A Theory of Ideology
Yale University Press
New Haven & London
Copyright © 1998 by Yale University. All rights reserved.
With the permission of Yale University
Press, the author has made an online version of this book available under a Creative Commons Noncommercial Sharealike (by nc
sa) license, accessible at the author’s website at http://www.balkin.com.
5 CONCEPTIONS OF IDEOLOGY
There are many different definitions of the concept ideology, and many different ways of approaching its study. In particular, a theory of ideology must consider the following questions:
1. What kinds of things (objects,
entities, mechanisms, or structures) are
we investigating? This is the problem of the proper object of study.
2. Do we define ideology in terms of its content (for example, distortion or mystification), the functions it serves (for example, furthering the interests of the ruling class), its causes (for example, cognitive bias, reduction of cognitive dissonance), or its effects (for example, creating or sustaining unjust relations of social power)? This is the problem of the proper mode of explanation.
3. What is our attitude toward ideology-pejorative, positive, or neutral? This is the problem of interpretative stance.
4. How does our theory handle the inevitable difficulty that the analysis of ideology may itself be ideological? This is the problem of self-reference.
The first two questions are the subject of the present chapter; the last two are the subject of Chapter 6.
The Object of Study
Some theories of ideology define their subject matter in terms of beliefs, held either by groups or by individuals. Jon Elster, for example, defines ideology as false or distorted conscious beliefs held by individuals about the social world. Other theories of ideology are concerned with linguistic or cultural products or social practices of meaning that do not necessarily exist in the minds of individuals but that individuals use in understanding the social world. Thus John Thompson and Clifford Geertz view the study of ideology as the study of "symbolic forms." This expression includes such diverse phenomena as linguistic utterances, symbols, literature, traffic signs, television broadcasts, and advertisements. In Geertz's case this choice is a deliberate attempt to externalize the object of study--to move from the investigation of internal mental processes to external observable entities like written symbols, linguistic utterances, artistic objects, and behavioral practices.
The theory of cultural software takes as its object of study tools of human understanding produced by cultural evolution. Symbolic forms play a key role in cultural evolution because they carry units of cultural transmission; hence the study of symbolic forms is crucial to the study of cultural software. When we study a symbolic form (such as an advertisement), however, we are interested in the ways of understanding that produced it and the effects that it has on the way that others understand the world. Hence we are interested in cultural artifacts and symbolic forms for four reasons. First, symbolic forms are effects of cultural software and therefore are evidence of the mechanisms of thought. Second, symbolic forms have reciprocal effects on individual cultural software. Third, symbolic forms are media through which minds communicate and by which they share meaning. Fourth, symbolic forms are a common terrain of negotiation and struggle over shared meanings. These negotiations and struggles, in turn, affect the cultural software of the individuals who engage in them.
Moreover, the object of our study is necessarily broader than conscious beliefs, at least if by this term we mean beliefs that can be expressed in the form of propositions, like "Jews are greedy," "Women don't make good pilots," or "The Holocaust never happened." To be sure, beliefs can be tools of understanding and can be used to create new tools. But more important objects of study are cognitive mechanisms that produce beliefs. Examples include the tendency to structure experience in terms of narratives, psychological methods of categorization, varieties of metaphoric and metonymic thinking, strategies for reduction of cognitive dissonance, heuristics and biases employed in making judgments under uncertainty, and understanding by means of networks of conceptual oppositions in the form "A is to B as C is to D." Propositional beliefs can be true or false, but cognitive mechanisms are neither true nor false. Rather, they are the ways in which attitudes and judgments are formed: they produce beliefs that can be true or false.
For example, consider the tendency, noted by many feminist writers, for people to think of the male as the standard case or unspoken norm of gender, so that the feminine is treated as an afterthought, an additional feature, or a special case. This tendency is produced by various mechanisms of understanding, some of which will be discussed in more detail in Chapters 10 and 11. Nevertheless, it is important not to confuse these mechanisms with the propositional belief "Men are the normal case and women are the exceptional case," or the directive "Think first in terms of men and then consider women later as an afterthought if it is brought to your attention." Individuals' practices of thought may be aptly summarized by such a propositional belief or such a directive, but this does not mean that the mind consciously employs such beliefs or rules in forming its judgments. Social understanding does not proceed exclusively or even predominantly at the level of such conscious propositional belief or conscious rule following. Understanding also occurs through various mechanisms of framing, narrative construction, characterization, and categorization. Nevertheless, like the study of symbolic forms, the study of conscious beliefs is important because through them we can attempt to understand the mechanisms of social understanding that produce them.
The Focus of Study
Once a theory of ideology has settled upon its objects of investigation, it can study these objects in many different ways. We can study them in terms of their content, their causes, their effects, or the social functions they serve. Often the way that ideology is defined leads to a focus on some of these aspects to the detriment of others. If we define ideology purely in terms of false or distorted beliefs, for example, the study of ideology becomes centered on the question of the content of beliefs. To study ideology is to study how certain beliefs are false or misleading; hence when the analyst has revealed this falsity or distortion, her task is largely completed.
Marxist theories of ideology often approach ideology in functional terms. They study how ideologies serve the interests of various classes. Michele Barrett summarizes the classical Marxist definition of ideology as "mystification that serves class interest." This definition is functional (although it might be restated in nonfunctional terms): ideologies are defined and studied in terms of the social interests they serve. A functional approach, however, is necessarily limited. It tends to explain the development and content of ideologies solely in terms of the class interests that they further, rather than offering evolutionary or other causal mechanisms that explain how and why ideological beliefs are produced. Put another way, functional accounts cannot reliably serve as causal explanations: even if a belief serves the interests of a particular class, it does not follow that the belief was the result of something that class did. Similarly, showing that something serves the interests of one class does not by itself explain how a belief was generated or held by another class. In contrast, Jon Elster rejects functional approaches to ideology entirely. His theory seeks to offer purely causal explanations of ideology. Thus, although he defines ideologies in terms of their content as false or distorted beliefs, he is specifically interested in how these beliefs are caused by various social psychological mechanisms.
Finally, we might define ideology in terms of its effects. For example, John Thompson defines the study of ideology as the study of how symbolic forms create or sustain relations of domination. Thus his approach is centered on what ideology does rather than what causes it or what interests it serves. In Thompson's view, approaches that focus on content are insufficient because the content of a particular symbolic form does not by itself tell us whether it helps sustain relations of domination.
The theory of cultural software has a twofold focus: First, it is concerned with how tools of understanding are produced through conceptual bricolage. Second, it is concerned with how these tools of understanding help create or sustain injustices in particular social contexts. The goal of this theory is not primarily functional explanation but causal or evolutionary explanation. It studies the tools of understanding in terms of the causes that produce them and the effects that they in turn produce. Although this study is obviously concerned with the content of beliefs and symbolic forms, that inquiry is subsidiary to the study of the effects produced. Moreover, because the theory focuses on just and unjust effects, its analysis is overtly normative as opposed to merely descriptive.
Under this approach there is, strictly speaking, no longer a single thing called ideology. The theory of cultural software, while dissolving the study of ideology into the larger study of cultural understanding, also breaks the study of ideology down into the study of ideological mechanisms and ideological effects. Ideological mechanisms are mechanisms of social cognition that produce ideological effects. Ideological effects are effects of cultural software that help create or sustain unjust social conditions, unjust social relations, or the unjust use of social power. Ideological thinking, in short, is employment of ideological mechanisms of cultural software that produce ideological effects. Note that symbolic forms produced through the use of cultural software can also have ideological effects through their effects on human understanding. For example, perfume advertisements can have ideological effects if they help to create or sustain unjust relations between men and women.
The phrase "help to create or sustain" in the definition of ideological effects must be understood in a limited way, for otherwise the definition is seriously overinclusive. If a person used statistical formulas to calculate the numbers of individuals who could be transferred to a concentration camp, we would not say that the mere skill involved in applying the algorithm was an example of ideological thinking, even though it would literally be a use of cultural software (a mathematical skill) that helped maintain unjust social conditions. Rather, cultural software has ideological effects when it creates ways of thinking about the social world or about others in the social world. Although the study of ideology is not concerned with the skills involved in statistical computation, it is concerned with the ways of looking at people that lead to the judgments that it is appropriate to apply these statistical methods to facilitate genocide. These judgments include, among others, that people should be shipped to concentration camps because they are inferior or that it is appropriate to think about people as commodities that must be efficiently shipped to the most efficient locations for the most efficient forms of slaughter.
These definitions of ideological effects and ideological mechanisms make what is ideological turn heavily on social context. Cultural software has ideological effects only when and only to the extent that it results in various forms of injustice. This means that in other contexts cultural software may have no significant ideological effects. Moreover, even when cultural software has ideological effects, these effects do not exhaust its social meaning, its content, or its usefulness. We can make a similar point about symbolic forms: a perfume advertisement is not merely a symbolic form with ideological effects; it is also, among other things, an advertisement for perfume. More generally, the tools of cultural understanding may have many other features and advantages and may serve many other functions apart from their tendency to produce ideological effects in certain circumstances.
In like fashion, ideological mechanisms are defined contextually. When mechanisms of social cognition produce ideological effects, one can speak of them--for this purpose and to this extent--as ideological mechanisms. But they are ideological not because of their inherent nature but because of the context in which they are employed and the effects that they have.
What distinguishes ideological thinking from mere fantasy or mistake is the social context in which belief occurs and the use that people make of it. An important consequence of this approach is its emphasis on the normative dimension of all ideological analysis. To understand what is ideological, we need a notion not only of what is true but also of what is just. False beliefs about other people, no matter how mistaken or unflattering, are not ideological until we can demonstrate that they have ideological effects in the social world. To demonstrate this, we must know something about the relationship between a person's thought and the existing conditions of social power, as these provide the necessary background for considering questions of justice and injustice.
For this reason, the study of ideology necessarily intersects with the study of how social power is created, sustained, and distributed, because one of the objects of this study, the ideological effect, is a highly contextual product of cognitive capability and social situation. To be sure, sometimes we may infer that a particular way of thinking--a white American's belief that all black people are lazy and immoral, for example--is so likely to produce or contribute to injustice that we may consider it presumptively ideological. But this is the case only because we already understand the social context in which this way of thinking occurs, the forms of behavior it is likely to lead to, and its place in a larger social system of race relations. Nevertheless, unjust social relations or unjust social power may be created or sustained in many different ways that are not always easily discernible from the content of a particular belief, especially when the context is unusual or unfamiliar.
Perhaps the best example of this principle is the bizarre phenomenon of Japanese anti-Semitism. There are very few Jews in Japan today and thus very few opportunities for discrimination against them. Nevertheless, anti-Semitic books and comments have appeared continually in Japan over the years, often repeating the most vicious claims of Nazi ideology and Eastern European antiSemitism. Especially popular are beliefs about a secret worldwide Jewish financial and media conspiracy of enormous scope and power. What is most amazing is that the very same libels that in the European context were part and parcel of a terrible social system of discrimination (and extermination) are combined in Japan with a peculiar form of philo-Semitism in which Jews are admired for their supposed shrewdness and business acumen.
All of this is not to claim that Japanese anti-Semitism has no ideological effects. Rather, my point is that we must not conflate this phenomenon with European anti-Semitism even though its beliefs and slogans appear to be similar in content and may even have their origins in European anti-Semitic literature. The ideological effects of Japanese attitudes toward Jews seem to have more to do with supporting and sustaining a larger system of beliefs about business and economic competition in Japan. These ways of thinking, in turn, may help sustain relations of unjust power not between the Japanese and a Jewish minority but within Japanese society itself, or between the Japanese and the outside world. Moreover, Japanese anti-Semitism also serves as a way of expressing anti-American sentiments, which have surfaced as Japan and the United States have increasingly become economic adversaries. Because Jews are portrayed as the hidden masters of American business and government, anti-Semitic rhetoric becomes another way of complaining about American culture and American trade policies.
Of course, if large numbers of Jews were to emigrate to Japan, existing antiSemitic attitudes might lead to unjust treatment of Jews, just as they did in Europe and America. This is yet another consequence of my basic point about the uses and effects of conceptual tools. When introduced into new social settings, the tools of understanding display different effects, benefits, and disadvantages. That is why the study of ideology cannot rest on content alone but must take into account the environment in which cultural software operates. Indeed, the view that the power of ideas lies in their content and not in their content in a particular context is itself a way of thinking that causes us to misunderstand social situations.
This approach sheds a somewhat different light on so-called beneficial ideologies. Suppose that the students in a particular elementary school classroom are falsely told that they are very bright and very able, indeed, much more bright and able than other students of their age. As a result, their test scores, as a group, actually begin to improve. The source of their esprit de corps is fraudulent, yet it seems to benefit them. This is an example of what Jon Elster calls the "benefits of bias." Such situations are important to explain in the Marxist tradition because ideology is often defined functionally in terms of what serves the interests of a particular class. Hence it follows that some ideologies, while false, may actually benefit the people who hold them--for example, the members of the bourgeoisie. Because I define ideology in terms of what is just rather than what is in a particular group's interests, my analysis of this example is quite different: we cannot yet even say that these students are engaged in ideological thinking until we study how justice might be affected by their views of themselves. First, these students may start to look down on students in other classes and other schools and to discriminate against them although the others have equal or greater abilities. Second, some students in the class may not be able to live up to their teacher's claims of superior ability, and they may engage in strategies of dissonance reduction to avoid this recognition: for example, they may be more likely to assume that people who criticize their work are simply mistaken, or they may come to think that their failures are due less to ability than to luck or sheer coincidence. This may harm them in the long run. Thus, a so-called ideology of superior achievement is not ideological thinking in my sense of the word unless and until it has particular effects, and then only to this extent. Moreover, the flip side of this claim is that even the most seemingly benign and beneficial forms of thinking can have unexpected and unfortunate effects as they are extended into new contexts and situations. It is at that point that they become forms of ideological thought.
Above all, this approach does not view ideology as something separate from cultural understanding. The mechanisms of what we call ideological thinking are no different in kind from the ordinary forms of thought. There is not a separate set of devices that constitute "the ideological" and another set that constitute "the nonideological." There are not mechanisms of social cognition that always produce ideological effects and other mechanisms that never do so. In particular, we must resist the natural tendency to think that ideology constitutes a separate, deviant form of social cognition that can readily be distinguished in terms of its operations from the supposedly normal, nonideological forms and mechanisms of thought that characterize everyday reasoning. The mechanisms of ideology are the mechanisms of everyday thought, which in particular contexts produce effects that are both unfortunate and unjust. Conversely, the mechanisms of everyday thought can become ideological mechanisms if they are employed in inappropriate contexts and situations.
This conclusion is consistent with our earlier discussion of conceptual bricolage. The tools of our understanding can be alternatively advantageous and disadvantageous as they are applied in new situations and new contexts. Among the many possible disadvantages that conceptual tools can have is their tendency to promote injustices; conversely, one of their many possible advantages is the relative lack of this tendency. Thus tools of understanding that are entirely benign in some circumstances may become malignant if too much is demanded of them or if the context in which they are employed changes sufficiently. Then their limitations become apparent in the same way that many other disadvantages of tools may suddenly surface.
The temptation to identify ideology with a sort of pathology may stem from the familiar notion that ideology is false or distorted belief. Given this assumption, it seems natural to think of falsity or distortion as a kind of illness or malady, especially if it has harmful effects. For example, we often speak of racism or anti-Semitism as a sickness or a disease. In fact, the metaphor of disease is not completely unreasonable, as I shall discuss momentarily. But identifying ideology with pathology simply because beliefs are false or distorted improperly focuses on content rather than mechanisms--or, to use the metaphor of disease, it focuses on symptoms rather than etiology or cause. From the standpoint of causal mechanisms, the question is whether the effects that people have traditionally assigned to the ideological are due to (1) a special mechanism different from the ordinary mechanisms of social cognition; (2) the extension or employment of cognitive mechanisms into contexts for which they are not well adapted; (3) a spontaneous malfunction in cognitive processes; or (4) the invasion of some external force into normally and properly functioning cognitive processes that causes them to malfunction. I reject (1) and suggest that many ideological effects are produced by (2).
This leaves cases (3) and (4), both of which explain ideological effects in terms of malfunctions. Obviously, there is some overlap between the notion of overextension and the notion of malfunction. Nevertheless, the concepts are not identical: we would not say of an airplane that it malfunctions because it is a poor vehicle for traveling on land. This is not malfunction but maladaptation. One could collapse the distinction between malfunction and maladaptation only if one assumed that our tools of understanding should be capable of understanding everything in all contexts. Then to the extent that they failed to do so, we would say that they were malfunctioning. This seems to ask too much of our tools of understanding, though; after all, no tool exists that is equally well adapted to all tasks.
Much of the distortion that we see in ideology involves the side effects of tools of understanding that become prominent and maladaptive in particular contexts. Ideological effects are usually the unexpected and unpleasant side effects of conceptual bricolage. I do not reject out of hand the possibility that some ideological effects are due to a genuine malfunction in cognitive processes. But this malfunction would have to appear in many individuals at once in order to qualify as an ideological phenomenon. A simultaneous malfunction by members of a culture is unlikely. This leaves the possibility that if some ideological phenomena are due to a malfunction in our cognitive processes, it is a malfunction brought on by some external force that affects many people at once. One possibility is that when individuals are placed in situations with which their cognitive systems cannot cope, they break down or malfunction, just as we say that a car malfunctions when it is forced to drive through water, or a vacuum cleaner malfunctions when it is forced to deal with too great a quantity of dust. If many individuals face the same type of experience, this malfunction would be similar for all of them. But it is hard to imagine that this explains most ideological effects. After all, human intelligence is quite adaptable, and many ideological effects, like racism or anti-Semitism, are long-lasting phenomena that occur over many generations. The idea of a long-term breakdown in cognitive processes seems implausible.
Instead, the theory of cultural software offers a somewhat different account of how relatively robust and long-term ideological effects can be produced by a malfunction due to an "external entity." This external entity is none other than cultural software itself, transmitted from other individuals and spread throughout a culture like a computer virus. A computer virus is just a special kind of computer software that is able to spread and reproduce itself in other computers. By analogy, cultural software may act like an informational virus that infects one node on a network and then, through the exchange of information, gradually infects all the others.
Under this model, long-lasting and widespread ideological effects are produced by informational or cognitive "viruses" that are passed from person to person and generation to generation. If so, we might think of racism or antiSemitism as a sort of socially spread informational virus or parasite that, while not totally debilitating subjects, affects their behavior and cognition for the worse.
In fact, this model of ideological effects is the model of memetic evolution through cultural communication. Memes are reproduced in individuals through a social network of communication and transmission. The spread of ideological viruses is merely a special case of the basic mechanism through which cultural software is written, transmitted, and modified. All cultural software can be thought of as a kind of informational virus, transmitted from person to person; or, put another way, what we might call an ideological virus is just another kind of cultural software. Our devices for understanding the social world are constituted in large part by idea-programs that were able successfully to be transmitted to us and absorbed into our cultural software. The complexes of memes that give rise to racism and anti-Semitism, in this sense, are no different from any other idea-programs--like those producing predilections for free speech or free mar-kets--that make use of our cognitive capacities to grow, spread, and develop, just as genes "use" bodies in an evolutionary system.
Hence what differentiates cultural software from a so-called ideological virus is the harmful effect that the latter produces in a particular social context. As the example of European versus Japanese anti-Semitism demonstrates, an ideological virus can produce very different effects when it is introduced into different environments. If an informational virus produces no such harmful effects--just as there are many viruses in the human body that are relatively benign or harmless--then it does not produce an ideological effect. Fantasies about people in far-off lands may be distorted and false, but they do not become ideological until there are conditions of justice between the two peoples-that is, until there is communication, trade, and the possibility of war, conflict, struggle, economic exploitation, or colonization. Then these fairy tales (which may already have had certain ideological effects within a culture) take on a more serious and harmful tone. Fantasy becomes ideology when justice is at stake.
This line of reasoning brings us back to our original hypothesis--that ideological effects are produced by ordinary mechanisms of thought that have harmful or maladaptive consequences in particular contexts and situations. Ideological effects occur when cultural software "goes wrong" in some important way. The power of ideology over our imaginations is a special case of the power that all cultural software has over our imaginations.
The power of ideology within this picture is quite different from the picture underlying a more traditional Marxist theory of ideology. In the traditional account, ideas have power because they present a distorted picture of reality to the minds of the persons holding them, causing these persons to act against their objective interests. From the standpoint of the theory of cultural software, the power of ideology is the power of the culturally produced capacities of our minds to shape social reality for us, and thus simultaneously to empower and to limit our imaginations.
This approach makes considerable use of concepts like usefulness, ade quacy, and suitability. But these concepts can hardly be considered inherent properties of the tools of understanding. Adaptability is a judgment made about the operation of a tool in a particular context. It is also a judgment made by an observer who assesses the operations and effects of mechanisms of thought. This means, among other things, that the study of ideology is necessarily an interpretive endeavor, although this fact makes it no less useful. Finally, because all cultural and social understanding makes use of cultural software, all ideological analysis-that is, all judgments about the existence and nature of ideological effects-involves judgments by an analyst that employ the analyst's cultural software. This raises problems of self-reference, which are discussed more fully in the next chapter.
What Kinds of Effects: Hegemony or Unjust Power?
The study of ideology necessarily has a normative dimension. It cannot be value free but must presuppose a view about what is good and bad, advantageous and disadvantageous, just and unjust. The analyst cannot describe and analyze ideological effects without reference to concepts like truth or justice. She must make interpretive judgments about what social conditions are like, and she must also make judgments about whether a way of thinking is adequate or inadequate to serve particular ends and whether social conditions are just or unjust. Ideological analysis does not end with a demonstration that a particular belief or symbolic form is partly or wholly false or distorted. It must ask how this falsity or distortion might create or sustain unjust social conditions or unjust relations of social power. Thus ideological analysis does not merely involve considerations of truth and justice; it is fundamentally a question of the relationship of truth to justice.
Because I define ideological effects in terms of actual or potential injustices rather than the presence of hegemony or domination, it may be helpful to contrast my approach with that recently offered by John Thompson. Thompson defines the study of ideology as the study of how symbolic forms create or sustain conditions of domination. He then defines domination in terms of systematic asymmetries in relations of power-that is, "when particular agents or groups of agents are endowed with power in a durable way which excludes, and to some significant degree remains inaccessible to, other agents or groups of agents, irrespective of the basis on which such exclusion is carried out." Under Thompson's definition, women in the United States would be dominated if we could show that they are disadvantaged vis a vis men systematically in many different ways, including jobs, income, status, education, economic opportunities, and other resources. Thus while Thompson argues that the essential feature of ideology is the creation or preservation of domination, I have argued that it is the creation or preservation of unjust power or unjust social conditions.
One reason for this difference is that Thompson's definition is underinclusive. Not every example of ideological thinking contributes to systematic asymmetries in social resources or power relations between groups. Consider, for example, the phenomenon of black anti-Semitism in the United States. AntiSemitic propaganda by black nationalist groups like the Nation of Islam does not contribute to or produce systematic asymmetries in resources or power relations between blacks and Jews or even between all Christians and Jews. Indeed, in contrast to blacks, Jews have been relatively successful in gaining access to social resources in the United States. For this and other reasons, Jews provide a convenient scapegoat for some members of the black underclass, just as blacks themselves have provided a convenient scapegoat for lower-class whites in the United States. Black anti-Semitism, like resentment and hostility among some blacks toward Asian Americans, is in part the result of competition between various minority groups; it is not a means by which blacks oppress Jews or Asians and systematically deny them access to social resources. Nevertheless, anti-Semitism and anti-Asian beliefs may in fact lead to particular injustices-acts of violence, for example-against Jews or Asians, either by blacks or by other groups. Thus a focus on systematic asymmetries in power defines ideology too narrowly; the study of ideology must be concerned with injustices produced by tools of understanding whether or not they stem from domination of a subordinated group by a dominant group.
To be sure, black anti-Semitism or anti-Asian sentiments may also contribute to the perpetuation of systematic asymmetries between blacks and whites, by diverting attention onto scapegoats and away from positive solutions to the challenges the black community faces. Similarly, prejudice against other racial minorities, like Asians or Hispanics, alienates potential allies who might otherwise fight together with blacks against white supremacy. Nevertheless, the ideological effects of black anti-Semitism or anti-Asian prejudice are not exhausted by their ability to hinder black economic progress and further white supremacy. Even if these prejudices did not harm the just interests of blacks, they would still be ideological, because they can and do lead to injustices between members of different minority groups.
Thompson's formulation suffers from these difficulties because it has not yet thrown off the shackles of a traditional Marxist model that envisions a dominant class, a subordinate class, and an ideology that justifies the subordination of the latter by the former. Systematic group domination by a dominated class over a subordinate class is the central concern; it follows that forms of social injustice or unjust social power that do not involve hegemony are not properly the concern of the theory of ideology. As Thompson himself stresses, "Ideology, according to this conception, is by nature hegemonic, in the sense that it naturally serves to establish and sustain relations of domination and thereby to reproduce a social order which favors dominant individuals and groups. Hence Thompson limits ideology to the study of "the ways in which meaning is mobilized in the service of dominant individuals and groups."
Unfortunately, this model is too simplistic to describe a large number of ideological phenomena, particularly in a country like the United States, where there are many different groups with varying degrees of social power and multiple and cross-cutting social identities. Antiblack prejudice by Korean Americans and anti-Korean prejudice by American blacks cannot easily be subsumed within a hegemonic conception of ideology. Nor does a hegemonic approach contemplate the possibility of simultaneous membership in groups that are dominant and subordinate-working-class white males who are homosexual, for example, or upper-class heterosexual women who are not physically disabled. One is perfectly free to limit the scope of the study of ideology in this way, of course, but the danger is that a large portion of what most people would consider ideological phenomena will be missed. Moreover, this limitation may have significant ideological effects on the analyst's own thought about ideology and social conditions.
For the theory of cultural software, the equation of ideology and hegemony is problematic for seven additional reasons. The first stems from the basic point that ideological mechanisms are the mechanisms of everyday thought about the social world. There is no reason to think that the kinds of cognitive mechanisms producing ideological effects that benefit dominant groups and harm subordinate groups are different in kind from those producing benefit and harm to other groups. It is likely that the mechanisms that produce prejudices between groups are fairly similar, although the results may differ because of the relative positions and histories of various groups in society. If we restrict our study of ideology to mechanisms producing beliefs that benefit dominant groups, we cut ourselves off from many examples of ideological thinking that not only shed considerable light on more hegemonic examples but are fully worth studying in their own right.
Second, the concept of a dominant ideology leads us to view ideology in terms that are too monolithic. What people usually think of as ideology is really the confluence of many different types of cognitive mechanisms. The ideology of patriarchy, for example, is not a single thing, or a coherent system, but rather a group of heterogenous and partly reinforcing ideological effects. This heterogeneity may be one cause of its adaptability as well as a source of its possible deconstruction and subversion.
Indeed, there is a notable tendency among theorists of ideology to confuse the pervasiveness and the wrongfulness of a worldview with its systematicity. Thus, Catharine MacKinnon, in a famous passage, has described patriarchy as a "metaphysically nearly perfect" system. This way of thinking may itself betray a certain ideological effect, because it conflates the powerful with the well-ordered. Ideologies in the larger sense that MacKinnon is concerned with are always the product of bricolage and memetic evolution. Hence they lack the characteristics of design: they always have conflicting and variegated elements, their seams always show, and loose threads are always dangling. Of course, this makes them no less powerful: an avalanche of motley elements is still an avalanche. But it does suggest that the theorist of ideology may be misled if she attempts to fit the entire phenomenon into a single, systematic analysis rather than looking for the confluence of various ideological effects and for their possible points of interaction and conflict. Indeed, the heterogeneity of cultural software is important precisely because it makes possible forms of resistance to received ways of thinking.
Third, the notion that ideology is concerned only with the preservation and maintenance of dominant ideologies neglects the importance of competition between various ways of thinking within a culture. This competition occurs at many different levels and at many different places in society; there are not simply two armies contending on the field, and those armies that do contend already are fragmented and partly divided against themselves. Within American society for example, many different and partially overlapping groups promote their ways of thinking about the social world; and many different currents and eddies of social power result from these encounters. Together these encounters produce heterogeneous matrices of social power, mixing together the just and the unjust in an atrocious and unpalatable stew. To see only some elements of this mixture as worthy of the title of ideology is itself ideological, for it hinders the identification and critique of the many forms of social injustice that do not correspond to the grand narrative of the "hegemonic."
Fourth, when we define ideology in terms of symbolic forms that benefit dominant groups, we risk sentimentalizing the attitudes and interests of other groups, in particular subordinated groups. We risk overlooking the possibility that the beliefs of subordinated groups can also be distorted, self-serving, and unjust to other groups, even including more dominant groups. There is no reason to think that self-serving or distorted views of the social world are confined to dominant groups. Prejudice tends to beget prejudice and hate tends to beget hate. Persecution can lead to persecution complexes. Moreover, if a group's opportunities and access to knowledge have been limited by its social condition and its comparative lack of social power, this may seriously affect its members' understanding of the social world, producing ideological effects in their thought.
Even when subordinated groups have a relatively adequate understanding of the social world, it by no means follows that what these groups believe to be in their interest is always just, or even that what is actually in their interests is always just. This is especially so, one might think, in a multicultural society in which many different subordinated groups scramble for social betterment and political power. An obvious example involves tensions between black and Hispanic communities in the United States over the drawing of district boundaries that effectively determine the result of elections to state and federal legislatures. Black and Hispanic communities may correctly recognize that drawing boundary lines one way rather than another would guarantee the election of a black or a Hispanic representative in Congress, and they may also correctly assess that this would further black or Hispanic interests. Nevertheless, it is entirely possible that furthering one group's interests in this way may be unjust to other groups. We are no longer in the Marxist world, where furthering the interests of the oppressed (the proletariat) necessarily furthers justice or the proper direction of history. The belief that something is just simply because it favors a subordinated group may itself, under some conditions, involve ideological effects.
Fifth, defining ideology in terms of what benefits members of dominant groups is problematic because subordination is not simply an on-off property of individuals or groups. There are different degrees and kinds of subordination among different groups, and individuals have multiple group identifications. Thus it is possible for an individual to be in a subordinated position with respect to some groups but in a privileged or dominant position with respect to others--consider the example of white middle-class heterosexual women. Pursuing the interests of white women may infringe on the just interests of black men, and vice versa. The endless possibilities for self-serving views of the social world between groups all of which can claim to be subordinated in one way or another--along with the concomitant injustices that may be produced by these views--shows how limited and simplistic a bipolar dominator-dominated model can be, and demonstrates the need to expand the notion of ideology beyond a hegemonic conception.
The traditional proletariat-bourgeoisie model avoids these problems, first because it tends to reduce the number of groups to two, and second because it assumes that what is in the proletariat's interest is necessarily just or at least follows the course of proper historical development. Within this model the problem, rather, is ensuring that the proletariat understands what is in its own interest-that is, ensuring that it develops an appropriately revolutionary consciousness. Nevertheless, in a society where injustices do not derive wholly from economic power, in a society that features many competing and partially overlapping groups, divided on the basis of race, ethnicity, religion, language, gender, sexual orientation, and disability, one can no longer employ such simplifying assumptions. The simple model of dominator and dominated itself threatens to become ideological because it obscures the complexity of social conditions.
Not only does a bipolar approach tend to neglect the many different kinds of subordinated groups, it also tends to collapse, homogenize, and demonize the interests, attitudes, and beliefs of whatever group is described as dominant. Such a homogenization may disguise fragmentation within the dominant group as well as the existence of relatively subordinate and distinct subgroups. To speak about hegemony by whites, for example, is to forget that some women are also white; to speak about the hegemony of white males is to neglect the fact that the interests of lower-class white men may be quite different from those of more affluent white males. Moreover, the homogenization of white males into a single group obscures the fact that some of the most vitriolic race hatred appears not among the most powerful members of white society but among the most disaffected and disenfranchised. The Ku Klux Klan and other white supremacist organizations have often found that the poor and uneducated are more promising recruits than the well educated and the well-to-do. One reason why such groups turn to rabid racism, anti-Semitism, and anti-Catholicism is that, given their economic and class subordination, they cling to their whiteness as a guarantee of social status. Thus even the supposedly simple case of prejudice against blacks is more complicated than a hegemony model of ideology suggests; like the case of black anti-Semitism, it involves competing ideologies among groups that suffer varying degrees and kinds of subordination.
Approaching the study of ideology in terms of hegemony rather than justice creates a sixth problem: the familiar but troublesome concept of "false consciousness." This concept focuses not on the question of what is just but on the relationship between individuals' thought and the (objective) interest of the class to which they belong. A person whose beliefs and preferences are contrary to that interest is said to suffer from false consciousness. Hegemonic conceptions of ideology lead inevitably to notions like false consciousness--whether or not they use that precise terminology--because members of subordinate groups often accept their lot and may even oppose political activity designed to undermine the hegemony of superordinate groups.
Inquiries into false consciousness are problematic for four reasons. First, they presuppose that a class can have a unified, objective interest and that the interests of each of the members of that class are not substantially in conflict with it. In other words, the concept of false consciousness assumes without further investigation a particular state of affairs about the benefits of collective action: it assumes that each individual member benefits sufficiently from promoting the group's interest. But even assuming that African Americans as a group have an objective interest, there is no reason to think that the individual interests of some African Americans might not conflict with that group interest. Indeed, by taking contrary positions they may realize considerably greater benefits personally than they would have if they had adhered to the "party line." It is hard to argue that such persons suffer from false consciousness--indeed, they may see what is in their interests more clearly than many other people.
Second, the concept of false consciousness tends to elide distinctions between long- and short-term interests, in part because it is premised on an underlying historical narrative of eventual liberation. But if one no longer believes in such a narrative--for example, the Marxist narrative of the inevitability of proletarian revolution--the multiple and conflicting interests of persons and groups reassert themselves forcefully. It becomes more difficult to state conclusively that a particular perspective is false consciousness. Rather, people may disagree simply because they balance long- and short-term interests, or group interests and individual interests, differently.
Third, accusations of false consciousness are normally directed at members of subordinated groups that dissent from the analyst's view of what is in their class's interest. But the same logic applied to superordinate groups leads to a paradoxical result: members of superordinate groups that support the dismantling of unjust hierarchies also suffer from false consciousness because they are working against their class's interests in maintaining hegemony. If women who oppose gender equality suffer from false consciousness, so too do men who support gender equality. This paradox arises from the fact that the notion of false consciousness is concerned not with the justice of a position but its relation to the interests of a class.
Fourth, the notion of false consciousness is problematic because it is a holdover from the bivalent oppressor-oppressed model of hegemony that I have just criticized. This model makes little sense in a world in which people have multiple and cross-cutting identities. Even assuming that African Americans and women have objective interests as a class, surely these interests can sometimes conflict. When they do, how can an African-American woman avoid a charge of false consciousness, regardless of the position she takes?
Indeed, accusations of false consciousness are often attempts by one portion of a social group to assert a unitary and objective interest that disadvantages or ignores the claims of another portion or subgroup. Working-class women may be accused of false consciousness by middle-class women when in fact their interests differ because of their class position. Similarly, the interests of African-American women may diverge in important respects from those of white women. Once internal divisions and cross-cutting identities are recognized, the notion of false consciousness threatens to become incoherent or at best self-serving.
The approach that I take in this book rejects the notion of grounding an analysis of ideology in the objective and unified interests of social groups. It asks instead whether cultural software tends to produce or sustain unjust effects. This does not eliminate inquiries into the interests of social groups. But it mediates them through the larger question of what is just for all concerned. Because our primary concern is justice, the notion of false consciousness becomes superfluous. An African American who takes positions that undermine the achievement of racial justice may be acting in his or her personal interests at the expense of the interests of other African Americans; but the important question is whether taking those positions promotes or hinders justice. Moreover, a focus on justice as opposed to objective group interest puts the conflicting claims of social groups in proper perspective, for justice does not consist in each group achieving its interests; it involves accommodating the just interests of all.
Ideological Analysis and Normative Commitment
The seventh and final reason to prefer a definition of ideology based on the question of justice rather than on the question of domination is that ideological analysis is essentially and ineluctably normative and interpretive. A definition of ideology in terms of "domination" tends to disguise the normative commitments of ideological analysis. What constitutes domination cannot be articulated in a purely factual way; it requires a view about what is just and unjust in a society. Moreover, the very concept of domination that one might use to distinguish the ideological from the nonideological is itself an object of ideological disputation.
Consider Thompson's definition of domination in terms of "systematic asymmetry" in power and access to social resources. Although this definition seems to rest on facts about society, it must also rest on a conception of justice. The concept of domination must also include a normative judgment about just and unjust treatment if it is to be of any use in a theory of ideology.
In fact, Thompson's definition would be seriously overinclusive if it rested only on the existence of systematic asymmetries in power between groups. Not all examples of systematic asymmetry in power relations involve unjust domination, and not all beliefs that justify or sustain systematic asymmetries between groups are ideological in a pejorative sense. Some systematic asymmetries between groups are in fact justified. Take, for example, the case of felons. Surely this group is systematically disadvantaged in the United States. Indeed, in the United States, we incarcerate felons and deny them the right to vote.
We would probably not say that say that felons suffer from social domination by the law abiding, although we might contend that particular felons suffer from social domination because they also belong to groups that are unjustly treated. The reason we do not claim that felons as a class suffer from domination is that we believe that the systematic disadvantages these people suffer on account of being felons are fully justified. We are justified in systematically disadvantaging rapists, murderers, and child molesters because they have seriously injured other people. That is why our judgments of social domination necessarily require judgments of just and unjust treatment. What differentiates a dominated group from a systematically disadvantaged but undominated one is the question whether the group's lot is due to some present or previous injustice.
Our judgments about social domination are inextricable from our judgments about justice. People of low intelligence are systematically denied many advantages in the United States, including entrance to elite educational institutions and employment in many high-paying occupations like medicine. We might also note the systematic disadvantages suffered by people who are lazy, disagreeable, shy, unambitious, and untalented. Does the mere fact of these systematic disadvantages mean these groups are also dominated? Not necessarily; it all depends on our theory of justice.
Under some conceptions of distributive justice, one might well conclude that people who are lazy, unintelligent, and untalented are oppressed by the industrious, the clever, and the talented. Suppose, for example, that our theory of distributive justice holds that people do not have rights to the fruit of their talents, and that inequalities produced by the use of these talents unfairly disadvantage those with lesser abilities. Or suppose that we think that purportedly negative qualities like laziness are produced by oppressive social structures and that these qualities would be differently produced, differently understood, and differently distributed if these social structures were altered. Finally, suppose that we believe that negative qualities like laziness are matters of social convention, regularly and opportunistically invoked to benefit certain identifiable social groups. Each of these theories of justice may be controversial in some respects. But they aptly demonstrate that our social judgments about domination and oppression are not judgments about facts but about facts mediated through underlying values. They are complicated appraisals of social meaning with ineluctably normative underpinnings.
Moreover, a systematically disadvantaged group may be unfairly dominated, but its unjust domination may not be coextensive with the full degree of its systematic disadvantage. Some of the disadvantages its members suffer may be unjust, but others are not. For example, it may be perfectly just to imprison certain types of criminals and to discriminate against them in all sorts of ways, but there is a point at which their punishment becomes oppressive and unjustified. If criminals are denied due process, tortured, or imprisoned under inhumane conditions, they may well suffer from domination or oppression. In addition, if all felons are lumped together in people's judgments, so that less culpable criminals like petty thieves are treated the same as serial killers, this may also lead to injustice toward and oppression of the former subgroup. Mentally retarded persons suffer systemic disadvantages in social power that can be justified to some degree by their limited mental capacity, but some of their disadvantages cannot be justified on these grounds. These disadvantages are oppressive, and ways of thinking that justify such oppressive treatment are the proper concern of a theory of ideology. Here too, we cannot base our definition of ideology on the bare fact of disparate treatment or systematic disadvantage alone. We need a conception of justice to distinguish those parts of a group's unequal treatment that involve unjust domination and oppression from the parts that do not.
In this chapter I have argued that a theory of ideology needs a conception of justice. By this I mean that to understand and describe ideology the analyst must bring to bear her sense of what is just and unjust. However, ideological analysis does not require that the analyst have a full-fledged philosophical theory of justice. Nor does this book offer a complete philosophical account of justice. Most people go through their whole lives without developing such theories, and they are nevertheless able to discuss and reason about questions of justice and injustice. Conversely, well-developed philosophical theories of justice are often too abstract to offer specific judgments about whether particular policies or social conditions are just or unjust.
Finally, as we shall see in the next chapter, the very act of engaging in ideological analysis can change our views about what is just and unjust. We must be open to such changes as a condition of our understanding. So the theory of ideology that I offer in this book is designed to be compatible with a wide variety of different philosophical theories of justice. Indeed, in Chapter 7 I will argue that justice is an indeterminate value that must be articulated through human culture. The many different philosophical theories of justice are but one form of this cultural articulation.
Nevertheless, throughout this book I offer examples that assume that certain positions and social conditions are relatively just or unjust. I do this to clarify my arguments about ideology through concrete examples. But these specific judgments are independent of the theory of ideology I present. And Iwould hardly be surprised if my own assumptions about what is just and unjust are not themselves possible subjects of analysis and criticism.
My conclusion that the study of ideology must rely on a conception of justice prefigures the answer to the third of the four questions with which I began this chapter-namely, the interpretive stance that we must take toward the object of our study. And it raises even more urgently the fourth question how to deal with the problem of self-reference, given that ideological analysis can also be applied to the analyst's own thought. These questions form the subject of the next chapter.
 Jon Elster, Making Sense of Marx (Cambridge: Cambridge University Press, 1985), 462-64.
 John Thompson. Ideology and Modern Culture (Stanford: Stanford University Press, 1991), 59 ("By 'symbolic forms' I understand a broad range of actions and utterances, images and texts, which are produced by subjects and recognized by them and others as meaningful constructs"); Clifford Geertz, The Interpretation of Cultures (New York: Basic, 1973), 212-15. The locus classicus of the term is Ernst Cassirer, The Philosophy of Symbolic Forms, vol. 1, Language (New Haven: Yale University Press, 1955).
 In contrast, Thompson does not seem concerned with a distinction between mental processes and symbolic forms existing in the social world. His approach is sociological rather than philosophical, and hence these questions do not concern him. On the other hand, in his essay "Ideology as a Cultural System," Geertz seems specifically interested in externalizing the study of ideology from internal mental operations to symbols. Geertz, The Interpretation of Cultures, 215.
It is interesting to note that Cassirer's original use of this concept was strongly Kantian in spirit. Cassirer argued that symbolic forms in science, language, myth, art, and religion constructed the world for us and enabled us to understand it; at the same time he viewed these forms as functions of mind that allowed people to conceive both the world and themselves and created a bridge between the two. Cassirer, The Philosophy of Symbolic Forms, 1: See Jon Elster, Sour Grapes: Studies in the Subversion of Rationality (Cambridge: Cambridge University Press, 1983), 16391.
 See, e.g., Catharine MacKinnon, Feminism Unmodified:: Discourses on Life and Law (Cambridge: Harvard University Press, 1987), 36, 42-43, 71.
 Michele Barrett, The Politics of Truth: From Marx to Foucault (Cambridge: Polity, 1991), 4.
 See Jon Elster, Sour Grapes: Studies in the Subversion of Rationality (Cambridge: Cambridge University Press, 1983), 163.
 Elster, Making Sense of Marx, 4, 27-29.
 Ibid., 18-27, 465-93.
 Thompson, Ideology and Modern Culture, 56.
 See Ben-Ami Shillony, The Jews and The Japanese: The Successful Outsiders (Rutland, Vt.: Tuttle, 1991), 216-22.
 Sheila K. Johnson, "Japanese and Jews: Intersection of Myths," Los Angeles Times, November 30, 1992. A good example of this mixture of admiration and negative stereotyping is a 1972 book by Fujita Den, the president of McDonald's of Japan. Entitled Jewish Trade Practices, it advises the Japanese to be more shrewd and unscrupulous like Jews in order to achieve business success. Shillony, The Jews and The Japanese, 217.
 Leslie Helm, "Japan Newspaper Ad Revives Fears of Anti-Semitism," LosAngeles Times, July 29, 1993.
 Elster, Sour Grapes, 157.
 Thompson, Ideology and Modern Culture, 56, 59.
 Ibid., 68, 73.
 For example, the psychological mechanisms that reduce cognitive dissonance may operate to produce self-serving justifications by members of subordinate groups toward each other and forms of wishful thinking that work to the advantage of members of relatively dominant groups, as discussed in Chapter 8.
 Catharine MacKinnon, Toward a Feminist Theory of the State (Cambridge: Harvard University Press, 1989), 116.
 Thompson's model of ideology does not specifically consider the competing ideologies of subordinated groups. He is concerned only with the counterideology of subordinate groups that he calls "incipient forms of the critique of ideology." See Thompson, Ideology and Modern Culture, 68.
 See, e.g., Elizabeth V. Spellman, Inessential Woman: Problems of Exclusion in Feminist Thought (London: Women's Press, 1988); Martha R. Mahoney, "Whiteness and Women, in Practice and Theory: A Reply to Catharine MacKinnon," Yale Journal of Law and Feminism 5 (1993): 217-51.