, but took its present direction at the end of the Depression era, when what is now called social psychology began to take on the more technical, experimental character it would later exhibit on so many fronts. Over the next few years emigré scholars in the United States such as Theodor Adorno and Kurt Lewin conducted and inspired new research on the relationship between personality and society, which was subsequently gathered up in Talcott Parsons's (1951) grand synthesis of the social and behavioral sciences.

The most important early figure in this discussion was probably Lewin, who is often considered the father of American social psychology. In 1941 (but see also Lewin, 1948) he explained what he considered Jewish "self hatred" — a matter of personal identity — as the product of the way Jews constructed their social identity. That is, they identified with the "more powerful" people who persecuted them and, in consequence, directed their hostility toward themselves rather than toward their oppressors. A few years later, Kenneth and Mamie Clark (1947; see also their earlier studies in Clark & Clark, 1939, 1940) used the same sort of from-social-to-personal explanation for their famous experiments with black children who disliked dolls most like themselves, preferring instead dolls with lighter skin color and other mainstream (i.e., white) features. About the same time, a similar account was given by S. W. Fernberger (1948) of women who internalized negative gender stereotypes. This way of understanding low self-esteem continued well into the 1970s (see Kardiner and Ovesey, 1951; Lynne, 1959; Brody, 1964; Herzberg, 1977). Of course there were variations on the basic theme: as time went on many social psychologists based their theories of personality (and their research methodologies) not on Lewin's quasi-mathematical "field theory" of cognition and emotion, which by 1960 or so had fallen out of favor, but on the more cognitive "looking glass theory" introduced much earlier by Charles Horton Cooley (1902/1964), according to which one's self-concept is directly and entirely a "reflection" of images of oneself picked up from significant others and from society at large.

Such identifications and reflections are undoubtedly important factors in the construction of one's self-concept, especially its evaluative dimension (self-esteem). Furthermore, variants of the looking glass theory of the self (e.g., G. H. Mead, 1934; Goffman, 1959) provided alternatives to substantival notions of the self that had dominated formal and informal personality theory since Descartes. But these alternatives, like the classical substantival notions of the self, were static and, when applied to issues of cultural identity, profoundly misleading. To appreciate this point, let us consider as a kind of intellectual case study a very illustrative and often-cited attempt to separate personal and group identity by making them two independent variables.

Cross's Discovery: A Story with the Story

In the 1970s the young psychologist William Cross, while working with Uri Bronfenbrenner and M. Cochran on the socialization of black children, began to suspect that social psychologists and personality theorists who had written on cultural identity over the preceding decades had seriously distorted the black experience by confounding the concepts of personal identity and group identity. The early psychological investigations of black children's identity formation was based on the use of dolls, puppets, and other nonverbal devices. The first and probably still best known of these are the above-mentioned studies by the Clarks, in which very young children were asked to choose between black and white dolls. Because in a variety of test tasks these children, most of whom were around 5 years old, preferred the white doll, those studies all concluded that the children had low self-esteem. However, subsequent studies using verbal devices, such as interviews and surveys rather than dolls, showed black children's self-esteem to be as strong, if not stronger, than that of white children (Wylie, 1979; Gordon, 1977).

For a while, the discrepancy between these findings was seen by psychologists as the result of faulty or incompatible methodologies (surveys are more cognitively complex than dolls, can be administered to more subjects, etc.). However, when Cross examined this literature more closely he discovered a much more basic source of confusion in the very idea of a "self-concept." Unfortunately he did not challenge its very legitimacy as a theoretical construct, but he did correctly complain that most of the work carried on under this rubric made no significant distinction between self-esteem (an affective construct — roughly speaking, how one feels about oneself) and self-concept in the narrow sense (a cognitive construct -- roughly speaking, how one thinks about oneself). Furthermore, in spite of their labels each of these two secondary constructs was used to refer to what one feels or thinks (or thinks one feels) about a group as well as about one's self. It was for this reason that in his Resolving Social Conflicts (1948), Lewin had treated group self-hatred and individual self-hatred as though they were a single variable. Sad to say, a half-century later such confusions still linger on in the literature and practice of multicultural education and, one suspects, in the public mind in general.

After reviewing 161 studies on black identity, Cross tried to sort out this confusion by distinguishing two component constructs or independent variables which together constituted the otherwise hopelessly vague construct of self-concept (SC). These were personal identity (PI) and reference group orientation (RGO). Thus SC = PI + RGO. Cross's review (1985) of the technical literature on black identity involved the usual psychometric apparatus of American social psychology (null hypotheses, Chi-squares and correlations, statistical significance, etc), but we need not go into such matters here. It is only necessary to understand the difference between the PI and RGO variables. Unfortunately, our task is complicated because both variables are "superordinate constructs" that refer to one or another cluster of related subordinate constructs rather than to a single construct or psychological measure. Cross was looking at a wide variety of studies, some of which tested for self-esteem, some for self-confidence, and so on. (See Table 1.) These he called studies of personal identity. In contrast, those studies which tested for racial attitudes, degree of identification with a group, etc., were what he discussed under the rubric of reference group orientation studies. However, many of the research programs used both sorts of measures, with the confusing results mentioned above. (8)

Table 1. Self-Concept Constructions in 161 Black Identity Studies

1a. Superordinate Construct:

1b. Subordinate Constructs:



2a. Superordinate Construct:

2b. Subordinate Constructs:

Reference Group Orientation (RGO)

Racial attitudes, group identity, race awareness, racial ideology, race preference, race evaluation, race-esteem, racial self-identification, race image, esteem for group, in-or-out group orientation, etc.

Personal Identity (PI)

Self-esteem, self-worth, self-confidence, personal-self, general personality, personality interpersonal competence, primary-self, personality traits, anxiety level, unconscious-self, hidden personality dimensions, etc.

(Adapted from Cross, 1985, p. 159)

For Cross, what the second set of subordinate constructs (those featured in PI studies) had in common was the assumption that PI can be measured in terms of certain universal human tendencies, traits, or behaviors, and that "cultural differences" — understood in a broad sense that included differences in race, ethnicity, and gender as well as traditions and symbol systems — could be excluded from the criteria used to assess personal identity. In the first set (the subordinate constructs featured in group identity studies), on the other hand, some allusion to cultural differences (usually specifically racial difference), such as the colors of the different dolls, was essential to the instrument used to measure the construct. Unfortunately, when both sorts of instrument were used in the same study, the corresponding subordinate constructs were not distinguished. Alternatively, authors of studies that used only one type of instrument — usually RGO studies -- summarized their results as though the study had actually assessed both sorts of constructs, including self-esteem or some other PI variable. As a result, the difference between disaffection for one's group and disaffection for one's own self was erased. Terms such as "self hatred" were used to refer to both sorts of alienation, à la Lewin.

In other words, the correlation between RGO and PI constructs is just that, a correlation, which is to say a statistical relationship. It was not a semantic or conceptual association, an entailment relationship, or even a causal connection. And it was most definitely not a "real relation," if by that term one means a mind-independent relationship rather than a social construction. What Cross correctly called the "long held Lewinian hypothesis" (Cross, 1985, p. 170) was that RGO predicts PI; however, he failed to recognize that the actual strength of this predictive power is itself unpredictable, since the degree of correlation between the constructs is a matter of history, not logic. What the data actually suggested was that from the beginning of World War II until well into the Civil Rights movement, many black people were indeed more "out-group oriented" than whites, as evidenced by the black children's preference for white dolls and at the adult level by blacks' preference for commodities, social policies, and personal behaviors associated with the white mainstream. When blacks of that era did prefer alternatives specifically associated with the black community, their pro-black preferences were normally mixed in with other preferences for white symbols. However, there was virtually no evidence collected during this time about how black children or adults felt about themselves personally. Strictly speaking, this lack of evidence means that, as Cross put it, "Conclusions about the negative or positive nature of the personal identity of blacks from 1939-1960 were purely speculative" (p. 161). Even so, studies carried out over the subsequent decades — which were more carefully designed and did test both PI and RGO — suggest that if one does speculate (as social theorists must), there is good reason to think that long before the Civil Rights movement, black people's personal self-esteem was normally not a problem — not even for those blacks least inclined to identify themselves with their racial group. Furthermore (as long as we are speculating), there seems to be no reason for us to expect other racial or cultural minorities to be different in this respect.

I suspect that white readers will be more surprised at these findings than blacks or other minority members. The hardships of minority membership, which of course include much more than sheer economic privation, can produce strong egos as well as damaged ones. Memories, autobiographies, and other accounts of growing up black (or chicano, jewish, or otherwise "other") are full of anecdotes about warm and nurturing child rearing practices, whose natural outcome would seem to be intact egos. Such anecdotal evidence supports the thesis that most people can and do have high PI, regardless of whether they think about their reference group in positive or negative terms or indeed in any terms at all. However, it does not follow from this thesis that the public perception (by whites or non-whites) of how members of a racial group affiliate necessarily reflects the individual attitudes of those members.

Going beyond Cross's Constructs: The Stories Merge

Over the last few years social psychology (unlike most multicultural education theory) has finally begun to recognize the difference between personal and group identity. For instance, the Dutch psychologist M. Verkuyten recently wrote: "Most recent studies using well-established scales and adequate control groups have shown no relationship between ethnicity and global personal self-esteem among youth in different Western countries .... Empirical research in general does not support the assumption that Afro-Americans have lower self-esteem" (Verkuyten, 1995, p. 156). However, empirical research continues to operate on the assumption that these constructs (i.e., "ethnicity and a global personal self-esteem") as well as the several subordinate constructs considered by Cross, are either natural categories or at least indispensable heuristic devices for uncovering relationships that are "always already there" in the social world. Consider Cross's concluding assertion that:

Blacks have had, and continue to have, a multifaceted reference group orientation such that black and white anchor points may determine behavior depending upon the situation being confronted. The Black Movement probably increased the number of black anchor points in a person's world view. (Cross, 1985, p. 170)

This assertion may seem quite reasonable to many readers, especially those dissatisfied with the sometimes excessive claims by proponents of identity politics. After all, it is easy to agree that cultural identities and other sorts of group orientation are not monolithic, and do not determine action in abstraction from concrete situations. There is, for instance, no one right way to be black, any more than there is a single way to be white. But this way of thinking about cultural identity seems reasonable only because like Cross and other social psychologists, most of us are all too willing to continue thinking of social groups and their psychological correlates (group identities) as entities. We only require that large conceptual monoliths be broken down into smaller ones, miniliths as it were, in rough analogy to the way capitalist governments have tried to dissolve large monopolies such as Standard Oil, AT&T, Microsoft, and other huge organizations. The point of this business analogy is that such dissolution is diversification (surface reform) not deconstruction (radical revision), since at the end we still have autonomous corporations. Similarly, Cross's "multifaceted" theorizing about group identity simply replaces a general reference group orientation with several more specific orientations, which are the relations a person has to several groups simultaneously albeit not in the same respects. And these several specific groups are just as reified as the general one is.

A simple logical point may help us appreciate why social groups should not be reified — bearing in mind that throughout this essay we are concerned primarily with those social groups variously referred to as cultures, races, and ethnics, or (to speak more carefully) cultural, racial, and ethnic groups. As the term itself indicates, a group is not only a plurality but a grouped plurality, in the sense that certain items have been singled out and designated as members of the group in question. Thus the abstract notion of a group is not quite the same as the mathematician's notion of a set (though the two notions can coincide), since a set is not a delimited collection — at least, not by its very definition. The set of all sets would not be a group, nor would "the universe," using the latter term in the widest possible sense. Furthermore, when a group is a distinctly human group, it is normally thought of in a special way: it not only comprises human beings the way a herd of cows contains cows, but also and much more importantly, the very grouping of the group is a cognitive act (some philosophers would rather say a propositional attitude) carried out by the group members themselves. That is, a city, tribe, community, etc., is a group — considered both generically and in particular (i.e., this city, tribe, etc.) — because its members recognize themselves as such. Thus the concept of a human group is reflexive.

To this conceptual point, I should add the commonsense observation that the reflexivity admits of degrees. A nursery of wailing newborns would be a minimally human group, not because the babies are minimally human (whenever that might mean) but because their awareness is utterly nonverbal, even though they are rudimentarily aware of each other's existence and of their common situation. It is a well documented fact that newborn babies cry empathically, such that a misplaced diaper pin in one baby triggers a collective empathic crying that is different from simple distress crying (see M. Hoffman, 1976). But no diaper pin has ever triggered a discussion among the babies themselves of their shared plight at the hands of clumsy nurses or led them to label themselves as, say, "we helpless and hapless newborns." Nor do preschoolers, another minimally reflexive group, abandon their solitary play in order to discuss their common status as children of working mothers, prospective kindergartners, etc. Even so, it is a truism in popular as well as scholarly discussions of social groups and group identities (see Barth, 1969, 1994), that the most important part of what makes an assemblage of human beings a group is their shared perception of themselves as (1) a collective which is (2) numerically different from other collectives (expressed indexically as: "We are not they") and also (3) qualitatively different, at least in most cases. Note, by the way, that when this last feature is absent individuals tend to construct such differences for themselves, with the eventual result that the group does in fact take on qualities that differentiate it from other groups.

The Moral of These Stories

Like personal selves, cultural groups are constructions that arise, not arbitrarily or subjectively, but from established discursive practices, practices that transcend individual desires and beliefs in the sense that these practices shape desires and beliefs. Most importantly, these groups are not entities in their own right. As with cultural differences such as religion or language, so also with the features that differentiate racial or ethnic groupings: they are conventional delineations, some of which are more useful than others as we pursue the business of living together. As philosophers and as consociates, we must avoid the tendency to think of such features — including even skin color — as brute facts or primordial properties rather than as conventional templates for interpreting the social world. As we saw above, this tendency, which has dominated social science since it became a separate discipline in the early 20th century, can be traced back to Descartes. However, there is an alternative philosophical tradition, whose forebear is Descartes' contemporary Giambattista Vico, which understands social groupings and the conceptualized differences between them as products of human agency and, because agency is generally goal-directed, as products of human interests. For these thinkers, social groupings are interpretations all the way down, to borrow the slogan of Clifford Geertz's (1973) hermeneutical approach to anthropology. Not only do these interpretations understand groups as radically historical (no essences here!), but in their better moments they also understand themselves, qua interpretation, as historical, contingent, fluid, multiperspectival, self-revising, and so on. Anthropologists know this quite well, having discarded the notion of culture as a determinate, frangible unity in favor of much more sophisticated ways of understanding "the other" (see, inter alia, Clifford, 1988; Rosaldo, 1989; Kahn, 1995). All this is consistent with the general hermeneutical conception of knowledge as self-adjusting. Some interpretations accent and other interpretations obscure differences and continuities within individual groupings. Some accent and others obscure influences, commonalities, and contrasts between social groupings. What Segal and Handler have said of cultures is actually true of all human groups: their parsing is "a matter of interpretation, notwithstanding that at any given moment, and even for long durations, some interpretations are enforced more than others" (1990).

In contrast, most social psychology studies of cultural identity, including Cross's, lack a true appreciation of the oft-stated but poorly understood idea that cultures and other sorts of reference groups are socially constructed. They are not natural entities which one "discovers," the way one might discover a lump under one's arm. But who constructs them? In the case of black identity, one answer to this question is that in the course of maintaining its own dominance the white mainstream constructed the pre-Civil Rights conception of black ethnicity, which was embodied in Booker T. Washington, and that emerging black leaders constructed a later and more militant conception, which was associated with figures such as W.E.B. DuBois. Another answer is that both of these constructions were by-products of social processes that transcended the intentions of individual agents, black or white, and this answer seems to me much more plausible though I cannot develop it in these pages. Even more plausible, though, is a third answer that includes but goes beyond the first two, namely, that the members of this reference group have constructed it themselves, by accepting and passing on to their children the collective names ascribed to them, the latest of which is the name "African American." To give Cross the credit due him, he gets it right when he says they "superimpose a black orientation" on their own experience. In this respect they are not different from any other social group, though it must be admitted that until recently black/African Americans had much less control over the way people referred to them than did other reference groups. The greatest significance of what Cross calls the Black Movement is not that it put black Americans in touch with their true or authentic selves (a fundamentally essentialist notion), but that it gave them a new set of interpersonal relationships that matched the changing political climate of the larger society. If one must use the language of discovery, the thing to say is that blacks discovered new solidarity, not race, which is to say that they created a new (but of course it was not altogether new) collectivity for themselves.

Cross's essentialism becomes even more problematic at the end of the essay I have drawn on, in which he moves from a primarily descriptive mode of discourse to a primarily prescriptive one having to do with the black community's collective need for members with strong black RGOs. It is one thing to say, as Cross did a few lines earlier in the same essay, that the Black Movement has increased the probability (which is not the same as desirability) of more — and more comprehensive — black RGOs. But in the concluding paragraph pp. 170-71 he goes much further. "The black community needs programs that increase [this] probability," he declares, and this need justifies efforts to develop black children into "adults who will identify and participate with other blacks in political action" (1985, pp. 170-71). At this point one wonders if the tail has not begun to wag the dog. Self-esteem and other PI variables of black children can be enhanced in many ways, not all of which involve other blacks. Regardless of its truth-status, to say, as Cross does, that the needs of the group "require that a critical mass of black people strongly identify with black people and black culture" makes sense only if the group is thought of, improbably, as a determinate, superorganic individual.

Far better, I think, is the view of group identity which the political philosopher David Ingram describes as emerging from a collective, collaborative dialogue among real people with "fluid, overlapping, but distinct perspectives and preferences" (2000, p. 212). This view, which guides the political efforts of minority rights advocates such as Lani Guinier and before her Martin Luther King, consolidates the notions of cultural and personal identity without confounding them. Instead of being conscripted into service of one's group in the fashion described by Cross, a liberated but nonetheless loyal member of a racial or any other cultural group will understand that it is all right to interpret that group's identity in his or her own way. Indeed, it is not only all right to do this, it is inevitable. We can only understand our group from our own individual situation in the moving stream of discourse that continually constructs and reconstructs our personal and cultural identities.


1. The concept of culture and, by implication, that of cultural identity, is no longer understood primarily in terms of worldview, shared history, or other aspects of a group's "cultural heritage" — at least not in the United States. For better or worse, in my part of the world culture has evolved into a sociological category, so that today one speaks of working class culture, gay culture, the deaf culture, and even the culture of the homeless. These are somewhat easier to understand as purely social constructions than are traditional ethnic cultures. However, I have argued elsewhere that in the final analysis all cultures are constructed, even those most closely associated with ethnicity. This change is nowhere more evident than in the discourse of multicultural education, where militant "critical multiculturalists" of the late 1990s accused the previous generation of multiculturalists, including Afrocentric educators of the 1980s and early 1990s, of using anthroplogical notions of culture to divert attention from more important class and racial differences in American society (see Gordon & Newfield, 1996; Kincheloe & Steinbert, 1997). Even so, readers from other countries may prefer to read "social" for "cultural" whenever I use the term "cultural identity," even though the latter term dominates the discussions of identity that I reference in the following pages.

2. See note 1.

3. The phrase "in the western intellectual tradition" is an important qualification. As Diane Hoffman (1998) has shown, the very concept of personal identity is culture-specific (as are the concepts of culture and cultural identity). This claim, which she defends quite persuasively, relativizes everything I am claiming in this essay, which of course is not the same as refuting or invalidating it.

4. For convenience, I use these two terms interchangeably in this essay, but see Rom Harré's Personal Being (1984) for a helpful fine-grained differentiation.

5. In nonwestern philosophy one finds various notions corresponding to our idea of the self, but they are very hard to describe in western terms. The present discussion makes no pretense of offering such a description. Furthermore, I have severely telescoped the story of the self in modern western philosophy, by speaking as though Descartes himself normally used terms such as "self" or "soi" in the way described here (which is why I have put the word "self' in scare quotes in this paragraph). In fact it was his great successors — Locke, Berkeley, Hume, Kant who used the term "self" and its cognates to express Descartes' basic idea of subjectivity as a thinking thing, as well as their own misgivings about that idea.

6. As I noted at the outset of this essay, readers not used to the broad sociological sense of "cultural identity" now commonplace in the United States may wish to mentally substitute "social identity" for that term in what follows.

7. In the following pages most of my comments about psychology and other domains of social science are, to my considerable regret, limited to what I know about the literature in the United States and Canada. Since this literature has been heavily influenced by European thinkers, it seems reasonable to hope non-American readers, at least those educated in the western tradition, will find these remarks of some use. However, readers of all nationalities deserve my apologies for the unmistakable, though because of my own limitations unavoidable, ethnocentric bias in this essay.

8. It is worth noting by the way that some of the instruments used in the PI studies contained the term "self-concept" in their titles, e.g., the Piers-Harris Children's Self-Concept Test and the Tennessee Self-Concept Scale. However, these were mainly measures of self-esteem and did not tap directly into the child's reference group orientation.


Barth, F. 1969. Introduction. In F. Barth (Ed.), Ethnic groups and boundaries: The social organization of cultural difference, 9-38. Boston, Little, Brown and Company.

Barth, F. 1994. Enduring and emerging issues in the analysis of ethnicity. In H. Vermeulen and C. Govers (Eds), The anthropology of ethnicity: Beyond Ethnic groups and boundaries, 11-32. Amsterdam: Het Spinhuis.

Bakhurst, D., and Sypnowich, C. (eds). 1995.The Social Self. New York and London: Sage Publications.

Brody, E. 1964. Marginality, identity and behavior in the American Negro: A functional analysis. International Journal of Social Psychiatry 10:7-12.

Chodorow, N.1978. Reproduction of mothering : Psychoanalysis and the sociology of gender. Berkeley: University of California Press.

Clark, K., and Clark, M. 1947. Racial identification and preference in Negro children. In T. M. Newcomb and E. L. Hartley (Eds.), Readings in social psychology, 602-611. New York: Henry Holt.

Clark, K., and Clark, M, 1939. The development of consciousness of self and the emergence of racial identification in Negro pre-school children. Journal of social psychology 10:591-599.

Clark, K., and Clark, M. 1940. Skin color as a factor in racial identification of Negro pre-school children. Journal of Social Psychology 11:159-169.

Clifford, J. 1988.The Predicament of Culture: Twentieth-Century Ethnography, Literature and Art. Cambridge, MA: Harvard University Press.

Cooley, C. H. 1902/1964. Human nature and the social order. New York: Schocken Books. Originally published 1902.

Cross, W. E. 1991. Shades of Black: Diversity in African-American Identity. Philadelphia: Temple University Press, 1991.

Cross, W. E. 1985. Black identity: Rediscovering the distinction between personal identity and reference group orientation. In M. Spencer, G. Brookins, and W. Allen (Eds.), Beginnings: The social and affective development of Black children, 155-171. Hillsdale, NJ: Erlbaum.

Descartes, R. 1641/1986. Meditations on first philosophy, trans. John Cottingham. Cambridge, Cambridge University Press. Originally published 1641.

Fernberger, S. W. 1948 . Persistence of stereotypes concerning sex differences. Journal of Abnormal and Social Psychology 43, 97-101.

Fuss, Diane. 1989. Essentially Speaking: Feminism, Nature, and Difference. New York: Routledge, Chapman, and Hall.

Geertz, C. 1973. The interpretation of cultures. New York: Basic Books.

Goffman, E. 1959. The presentation of self in everyday life. Garden City: Doubleday.

Gordon, V. V. 1977. The self-concept of Black Americans. Washington, DC: University Press of America.

Hall, S. 1996. Who needs 'Identity'?. In S. Hall and P. du Gay (Eds.), Questions of Cultural identity, 1-17. London: Sage Publications.

Harré, R. 1984. Personal being. Cambridge, MA: Harvard University Press.

Herzberg, J. 1977.Self-excoriation by young women. American Journal of Psychology 134, 320-321.

Hoffman, M. 1976. Empathy, role-taking, guilt, and development of altruistic motives. In T. Lickona (Ed.), Moral development and behavior: Theory, research and social issues, 124-143. New York: Holt, Rinehart and Winston.

Hoffman, D.1998. A therapeutic moment? Identity, self, and culture in the anthropology of education. Anthropology and Education Quarterly. 29:3, 324-346.

Hyman, J. J., and Singer, E. 1968. Readings in reference group theory and research. New York: McGraw-Hill.

Ingram, D. 2000. Group rights: Reconciling equality and difference. Lawrence, KS: University of Kansas Press.

Kahn, J. 1995. Culture, multiculture, postculture. London: Sage Publications.

Kardiner, A. and Ovesey, L. 1951. The mark of oppression. New York: Norton.

Kincheloe. J., and Steinbert, S. 1997. Changing multiculturalism. Buckingham/Philadelphia: Open University Press.

Lewin, K. 1941. Jewish self-hatred. Contemporary Jewish Record 4:219-232.

Lewin, K. 1948. Resolving social conflicts. New York: Harper and Row.

Lynne, D. B. 1959. A note on sex differences in the development of masculine and feminine identification. Psychological Review 66, 126-135.

Mead, G. H. 1934. Mind, self, and society. Chicago: University of Chicago Press.

Parsons, T. 1951. The social system. New York: Free Press.

Rosaldo, R. Culture and truth. Boston: Beacon Press, 1989.

Handler, R., and Segal, D. 1990. Jane Austen and the fiction of culture: An essay on the narration of social realities. Tucson: University of Arizona Press.

Shweder, R. 1991. Thinking through cultures: Expeditions in cultural psychology. Cambridge, MA: Harvard University Press.

Shweder, R., and LeVine, R. (eds). 1984. Cultural theory: Essays on mind, self and emotion. Cambridge: Cambridge University Press.

Taylor, C. 1994. The politics of recognition. In A. Gutmann (Ed.), Multiculturalism: Examining the politics of recognition, 25-73. Princeton: Princeton University Press.

Verkuyten, M. 1995. Self-esteem, self-concept, stability, and aspects of ethnic identify among minority and majority youth in the Netherlands. Journal of Youth and Adolescence 24:2:155-175.

Wren, T. 1990. Caring About morality: Philosophical issues in moral psychology. Cambridge, MA: MIT Press/London: Routledge.

Wylie, R. C.1979. The self-concept. Vol. 2: Theory and research on selected topics. Lincoln, NB: University of Nebraska Press.