A Theory of Ideology
J.M. Balkin

Yale University Press
New Haven & London

Copyright © 1998 by Yale University. All rights reserved.

Online version

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 JackBalkin.Yale.edu



Because language and symbol are the most pervasive forms of cultural transmission, they offer a rich trove of cultural software and ideological effects. This chapter examines the cultural software that is used to form social meanings transmitted through language in the form of conceptual oppositions and networks of conceptual associations.

Much political and social reasoning draws on conceptual oppositions. A simple and powerful example concerns American attitudes about race. Ameri­can culture understands whiteness and blackness as opposites, even thoughthere are several races in the United States, and the boundaries between those groups are hardly distinct. More important, the opposition between whiteness and blackness is understood in terms of a network of evaluative conceptual oppositions: law abidingness as opposed to criminality, morality as opposed to immorality, higher intelligence as opposed to lower intelligence, knowledge as opposed to ignorance, industry as opposed to laziness, and so on.[1] This network of associations is an important ideological mechanism in producing an image of black Americans as an inferior Other onto whom all manner of unsavory characteristics are projected. Moreover, this network of cultural associations helps sustain unjust stereotypes about whites and blacks, wrongly presents them as separate, homogeneous, and unified groups, and helps perpetuate misleading assumptions about American society. For example, many American whites as­sociate drug use, welfare dependency, and crime with blackness, and politicians regularly play on these associations, sometimes subtly and sometimes not so subtly. Nevertheless, a majority of drug users, welfare recipients, and criminals are white.[2] Thus the network of cultural associations projects criminality and immorality onto blacks and away from whites. This has a dual ideological effect, identifying blacks as the cause of America’s moral problems, and making white criminality and immorality relatively invisible.[3]

Semiotic Systems as Cultural Software

To explain how networks of association operate, I will use a well-known theory of cultural meaning-structuralism. Structuralism argues that cultural meaning is produced when subjects understand the social world through conceptual op­positions. These oppositions may be between things (sun and moon), directions (left and right), abstract concepts (reason and passion), degrees or qualities (higher and lower), classifications (male and female), and groups of persons organized by gender, race, ethnicity, or social class (men and women, blacks and whites).

People make sense of the cultural world not through isolated conceptual oppositions but through networks of linked conceptual oppositions. When peo­ple understand two things or concepts as opposed, they relate this opposition to other oppositions they are already familiar with. The difference between A and is understood in terms of the difference between C and D. A’s association with C is mirrored by the association of with D. New conceptual oppositions are fashioned by analogy to, and understood through association with, previous conceptual oppositions. The process of concatenation and collation of concep­tual oppositions goes on indefinitely. In this way, a huge network of associations develops that produces rich sources of cultural meaning.

This approach to cultural meaning is the basic insight of structuralist the­ory; it also underlies the semiotics (or semiology) of later thinkers like Roland Barthes.[4] For this reason, I shall speak of structuralist and semiotic analysis interchangeably. By each I refer to the basic strategy of studying cultural mean­ing in terms of networks of differences and conceptual oppositions.

Nevertheless, structuralism is also often associated with two more contro­versial claims about human cognition. The first is belief in innate ideas (asso­ciated with Noam Chomsky); the second is belief in universal structures of the human unconscious (associated with Claude Levi-Strauss). The idea that cul­tural meaning is constructed through networks of conceptual oppositions, how­ever, requires neither of these assumptions, and the theory of cultural software is committed to neither. To the contrary, I shall argue that the structuralist or semiotic method does not identify stable or permanent conceptual oppositions, though structuralist theory often presents them as such. Rather, what this anal­ysis identifies are the traces or marks of human cultural software as it evolves and combines in different cultures.

When Levi-Strauss offered his views about innate human ideas, he at­tempted to solve the problem that faces all accounts of cultural understanding:  to explain why and how shared meanings are shared. His solution, likeChomsky’s, was Kantian-style: every person possesses the same unconscious structures; hence each has the same tendencies to understand the social world in terms of networks of conceptual oppositions. Nevertheless, Levi-Strauss’s account does not really explain why particular substantive ideas are opposed to each other, why we find different combinations of ideas opposed or connected to each other in different cultures, or why people in the same culture might have different sets of cultural associations. Nor can his account explain how networks of cultural associations might change over time. At best he offers a formal and ahistorical account of cultural understanding.

Moreover, his assumptions are largely jettisoned in later versions of semi­otics and semiology: semiotic studies of advertisements and the fashion system, for example, do not presuppose universal unconscious structures. They are concerned with codes and meanings that change historically, often quite rap­idly.[5] But when semioticians ignore Levi-Strauss’s assumptions about the hu­man unconscious, they simply beg the question of how these cultural codes are created and shared.

A distinct advantage of the theory of cultural software is that it can offer answers to these questions. The human mind readily absorbs and memorizes conceptual oppositions and networks of conceptual association, just as it does in the case of narrative structures. This tendency may be innate, or it may be a side effect of other very basic forms of cultural software. Whatever the cause, networks of association can be and are widely dispersed through communica­tion and social learning. For example, the cultural associations of whiteness and blackness that I mentioned above are stamped into innumerable cultural sym­bols and messages that Americans absorb from their very earliest years. They are woven into countless sets of expectations and social meanings that we de­pend on in our daily encounters with others. That is one reason why racist forms of thought are so powerful, pervasive, and difficult to eradicate. Never­theless, these associations have not always existed; they are the products of memetic evolution. And one day, we may hope, they will again cease to exist in the minds of human beings.

The theory of cultural software also accounts for the ubiquity of certain conceptual oppositions despite cultural change. Very basic conceptual opposi­tions and associations (like male/female, day/night, or sun/moon) are likely tobe found in some form in almost every culture, though each may have different associations. Like other forms of cultural software, their ubiquity can be ex­plained in several ways: These basic oppositions may be dispersed to manydifferent civilizations through memetic descent. Or they may have been inde­pendently created because people in many different cultures have faced similar experiences and dealt with them in similar ways.

Conversely, a memetic approach allows us to account, in ways that a Levi-Straussian cannot, for diversity across cultures and dissensus within cultures. There is no reason that particular conceptual oppositions would be equallysalient for all persons in all cultures. For example, we would not expect that oppositions like right/duty or hardware/software would exist in every culture. Over time the kinds of associations that spread widely among members of a culture might change, with older ones transformed and newer ones emerging. Finally, if networks of association are a form of cultural software, they will exist in populations, with slight differences in each person’s tools of understanding. Although communication and social learning can harmonize the understand­ings of a culture’s members, no two persons will have exactly the same set of cultural associations.

In this way we can explain convergence of understandings without having to assume Levi-Strauss’s universal structures of the unconscious. And we can explain the forms of divergence, dissensus, and historical change that his theory could not account for. In short, the theory of cultural software shows how we can reinterpret structuralism in terms of individual thought and belief. Struc­turalism can be freed from its questionable metaphysics, and its insights can finally be appreciated by even the methodological individualist.

Structuralist Homologies

The central focus of structuralist analysis is conceptual homology. A homology is an association of conceptual oppositions. For example, given oppositions between and and between C and D, we might have the conceptual ho­mology “A is to as C is to D,” or A : B:: C : D. Although this form of analysis is best known through its application to the culture of primitive societies, I shall offer an example taken from contemporary American law: the contrasting roles of the judge and jury in American legal thought.

An American trial involves both questions of law and questions of fact. In the American legal system, citizen juries are the triers of fact, while the judge is responsible for determining the law to be applied to those facts. At the endof the trial, the judge offers instructions to the jury about the law that they must apply to the facts that they find. Reliance on lay juries to find facts and apply law in many different types of legal controversies is a peculiar feature ofAmerican legal culture. In England, for example, the jury has been eliminated in most civil cases (except, interestingly, in cases of defamation). In civil law jurisdictions, the lay jury has never had the importance it has had in America.

Gerald Torres and Donald Brewster have noted that the common under­standing of the division of labor between judges and juries in the American legal system is linked to a number of conceptual oppositions: these include not only law versus fact but also reason versus passion.[6]  Thus judges, who construe the law, are associated with reason, while jurors, who must sort out the partic­ulars of the factual situation, judge the credibility of witnesses, and act as the conscience of the community, are associated with passion and emotion. The judge is a learned professional who understands legal doctrine, while the jury is unschooled in the law, having only a “brute” sense of justice. The judge presides over many different cases and issues rulings of law that are entitled toprecedental effect. A particular jury sits on only one case and is dissolved after doing its work. It does not create durable rules of law but offers a ruling on only one factual situation; its finding has no general precedental effect (al­though specific findings of fact can bind the identical parties in future litiga­tion).

This analysis produces the following homologies: judges are to juries as law is to fact, as reason is to passion, as formal justice is to informal justice, as the permanent is to the transient, as justice through application of general rules is to justice in the particular case. The association of these conceptual oppositions is a tool of cultural understanding. Like all other such tools, it both enables and misleads understanding. This network of associations allows people to un­derstand, describe, and make judgments about the respective roles of judges and juries, but it simultaneously produces prejudgments, prejudices, stereo­types, and overgeneralizations about judges and juries. In this way the system of networked associations both assists cultural understanding and creates ide­ological effects.

The above homology explains a surprising amount about how people con­ceptualize the respective roles of judge and jury. Because the jury is associated with passion, for example, the jury, and not the judge, needs to be controlled.The emotional and discretionary functions of the jury are necessary to justice. Yet the jury must be disempowered, held back, for its own good. Otherwise, it will be a “runaway” jury (a term that connotes a wild beast–the symbol of dangerous passion). Hence American law contains a set of rules of evidence that are said to “constrain” jury discretion (as one would constrain a wild animal or a capricious child). Evidence is excluded when it would unduly prej­udice the jury (the evidence of a rape victim’s prior sexual history, for example) or inflame the jury (particularly gruesome evidence of a murder). Failure to abide by these rules can be reversible error. Thus to the previous homology of conceptual oppositions we should add the additional opposition of constraining: requiring constraint.

Our semiotic analysis of the way people think about and talk about judges and juries should not be confused with an assertion or a proof that judges really are more rational than juries or that juries really are more emotional than judges. Nor is it a claim that this is the proper or morally appropriate way of thinking about judges and juries. Rather, we are trying to identify patterns of existing cultural thought about judges and juries, and it is entirely possible that this way of thinking may be misleading or unjust.

Moreover, this analysis does not offer us a series of rules that people con­sciously follow when thinking about judges and juries. People do not recite the homology “judges are to juries as reason is to passion” to themselves and then apply it. Rather, the homology that we discover in symbolic forms is evidence of a cognitive construction that has already occurred, which the semiotician codifies into a set of organizing rules or principles after the fact. The semiotic analysis shows us the results of a series of generative tools applied to a particular aspect of social life. Because of the network of associations that people use in thinking about judges and juries, they construct what they consider appropriate roles for the social institutions of judges and juries, and their associated social institutions. Thus, it is not surprising that semiotics finds a set of structures in what was produced according to these generative tools. Nor is it surprising that, when asked to defend the institution of judges and juries, or the particular features of this institution, people will respond in terms of the network of oppositional categories that structuralism discovers.[7]

Although homologies are not rules consciously followed, they do seem to have an important hermeneutic function. Such analogies offer people a way of understanding conceptual oppositions and the opposed concepts themselves. The homology A : B :: C : D not only links and C (or and D); it also helps us to understand the nature of in terms of C, and the nature of C in terms of A. It also helps us understand conceptual opposites (like and B) in terms of each other. A homology of conceptual oppositions is mutually explanatory. It sheds light on the thing to be explained but also reflects light back onto that which is used to explain.

Mediation, Subcategorization, and Nesting

Although people employ conceptional oppositions to understand the world, the world does not always easily conform to on/off categories. Many situations arise that fall between the poles of existing conceptual oppositions; many situations resemble both sides of a given opposition, depending on how they are described or understood.

In fact, many legal situations mediate between the network of oppositions of judge and jury, law and fact, reason and passion. For example, judges in the American system are sometimes involved in factual issues. If judges can be associated with law rather than fact, how does the conceptual system that we have just discussed comprehend this situation? Often the mediation is treated as a subcategorization of one of the terms of a conceptual opposition, resulting in a new conceptual opposition. Instead of thinking of judicial fact-finding as a mediation between or mixing of the roles of judges and juries, for example, we might think of it as splitting the category of fact-finding into two catego­ries-facts found by judges and facts found by juries. This produces a new opposition of judicial fact-finding and jury fact-finding. The mediation of op­positional categories, or the subcategorization of one term in a conceptual opposition into two opposed terms, produces a second-level opposition, an opposition within an opposition. This phenomenon is called nesting.[8]

I have noted that people understand new conceptual oppositions in terms of older oppositions in a cultural system. The same phenomenon applies to oppositions created from subcategorization. A second-level opposition repli­cates the associations that are linked with the first-level opposition. Put sym­bolically, if A : B:: C : D, thena’: b’:: A : B:: C : D. If the judge is associated with reason, law, and restraint of passion, judicial fact-finding will be charac­terized according to these criteria in opposition to the fact-finding of juries (who are associated with undisciplined passion). That is to say, judicial fact­-finding is to jury fact-finding as judges are to juries, as reason is to passion, and as law is to fact. Judges engage in a more “reasonable” and disciplined type of fact-finding, or a fact-finding that cultivates reason and restrains pas­sion. Sometimes, for example, the parties waive their rights to a jury trial, and the judge acts as the trier of fact. In such cases, judges often dispense with many of the rules of evidence that they use to shield evidence from juries, on the grounds that unlike juries, they will not be unduly swayed (that is, that they are more reasonable). Moreover, because judges already know the law, they can easily sort out the admissible from the inadmissible evidence.

Even in jury trials, the judge engages in some kinds of fact-finding. Often, for example, the judge has to rule on whether certain evidence is admissible by reference to the facts. The judge’s fact-finding in these circumstances is in aid of reason because it helps restrain the jury’s unreasoning passion. In addition, a judge in the American system must follow the evidence presented at a jury trial in order to act as a check on jury discretion. If the jury’s verdict is too unreasonable given the weight of the evidence, the judge may order a new trial or, in extreme cases, direct a different result notwithstanding the jury’s verdict. Finally, the judge may remove a certain question from the jury if the testimony presented by the parties is such that no reasonable jury could come to a con­trary conclusion. Thus the judge finds facts in order to police the boundaries of reasonableness. Not surprisingly, such questions are called questions of law in the American system, even though they clearly involve factual inquiry.[9]

The concept of oppositions nested within oppositions is a corollary of a fundamental structuralist tenet: the meaning of a cultural event or artifact comes from its opposition to other cultural events or artifacts–in short, the context in which it is considered. The identification of judges with reason and lack of prejudice arises only in the context of an opposition with juries. Al­though judges are associated with reason in comparison with juries, there isgreat concern in jurisprudential debates over the possibility that judges them­selves might be swayed by passion and prejudice. In this case, the reason/passion opposition is no longer identified with judges as a group versus juries as a group; it is now used to understand a division within the category of judges themselves. The context has changed, and with it, the associations between conceptual oppositions. Nevertheless, the terms of this debate are a variation on the reason/passion homology that we have been studying. Indeed, we might even understand it as a subcategorization or mediation of the opposition be­tween judges and juries. The prejudiced judge is to the nonprejudiced judge as juries are to judges. The prejudiced judge is lawless, and therefore needs con­straint to perform her proper function; hence she must submit to the con­straints of the larger reason of the law. The law itself, which the judge applies, therefore acts as a constraint on the judge; it ensures that her actions are inaccordance with the reason of the law. Thus we have a mediation of the original homology. Judge constrained by law :judge unconstrained by law :: reason : passion. Once again, passion is dangerous and in need of control. Once again, reason is in charge of restraining passion.

Homologies and Hierarchies

We thus see two basic features in the structuralist theory of cultural meaning. The first is the notion of clusters of linked conceptual oppositions: judge/jury, reason/passion, law/fact, restraining/needing restraint. The second is the no­tion of oppositions within oppositions-the idea that mediating categories or subcategories reproduce prior associations in a new form.

Clusters or networks of oppositions do not merely differentiate situations and things. They also have evaluative significance. A differentiation can be a comparative evaluation or a statement of comparative importance. It can also assert a hierarchy. This hierarchy can be one of comparative value, of existing power or status, or of legitimate power or status. Thus the division between judges and juries may carry with it unspoken assumptions about the appropriate distribution of power and authority between them. That is because we are not neutral about the relative value of reason and passion, their relative importance in our lives, or the relative authority that each should have.

If we prefer reason to passion, then we also are likely to prefer that which is associated with reason to that which is associated with passion. We will think it important that reason be in control of passion, and hence that things asso­ciated with reason should be in control of those things associated with passion.  Passion is at its best when it serves its appropriate function and is in its appro­priate place. That is when it is subservient to reason or less powerful than reason (as in the case of judicial control of juries), when it resembles reason (as in the case of judicial fact-finding or a jury not swayed by unnecessary emotion), or when it acts to further the goals of reason (as in the case of juries supervised by judicial control).

Conceptual oppositions that form hierarchies of comparative value, status, power, or authority can be associated with other hierarchies of comparative value, status, power, or authority. Thus the hierarchy of judges over juries can be buttressed by analogies to other accepted or prevalent forms of hierarchy or comparative evaluation. Put more generally, the hierarchy of over can be supported by analogy to the hierarchy of C over D. Thus the homology A B :: C : D not only explains or clarifies the nature of and but also supports the comparative evaluation between them.

Torres and Brewster, for example, have suggested that the way that people talk about juries is also the way they stereotype women.[10] Juries are said to be capricious and unpredictable, easily swayed by emotion, and yet, on the other hand, intuitive, perceptive, and merciful; while judges are associated with the ”masculine” stereotypes of reason, law, rules, and order. This analysis does not claim that men and women are actually the way that masculine and feminine stereotypes portray them to be. It assumes only that these stereotypes exist and that they occur widely in American culture. Thus, the use of “feminine” meta­phors about the jury establishes a connection between the dominance of judges and the cultural hierarchy of patriarchy–the ideology that values men andthings associated with men over women and things associated with women. Stereotypes that justify one kind of hierarchy become linked to other hierar­chical oppositions and serve to justify them as well. In this way, the associated hierarchical oppositions mutually reinforce each other.

Indeed, patriarchy itself is supported in part by a series of analogies to other oppositions in culture and nature. The hierarchical relation of men to women is explained by, understood by, and justified by the relation of this opposition to other conceptual oppositions. These oppositions, in turn, are linked to still further ones, and so on indefinitely.[11]

We must be careful not to infer from the above example that the judge/jury relationship directly reproduces or supports patriarchy, or that the judge­/jury system is an inherently “male” institution. This misunderstands the natureof the ideological effect. People use hierarchies they are already familiar with in order to explain other hierarchies. A hierarchy is made to seem more natural by analogizing it to another hierarchy that already has some cognitive force. Moreover, two conceptual hierarchies may mutually reinforce each other by being associated. Nevertheless, the effect is not perfectly symmetrical, because some conceptual oppositions are more basic, or more powerful, or more central to our thought than others are. Thus, in theory, the opposition of judges and juries might reinforce the hierarchy of male and female or the opposition be­tween reason and passion just as the latter two conceptual hierarchies tend to explain, justify, and support it. In practice, however, it is more likely that the power of explanation runs largely in one direction. The hierarchical relations of male and female stereotypes and of reason and passion are probably more deeply rooted in our culture than the relation of judges and juries and thereforeoffer more support to the reproduction of the hierarchy between judges and juries than that opposition does to either of the other two.

The “feminization” of discourse about the jury is more than merely a de­scription and more than an evaluation. It also assigns roles of comparative authority and power. One way of establishing and justifying the superior power of judges over juries and reason over passion is to link this opposition with other oppositions in which one term has greater power over the other, or with oppositions in which one term should have greater power over the other.

In the standard conception of judges and juries, the jury and passion must be associated with lesser power because passion is not necessarily less powerful than reason in all of its manifestations. Passion can sometimes be more pow­erful than reason; consider a mob or a violent animal. Thus, the jury is “fem­inized” in discourse in order to associate it with a subordinated, less powerful version of passion-to avoid the dangerous implications of a passion that could be more powerful than reason. This discourse thus trades upon (or rather as­sumes) the comparative powerlessness of women in order to establish the sub­ordinate role of passion (or its representative, the jury) in a system of law. (More to the point, it trades on or assumes the patriarchal attitude that it is appropriate that women should be less powerful than men.) Thus when Torres and Brewster claim that the jury is feminized and therefore thought less pow­erful, they borrow an insight from feminist scholars that feminization of a concept is a way of dominating it.[12]

The association of women with less power than men is a standard patri­archal assumption. Even so, I must stress that the concept of women’s power is a problem for patriarchal ideology. It is more correct to say, from the stand­point of patriarchal thought, that whether or not women are less powerful, they must be associated with lesser power. The association of women with weakness or with lower hierarchical status must be established and reinforced and reproduced just as much as the association of lesser power for juries must be established, reinforced, and reproduced in the ideological construction of judges and juries. Thus patriarchal thought has many strategies for handling potential associations between women and power. Powerful women are seen as dangerous and unfeminine. Images of powerful women are generally unflattering stereotypes of deviance: witches, for example, are women who possess the power of magic. The power that women are permitted within patriarchal con­ceptions must be limited to specified areas, must never threaten the larger power of men, and must ultimately be subservient to male authority. These ideological strategies are never fully successful; even so, they may still be per­vasive.

Once again, the structuralist analysis does not assume that the stereotypes implicated in the homology

masculine : feminine :: more powerful : less powerful

necessarily reflect current social realities about men and women. Rather, the assignment of stereotypical male and female characteristics is one of the ways in which subordination of things associated with the feminine is justified ormade to seem natural. This assignment is a source of ideological power. The consistent use of feminine metaphors to describe a nonsexual concept like a jury is not so much evidence of differences between men and women as evi­dence of an ideological strategy of justifying particular relations of power. In the story of power relations, the one who plays the part of “the girl” is the one who is subordinated and whose power or dangerousness must be elimi­nated. (Indeed, the identification of woman with “girl”–a child–itself reflects a homology of greater power/lesser power.) Thus as a result of this analogical strategy, passion has its proper place within a system controlled by reason. A properly functioning jury can be emotional only as long as it is assigned the sort of values that patriarchy stereotypically associates with women who “know their place”–sympathy, mercy, intuition, and so on. The feminization of the jury thus solves the problem of keeping the jury in its proper place.

Note, moreover, that such analogies work in the opposite direction as well. The identification of the jury with passion and with feminine stereotypes not only justifies its constraint by the judge and the need to subordinate its passion and rough sense of justice to the judge’s reason and knowledge of law; it also establishes the connection between the judge and reason, thus justifying the judge’s superior position. Just as the jury is feminized (emasculated) in dis­course, the judge is associated with the “male” metaphors of reason, authority, and justice.

The Economy of Oppositional Logic

I have argued that networks of oppositions evaluate by combining conceptual oppositions that are evaluative or hierarchical. In our example, the hierarchy of judges over juries was supported by the hierarchy of reason over passion. This discussion assumed a general preference for reason over passion. Nevertheless, in conceptual oppositions like that of reason and passion, we do not always privilege one term over the other, and hence the two terms do not have a unitary set of hierarchical or evaluative associations. Each opposition is a conceptual tool that is used over and over again in many different contexts; through this usage, it becomes linked to many different oppositions, and it forms many different networks of association. Thus, although conceptual op­positions are used to buttress hierarchies or comparative evaluations, they can do so in many different ways, for each has many different associations con­nected to its terms, and these associations display the terms of the opposition in many different lights.

There are many good or superior associations connected with reason in its opposition with passion, for example, as well as many bad or inferior associa­tions. There are contexts in which reason is viewed as better or more powerful than passion, of greater authority than passion, or needing to be in control of passion, and other associations where the reverse is true. That is because we do not in every case prefer reason to passion or think that reason should always be in control of passion. In matters of artistic expression, a cold, logical, and unfeeling person may be thought to be at a comparative disadvantage with a person who we say is “in touch with” her emotions. In physical activities, people often report that it is better simply to act according to their feelings than to think consciously about what they are doing. To return to the example of legal controversies, legal decisions must be tempered with qualities like mercy, sympathy, and intuition if the law is truly to be just.

Similarly, in the many different contexts in which the conceptual opposition of mind and body is used, mind is sometimes the favored term, while body is sometimes more favored. There are both good and bad associations with mind in its opposition to the body, as well as good and bad associations with the body in its opposition to the mind. Thus we favor body over mind when we prefer the solid, practical, down-to-earth person to the impractical dreamer, the effete intellectual, or the pedant. Conversely, we may favor mind over body if the body is associated with the dangerous elements of passion, or with earthly troubles, while the mind is associated with reason and the more honored “life of the mind.” Under this network of associations, the philistine or savage is unfavorably compared to the person of culture, and the grasping merchant to the dispassionate scholar. It follows that the same conceptual opposition can have quite different meanings depending upon the context of associations com­bined with it. The savage can be noble or brutish, the person of culture wicked or civilized, depending upon the play of associations. The power of ideology, then, seems to come from its ability to emphasize some of these conflicting associations to the exclusion of others.

Let us call the preferred or privileged term in a conceptual opposition the dominant term and the other term the subordinate term. Let us call associa­tions of superior authority or superior value of a term in relation to its opposite”associations of superiority” or “superior associations” and associations in which the term has lesser authority or lesser value in its relation to its opposite ”associations of inferiority” or “inferior associations.” We can thus state our point more generally: The terms in a conceptual opposition have both superior and inferior associations, depending upon their place in different networks of conceptual oppositions. When a conceptual opposition is used in a network of evaluative or hierarchical oppositions, one term is dominant and the other sub­ordinate. The superior associations of the dominant term are emphasized and the inferior associations are deemphasized or suppressed. At the same time, the inferior associations of the subordinate term are manifested and the superior associations are deemphasized or suppressed. Even though the other associa­tions are forgotten in the process of buttressing a particular hierarchy or com­parative evaluation, they do not disappear. The deemphasized associations of the dominant and subordinate terms do not vanish. They are simply suppressed or forgotten in the particular associational network. They will emerge again in some other cultural context. They can also be used to attack or to deconstruct the evaluative hierarchy. Thus the networks of association that the structuralist identifies have a certain ideological force. But this force is hardly immune from attack. It can always be undermined by bringing into play an alternative net­work of associations in which some of the previously suppressed associationsare brought to light.

Conceptual oppositions like those of mind and body or reason and passion appear in culture in many different ways that look conflicting and contradictory if they are examined together. Traditional patriarchal attitudes provide a good example of how the same conceptual oppositions are employed in quite differ­ent ways. Patriarchy generally privileges the male over the female, although, as I shall discuss in a moment, this general privileging is accompanied by priv­ileging the female in distinct and limited areas of cultural life. The conceptual opposition of mind versus body is used in a patriarchal ideology in two ways–­one that connects the superior associations of the mind with the male and one that connects the superior associations of the body with the male. The first way simultaneously connects the inferior associations of body with the female, while the second does so for the inferior associations of mind.

Thus, if we take the homology male : female :: mind : body :: reason : pas­sion, we obtain a familiar set of patriarchal stereotypes: Men are reasonable while women are too emotional. Men can distance themselves from a situation and consider its merits abstractly, while women are too contextual and too caught up in the feelings generated by a situation. Men think, women feel. In each of these stereotypes, the male association is preferred, and thus the associations of mind are preferred to those of the body. Yet if we now consider the homology male :female:: body : mind:: active : passive, we can produce another set of patriarchal stereotypes in which the male association is still fa­vored, even though it is now conjoined to the body. Thus women cannot do jobs that require the strength of men, men are practical while women are impractical, men act while women worry, men work while women gossip, and so on. Here the associations linked to the body are preferred to those of themind. Jeanne Schroeder has summed up this phenomenon aptly when she states that in patriarchal thought, a thing is privileged not because it is male, but is called “male” because it is privileged.[13]

One might wonder how patriarchy can operate successfully if it employs a conceptual opposition like mind and body in completely contrary ways. Why doesn’t the apparent contradiction of privileging and then deprivileging the same term in different contexts lead to the intellectual incoherence of patri­archal thought and thus destroy its power over our imaginations? This situation is puzzling only if we assume that a way of thinking is powerful because it is coherent or orderly. But as I have argued, cultural software is the product of cumulative conceptual bricolage. It is not a rationally designed structure of conceptual relationships, but a historical jerry-built product. Patriarchy, like the panda’s thumb, is designoid rather than designed. It may have the surface appearance of design, but closer inspection reveals its historical and adventi­tious development.

The network of conceptual oppositions that we identify with patriarchy is the product of repeated applications of a relatively small number of conceptual oppositions in new and different situations. Conceptual oppositions can be used in many different ways and contexts because their terms have many different features and many different associations. Conceptual oppositions, in other words, are like versatile tools that can be used in many different ways for many different tasks, with correspondingly different results, and corresponding ad­vantages and disadvantages. Oppositions like mind and body, male and female, dark and light are employed over and over again to understand various features of the cultural world. As a result they are repeatedly stamped, like a marker, onto many different aspects of human culture, and the products that they ex­plain carry these associations with them as they are in turn used to explain or understand further aspects of culture. In this way, conceptual oppositions like mind and body or male and female become ubiquitous through recursion.

Thus our cultural understanding of gender is the result both of the many different conceptual oppositions used to understand the opposition of male and female, and of the many uses of the opposition of male and female to under­stand other aspects of culture. This process will evidence no grand design, although in hindsight people may see pattern and order; it may even look as if culture is the result of a grand patriarchal conspiracy. Yet once we recognize the historical construction of patriarchy, we should be surprised by its relative coherence and consistency, rather than by any lack of these features.

Nested Privileging

We have just seen that conceptual oppositions can be used in conflicting ways to buttress a comparative hierarchy or evaluation. We should now note the converse phenomenon: a network of conceptual oppositions does not privilege the same elements to the same degree in all cases. Consider once again the example of judges and juries. The cultural characterization of judges and juries that we have been examining generally privileges the judge over the jury to the extent that it views juries as ruled by passion and in need of judicial control. Yet this general ideological privileging is not a claim that judges are superior to juries in every respect. Indeed, ideological privileging often works by des­ignating particular areas in which the subordinated term appears to be domi­nant or privileged in some way. A large-scale ideological privileging of judges over juries contains smaller-scale areas or pockets where ideology attributes superior associations to juries.

The system of cultural beliefs about American judges and juries that we have been examining assumes that juries and judges have their proper roles and their proper places in a scheme of adjudication. Although judges need to su­pervise juries, juries have specific tasks to perform in the American trial system. The system relies on the jury to resolve conflicts of testimony, to judge the credibility of witnesses, and to apply concepts like sympathy, intuition, com­mon sense, and “rough justice.” These tasks are considered inappropriate for the judge unless the parties have agreed to a bench trial. If the parties have not so agreed, it is improper for the judge to usurp the jury’s role, even though the judge is in charge of controlling the jury and ensuring that it conforms to reason and law. If a judge improperly takes a factual question away from a jury and substitutes her own judgments of fact, for example, a higher court will usually reverse her. In short, within certain specified boundaries, the jury has a role that it performs better than the judge, or, more accurately, in which it has greater moral, political, or institutional authority to act than the judge.

It is therefore incorrect to say that the privileging of judges over juries is systematic in every area of the legal system. Juries and not judges are the privileged institution in some aspects of the trial process. This privileging islimited, however; it is of a secondary or subsidiary order. Although juries have positive associations in particular areas of the system, these areas of the system are in turn subordinated to the areas where judges are privileged over juries. The subordinated term, in short, is awarded first prize, but only in second-class contests. The jury is celebrated as long as, and to the extent that, it knows its place and performs the tasks assigned to it under the larger ideological schema. An ideological schema therefore works not only by evaluation but also by distribution–that is, it assigns particular places or situations in which the subordinated term will receive positive evaluations or some limited degree of hierarchical position or authority.

One can see the process of distribution at work in the context of traditional patriarchal stereotypes about men and women. Traditional patriarchy associates the opposition of male and female with the oppositions of the market and the family, or the public world of work and the private world of the home. This reflects the homology male :female:: market : family:: public : private:: out­side : inside:: marketplace : home. But in traditional patriarchy, women are not devalued in all areas of life. Rather, they are specifically valued for their con­tributions to the maintenance of the home and the family. Moreover, under traditional stereotypes, men are thought incompetent at, or at least not as skilled as women in, the tasks to which women are assigned–housework andchild rearing. According to traditional patriarchal views, married men who be­come full-time “househusbands” are devalued accordingly. They are subjected to social obloquy because under traditional stereotypes they are regarded as lazy, _unambitious, parasitic, or effeminate.

Men, of course, do work in the house even within traditional patriarchal stereotypes. This is reflected in a mediation or subcategorization of the original homology. Thus male : female:: public world of work : home:: men’s house­work : women’s housework. Men can be “handy around the house,” but this does not refer to child care or to cleaning and cooking but rather to a type of ”housework” that is appropriate for men–tasks like repairing machines or electronic equipment, carpentry, plumbing, mowing the lawn, and so on. (In­terestingly, the expression often used for this is work “around the house” rather than work in the house, which uncannily reflects the homology of male : female :: public : private:: outside : inside.) Under traditional stereotypes, it is thought unusual or even unfeminine for women to engage in this type of activity, even though such tasks are also clearly “housework” in the most literal sense.

When men engage in traditionally female housework, they are subjected to a “double whammy.” This double whammy also involves a form of conceptual mediation or subcategorization, but it works in the opposite way from the example of the “handyman.” The idea of being handy around the house is a conceptual mediation that preserves male privilege by conceptualizing certain work as male even though it is associated with the home. On the other hand, the phenomenon of the househusband, who deliberately undertakes tradition­ally “female” housework, is understood as a threat to traditional patriarchal male roles. The mediation is therefore viewed as a bad example, deviant behavior or incompetent performance in two directions. First, househusbands are viewed as working in a sphere inappropriate to their talents. So they are as­sumed to be incompetent at housework and child rearing, or at least less com­petent than women are. Second, because they adopt “women’s work,” they areacting in an “unmanly” fashion. They are assumed to lack manly virtues or to be failures at living up to expectations of male identity. The double whammy is the double penalty for crossing a boundary established by an association of two conceptual oppositions-it involves a penalty assessed from both sides of the association. In this case, it is the belief that a man who does women’s work is a comparative failure both at doing this kind of work and at being a man.

The double penalty works equally in the other direction. Women who adopt roles, traits, or attitudes that are assigned to men are subject to criticism in two directions. First, their performance in the male occupation and role is suspect. Second, they are criticized as unfeminine. Moreover, even their traits and attitudes are interpreted in negative terms. This is possible because traits and attitudes have both positive and negative associations. Consider, for ex­ample, a woman who is relatively unemotional. This trait violates the homology male : female:: reason: passion. Thus she may be criticized as being unfemi­nine. But she may also be assigned the negative traits associated with lack of emotion: coldness, dullness, ruthlessness, and secretiveness, whereas a male with a similar countenance might be assigned the positive associations of sta­bility, reasonableness, dependability, and depth.

Nested forms of privilege and evaluation are also present in American race relations. At the beginning of this chapter I noted the homology white : black :: knowledgeable : ignorant. The existence of knowledgable, intelligent blacks presents a problem for this homology, which is solved through a nested priv­ileging. Knowledgeable and intelligent blacks are viewed as more “white” than other blacks. Nevertheless, they are also regarded as somehow less accom­plished than the most accomplished whites. Stephen Carter has described a “Best Black” syndrome, under which black professional and educational ac­complishments are routinely judged against the standard of other blacks but not against those of whites.[14] A successful black scholar in a particular field thus becomes the best black but not the best scholar. This phenomenon offers black intellectuals some degree of status, but it is an inferior grade of status that ultimately suggests that blacks cannot compete with whites.[15] Blacks areconsidered successful, but only within a greatly circumscribed arena of com­petition. Carter argues that this phenomenon is exacerbated by race-conscious affirmative action programs, but it probably existed long before these programs arose.

More generally, black accomplishments and culture are often valued and appreciated in relatively narrow segments of society (sports and music, for example) which are subordinated to more “serious” concerns. Association of whiteness with the mind and blackness with the body leads to expectations that blacks will naturally excel in things having to do with the body or with the emotions, whereas whites will excel in things having to do with the mind, with intelligence, and with mental discipline.

This opposition between mind and body is reproduced even in the world of American sports, where black talent is particularly valued. Black athletes are thought to display “natural talent,” but white athletes are said to become ac­complished because of hard work, emotional maturity, and mental toughness. In this way black accomplishments and black culture are permitted to reign, but only in second-class fiefdoms.

Traditional attitudes toward homosexuality offer excellent examples of the ideological power of conceptual homologies and the operation of the double penalty. Many theorists argue that discrimination against homosexuals is linked to the preservation of traditional gender roles and stereotypes, which are both heterosexual and patriarchal.[16] Masculinity and femininity are defined in terms of attraction to the opposite sex. This produces the homology male : female :: attracted to women : attracted to men:: attractive to women : attractive to men. This homology, in turn, is linked with the homology male : female:: manly feminine:: dominant : subordinate. Homosexuality, and especially male ho­mosexuality, threatens this conceptual order because it undermines the clarity of traditional male and female gender identities and hence the clarity of ap­propriate male and female social roles, authority, and power. By confusing the network of gender associations, gay men appear both to surrender their mas­culine privileges and threaten the masculine privileges of other males. In like fashion, lesbians threaten the conceptual order because they refuse their roles as wives and mothers within a traditional heterosexual family.[17]

Thus homosexuality mediates between the conceptual network of associa­tions that define femininity and masculinity. But because this mediation is seen as destabilizing, it is interpreted pejoratively and assessed a double penalty. Homosexuality is viewed by heterosexual culture both as a failed case of het­erosexuality and as a failed attempt at gender crossing.[18] This is because gen­ders are defined in heterosexual culture in terms of desire for the other gender, yielding the conceptual homology heterosexual male : homosexual male:: mas­culine : feminine:: better example of masculinity : worse example of masculin­ity. A classic stereotype of homosexuals, for example, is that they display cross-gender behavior: lesbians are viewed as aggressively masculine, and gay men are viewed as effeminate. Homosexual men are stereotyped as wanting to dress in women’s clothing or as seeking “feminine” jobs. Because this is not a man’s “true nature” (which is to desire women and be associated with “manly” things) the homosexual man is doubly a failure. He is a failure as a man (because he does not desire women and attempts to behave like them); and he is a failure as a woman (because he is a man).

As part of the double penalty, homosexual men and women are also as­signed the negative stereotypes of the opposite sex. Thus gay men are not only viewed as failures as men but are also described in terms of the negative ster­eotypes of women-as being bitchy or passive, impulsive or overly emotional. Conversely, lesbians are not only viewed as failures at femininity but are also assigned negative associations of masculinity, like aggressiveness and coldness.

Categories as Nested Oppositions

Structuralist analysis shows us that networks of conceptual oppositions produce ideological effects because hierarchies of value, power, or authority are justified and sustained by their association with other hierarchies. A second type of ideological effect stems from the division of the social world into exclusive and opposed categories that are then opposed hierarchically.

To discuss these ideological effects, I will use deconstruction. Many kinds of critiques have come to be called deconstructive. In this chapter, however, I refer to my own theory of deconstruction, which, as I have argued elsewhere, makes the most sense out of many of the textual techniques employed by Der­rida and other deconstructionists.[19] This approach to deconstruction is called the theory of nested oppositions.

A nested opposition is a conceptual opposition in which the two terms bear a relationship of mutual dependence as well as differentiation. Put metaphorically, it is a conceptual opposition where the opposed terms “contain each other.” This relation of containment or mutual dependence may take one of several different forms, including similarity to the opposite; being a special case of the opposite; conceptual overlap with the opposite; historical, conceptual, or ontological de­pendence on the opposite; generation from the opposite; or transformation into the opposite over time.[20] These relations do not have a single property in com­mon; rather they have a Wittgensteinian “family resemblance.” The idea that conceptual opposites are nested is very old-mystical versions of this insight may be found in pre-Socratic thought in the West and in Taoist philosophy in the East.[21] The idea of a nested opposition is vividly captured in the symbol of yin and yang, which are traditionally portrayed as tearshapes of opposite colors thatseem to grow out of each other. In many portrayals of the yin/yang symbol, a small dot of the opposite color is placed in each figure in order to emphasize the mutual similarity and dependence of opposites.

The theory of nested oppositions can be stated simply: every conceptual opposition can be reinterpreted as some form of nested opposition. To decon­struct a conceptual opposition is to view that opposition as a nested opposition.  It is simultaneously to recognize the similarity and difference between opposites or the mutual dependence and differentiation of opposites. The idea of mutual dependence and differentiation between concepts underlies Derrida’s notion of Difference.[22]

Although the theory of nested oppositions claims that every conceptual opposition can be viewed as a form of nested opposition, it does not assert that all nested oppositions are false oppositions or false dichotomies. The fact that two concepts are thought to be similar or mutually dependent in some context does not mean that they are identical in all contexts. This is a confusion of similarity and identity; it betrays an insensitivity to the many changes in context that are the primary concern of deconstruction.

The analysis can work in the opposite direction: we can also find difference among things normally judged to be similar. Instead of discovering that the terms of a conceptual opposition bear a relation of similarity, we can discover a conceptual opposition among things that were previously judged to be similar. In some contexts of judgment we do not notice any conceptual opposition at all but see only similar things. The theory of nested oppositions reminds us that this similarity can become a difference-and hence a conceptual opposi­tion–if the context of judgment is sufficiently altered. Just as deconstruction does not show that all conceptual oppositions are false dichotomies, it does not show that all similarities are false unifies. The latter conclusion would also beinsensitive to the subtleties of contextual judgment.[23]

Although every conceptual opposition can be seen as a nested opposition, we often do not recognize the nested nature of such oppositions. Thus an important ideological mechanism is the suppression of similarity within con­ceptual oppositions and the suppression of difference within categories. Indeed, to some degree all thought necessarily suppresses the nested nature of con­ceptual oppositions. The very act of categorization involves the creation of a conceptual opposition (things inside versus outside the category) and the sup­pression of its nestedness (differentiation within the category and similarity or dependence across the conceptual boundary). This is yet another example of how an ideological mechanism makes use of the same conceptual tools as so-­called ordinary or nonideological thought. Indeed, we might even say that the opposition of ideological and nonideological thinking forms a nested opposition whose nestedness has been forgotten.

Suppression and Projection

Let us apply the concept of a nested opposition to the example of judges and juries. In our previous discussion of structuralist analysis, we noted that the homology of judge : jury :: reason : passion :: law : fact :: rule of law justice in the individual case characterized the respective roles of judges and juries. We can reveal the nested nature of these oppositions in many ways. One way is to reverse the associa­tions. We can deconstruct the homology of judge :jury:: reason : passion by showing how the judge, the law, and the rule of law itself can be the embod­iment of unreasoning passion, prejudice, and partisanship, while the finder of fact, and the purveyor of justice in the individual case, the jury, is actually the embodiment of reason and fairness.

It is obvious that as judges are human, they can sometimes be prejudiced, and the rulings that they make will reflect their biases. The judge can therefore be prejudiced or swayed by passion. More fundamentally, however, the law itself can sometimes embody prejudice or unreason. This may seem strange at first because we normally associate passion and prejudice with volatility and changeableness (consider the homology reason : passion :: cool :hot:: stable volatile). Yet the prejudice or the unreason of the law is a prejudice and an unreason in a relatively fixed medium.

How can the law reflect unreason or prejudice? Sometimes the law is too harsh: it fails to take into account the special circumstances of cases, or it lumps dissimilar cases together mechanically and insensitively. In such a case, the jury may ameliorate the harshness of the law by “finding” facts that, when applied to the law, produce a less harsh result. During the era of industrialization in Britain and America, for example, the courts established tort doctrines that allowed defendants to escape liability for negligence (for example, manufactur­ing defective products or maintaining unsafe working conditions) if the plaintiff also contributed to her injury in any way. These doctrines had such a draconian effect that eventually juries would often “find” that the plaintiff was not neg­ligent at all–even though the evidence indicated the contrary–and would award her a slightly reduced amount of compensation. The sub rosa manipu­lation of facts by juries might thus be compared to the Solomonic judgment­the judgment that at first looks like an attack of unreason but actually conceals a deeper rationality.

Indeed, sometimes the law may be not simply clumsy and insensitive but actively unjust–it may be the product of bigotry or prejudice or a denial of human rights. In such cases a jury may refuse to enforce the law as written, a process called jury nullification. One of the most famous trials in American history involved the prosecution during the colonial era of the newspaper editor Peter Zenger for seditious libel–that is, for accusing a government official of improper conduct or otherwise holding the official up to public scorn. Under the existing colonial law, the crime of seditious libel was so harsh that one could not even defend oneself on the grounds that the matter asserted was true.Over repeated threats by the sitting judge, who was determined to force a conviction, the defense counsel argued to the jury that the concept of free speech required the right to speak the truth without fear of prosecution, not­withstanding the positive law of seditious libel. The jury refused to enforce the law and acquitted Zenger. The origins of free speech doctrine in America thus began with an act of jury nullification. Indeed, Akhil Amar has argued that the right to trial by jury was later placed in the American Bill of Rights as a rec­ognition of the importance of citizen juries as defenders of human rights through jury nullification.[24] In these circumstances, then, the jury may seem to be a more reasonable institution than the judge or even the law, and it may even be necessary for the jury to exercise control over the law. The possibility of justified jury nullification or sub rosa manipulation of the facts subverts the homology of reason : passion :: judge : jury :: law : fact :: controlling : needingto be controlled. It does so by offering a new homology in its place: reason passion :: jury : judge :: fact : law ::controlling : needing to be controlled.[25]

Our deconstruction suggests an important feature of conceptual homolo­gies. The identification of judges with law and reason and juries with facts and passion is not only an explanation, evaluation, or hierarchization. It is also asuppression and a projection. Associating the judge with reason and things associated with reason (for example, the rule of law) simultaneously suppresses elements of judicial behavior that might be unreasonable or “passionate,” and projects them onto the judge’s opposite, the jury. In psychoanalysis, the term projection is used to describe the patient who sees in others traits, characteristics, or desires that she does not wish to acknowledge in herself. The homology

judge : jury :: rule of law : justice in the individual case :: reason : passion

involves a different kind of projection. It is a suppression or a forgetting of certain traits or characteristics that might apply to one side of the opposition by assigning them to the other side of opposition. By identifying the judge and the law with reason in opposition to passion, the homology forgets, deem­phasizes, or suppresses the passion or the unreason that inheres in judges and in the law itself. We might also call this process a distribution: it distributes reason to one side of the homology and passion to the other.

Suppression and projection help to maintain a hierarchy or comparative evaluation in a conceptual opposition. The reason/passion opposition subor­dinates the latter term to the former. Passion, in its more powerful and dan­gerous forms, is a threat to reason; if it is to be useful and nonthreatening to reason, it must be seen as under reason’s control. In order to be seen as under control, passion must be externalized and projected onto an Other that can be disciplined, subordinated, or supervised. Control through externalization requires that passion (unreason) must be located outside of the materials of the law and the system of rules. Hence the prejudice that might be located in the judge (or the legislature) is projected instead onto the jury, where it can be supervised and subordinated. This control can take the form of actual legal doctrines and institutions, or it can simply rest in the general sense that it is juries and not judges who present a problem that needs addressing. Without this projection, the nonrational element might be seen as emerging from the judges themselves or the law itself; this would destabilize and delegitimate the view of law as reasoned, ordered, and impersonal; it would undermine thecomparative evaluation or hierarchy that supports judges and the law.

The deconstructive analysis of homologies has a curious consequence. We have deconstructed the homology of

judge : jury :: reason : passion :: rule of law : justice in the individual case

by substituting a counterhomology. It follows that this new homology, precisely because it is a homology, will also involve its own form of suppression and projection. Thus our attempted subversion is as deconstructible as the original homology it attempted to subvert. We can deconstruct our deconstruction by noting that the portrait of the wise and rational jury it paints is much too optimistic. Few subjects can arouse the concern and indignation of lawyers and judges more than jury nullification. Judges are particularly incensed by lawyers who dare to argue, as did the defense counsel in the Zenger case, that the jury may disregard the law and substitute its own judgment about what is reason­able. Such an argument, they might insist, strikes at the very heart of the concept of a rule of law; more practically, it also strikes at the heart of the control of judges and lawyers over the lay jury. It is no accident that judges make clear repeatedly in their jury instructions that jury members have sworn an oath to obey the law, and that they must apply it whatever their private reservations about the case. (Note the homology implied here of public law private opinion:: reason : passion.)

The nullifying jury is the epitome of the “runaway” jury, the jury that takes the law into its own hands. Such a jury is the embodiment of unreasoning passion; it is an invitation to anarchy. If a jury can nullify a verdict, to protect human rights, it can also nullify a verdict to satisfy its own racial or religious prejudice. If a jury can find Peter Zenger innocent of seditious libel, it can also nullify a homicide charge against a white racist who murders a black civil rights protester. Only adherence to the judge’s instructions, and hence to the rule of law, can constrain the prejudices and passions of the jury. If the original ho­mology projected passion and unreason away from the judge and the law onto the jury, our deconstruction of it projects and distorts in the opposite direction. The jury is now viewed romantically as the guardian of rough justice and common sense, while the unreasoning and dangerous elements of passion are pro­jected onto legal administrators and the content of the law itself.

Neither the deconstruction of the original homology nor its reassertion captures the whole truth about the relation of judges and juries. Each homology involves a forgetting, a suppression, and a distribution of different sides of an opposition to different elements within the homology. Nor is this an accidental feature of this example. Indeed, I claim that this forgetting, this suppression, and this distribution constitute a commonplace ideological mechanism.

A homology distributes opposed properties or concepts (like reason and passion) between two other opposed properties or concepts (like judge and jury). This distribution is always both a projection and a suppression. Thehomology A : B :: X : Y neglects the Y-ness of and assigns (distributes) it to B; it forgets the X-ness of and conveys it to instead. This process works in reverse as well: the A-ness of is forgotten and distributed to X,while the B-ness of is similarly forgotten and distributed to Y. When one finds linked together a large network of associated oppositions-for example, that of

judge : jury :: law : fact:: reason : passion :: rule of law : individuated justice

and so on–it becomes clear that many different forms of suppression and projection are involved and hence many different types of deconstructions be­come possible.

Indeed, the possible deconstruction of the homology we have been work­ing with seem virtually limitless. For example, our discussion so far has left unquestioned the assumption that passion is a dangerous thing that must becontrolled by reason. We merely argued that the dangers of unchecked passion can be present in judges and in the law itself as much as in juries. Thus we still were contrasting undesirable elements of passion to desirable features ofreason. But we might challenge these assumptions about reason and passion. We can do this in several ways.

First, we might note that reason in the law can take undesirable forms­mechanical reasoning, rigid formalism, and “logic chopping,” for example­that need to be checked by desirable forms of passion, qualities like mercy, sympathy, and equity. The development of courts of equity to supplement and supervise courts of law is a historical embodiment of this recognition. (We should note that historically courts of equity became just as rigid and technical in their doctrines as the law courts that they were designed to supplement. This suggests that law is continually in need of equity, even in those institutions and doctrines originally designed to promote it.) We thus might argue that reason has dangerous and deleterious aspects that need to be checked by the virtues of passion. This suggests that the just judge, the one who performs her proper function, does not put aside passion and the qualities associated with it but rather incorporates them into her judging.[26] Similarly, as justice Oliver Wendell Holmes believed, law gains its ultimate legitimacy and intelligibility from its reflection of the sentiments of the larger community.[27] This decon­struction thus subverts the earlier homology by uncovering the good or praise­worthy aspects of passion in the judicial function and in the law and asserting their centrality.

Second, we could distinguish between the idea of reason, which we continue to hold a good and desirable thing, and its concrete historical embodiment in institutions like law. We can argue that what passes for reason in law is some­times not very reasonable-that it conceals its own forms of unreason, bias, and prejudice. In other words, although reason in the abstract may be desirable, its concrete manifestations often possess significant elements of undesirable passion. Conversely, what people think is merely the exercise of passion and emotion by juries and those entrusted with discretionary decisionmaking often turns out to have sound principles of reason behind it. This type of decon­struction works by driving a wedge between an abstract ideal and the concrete embodiment of the ideal. It also attempts to uncover a kind of suppression and projection. By identifying a concrete phenomenon with an ideal (identifying, for example, an actual legal institution with the idea of reason) people downplay the ways in which the concrete phenomenon fails to live up to the ideal model.

Third, we have so far assumed that reason and passion are separate entities. We can also show that the opposition between them is nested by showing that reason and passion depend upon each other, or that concrete examples of rea­son and emotion have a great deal in common that goes unnoticed .[28]

Finally, we can show that the opposition between reason and passion is nested by showing how the two terms produce each other or are transformed into each other over time. The reason embodied in law, for example, becomes a kind of passion (prejudice) unless it is made flexible and is subject to ques­tioning and reconceptualization over time. The same is true of beliefs held for a long period of time without question, or of unquestioned obedience to an authority that represents reasoned inquiry.[29]

Conversely, the passion embodied in a jury’s act of nullification–as in the Zenger trial–may cause people to change their minds and see the greater reason in the free speech position espoused by the jury. We see the beneficial effect of passion on reason in many other situations as well: protest and civil disobedience are often necessary to bring home to the audience the injustice of current conditions and laws. Often a highly charged emotional experience (like witnessing a person who is helpless, injured, or suffering) allows people to “see reason” and to change their minds about matters they had been firmly committed to. Thus an experience that touches us emotionally may melt a congealed prejudice that goes by the name of reason and lead to a more reasonable outlook. These examples suggest both that reason can degenerate into prejudice and that passion or emotion can be the wellspring of a more just reason.

Conceptual oppositions have power over human thought because they fa­cilitate thought; and this facilitation is the source of their ideological effects. Conceptual oppositions divide the world into comprehensible categories, but they also suppress similarities and distribute characteristics onto an opposite or an Other. Networks of oppositions help us understand and evaluate the world, but they also reinforce unjustified hierarchies of value and authority. Concep­tual oppositions are both necessary and deceptive, ineluctable and illusive; they are consummate examples of the equivocal nature of our cultural tools.

[1] See, e.g., Kimberle Williams Crenshaw, “Race, Reform, and Retrenchment: Transformation and Legitimation in Antidiscrimination Law,” Harvard Law Review 101 (1988): 1331-87, at 1370-74.

[2] See Office of Applied Studies, U.S. Department of Health and Human Services, Preliminary Estimates from the 1994 National Household Survey on Druge Abuse (Advance Report Number 10, 1995), 76 (estimating that 76 percent of illicit drug users are white, compared with 13.67 percent who are black); Martina Shea, U.S. Department of Com­merce, Dynamics of Economic Well-Being: Program Participation, 1990-1992 (1995), 13 (reporting that 64.5 percent of recipients in major means-tested aid programs are white, while 31.3 percent of such recipients are black); Kathleen Maguire and Ann L. Pastore, eds., Bureau of Justice Statistics, U.S. Department of Justice, Sourcebook of Criminal Justice Statistics (1995), 408 (reporting that 66.6 percent of all criminal arrestees in 1994 were white, while only 31.3 percent of arrestees were black). Although it is true that the percentages of blacks in these groups is higher than their representation in the U.S. population, these statistics nevertheless undercut the all too prevalent assumption that blacks are responsible for the vast majority of these social problems.

[3] See Patricia A. Williams, The Alchemy of Race and Rights: The Diary of a Law Professor (Cambridge: Harvard University Press, 1991).

[4] Roland Barthes, Mythologies (New York: Hill and Wang, 1972).

[5] See, e.g., Robert Hodge and Gunther Kress, Social Semiotics (Ithaca, N.Y.: Cor­nell University Press, 1988).

[6] Gerald Torres and Donald Brewster, “Judges and Juries: Separate Moments in the Same Phenomenon,” Law and Inequality 4 (1986): 171-88.

[7] Thus, although structuralism often claims to look for the underlying “grammar” of culture, the analogy is partly misleading. Structuralist and semiotic analyses do not so much reveal a set of articulable rules consciously used by subjects as the results of practical cognitive work that employs networks of association. Semiotic analysis studies what individuals produce using generative tools of understanding rather than articulable ”instructions” of cultural software itself.

This distinction is central to Pierre Bourdieu’s critique of structuralism. See Pierre Bourdieu, Outline of a Theory of Practice (Cambridge: Cambridge University Press, 1977); Pierre Bourdieu, The Logic of Practice (Cambridge: Polity, 1990). In Bourdieu’s words, we must not confuse the opus operandum with the modus operandi; that is, we must not con­fuse the structural features of the cultural artifacts produced with the cognitive capacities that produced them. Our conceptual tools create cultural artifacts with a particular con­ceptual structure. The thing produced is evidence of the generative capacities that pro­duced it; it is not, however, identical with those capacities. Rather than the compiler of a social grammar or a social etiquette book, the semiotician should think of herself as a fo­rensic scientist who, presented with the victim’s body, tries to imagine the murder weapon and the means and opportunity for the crime; or as a geologist who, presented with the sedimentary layers of culture, tries to discern the processes that led to that sedimentation.

[8] J.M. Balkin, “Nested Oppositions,” Yale Law Journal 99 (1990): 1669-1705.

[9] The standard of appellate review of a jury’s verdict also reflects the opposition between law and fact. Questions of law are reviewed de novo; that is, the higher court judge may substitute her own view of the law for that of the lower court judge. On the other hand, an appellate court is loath to reverse factual findings by a jury unless they are completely unreasonable, or, in the case of a trial by a lower court judge, clearly erroneous. This suggests a further homology between appellate court and trial court judges, on the one hand, and judges and juries on the other. Appellate courts are usually called “higher” courts, trial courts are usually called “lower” courts. One appeals to ahigher court from a lower court. The work of the higher courts is concerned almost exclusively with matters of law.

[10] Torres and Brewster, “Judges and Juries,” 181.

[11] The study of conceptual oppositions does not exhaust the nature of patriarchal thought. As noted in Chapter 5, patriarchy is the result of a jumble of variegated and heterogenous mechanisms. Here I focus on only one of the many different ideological mechanisms of patriarchy.

[12] See, e.g., Catharine MacKinnon, “Difference and Dominance,” in Feminism Un­modified: Discourses on Life and Law (Cambridge: Harvard University Press, 1987), 32­45.

[13] Jeanne L. Schroeder, “Abduction from the Seraglio: Feminist Methodologies and the Logic of Imagination,”Texas Law Review 70 (1991): 109-210.

[14] Stephen L. Carter, Reflections of an Affirmative Action Baby (New York: Basic, 1991),47-69.

[15] Ibid., 50-52.

[16] Andrew Koppelman, Antidiscrimination Law and Social Equality (New Haven: Yale University Press, 1996), 146-76; Silvia A. Law, “Homosexuality and the Social Meaning of Gender,” Wisconsin Law Review 1988: 187-235; Marc A. Fajer, “Can Two Real Men Eat Quiche Together? Storytelling, Gender-Role Stereotypes, and Legal Protection for Lesbians and Gay Men,” University of Miami Law Review 46 (1992): 511-651.

[17] J. M. Balkin, “The Constitution of Status,” Yale Law journal 106 (1997): 2313­74, at 2361-62.

[18] Fajer, “Can Two Real Men Eat Quiche Together?” 607-29.

[19] Balkin, “Nested Oppositions.”

[20] Ibid., 1676.

[21] On the Western tradition, see G. E. R. Lloyd, Polarity and Analogy: Two Types ofArgumentation in Early Greek Thought (Cambridge: Cambridge University Press, 1966); on the Eastern tradition, see Lao-Tzu, Tao Te Ching, Thomas Cleary, trans. (London: Shambala, 1983).

[22] See Jacques Derrida, “Differance,” in Margins of Philosophy (Chicago: University of Chicago Press, 1987), 1-27. Derrida describes his argument in terms of presence and absence of concepts. He argues that “the signified concept is never present in and of itself, in a sufficient presence that would refer only to itself” (11). A concept’s lack of full self-presence is its conceptual dependence upon other concepts. Derrida’s “differ­ance” also adds a temporal dimension to the notion of conceptual dependence, for he argues that concepts depend on their predecessors and their successors.

[23] These claims are defended at greater length in J. M. Balkin, “Transcendental

Deconstruction, Transcendent Justice,” Michigan Law Review 92 (1994): 1131-86.

[24] Akhil Amar, “The Bill of Rights as a Constitution,” Yale Law journal 100 (1991):

1131-1210, at 1195.

[25] Amar points out that the constitutional role of juries as populist protectors of human rights was not simply to nullify in sympathetic factual cases, but to apply the higher law of the constitution to control the ordinary positive law. Ibid., 1191. This suggests the homology reason : passion :: jury : judge :: higher law : positive law :: con­trolling :needing to be controlled. Indeed, Amar notes, “If we seek a paradigmatic image underlying the Bill of Rights, we cannot go far wrong in picking the jury” (1190). This role, he argues, has been suppressed over time, as judges have assumed for themselves the mantle of protectors of constitutional rights, regarding the jury as unlearned and thus as a threat to constitutional values. See 1192-95.

[26] As argued by former Supreme Court Justice William Brennan in William Bren­nan, “Reason, Passion, and `The Progress of the Law,’ “ Cardozo Law Review 10 (1988): 3-24.

[27] Oliver Wendell Holmes, Jr., “The Path of the Law,” Harvard Law Review 10 (1897): 457-78.

[28] For two quite different versions of this point see Andrew Ortony, Gerald L. Clare, and Allan Collins, The Cognitive Structure of Emotions (Cambridge: Cambridge University Press, 1988); and Robert H. Frank, Passions Within Reason: The Strategic Role of the Emotions (New York: Norton, 1988).

[29] An example would be the medical profession’s opinions on proper medical procedures for women.