Chapter One—
The Search for Conflict

In his well-known essay on "The Subjection of Women," John Stuart Mill (1869) made the following argument about what he regarded as the special character of the unequal relationship between men and women:

They [women] are so far in a position different from all other subject classes, that their masters require something more from them than actual service. Men do not want solely the obedience of women, they want their sentiments. All men, except the most brutish, desire to have, in the women most nearly connected with them, not a forced slave but a willing one, not a slave merely, but a favority. They have therefore put everything in practice to enslave their minds. The masters of all other slaves rely, for maintaining obedience, on fear—either fear of themselves, or religious fears. The masters of women wanted more than simple obedience, and they turned the whole force of education to effect their purpose. ([1869] 1970, 141)

Mill's argument is important, both for what it says and for what it does not say. First, Mill identified the clear advantages to a dominant group of obtaining the unthinking compliance of subordinates and the significance of ideology ("the whole force of education") as a weapon to achieve that effect. But, second, Mill assumes that relationships of inequality in general are not governed by such considerations. He argues that male-female relations are an exception (and, as I shall show later, only a partial exception at that) to the general rule of overt hostility between unequal groups and the forced compliance of subordinates. The first point is an important insight into the ideological dynamics of dominance and subordination. The second point is an equally important oversight that undermines the potentially broad significance of Mill's analysis.

Mill's oversight has been endemic to the disparate literature on the

ideology of inequality. The attitudes that are espoused by groups who are engaged in a relationship of inequality have been analyzed and dissected by students of class relations, racial prejudice, political tolerance, and gender relations. Each of these bodies of literature has operated independently, each caught up in its own specialized language and research agenda. But what each corpus of research does share is the pivotal, implicit assumption that conflict and hostility are the exclusive hallmark of problematic social relations. I turn now to an overview of the motivating concerns of each of these bodies of literature on the ideology of inequality. Such an overview reveals both the discrete concerns of each corpus of research and the common thread—expressed in various ways—of the assumption that there is an indefatigable link between problematic social relations and conflict.

Class and Class Conflict

Karl Marx's analysis of class relations put conflict squarely at the center stage of debate about the existence and meaning of social class. The abiding role that Marx assigned to conflict in the expression and resolution of exploitative social relations is made clear in Engels' succinct summary of Marx's approach in his introduction to "The Manifesto of the Communist Party":

The fundamental proposition . . . is that in every historical epoch the prevailing mode of economic production and exchange and the social organization necessarily following from it form the basis upon which is built up, and from which alone can be explained, the political and intellectual history of that epoch; that consequently the whole history of mankind (since the dissolution of primitive tribal society, holding land in common ownership) has been a history of class struggles, contests between exploiting and exploited, ruling and oppressed classes; that the history of these class struggles forms a series of evolutions in which, nowadays, a stage has been reached where the exploited and oppressed class—the proletariat—cannot attain its emancipation from the sway of the exploiting and ruling class—the bourgeoisie—without, at the same time, and once and for all, emancipating society at large from all exploitation, oppression, class distinctions, and class struggles. (Marx and Engels [1888] 1959, 4)

Capitalism would bring to a climax the evolutionary history of exploitation and conflict, and it would therefore intensify the strains that had been the catalyst for change throughout history. The language of conflict and struggle abound in Marx's analysis of capitalism:

The modern bourgeois society that has sprouted from the ruins of feudal society has not done away with class antagonisms. It has but established new classes, new conditions of oppression, new forms of struggle in place of the old ones.

Our epoch, the epoch of the bourgeoisie, possesses, however, this distinctive feature: it has simplified the class antagonisms. Society as a whole is more and more splitting up into two great hostile camps, into two great classes directly facing each other: bourgeoisie and proletariat. (Marx and Engels [1888] 1959, 8)

The Debate about Marx

Marx's thesis drew the first powerful lines of the framework that was to lock in the debate about social class. Those who wished to promote the view of class as exploitative felt compelled to show evidence of its injurious effects on social harmony. Those who sought to portray inequality as benign pointed to the lack of strife along class lines as vital evidence. Indeed, in the ensuing "conflict versus consensus" debate, the concept of class itself became tied-to the inseparable pair, exploitation and conflict. To claim no evidence of class conflict was tantamount to denying the existence of class at all as an organizing factor in social life.

The powerful impact of Marx's analysis on the concept of social class reflects not only the radical nature of the theory itself, but the particular quality of the criticisms it engendered. Although some analysts disputed the exact basis of class organization (for example, authority relations at work or type of work rather than relationship to the means of production), the most threatening criticisms that defined the other extremities of the framework for debate did not challenge the internal logic of Marx's theory or the logic of its empirical deductions. Attempts to discount Marx's analysis primarily took the form of working backwards from the ultimate empirical prediction, class conflict: using the absence of class conflict as their logical starting point, critics sought to establish that the notion of class exploitation on which conflict was predicated was therefore wrong.

Criticisms of Marx came in a series of waves. Indeed, until the relatively recent advent of neo-Marxist thought, Marx's influence on the literature on class inequality seemed to be primarily indirect, through the counterarguments that it spawned. The first enduring criticisms came from Max Weber. A second line of dispute came from Emile Durkheim and the subsequent development of the functionalist school of thought. Further criticisms were developed by scholars from the pluralist and postindustrial/postmaterialist schools of thought. Without giving a full account of each of these waves of criticism, I would like to draw attention to those aspects that reinforced the equation of exploitation with conflict and hostility.

Weber

Max Weber took issue with Marx on both methodological and theoretical grounds. In his advocacy of the "ideal type" methodology, he held that theory construction was a misleading tool for the analysis

of social life. He did not want to argue that Marx's theory was wrong in the sense that an alternative theory would be better. He argued that any theory of social life would be limited both by the elusively changeable nature of social phenomena themselves and by the baggage of personal values that the social scientist brings to the world in which he is both a participant and an observer (Weber 1949). The analytic goal of Weber's "ideal type" methodology was to construct an abstraction of some social phenomenon and then to use that abstraction, not as a lens through which to observe and interpret social reality, but as a reference against which reality could be compared. In this way, he hoped to identify the regularities of social life as well as to highlight the idiosyncracies. Weber's purpose, therefore, in criticizing Marx was not to advocate an alternative theory but to treat Marx's theory as though it were an "ideal type" and to elucidate those aspects of social life that did not square with Marx's model of it. Because of this methodological slant, Weber's legacy was more a series of questions about various pieces of Marx's theory than the development of an alternative conception of the dynamics of social inequality.

Weber raised numerous questions about Marx's theory, but the statement that had the most profound effect on the shape of the literature on social inequality was his fragmentary essay, "Class, Status, Party" (Weber 1946a ). In that paper, he argued that classes may be nothing more than bare, economic categories devoid of the sense of community or shared purpose that come out of status distinctions. Pointing to social distinctions that have no direct basis in economic inequality (and especially racial and ethnic distinctions), Weber argued that they "interfered with the sheer market principle" by providing alternative, more socially salient magnets for social exclusiveness and intergroup antagonism.

In the subsequent desire to move away from a class analysis of society, social scientists seized upon this idea of Weber's. Specifically, they attempted to demote the analytic importance of class by arguing that other culturally based distinctions (especially race and ethnicity) evidenced more conflict and hostility than did class. In this way, Weber's legacy was to heighten even further the significance attached to social conflict in the analysis of social inequality.

Durkheim

Durkheim's ideas about social inequality reinforced the linkage of conflict with problematic social relations more directly. Durkheim viewed social differentiation, not as the product of exploitative relationships, but as a benign and functionally integrative feature of organized social life (Durkheim [1897] 1951, [1893] 1965). At the same time, he denied the normality of conflict—he acknowledged its contemporaneous existence but portrayed it as an aberration from normal social intercourse brought about only by minor maladjustments in a complex

but overwhelmingly balanced, integrative social system. Social inequality was functionally integrative, and under "normal" circumstances this would be tacitly demonstrated in its consensual endorsement (Durkheim [1895] 1964). Thus, whereas Marx depicted conflict as intrinsic to the course of human history, Durkheim regarded conflict as abnormal. For Durkheim, consensus and harmony typically govern social intercourse, and this was taken as a manifestation of the functional integration of the underlying structure of society. Outbreaks of conflict were taken as a signal that things had slipped temporarily out of kilter, indicating that the system required repair to nudge it back to its normal state of integrated balance. Thus, Durkheim preserved the logic of a link between pervasive social struggle and exploitative social inequality. Even sporadic outbreaks of conflict were taken as evidence of a social problem, albeit a lapse capable of repair rather than an endemic structural flaw. In Durkheim's view, absence of conflict was the normal condition of human society and it demonstrated the absence of exploitation and even the absence of any divisive social problem.

Davis and Moore

In their functional theory of stratification (1945), Davis and Moore promoted Durkheim's line of reasoning, and, in so doing, they sealed the course of the debate about the meaning of social inequality into a choice between conflict (symptomatic of exploitative relations) and consensus (indicative of functional differentiation). Davis and Moore argued that socioeconomic inequality, far from being a weapon that frustrates human creativity and freedom, is the necessary instrument by which the potential of both society and its individual constituent members is realized. An unequal distribution of rewards does not indicate that one group is taking something from another but rather an implicit social contract in which all participants mutually understand their respective contributions to the common good. Tasks that contribute more are awarded higher benefits, especially if their fulfillment requires a greater personal investment.

Such an argument rests implicitly on the idea of consensual social relations: if one argues that social inequalities are freely and spontaneously entered into, it is inconvenient to find those with fewer rewards restive and complaining that something has been taken from them. The conception of society as a collective whole further encouraged an emphasis on consensual social behavior, rather than on discord with its implication of disparate social groups. Indeed, the holistic view of the functionalists made the very concept of discrete classes inappropriate and encouraged instead the perception of an unbroken continuum of inequality: the constituent functions and rewards of society provided no basis for the coalescence of distinct socioeconomic groups. The debate

over the meaning of social inequality was now shaped into well-demarcated positions that hinged critically on the constancy or infrequency of group conflict.

Pluralism

The inevitability of the connection between entrenched relations of exploitation and sustained group conflict was also taken as a given by scholars in the pluralist school of thought. Asserting that American society was characterized by an absence of repeated conflict along class lines, pluralists argued that the critical ingredient of sustained class exploitation was missing (for example, Coser 1956; Polsby 1980). Rather than seeing the political system as reflecting a reiterative bias in favor of one segment of society at the expense of another, pluralists conceived of the political system as fragmented and fluid (for example, Dahl 1961; Rose 1967). Issues on the political agenda were seen as constantly shifting and as relatively specialized in their import. Allegiances that might coalesce around one issue would disband for the next and possibly be transformed on opposing sides of a later issue. No group was seen as the systematic loser. Instead, the system would deal with each issue, one at a time, responding to the investment of effort of contending parties without any accumulation of bias from the resolution of previous issues. Such a system prevented the buildup of a sustained social cleavage along class (or other) lines. Individuals belonged to multiple groups with shifting and sometimes conflicting interests (for example, Coser 1956; Hodge and Treiman 1968; Polsby 1980). The allegiances of individuals were too transitory and fragmented to provide any basis for the development of sustained conflict between groups.

Embourgeoisement, Postindustrialism, Postmaterialism

In a variant on the pluralist theme, some analysts have argued that the idea of riveting class conflict has been rendered obsolete by the increased level of affluence and altered occupational structure in advanced industrial societies. In an era of postindustrialism, the embourgeoisement of the working class, the expansion of the middle class, and the growth of the service sector of the economy were held to move society away from traditional class distinctions into more individualistic orientations toward personal betterment and into increasing concern with "postmaterialist" issues bearing on noneconomic aspects of the quality of life (see, for example, Dahrendorf 1964; Wilensky 1970; Bell 1973; Inglehart 1977, 1990; Lipset 1981). This line of thought, too, has assumed that if an exploitative relationship between the economic have's and have-not's were the principal organizing factor of social life, that relationship must find expression in passionate, conflict-laden ideologies. If the latter have wilted, the cause must lie

in changed economic relationships that have muted exploitation and thrown open multiple, competing foci for political energy.

Attempts to Salvage the Class Concept

On the other side of the debate about class are scholars in the "neo-Marxist" and "interest-group" schools of thought who wish to promote the concept of socioeconomic exploitation, either in the form of Marx's class categories or in some other class framework. These scholars have not strayed from the original conception of class as a source of conflict. This has confronted them with some awkward choices as they go about interpreting empirical data on class ideology which have repeatedly shown an embarrassing lack of outright conflict. Indeed, the lack of blatant conflict in contemporary class relations has been the most pressing theoretical challenge facing analysts of class ideology. Various resolutions of the problem have been pursued.

Accentuating Class Conflict

The simplest strategy has been to interpret data in a way that highlights the disparity of ideological positions between socioeconomic groups rather than dwelling on the extent to which the views of the groups converge (for example, Centers 1949; Huber and Form 1973). Meanwhile, analysts who are less sympathetic to the concept of class have interpreted similarly modest differences between groups as clear evidence of the nonexistence or demise of exploitative class relations (for example, Dahrendorf 1964; Nisbet 1970; Wilensky 1970; Polsby 1980; Lipset 1981). These disagreements tend to lock empirical debate into a dispute over "half-empty versus half-full," as one side bravely emphasizes the evidence of dissension between classes while the other side draws attention to the evidence of consensus.

Hidden Resistance

A second attempt to salvage conflict from the jaws of consensus has been to claim that oppositional orientations lie simmering beneath the surface among subordinate classes (for example, Scott 1985, 1990). Known usually as "hidden resistance," this argument has been very appealing to students of class. It postulates that the governed, although they appear to have acceded to their position, nevertheless are cognizant of their exploitation; and that although they refrain from engaging the dominant class in major, organized confrontations, subordinates routinely engage in small, personal acts of resistance that are hidden in the everyday practice of life. This argument rightfully calls attention to more subtle forms of behavior and reminds us that, as with all things social, we should not rely on outward appearances alone to decipher the levels and nuances of human interaction.

Unfortunately, however, the search for hidden, private forms of

resistance involves the application of an awkwardly weighted accounting procedure. Any action that might be interpreted as less than fully compliant, no matter how small, private, or inconsequential, is given disproportionate weight in making inferences about the state of consciousness of subordinates—meanwhile, unambiguous and socially significant acts of compliance are discounted as instrumentally contrived. This weighting procedure involves an awkward presumption about which actions are true reflections of someone's state of mind and which are either "false" or instrumentally contrived.

The hidden resistance argument also uses ambiguous rules of evidence, making it impossible to draw a clear empirical line between resistant and accommodative acts. Indeed, in Genovese's sensitive interpretation of day-to-day resistance among African-American slaves in the antebellum South (1974, 597-621), he discusses the ambiguity inherent in such behaviors as lying, stealing, dissembling, and shirking, and he refers to them as "simultaneous accommodation and resistance" (1974, 597). Although such behaviors may buffer subordinates from the worst exactitudes of their status, they also pander to the prejudices and foibles of the dominant class, thereby helping to perpetuate the inequality between the classes by feeding the dominant class's preexisting conceptions of the limitations of their subordinates. For these reasons, Genovese argues that hidden resistance qualifies "at best as prepolitical and at worst as apolitical" (1974, 598).

"False Consciousness" and Cultural Hegemony

There has been a third general attempt to grapple with the lack of class conflict within a conflict model, and that has been to attribute the passivity and conservatism of the working class to an artificially induced state of "false consciousness." There are slightly different variants of the false-consciousness argument, but their common feature is to credit the dominant ideology with a pervasive power that pummels the thinking of subordinates. The idea originates with Marx and Engels' famous dictum, laid out in The German Ideology in 1846:

The ideas of the ruling class are in every epoch the ruling ideas, i.e., the class which is the ruling material force of society, is at the same time its ruling intellectual force. The class which has the means of material production at its disposal, has control at the same time over the means of mental production, so that thereby, generally speaking, the ideas of those who lack the means of mental production are subject to it. (Marx and Engels [1846] 1970, 64)

As Abercrombie and Turner (1982) point out, this postulate flatly contradicts Marx's primary prediction that people's consciousness will be

shaped directly by their experiences, that is, that people rationally perceive and interpret their experiences and that from this ability stems the inevitability of eventual class conflict. According to the dictum of false consciousness, the working class, which is designated by Marxist theory as the key agent of change in the capitalist dialectic, fails to perceive its true interests and hence lapses into a deceived acceptance of its own oppression.

Analysts have explored this idea in slightly different ways. Mann (1970) argued that, because subordinate classes display stronger support for the tenets of dominant-class ideology when they are asked about relatively abstract principles than when they are asked about specific applications of those principles, this constitutes evidence that the conservatism of subordinate classes is "false." Mann's logic is that specifics are more likely to be perceived in their context and are thus perceived with the clearer vision that comes from direct experience, whereas abstractions are imposed by dominant ideology and represent an incomplete socialization attempt. Similar arguments about the disjuncture between the abstract and the applied perspectives of the working class have been made by Parkin (1971) and Huber and Form (1973), but without the terminology of "false consciousness."

Perhaps the most sophisticated treatment of false consciousness is to be found in the concept of "cultural hegemony" developed by the Italian communist, Antonio Gramsci (Gramsci 1971). Because it offers an interpretation of the absence of class conflict within a conflict perspective, the concept of cultural hegemony has proved alluring to an increasing number of scholars. And the complexity of Gramsci's ideas and the lack of precision with which he expressed them has left room for a variety of interpretations of the concept (see, for example, G. Williams 1960; Marcuse 1964; Femia 1975; Anderson 1976–1977; R. Williams 1980; Lears 1985; Bocock 1986). The essential argument is that all the major institutions of society are infiltrated by the dominant ideology, blunting the working class's perceptions and molding its values so that the class is rendered impotent as an agent of change. The cultural hegemony of the dominant class prevents the working class from realizing its true interests or, indeed, from developing a value system that would provide the ideological infrastrcture for revolutionary change. From Gramsci's point of view, as a political activist, this meant that changing working-class values was to be an important priority in the political agenda; without a concerted effort to alienate working-class values from those of the cultural hegemony, a true revolution (that would abandon capitalist values) could not be realized.

The concept of cultural hegemony, like that of hidden resistance, offers important insights into the dynamics of ideological control. But,

like the concept of hidden resistance, cultural hegemony is hamstrung by its confinement within a conflict model of class relations. Although various interpretations of cultural hegemony have tackled the concept slightly differently, none of them has been able to resolve the intrinsic difficulties. First, the depiction of working-class consciousness as "false" involves an a priori decision about what the true interests of the working class are (is this a form of analytic paternalism for scholars to pronounce what the working class's interests are, irrespective of what people in the working class themselves believe?), and it leads to inconclusive debates about how to assess the "true" interests of a group (see, for example, Connolly 1972). A related problem (Mann's efforts notwithstanding, 1970), is that the concept of false consciousness is not empirically falsifiable—the falseness of a particular ideological postion can only be determined on a priori grounds. A second and more fundamental problem is the intrinsic contradiction noted above between the ideas of material determinism and cultural ascendancy. If the dominant class prevents subordinates from perceiving their true interests by the imposition of a cultural hegemony, this implies a persistent lack of rationality among subordinates—but the primary prediction of class conflict rests pivotally on the assumption that humans process and act on their experiences rationally. Thus, in an attempt to salvage the prediction of class conflict, some analysts have undermined the rationality assumption on which the prediction of conflict rests.

Exploitation and Conflict

Serious interpretive ambiguities or conceptual contradictions plague all three attempts to preserve the concept of class conflict in the face of class quiescence. It is more than a little awkward to deal with the empirical absence of open conflict within a theoretical approach that is driven by the expectation of conflict as the rational expression of exploitative social relations.

Within the confines of the established debate about class, the cleanest interpretation for an observed lack of conflict remains a lack of exploitation. And, indeed, an analyst's interpretation of the extent of class conflict in a society is generally an accurate barometer of his claims about the significance of class organizationally. Thus, for Nisbet, class conflict is absent and therefore class does not exist (Nisbet 1970). For Lipset, class exists in muted form as one of many weak divisions in the industrial democracies, and thus the "democratic class struggle" presents a picture of muted class conflict crosscut by a variety of other political issues (Lipset 1981). For Centers, classes are enduring, socioeconomic interest groups, and this is manifested in significant ideological conflict between socioeconomic groups (Centers 1949). For Wright, classes are carved out by the

relations between different positions in the means of production, and herein lies the fuel that energizes class conflict (Wright 1979, 1985).

Empirical Questions and Measures

The lines of debate about class and class conflict have molded the kinds of questions that have been central to empirical analyses. A large literature developed around the assessment of subjective class identification (for example, Centers 1949; Hodge and Treiman 1968; Jackman and Jackman 1973, 1983; Kleugel, Singleton, and Starnes 1977), as analysts sought to determine whether and in what form class exists in the public consciousness. Note that there has been no comparable empirical effort to measure baseline identification with race or gender groups. For class, however, the underlying issue has been that if people do not identify subjectively with social classes, then classes do not exist. Considerable empirical effort has also been devoted to the assessment of people's perceptions of the economic system and their economic policy dispositions, to determine whether people seem to define and pursue conflicting class interests (for example, Centers 1949; Landecker 1963; Wilensky 1970; Mann 1970; Huber and Form 1973; Jackman and Jackman 1983).

In short, the very existence and definition of class has come to be linked critically to the notion of class conflict. Marx's conception of politics and culture as epiphenomena reflecting underlying economic relationships has been commonly accepted in its most literal, direct sense. If classes have conflicting economic interests, it is assumed that they must eventually find expression in ideological conflict.

Racial Prejudice As Irrational Antagonism

The analysis of racial prejudice in the United States drew strength from two main sources, one pragmatic and the other theoretical. On the pragmatic side, the analysis of racial prejudice was propelled by the clear, in fact pressing, existence of a racial cleavage in American society between whites and blacks. As Pettigrew brings out in his excellent history of American scholarship on race relations (Pettigrew 1980), from the early years of this century the starkness of the racial problem has spurred a social-reformist approach among analysts of prejudice. The overriding concern of these scholars has been to find an answer to the policy question: how can whites' hostility to blacks be reduced, possibly even eliminated?

The analysis of racial prejudice gathered theoretical legitimacy primarily from Max Weber's argument that status differentiation should be separated from the issue of economic differentiation and that the

former is a highly salient independent factor in the formation of social cleavages (Weber 1946a ). The way was thus opened for the analysis of racial prejudice as an important field of inquiry in its own right, not merely as a subsidiary of the structural problem of class relations. As I argued above, Weber's legacy was even to encourage the conception of racial prejudice as an alternative problem to class relations, rather than something that interfaced with it.

The result was that the analysis of racial prejudice generally became divorced from the analysis of class—each corpus of research has operated under its own assumptions about the primacy of its subject matter and the critical dynamics of intergroup ideology. There have, of course, been some important studies in race relations that have examined the intersection of class and race stratification (for example, Dollard's classic book, Caste and Class in a Southern Town , 1937; Bonacich's work on split labor markets in the United States, 1972, 1975, 1976; more recently, William J. Wilson's influential books, The Declining Significance of Race , 1980, and The Truly Disadvantaged , 1987; and the subsequent flourishing line of research on the dual effects of economic pressures and racial discrimination on the quality of life for urban, poor blacks, such as Jencks and Peterson 1991; Massey 1990; Massey and Eggers 1990; and Massey and Denton 1993). Scholars have also occasionally argued that much of race prejudice may be due to class prejudice (for example, Bayton, McAlister, and Hamer 1956; Blalock 1967; Smedley and Bayton 1978). However, these studies have had little impact on the main thrust of theory and research about racial prejudice. The conception of racial prejudice as a product of independent cultural forces has been preserved.

If racial and ethnic cleavages did not have their origins in the class structure of society but in exogenous cultural factors, the human mind could be credited with an inventive capacity that Marx had not acknowledged. While the debate about class relations pertained to the structure of society, students of racial prejudice departed on their own path and came to define their problem in social psychological terms.

The Concept of Prejudice

The two propellants, pragmatic and theoretical, determined the definition of the problem and the approach to its analysis. Racial prejudice reflected an error of the human mind, a cultural anachronism that was out of keeping with democratic and scientific thinking. Gunnar Myrdal encapsulated this thinking when he posed the problem as "An American Dilemma"—the parochial anachronism of racial prejudice and discrimination embedded within a society that was governed by democratic, egalitarian ideals (Myrdal 1944). Analysts assumed that all racial and ethnic groups share the same, universal human values and that racial and ethnic

hostility must be driven by misperceptions and misunderstandings that grew out of ignorance and parochialism. Allport, in his seminal work The Nature of Prejudice ([1954] 1979, 265), emphasized this reasoning with a parable:

See that man over there?
Yes.
Well, I hate him.
But you don't know him.
That's why I hate him.

From ignorance came misunderstanding, and from misunderstanding came hatred. This framework led Allport to conclude optimistically that "knowledge and acquaintance are likely to engender sounder beliefs concerning minority groups, and for this reason contribute to the reduction of prejudice" (Allport [1954] 1979, 268).

The task of progressive social reform was to eradicate parochialism and thereby to break down the walls of hostility that are a barrier to human understanding and social harmony. These themes were forcefully expressed by Marshall Field in his foreword to Deutsch and Collins' landmark study, Interracial Housing: A Psychological Evaluation of a Social Experiment :

There are barricades in the United States restricting and confining Negroes, Mexicans, Orientals, and other segments of our population to ghettos which are as real as the walls that surrounded the Jews in Warsaw. They are intangible; they exist in the customs of society and in the minds of men . . . . The walls of custom and belief which keep Negroes segregated from whites are not only a blot on our own national, democratic ideals but they are also a serious blow to our reputation among the nations. . . . This study . . . may be of great value to those who would replace superstition with science in making up their minds about interracial housing. (Deutsch and Collins 1951, v, vi, emphasis added)

The issues that motivated the literature on racial prejudice are captured in Allport's classic definition of prejudice:

Ethnic prejudice is an antipathy based upon a faulty and inflexible generalization. It may be felt or expressed. It may be directed toward a group as a whole, or toward an individual because he is a member of that group. ([1954] 1979, 9)

Allport emphasized two ingredients that identify prejudice: first, definite hostility and rejection, and second, rejection based on categorical (and therefore erroneous) criteria (Allport [1954] 1979, 5). Many other definitions of prejudice have been offered in the literature, but almost all share Allport's emphasis on irrationally founded antipathy as the core of the problem (see, for example, Newcomb, Turner, and Converse 1965,

430–431; Harding et al. 1969, 3–6; Ashmore 1970; Simpson and Yinger 1972, 24; Berry and Tischler 1978, 235; Kinder and Sears 1981; Pettigrew 1982, 1–5; Dovidio and Gaertner 1986, 2–3; Marger 1994, 74–75).

Prejudice As Hostility

The concept of prejudice puts racial hostility in center stage. Whereas the significance of conflict for analysts of class relations was as a reflexive indicator of underlying structural disunity, students of racial prejudice focused on hostility as the definitional core of the problem itself. This was the form in which race relations manifested itself as an ugly social problem, and the social problem was what motivated the analysis of race relations. If there were no intergroup hostility, there would be no problem that required analysis. That spirit is illustrated in an early bulletin published by the Social Science Research Council on The Reduction of Intergroup Tensions (Williams, Jr. 1947), which was "Prepared under the direction of the Committee on Techniques for Reducing Group Hostility." The mandate of that prestigious committee has continued to dominate the agenda of research on racial attitudes (see, for example, Stephan's chapter on "Intergroup Relations" in the third edition of The Handbook of Social Psychology , 1985).

Prejudice As Blatant and Unitary

By the same token, prejudice was assumed to have a unitary character. Because prejudice was a purely expressive phenomenon (rather than a politically motivated articulation), the irrationally founded hatred on which it was based might manifest itself in a variety of modes—beliefs, feelings, social avoidance, discrimination. In the prejudiced individual, hostility would dominate all his dispositions toward the group in question or toward individuals from that group. Measurement might rely on any of these indicators, and assessment of the general extent of the problem in society could draw on any measure that reflected intergroup negativism. And because they were a product of parochialism, the personal feelings of antipathy that drove prejudice were also expected to find blatant and unsubtle expression. Prejudice was conceived as an uncontained expression of parochial negativism.

Prejudice As an Individual Phenomenon

Finally, because prejudice was divorced from the question of the economic structure of society, it was conceived as a phenomenon that resided in people's heads and therefore as a property of individuals. The way to understand the problem was to analyze the variance across individuals in expressed racial hostility and to match this with individual differences in personality (for example, the theory of the "authoritarian personality") or cultural experience (for example, the theory of "working class authoritarianism" and the "contact

theory"). As others have observed, this conception of the problem in individually based psychological terms also carried over into the analysis of ethnic relations more generally because of the prominent position occupied by the concept of prejudice in the literature on race and ethnic relations (Schermerhorn 1970, 6–8; Wilson 1973, 3–5; Wellman 1977, 20–22; Pettigrew 1980, xxxi-xxxii; Giles and Evans 1986).

Dominant Hostility and Subordinate Passivity

The hostility manifested in prejudiced attitudes has been regarded as an undesirable outcome by those who study it. This is, of course, implicit in what I have said in the preceding pages, but it deserves explicit note, because it contrasts sharply with the literature on class consciousness. Students of socioeconomic inequality have been divided as to whether they regard that phenomenon as functionally integrative or exploitative. Only the latter group have seen conflict as an integral part of the phenomenon, and they have viewed class hostility as a desirable outcome, the fuse that would tear apart the current structure of society and forge positive social change. Students of prejudice, however, have been unified in their perception of the existence of racial hostility. And because they have viewed the hostility as lodged in people's heads (rather than being driven by structurally induced differences in interests), it is viewed as an undesirable property that has damaging effects on everything that it touches, from the moral certitude of society as a whole to the reduced opportunities and self-esteem of the unfortunate target group.

Those who regarded socioeconomic inequality as the product of exploitative relations assumed hostile attitudes on the part of those above toward those below, and they eagerly sought evidence of a reciprocated hostility from those below. Marx's description of capitalist society as divided into "two great hostile camps" exudes an excitement that has never been shed in the search for hostility among the victims of exploitation in the class relationship. The development of the prejudice literature, however, reflects an overwhelming concern about the degree of hostility in the racial attitudes of the dominant group, whites. The victims, blacks, were generally regarded in passive terms. Whites constituted the most probable agents of change, in part because the locus of the problem was in the minds of whites and in part because blacks had been rendered relatively powerless by the discriminatory practices that white prejudice had installed.

Students of socioeconomic inequality viewed the lower echelons in more active terms and focused primarily on their attitudes, as indicative of either spontaneous approval of inequality or assertive decrying of the system. By contrast, students of prejudice viewed the essential energy as coming from the dominant group, whites. To the extent that blacks'

attitudes were investigated, it was primarily in passive terms, to assess the damage wrought by whites on the black psyche (see, for example, Clark and Clark's classic study of ethnic self-devaluation among black schoolchildren, [1947] 1958; and Brand, Ruiz, and Padilla 1974). Studies by Johnson (1943) and Williams, Jr. (1964) were important early exceptions that examined the racial attitudes of both whites and blacks. For most scholars, it was not until the black urban riots of the late 1960s that there was a shift in emphasis, and that shift has been only partial and gradual. Interest in blacks' racial attitudes was initially propelled by questions about the new social problem of mass black violence (for example, G. Marx 1967; Campbell and Schuman 1968), but increasingly black identity and black consciousness emerged as a legitimate field of inquiry in its own right (see, for example, Schuman and Hatchett 1974; Gurin and Epps 1975). The realization has crept into the field that a full investigation of racial attitudes requires the inclusion of blacks' as well as whites' perceptions and attitudes (Schuman, Steeh, and Bobo 1985; Sigelman and Welch 1991). At the same time, however, inclusion of blacks in analyses has not led to any fundamental recasting of the dynamics of racial attitudes in more politically interactive terms. The concept of prejudice has not been dislodged and the powerful core of the problem is still seen as residing with whites. Assessments of racial prejudice among whites have continued to flourish as a discrete line of inquiry (for example, Selznick and Steinberg 1969; Campbell 1971; Katz 1976; Fairchild and Gurin 1978; Pettigrew 1979; Sears, Hensler, and Speer 1979; Kinder and Sears 1981; Apostle et al. 1983; Dovidio and Gaertner 1986; Sniderman et al. 1991).

Realistic Group Conflict Theory

A divergent strand in the prejudice literature, usually identified as realistic group conflict theory, has been attracting growing interest in recent years. This loose school of thought warrants separate discussion because it represents a different attack on the problem of racial attitudes which abandons many of the central tenets of the prejudice literature; at the same time, it does not abandon the primary concern with conflict as the core factor in intergroup attitudes. The exact arguments made by scholars using the realistic group conflict perspective vary somewhat, but the perspective may be described as roughly analogous to the interest-group approach in the class literature: racial groups are seen at least in part as having distinctive interests that engender conflict. LeVine and Campbell (1972, 29) provide a clear summary of this approach:

This theory assumes that group conflicts are rational in the sense that groups do have incompatible goals and are in competition for scarce re-

sources. Such "realistic" sources of group conflict are contrasted with the psychological theories that consider intergroup conflicts as displacements or projective expressions of problems that are essentially intragroup or intraindividual in origin. . . . Not all of them eschew psychological explanations. . . . But for all, realistic sources of group conflict are a primary emphasis.

Blumer's essay on "Race Prejudice as a Sense of Group Position" (1958) provides the roots (albeit rather tenuously) for this school of thought. Blumer made the case that prejudice should not be seen as a property of individuals, but of groups seeking to protect their position against possible incursions by subordinates. He introduced the idea of prejudice as motivated by feelings of threat, but he held back from locating those feelings in tangible advantages enjoyed by the dominant group. Blumer's "sense of group position" was a subjectively shared, group-level "sense" rather than a response to tangible, group-level interests (see also his discussion of the impact of industrialization on race relations, 1965). The small school of thought that has grown from Blumer's argument has generally assigned a clearer role to tangible group interests (see, for example, Wellman 1977; Giles and Evans 1986), although an irresolution about the extent to which prejudice derives from irreconcilable, mutually opposed group interests or merely subjectively felt feelings of threat characterizes some of the work in this approach (see, for example, Bobo 1983). The prevailing social-reformist goals of the prejudice literature have pervaded even this school of thought to some degree, constraining espousal of an interest-group approach with its attendant enthusiasm for conflict as a positive rather than as a negative outcome. Doubtless an influential factor here has been that, although realistic group conflict theory implicitly views subordinates in more active terms as part of a power relationship, empirical analysts in this framework have shared with most other analysts of prejudice a primary concern with the attitudes of the dominant racial group, whites (see, for example, Wellman 1977; Smith 1981; Bobo 1983; Giles and Evans 1986). When viewing the attitudes of the more powerful group, expressions of hostility have connotations that analysts find harder to endorse.

Thus, this small but growing school of thought has moved away from the traditional conception of prejudice as an entirely irrational phenomenon lodged in individuals' heads and has instead stressed the need to be attentive to group-level pressures and interests. However, scholars in this school have continued to be attentive primarily to the attitudes of the dominant racial group, whites. Perhaps for this reason, their view that conflict follows inevitably from opposed group interests is accompanied by less enthusiasm than it is among interest-group theorists of class, who focus more on the attitudes of subordinates.

Reassertions and Modifications of the Prejudice Concept

The traditional concept of prejudice has remained essentially intact over the years, and some of its elements have even been reaffirmed. However, there has been a growing sense among students of prejudice that the concept needs some modification and refinement to address the complexities of contemporary racial prejudice.

The Cognitive School

The conceptual proclivities of the traditional prejudice literature have been reasserted in the recent explosion of research on intergroup attitudes conducted within what is broadly termed the cognitive school. Experimental research on the "minimal group paradigm" has promoted the idea that the creation of even a "minimal group" (on the basis of an arbitrary criterion or even random assignment, with no personal interaction among fellow "group" members) generates own-group bias and out-group discrimination, in terms of both evaluations and allocation of rewards (see, for example, Tajfel 1969, 1978; Billig 1976; Brewer 1979).

A related, rapidly growing body of research has focused on stereotyping and discrimination as an outgrowth of normal social categorization and social comparison processes (see, for example, Tajfel 1969; Hamilton 1979, 1981; Pettigrew 1979; Stephan 1985). The assumptions and predelictions of that literature are captured in an empirical question posed by Taylor (1981, 98): "Given that stereotyping appears to be rooted in basic categorical processes, what are the individual differences that would influence the propensity to categorize or the ability to make within-category discriminations?"

Although some of this research has been attentive to the influence of power or status differentials on the shaping of individuals' perceptions (see, for example, Tajfel and Turner 1979; Tajfel 1982), the governing premise has been psychologically reductionist: that intergroup hostility is rooted in cognitive and motivational mechanisms that are intrinsic to individual behavior. Like Allport before them, these scholars have conceived of the problem in terms of affectively constrained cognitions, and they have directed their energies toward specifying the mechanisms that convert basic cognitive processes into prejudice and discrimination.

"Modern Prejudice"

At the same time, there has been an increasing appreciation over the years that the traditional conception of prejudice is inadequate to reflect the complexities of contemporary racism. Particularly troubling to investigators has been the repeated observation of a sharp discrepancy between, on the one hand, whites' apparent progress toward racial liberalism on traditional measures of prejudice, and, on the other, their continued resistance to specific policies of affirmative

racial change (see, for example, Jackman 1978, 1981a ; McConahay, Hardee, and Batts 1981; Schuman, Steeh, and Bobo 1985; Schuman and Bobo 1988). The ensuing debate about whether white racial prejudice has "really" declined quickly devolved into disagreements about what the critical elements of interracial attitudes are.

I have suggested elsewhere that the concept of prejudice be abandoned in favor of a conception of interracial attitudes that views them as politically motivated communications to defend group interests rather than as expressions of parochial negativism. In the context of political challenge from blacks, many of the symbolic messages in whites' racial attitudes have become more moderate and subtle, but whites continue to defend their privileged position by opposing specific policies for affirmative racial change (Jackman and Muha 1984). This approach puts primary emphasis on whites' support for specific affirmative racial policies as a gauge of their dispositions toward blacks (Jackman 1978, 1981a , 1981b ; Jackman and Muha 1984). As whites seek to maintain their privileged position in the face of challenge, they have moved away from inflammatory racial epithets and learned to rely instead on the seemingly more neutral principle of individualism.

Some other scholars have shared my position that whites' racial attitudes have become more subtle and nuanced, rather than clearly more positive, but they retained the basic conception of prejudice as an irrationally founded antipathy. The theories of "symbolic racism" (McConahay and Hough 1976; Kinder and Sears 1981; Kinder 1986) and "modern racism" (McConahay, Hardee, and Batts 1981; McConahay 1986) have maintained that prejudice lives on in the white American psyche but that it has found a more subtle expression that is more in keeping with contemporary cultural norms. The expression of prejudice becomes manifest in specific policies that bear on race relations: these policies are thought to symbolize whites' longstanding repugnance for people who are seen as violating the old-fashioned American values of individualism and self-reliance.

Both these arguments place considerable emphasis on the norm of individualism in the shaping of whites' racial attitudes, although the precise way that individualism and racial attitudes are thought to intersect is different. Individualism is represented either as a politically driven rationale for opposition to affirmative policy interventions (Jackman and Muha 1984) or as an intrinsic part of the prejudice syndrome (e.g., Kinder and Sears 1981). Still other scholars have rejected both these interpretations and have maintained that individualism is an independently held American value that unfortunately confounds policy interventions on behalf of blacks even as whites have become genuinely less prejudiced (for example, Sniderman and Hagen 1985).

Indeed, several scholars have voiced considerable irritation with the new conceptions of racial attitudes. For example, Sniderman and Tetlock conclude their critique of "symbolic racism" with an ardent allusion to "genuine prejudice—a deep-seated, irrational insistence on the inferiority of blacks, and contempt and hostility toward them" (1986, 186). Nonetheless, whatever their predilections, scholars have begun to distinguish in their terminology between "traditional prejudice" and "modern prejudice" (or "new racism"), the former being more blatant and unitary and the latter more subtle and nuanced (see, for example, Firebaugh and Davis 1988; Bobo and Kluegel 1991; Sniderman et al. 1991). And the role of individualism in the formation of racial attitudes remains prominent but uncertain in its meaning.

Empirical Questions and Measures

The questions and measures that have occupied a central position in empirical research on prejudice reflect the themes I have discussed. Analysts have probed into what aspects of individuals' backgrounds and experiences might account for variance in their racial prejudice. The emphasis on prejudice as an expression of parochialism has led to a continuing interest in the effects of formal education in reducing prejudice (see, for example, Lipset 1960; Selznick and Steinberg 1969; Jackman and Muha 1984; Bobo and Licari 1989). The individual, psychological cast of the prejudice concept led some researchers to investigate prejudice as an aspect of personality (most notably, Adorno et al.'s theory of the authoritarian personality, 1950). The casting of prejudice as a symptom of authoritarianism has continued to hold an allure for empirical scholars, especially among those who have seen education as a positive influence (for example, Lipset 1960; Selznick and Steinberg 1969; Bobo and Licari 1989). The question continues to hang over research on prejudice about the extent to which it is a product of individual personality or individual experience (for example, Pettigrew 1958; Allport 1962; Selznick and Steinberg 1969; Kinder and Sears 1981; Yinger 1983). The social-reformist spirit of prejudice research has flowered in the longstanding prominence of the contact theory of prejudice, which postulates that increased intergroup contact, under the right conditions, reduces prejudice (for example, Deutsch and Collins 1951; Wilner, Walkley, and Cook 1955; Miller and Brewer 1984; Jackman and Crane 1986; Kinder and Mendelberg 1991). Social-reformist concerns have also produced a long tradition of studies devoted to monitoring the trends in whites' racial attitudes over time (for example, Hyman and Sheatsley 1956, 1964; Greeley and Sheatsley 1971, 1974; Schuman, Steeh, and Bobo 1985; Firebaugh and Davis 1988).

The most common measures that have been used in analyses of preju-

dice are those reflecting stereotypical beliefs, feelings of social distance, and support for racial integration and other affirmative racial policies. Stereotypes were conceived as irrational, negative, categorical beliefs about a group. Because of their propinquity to the essence of prejudice itself, stereotypes assumed a preeminent importance, and a vast literature developed that was devoted to their conceptual and empirical analysis (see, for example, Lippmann 1922; Katz and Braly 1947; Vinacke 1957; Karlins, Coffman, and Walters 1969; Brigham 1971; Mackie 1973). Interest in stereotypes has continued to thrive, promoted by social psychologists in the cognitive school (for example, Tajfel 1969; Hamilton 1979, 1981; Hamilton and Trolier 1986) and pursued by other students of prejudice as well (for example, Selznick and Steinberg 1969; Dovidio and Gaertner 1986: Bobo and Kluegel 1991). There has also been a growing interest in assessing people's explanations of the causes of racial differences: do people think racial differences are physically inherent, cultural, or caused by societal discrimination and bias? (See, for example, Pettigrew 1979; Apostle et al. 1983; Kluegel and Smith 1986.)

The concept of social distance was devised by Bogardus (1925). In a historical era when racial inequality was marked by physical segregation, the notion of personal acceptance was readily adopted as a vital signal of the breakdown of negative attitudes (see, for example, Deutsch and Collins 1951; Greeley and Sheatsley 1974).

For the same reason, rejection of racial segregration became widely interpreted as laden with significance for the breakdown of negative racial attitudes (see, for example, Hyman and Sheatsley 1956, 1964; Campbell 1971; Greeley and Sheatsley 1971; Taylor, Sheatsley, and Greeley 1978). And in recent years, analysts have become increasingly interested in probing whites' support for various affirmative racial policies, as they have tried to grapple with the discrepancy between the linear decline in whites' subscription to traditionally racist ideas and their continued resistance to affirmative policies of racial change (for example, Jackman 1978; Schuman, Steeh, and Bobo 1985). That discrepancy has engendered an increasing debate about how best to measure prejudice in order to capture the nuances of its contemporary expression (for example, Jackman 1977, 1978, 1981b ; Crosby, Bromley, and Saxe 1980; McConahay, Hardee, and Batts 1981; Pettigrew 1985; Kinder 1986; McConahay 1986; Schuman and Bobo 1988).

Political Tolerance As Hostility Defused

The literature on political tolerance developed independently of the two currents of research that we have just considered on class conflict and racial prejudice. Some familiar strands reappear in research on

tolerance, but they are fused uniquely to produce an alternative approach to the issue of intergroup conflict.

Like some analysts of class, students of tolerance accepted as a given that society was comprised of groups with contending interests. However, students of tolerance did not concern themselves with the root causes of those competing interests, whether they reflected deep rifts in the organizational structure of society or mere differences in beliefs, lifestyle, or skin color. Like analysts of prejudice, they did not view the unbridled expression of intergroup hostility as a progressive outcome, and they were especially concerned about the attitudes of more powerful groups toward political, religious, and ethnic minorities. But unlike students of prejudice, those who studied tolerance were not concerned with how to eliminate hostility—they concentrated, instead, on how to defuse it. Tolerance does not imply positive feelings toward a group—it means the equal treatment of those whom we dislike. In the words commonly attributed to Voltaire, "I disapprove of what you say, but I will defend to the death your right to say it" (Bartlett 1968).

Tolerance and Democracy

The analysis of tolerance evolved within the literature on empirical democratic theory. In that setting, democratic society was viewed as an unwieldy amalgam of contending groups: the stability of democratic institutions depended on a delicate balance between the dual principles of majority rule and minority rights. The first principle provided for the peaceful rule of the declared winner in political competition; the second principle provided for the preservation of political competition by granting full civil liberties to the losing side (Dahl 1956). Within this context, tolerance of those with whom we disagree (whatever the basis of the disagreement) is vital to democratic stability.

The initial concern was with the issue of maintaining freedom of speech so that majority rule would not deteriorate into the tyranny of the majority. The empirical literature on tolerance was launched in the aftermath of the McCarthy era with Stouffer's book, Communism, Conformity, and Civil Liberties (1955). The title to the book reflects the concern that many liberals felt at that time over the threat that McCarthy's anti-communist witch-hunt posed for the democratic principle of freedom of speech. Stouffer focused on the extent to which political leaders and members of the mass public were inclined to restrict the civil liberties of political and religious nonconformists—communists, socialists, and atheists. Most of his questions centered on freedom of expression directly, but he also addressed the broader civil libertarian issues implied in hiring and firing decisions. In a historical era when the extension of full civil rights to blacks became a prominent public issue, the broader

significance of tolerance was highlighted, and some analysts applied the concept to the arena of black civil rights (for example, Prothro and Grigg 1960; Jackman 1977, 1978). The primary concern of the literature has remained with attitudes that bear specifically on the openness of political life (see, for example, Sullivan, Piereson, and Marcus 1982; McClosky and Brill 1983), but tolerance has assumed a more general application to the broad democratic principle of "equality of opportunity, regardless of race, color, or creed."

The ability to disconnect one's behavioral dispositions from one's personal feelings toward a group was regarded as a uniquely democratic quality. Thus, students of tolerance, like their counterparts who studied prejudice, implicitly regarded the primary institutions in democratic society as a progressive force that offered the promise of a more enlightened existence. The more people were exposed to democratic institutions, either through formal education (for example, Stouffer 1955; Lispet 1960; Prothro and Grigg 1960) or through active political participation (for example, Stouffer 1955; McClosky 1964; McClosky and Brill 1983), the more eroded would become the raw tendency to take action against groups who are the object of dislike. Democratic institutions were seen as mechanisms that socialize people into the benefits of diversity and teach them the reciprocal advantages of removing positive personal feelings as a requirement for positive treatment.

The Resolution of Conflict through Consensual Rules of the Game

Although analysts of tolerance have not explicitly concerned themselves with the issue of what underlies the divisions between social groups or the nature of the relationship between unequal groups, implicit assumptions about these questions do provide the contextual logic for the concept of tolerance. It is important to remember that the issue of tolerance was raised within the framework of empirical democratic theory. That body of theory specifies that democratic stability requires a fundamental working consensus among the populace on rules of the game; such a consensus is not considered viable in a society marked by deep cleavages (e.g, Dahl 1956; Almond and Verba 1963). Thus, although analysts of tolerance deviate from those who study prejudice in their acceptance of group conflict as a normal part of life, this deviation is not as great as it might seem, because the basis for the conflict must be sufficiently shallow to prevent intense feelings from being aroused.

But if the inequality between groups is in fact expropriative, where one group's gain is the other's loss, there might be too much at stake for tolerance to be practiced with ease. Add to this the awkward point that the most salient rifts in society do not involve relationships among equals who are buffeted about in a political game of give and take, trading places

from time to time as winners and losers. In the most serious cleavages that American society faces—those based on race, class, and gender—it is not too difficult to identify which groups are repeatedly winners and losers.

The concept of tolerance implicitly neutralizes the issue of social inequality. By directing attention to the defusing of conflict without regard to its underlying causes, it places more weight on societal stability than on the substantive resolution of group differences. What takes precedence is the procedural issue of how the game will be played, rather than the substantive issue of what the game is about. The libertarian concern with freedom of speech and equality of opportunity more generally does not address the problem of the unequal positions from which different groups may avail themselves of those opportunities (Schaar 1967; Parkin 1971; Jackman and Muha 1984). Instead, the concept of tolerance lends itself more easily to the pluralist view of a society unmarked by deep rifts between have's and have-not's. And, indeed, some of the most prominent contributors to our empirical understanding of tolerance have promoted such a view (for example, Stouffer 1955; Lipset 1960; McClosky 1964).

Rising above Hostility

Like students of prejudice, those who study tolerance are interested in getting society beyond conflict. Students of prejudice implicitly assume that there are no hard substantive grounds for racial conflict, but that intense interracial dislike can be generated by malleable cultural forces. Thus, the critical problem for them is to undermine the negative intergroup feelings that drive discrimination. Students of tolerance implicitly assume that intergroup hostility is not intense, because the gap in interests between contending groups is neither profound nor irreconcilable. If there are no grounds for intense emotional rift, one need not worry about feelings per se. Instead, one should concentrate on the issue of dislodging behavioral predispositions from their affective base. The task is not to eliminate intergroup hostility (as would students of prejudice) or to debate the substantive grounds for its existence (as would students of class relations). The normality of some intergroup disparity is accepted without question—the problem is how to defuse it so that it does not escalate and disturb the peaceful practice of democratic life.

Empirical Questions and Measures

The central empirical questions driving analyses of tolerance have been the assessment of its prevalence among the democratic citizens of the United States (for example, Stouffer 1955; Sullivan, Piereson, and Marcus 1979; Abramson 1980), the measurement of change in levels of tolerance over time (for example, Sullivan, Piereson, and Marcus 1979;

Abramson 1980), and examination of what factors (personality, democratic socialization, education, or demographic characteristics) produce political tolerance (for example, Stouffer 1955; McClosky 1964; R. Jackman 1972; Sullivan, Piereson, and Marcus 1982; McClosky and Brill 1983). Empirical analyses of tolerance have thus been governed by very similar kinds of questions to those found in research on racial prejudice. And, to the extent that analyses of prejudice have been pervaded by Myrdal's portrayal of prejudice as an anachronism that jars against the American commitment to democratic values and ideals (Myrdal 1944), much of the tone of analyses of tolerance also resembles prejudice research.

Empirical measures of tolerance, however, have developed along very different lines than measures of prejudice. Analysts of tolerance have usually relied exclusively on indicators of people's policy or behavioral dispositions, with positive responses being interpreted as evidence of tolerance (for example, Stouffer 1955; McClosky and Brill 1983). Students of prejudice, with their more holistic approach to attitudes, might interpret the same response as reflecting an overall favorable attitude toward the group—the elimination of prejudice. This disjuncture between the concept of tolerance and its empirical measurement has been noted by Jackman (1977, 1978) and Sullivan, Piereson, and Marcus (1979, 1982). Sullivan and his colleagues found that when one takes explicit account of people's feelings toward specific groups, the granting of full civil rights to groups that are disliked is a rare phenomenon—considerably less common than what is ostensibly required for the democratic functioning of society.

Gender Relations As Hostility Revealed

Relations between men and women were largely neglected by scholars until relatively recently, perhaps because the manifest docility of gender relations did little to excite observers' attention (and the fact that scholarly observers tended to be male doubtless exacerbated the invisibility of anything problematic in relations between men and women). Gender relations were the subject of passing commentary from time to time, but a sustained literature on gender relations is a comparatively recent development, emerging in the shadow of already well-established bodies of research on class and race relations. If the relationship between men and women had failed to draw attention to itself by offering any blatant prospects of societal disruption, the task of those who sought to establish its significance as a field of inquiry was to reveal its dynamics in terms that analysts of class and race would clearly recognize—friction, violence, emotional disparity.

At the beginning of this chapter I quoted a passage from John Stuart Mill in which he astutely noted that gender relations are distinguished by a lack of reliance on force. Despite that insight, even Mill sought to engage the skeptic by highlighting elements of hostility and conflict in those relations. Indeed, the latter theme is the prevalent one in his essay on "The Subjection of Women":

In struggles for political emancipation, everybody knows how often its champions are bought off by bribes, or daunted by terrors. In the case of women, each individual of the subject-class is in a chronic state of bribery and intimidation combined. . . . If ever any system of privilege and enforced subjection had its yoke tightly riveted on the neck of those who are kept down by it, this has. . . . But, it will be said, the rule of men over women . . . is accepted voluntarily; women make no complaint, and are consenting parties to it. In the first place, a great number of women do not accept it. Ever since there have been women able to make their sentiments known by their writings (the only mode of publicity which society permits to them), an increasing number of them have recorded protests against their present social condition: and recently many thousands of them have petitioned Parliament. . . . The claim of women . . . is urged with growing intensity, while the demand for their admission into professions and occupations hitherto closed against them, becomes every year more urgent. . . . How many more women there are who silently cherish similar aspirations, no one can possibly know. . . . There is never any want of women who complain of ill usage by their husbands. There would be infinitely more, if complaint were not the greatest of all provocatives to a repetition and increase of the ill usage. (J. S. Mill [1869] 1970, 136-137, 139-140)

I have quoted from Mill at length because the language of his essay demonstrates so clearly the implicit assumption that if there were no conflict between the groups, there would be no problem worthy of analysis. In his effort to draw attention to the relationship between men and women, Mill portrayed it as ridden with violence and intimidation on the part of the oppressors and protest and simmering resentment on the part of the oppressed. He conceded, by implication, that if women were found to be "consenting parties," his case would be weaker, and he set about revealing the intensity and urgency of women's discontent. Later in the essay, Mill acknowledged the affection that many men feel for their wives, and he regarded this as something that mitigates the tyranny of men over women (Mill 1970, 161-162). The affection that women may feel for their husbands, he dismissed, however, as lodged in fear: "It is part of the irony of life, that the strongest feelings of devoted gratitude . . . are called forth in human beings towards those who, having the power entirely to crush their earthly existence, voluntar-

ily refrain from using that power" (Mill 1970, 162–163). Mill went on to appeal to the reader that it is unreasonable to judge any system of oppression by "its best instances":

Who doubts that there may be great goodness, and great happiness, and great affection, under the absolute government of a good man? Meanwhile, laws and institutions need to be adapted, not to good men, but to bad. . . . The tie of affection and obligation to a wife and children is very strong with those whose general social feelings are strong, and with many who are little sensible to any other social ties; but there are all degrees of sensibility and insensibility to it, as there are all grades of goodness and wickedness in men. . . . The vilest malefactor has some wretched women tied to him, against whom he can commit any atrocity except killing her. . . . When we consider how vast is the number of men, in any great country, who are little higher than brutes, and that this never prevents them from being able, through the law of marriage, to obtain a victim, the breadth and depth of human misery caused in this shape alone . . . swells to something appalling. (Mill 1970, 163–164)

The hapless reader might have drawn comfort from the apparent lack of friction in gender relations. Mill's strategy was not to question the logic of that derivation, but to undermine the validity of the empirical observation on which it is based. He endeavored to reveal in gender relations the same hallmark symptoms as those emphasized in more commonly recognized systems of oppression—violence and fear, hostility and resentment.

The Racial Analogy

In his essay, Mill drew several explicit analogies between gender relations and racial oppression. The same device was used, much later, by such varied advocates of the analysis of gender relations as Gunnar Myrdal (1944), Helen Hacker (1951), and Simone de Beauvoir (1953). In appendix 5 to his epic analysis of race relations in the United States, An American Dilemma (1944), Myrdal presented "A Parallel to the Negro Problem," in which he argued that women and blacks alike are oppressed by the lagged presence of a paternalism that was once nourished by a preindustrial economy but which is now but an atavistic element destined to be discarded in the contemporary, industrial world. Hacker, in her classic article, "Women as a Minority Group" (1951), rested her case on an enumeration of the parallels between the condition of women and blacks (highlighted by a chart entitled "Castelike Status of Women and Negroes" with parallel columns for the two groups). Where Hacker found an exact parallel between the condition of the two groups, she drew evidence for the case that women constitute a "minority group"; where she found the condition of women to deviate from that of blacks, Hacker

inferred that the minority status of women is mitigated somewhat. And de Beauvoir, too, drew on parallelism to establish her case:

Whether it is a race, a caste, a class, or a sex that is reduced to a position of inferiority, the methods of justification are the same. "The eternal feminine" corresponds to "the black soul." . . . There are deep similarities between the situation of the woman and that of the Negro. (de Beauvoir [1953] 1973, 683)

The analysis of the ideology of gender relations was thus established in a framework based on parallelism with race. The same concepts that dominated the analysis of racial attitudes were applied to gender attitudes: hostility, stereotyping, prejudice, and low self-esteem among subordinates. Simone de Beauvoir stated the case boldly:

It is easy to see that the duality of the sexes, like any duality, gives rise to conflict. . . . All oppression creates a state of war. And this is no exception. (de Beauvoir [1953] 1973, 681, 690)

She portrayed the relationship between the sexes as one of chronic animosity, punctuated by stereotyping of women and aggressive assertions of feminine inferiority on the one side and festering resentment and frustration on the other.

A similar argument is made by Rubin in her analysis of working-class family life, Worlds of Pain (1976). She portrays that life as infested with pain, not only in its economic constraints, but also in its cultural dictate of disparate worlds for husbands and wives. Those two worlds are intersected only by the uncaring negligence of men and the angry resentment of their wives. She concludes thus:

Are there good times? Yes, a birthday remembered with joy, a happy Christmas, a loving and tender moment between wife and husband. But they stand out in memory as unique and treasured events, monumentally important because they happen so seldom, because they are so little a part of daily experience. . . . As people talk about their lives, such small events become insubstantial, slipping away before a more compelling reality. (1976, 215)

Gender Violence

The theme of dreariness and disparity in gender relations has been carried a step further by some radical feminist scholars, who portray gender relations as ridden with blatant hostility and violence. The criminal act of rape is identified as epitomizing the general quality of the relationship between men and women—a relationship in which men are depicted as subjugating women to their will through sexual terrorism, backed up by physical and institutional force (see, for example, Millett 1970; Griffin 1980; Cronan 1984). Millett states:

Patriarchal force also relies on a form of violence particularly sexual in character and realized most completely in the act of rape. . . . In rape, the emotions of aggression, hatred, contempt, and the desire to break or violate personality, take a form consummately appropriate to sexual politics. . . . Patriarchal societies typically link feelings of cruelty with sexuality, the latter often equated both with evil and with power. (1970, 44)

Heterosexuality is depicted as intrinsically misogynist, an instrument purely to subjugate women (see also Rich 1980). In another gender-race comparison, Millett compares the crimes of rape and lynching, and she argues that "unconsciously, both crimes may serve the larger group [whites or men] as a ritual act, cathartic in effect" (Millett 1970, 45). Prostitution and marriage are described in like terms as sexual and political slavery for women (see also Goldman [1910] 1969; Barry 1979; Cronan 1984). Millett sums up the relationship between men and women thus:

The history of patriarchy presents a variety of cruelties and barbarities. . . . The rationale which accompanies that imposition of male authority euphemistically referred to as "the battle of the sexes" bears a certain resemblance to the formulas of nations at war, where any heinousness is justified on the grounds that the enemy is either an inferior species or really not human at all. (1970, 46)

Women's Resistance—Or Lack Thereof

The lack of concerted resistance among women has left scholars who depict gender relations as conflictive with an empirical shortfall. As in some accounts of class and race consciousness (Genovese 1974; Scott 1985), some scholars of gender relations have attempted to reconcile the discrepancy by emphasizing the many ways in which women engage in personal behaviors that constitute hidden or private forms of protest (Cloward and Piven 1979; Anyon 1984). The same interpretive difficulties plague this argument in its application to gender relations as I noted in the earlier section on class ideology. As with the debate about class consciousness, the theoretical significance attached to the presence or absence of conflict is so great that some scholars have felt pressed to scour the minutiae of the relationship to uncover behaviors among women that might be construed as showing evidence of alienation and resistance.

Conservative Theoretical Pressures

The search for conflict in gender relations reflects not only the habits learned from the analysis of class and race relations. It has also been spurred, perhaps even more urgently than in the class literature, by the presence of powerful conservative arguments that gender differentiation

reflects a "natural order" that is manifested in its consensual endorsement by both men and women. Parsons and Bales (1955) argued that gender differentiation is socially functional, serving to integrate the family unit with the larger society while providing for the nurturance and socialization of society's youngest members. The specific "expressive" role of women and "instrumental" role of men were explained as the natural outcome of differences between the sexes in their biological contributions to procreation.

There has also been a large body of literature that promotes the ideas of biological determinism in gender relations. Numerous analysts have argued that there are natural differences in male and female personality types stemming from physically complementary functions in procreation: for example, that genital penetration of the female by the male requires that the male be assertive and the female submissive, that the length of the pregnancy requires nurturant feelings and emotional commitment from the mother whereas the father's physical contribution is fleeting, and so on (see, for example, Storr 1968; Tiger and Fox 1971; Wilson 1978; Bettelheim 1984).

Perhaps because of the readily identifiable physical difference between men and women, combined with the pervasiveness of gender differentiation at the heart of everyday life, conservative arguments based on functionalism and biological determinism have had a deeper impact on the shape of the literature on gender relations than they have on race relations or even class relations. There has been no serious scholarly attempt in the past few decades to explain racial differentiation as functional. In the early years of this century, biological determinism had a central position in American scholarly debate about race (see, for example, the debate published in the American Journal of Sociology in 1908 between Stone and Willcox et al., as well as Pettigrew's history of American scholarship on race relations, 1980). But the scholarly impact of biological determinism on research about race has waned over this century. A few analysts have persisted with the question of whether there are genetic differences between whites and blacks in intelligence (Jensen 1969; Shockley 1971, 1972a , 1972b ; Herrnstein 1990), but such arguments have had little impact on the main thrust of research on either racial inequality or racial prejudice (see, for example, the comprehensive report on the status of race relations in the United States sponsored by the National Academy of Sciences, Jaynes and Williams, Jr. 1989). Indeed, scholars of race relations have tended to dismiss arguments about immutable racial differences as Victorian or fascist. In the explanation of socioeconomic inequality, functionalism had a major impact, although still perhaps not quite to the extent that it has in the gender literature. And more blatant biological determinism, in the form of social Darwin-

ism (Spencer [1862] 1958; Sumner 1883), made only a brief visitation to the literature on class. Yet biological determinism continues to be a major influence on the debate about gender differentiation (see, for example, Maccoby and Jacklin 1974; Rossi 1977a ; Gross et al. 1979). A "natural order" is expected to manifest itself in spontaneous consensual endorsement: those who would depict gender inequality as other than "natural" or functional have thus taken the reflexive counterposition that gender relations are fraught with tension and conflict.

Empirical Questions and Measures

The empirical analysis of gender attitudes has been shaped by the theoretical currents outlined above, although not always explicitly. Feminist theorists with a more critical perspective have often expressed philosophical objections to systematic, public-opinion research in the "positivist" tradition (see, for example, Stacey and Thorne 1985), and survey researchers interested in gender attitudes have generally returned the compliment by ignoring theoretical work that depicts gender relations in political terms. Indeed, public-opinion studies of gender attitudes have usually had an empirically driven agenda, proceeding to the empirical business with little or no discussion of theoretical issues. However, the kinds of empirical questions that have been asked primarily reflect the influence of the liberal tradition, exemplified by John Stuart Mill and Gunnar Myrdal, juxtaposed with the functionalists and biological determinists. In more recent years, there has been an increasing tendency for empirical analysts to draw on theories that depict gender relations as expropriative (for example, Gurin 1985; Kane 1989). The resulting blend of empirical questions in research on gender attitudes shares some features with the literature on class consciousness, and some with the literature on prejudice.

Students of social class invested particularly in the attitudes of the working class. Both Marxist and functionalist theories put a heavy onus on the have-not's either to challenge or to endorse socioeconomic inequality, and this is reflected in the empirical literature on class consciousness (for example, Lane 1962; Leggett 1968; Goldthorpe et al. 1969; Hamilton 1972; Huber and Form 1973; Newby 1977b ; Jackman and Jackman 1983). So too have the gender attitudes of women been the object of particular empirical scrutiny. There are numerous studies of women's gender attitudes taken alone (for example, Goldberg 1968; Horner 1972; Parelius 1975; Mason, Czajka, and Arber 1976; Thornton and Freedman 1979; Gurin 1985), along with an increasing number that compare men's and women's attitudes (for example, Sherriffs and McKee 1957; Broverman et al. 1972; Cherlin and Walters 1981; Simon and Landis 1989; Kane 1989).

Most of these studies have focused on attitudes about "sex-roles," rather than on attitudes about power relations between the sexes, suggesting an almost exclusive influence of functionalist ideas. The net result was that gender-attitudes research shared the emphasis of class-ideology research on the attitudes of subordinates, but without the attendant conception of gender relations as exploitative. Recently, however, there has been a growing awareness of this deficiency in gender-attitudes research, and an increasing number of studies are incorporating measures that show sensitivity to the power dynamics of gender relations (see, for example, Gurin 1985; Kane 1989; Simon and Landis 1989; Kane and Sanchez 1992).

There has been little tendency to follow the lead of the prejudice literature by focusing exclusively on the attitudes of the dominant group (Goode's essay on "Why Men Resist" [1982] is a rare example of a discrete discussion of men's gender attitudes.) At the same time, the liberal, social-reformist sentiment of the prejudice literature is shared by many students of gender attitudes, and that tradition too has shaped the kinds of questions that have been asked about gender attitudes—and especially women's gender attitudes.

These dual influences from research on class consciousness and racial prejudice have resulted in two broad themes of empirical inquiry about gender ideology. The first of these is a focus on prejudice against women and on women's self-esteem. A repeated question has been whether women suffer from low self-esteem and poor achievement orientation. There have been numerous studies of gender stereotyping, devaluation of women, and low achievement orientations among women. The language of these studies parallels that of the literature on racial prejudice. For example, Sherriffs and McKee state that "both men and women esteem men significantly more highly than women" (1957, 451); Broverman et al. conclude that "sterotypically masculine traits are more often perceived to be desirable than are sterotypically feminine characteristics" (1972, 75); Goldberg answers "yes" to the question "Are women prejudiced against women?" (1968); and Horner declares that women have a "fear of success" because their self-conceptions are tied to low status (1972).

A second major line of empirical inquiry has been the extent to which people, and especially women, endorse the traditional division of labor between the sexes and, more recently, the division of power and economic opportunities between the sexes (for example, Mason, Czajka, and Arber 1976; Thornton and Freedman 1979; Spitze and Huber 1980; Cherlin and Walters 1981; Helmreich, Spence, and Gibson 1982; Thornton, Alwin, and Camburn 1983; Gurin 1985; Mason and Lu 1986; Simon and Landis 1989; Kane 1989). Analysts have been interested both in

monitoring the overall level of support for "nontraditional" gender arrangements and in assessing what individual factors (personality, socialization, personal experience, or demographic characteristics) affect gender attitudes. And as in the prejudice literature, there has been a special interest in monitoring change in gender attitudes over time, as an indicator of the wave of liberal progress. But unlike students of racial attitudes, who have been largely unplagued by notions of a "natural" or functional order of racial differentiation, students of gender attitudes have not been free to lay the onus of progress unequivocally on to the dominant group. Many analysts of "sex-role attitudes" appear to base their inferences about the pace of change on the attitudes of women alone, implying that women's attitudes are the prime source of energy in determining the form of their relationship with men. Other analysts draw on concepts about class consciousness to view women's gender attitudes within the political context of an expropriative intergroup relationship in which men are the more powerful group.

Thus, the empirical analysis of gender attitudes reflects a blending of concerns from the theoretical and empirical traditions of class consciousness and racial prejudice. The resulting depictions of women are pervaded by two conflicting undercurrents: the female as victim (like the black), beaten by prejudice into abject submission, and the female as a conjectured collaborator (like some images of the poor and working class), contentedly endorsing her assigned position in life or moving assertively to alter it. In this context, women are held to have some complicity in their own fate in a way that blacks are not. It is implied that the onus is on women to indicate whether their station in life has become a painful anachronism or continues to be the fulfillment of their biological proclivities.

Summary

I have analyzed the prevailing concerns and assumptions that have governed the study of class, race, and gender attitudes. My purpose has not been to elucidate the full variety of ideas that are to be found in each corpus of research—such a purpose would require volumes in itself. My purpose, rather, has been to identify major themes that have shaped the development of each literature.

I have attempted to draw out two important points. First, from a common family of ideas, research on class conflict, racial prejudice, political tolerance, and gender ideology broke apart in disparate networks of development. Second, the abiding focal point for each network has been conflict and hostility. The result has been that, as each body of literature followed its own course, it developed a distinctive welding of ideas to

address a common and enduring concern with conflict versus consensus in the assessment of unequal social relations.

In the analysis of social class, the legacy of Karl Marx was to weld conflict to the very conception of class as the basis for expropriative social relations. According to Marx's model, the working class signals its realization of the exploitative basis of its relationship with capitialists by becoming ideologically alienated from the capitalist system and organizing to overthrow it. Those who debated with Marx or his ghost did nothing to dislodge the theoretical expectation that exploitation would lead to conflict, and, indeed, they reinforced that expectation as they argued backwards from the lack of empirically observed class conflict to claim that class was either not exploitative or that it did not exist at all. In the face of the empirical shortfall in class conflict, some scholars have attempted to salvage the central elements of Marx's class analysis by accentuating evidence of dissension among classes, by arguing that subordinate classes engage in "hidden resistance," or by depicting subordinates' consciousness as "false." I have argued that all these attempts contain important insights, but they are weakened by their ensnarement in the conflict model. That ensnarement causes analysts to stretch their arguments out onto a thin ledge where rules of evidence lose their clarity or contradictory assumptions undermine their theoretical consistency. Empirical analyses of class ideology continue to revolve around issues pertaining to the existence and importance of class as a basis for exploitative and conflictive relations, with particular attention to the political consciousness of subordinates.

The analysis of racial prejudice has been driven by the spirit of social reform. Students of prejudice have sought to eradicate racial discrimination by developing models that approach the phenomenon in terms of individual psychological dispositions. Analysts have focused on the psychological construct of prejudice—an irrational, parochial antipathy based on faulty, categorical perceptions of the target group. The intergroup beliefs, feelings, and behavioral dispositions of prejudiced individuals in the dominant group are thought to be locked together in a negative position. Some analysts have diverged on a theoretical strand that posits that prejudice is motivated by a sense of group position (rather than by irrational parochialism), and some analysts have argued that prejudice is being expressed in more subtle and nuanced ways than the traditional prejudice concept stipulates. But all conceptions of prejudice have retained the central emphasis on intergroup hostility as the core of the problem. With relatively little attention to the dispositions of subordinates, students of prejudice have focused on trying to monitor and explain the racial animosity of the dominant group, whites.

The concept of political tolerance grew out of democratic theory. It

was held that democratic systems require citizens to have consensus on rules of the game in order to resolve conflicts peacefully. According to this perspective, hostility and conflict are unavoidable aspects of social life, but democratic citizens must learn to respect the civil rights and civil liberties of all other citizens, even when they personally disagree with them or harbor animosity toward them. The problem for analysts of tolerance has been to assess how successfully democratic citizens have disconnected their behavioral dispositions from their personal feelings toward a group. Interest has focused primarily on the tolerance of dominant-group members toward political and racial minorities.

The relatively recent growth of research on gender relations has fed from themes that were developed in the analysis of class and race relations. In the attempt to draw attention to gender as a significant and problematic social relationship, analysts have depicted gender relations as ridden with the commonly recognized symptoms of oppressive systems—force, violence, and hostility on the one hand, and fear, resentment, and low self-esteem on the other. Even more than in the class literature, the analysis of gender relations has been wedged in by a theoretical debate that includes staunchly conservative arguments depicting gender differentiation as a "natural order" that is endorsed consensually by men and women. In this context, the manifest passivity of women has created serious theoretical difficulties for scholars who wish to depict gender relations as expropriative. And as in the literature on prejudice, many analysts of gender attitudes have been motivated by the liberal spirit of social reform, but their primary focus on the gender attitudes of women has implied that it is the subordinate group (as in class relations) that must bring about change, rather than the dominant group (as in race relations). This merging of somewhat incompatible themes has often left empirical analyses of gender ideology in something of a theoretical void, as analysts have proceeded with an empirically driven agenda to monitor and explain popular endorsement or rejection of traditional gender arrangements.

All these lines of research tend to depict intergroup attitudes as primarily the property of individuals and as being purely expressive, rather than as political communications. That is to say, the common working assumption is that attitudes are raw acts of individual expression, articulating the inner thoughts, feelings, and reactions that individuals generate as they march through life.

At the same time, different assumptions have been made about whether humans express themselves in ways that suggest rationality or irrationality. Analysts of prejudice have been explicit on the point that prejudice is an irrational phenomenon that bears no relation to a rational processing of valid information about the target group. In the literature

on class conflict, however, it has been taken as a given that people do rationally process information and personal experiences. In the Marxist perspective, individuals are presumed to perceive and act rationally on their own behalf. Rationality in that model refers specifically to the ability to perceive one's economic interests and to act in concerted ways to further those interests, either by defending current economic arrangements (among those who benefit from them) or by rejecting those arrangements and working to alter them (among those who are hurt by them). Research on political tolerance is less clear on the issue of rationality, but it seems to assume that people rationally perceive competing interests in society: the task for democratic citizens is to learn to disregard the feelings of antipathy that arise from those perceptions as they formulate their policy dispositions toward competing groups. Research on gender ideology has embroiled competing strains of thought from theories of class conflict and prejudice, and thus no clear set of expectations about human rationality can be deduced from this line of research.

In the next chapter, I begin to raise questions about these various facets of ideology. Are attitudes raw acts of individual expression or politically motivated communications? Do intergroup attitudes originate with the individual or with the group to which he or she belongs? Are intergroup attitudes rational or irrational? How are these attitudes formed? I start by proposing a specific conception of rationality and considering some of the constraints that are present in intergroup relationships of inequality. Those considerations lead me to suggest that the preoccupation with conflict and hostility in the analysis of class, race, and gender relations is misplaced. My purpose is not to argue that such factors are devoid of significance for unequal social relations, only that their importance has been seriously overestimated. As we consider the role of ideology in social control, we will see many pressures that work against conflict and the expression of hostility. By using hostility as the flag for the inequality, students of class, race, and gender alike have misread the character of their material.