Covering

Kenji Yoshino

Contents

I.    Introduction....................................................................... 771

 

II.   Concept

A.  Gay Conversion

1.   Cultural Contexts

a.   The Freudian Period (1870-1938)

b.   The Gilded Age of Conversion Therapy (1938-1969)

c.   The Post-Stonewall Period (1969 to the Present)

2.   Legal Contexts

B.   Gay Passing

1.   Cultural Contexts

2.   Legal Contexts

C.   Gay Covering

1.   Cultural Contexts

2.   Legal Contexts

a.   Civil Service

b.   Custody and Visitation

c.   Precocity of These Contexts

D.  The Performative Turn

 

III. Convergence

A.  The Antidiscrimination Schism............................................... 784

B.   Race-Based Covering............................................................ 787

1.   Cultural Contexts............................................................ 788

2.   Legal Contexts................................................................ 790

a.   Grooming.................................................................. 790

b.   Language.................................................................. 790

3.   The Performative Turn in Race....................................... 790

C.   Sex-Based Covering.............................................................. 790

1.   Cultural Contexts............................................................ 790

2.   Legal Contexts................................................................ 790

a.   Pregnancy................................................................. 790

b.   Demeanor and Grooming—The Double Bind

      Revisited.................................................................... 790

3.   The Performative Turn in Sex.......................................... 790

D.  Synthesis............................................................................... 790

 

IV. Critiques

A.  The Questionable Primacy of Orientation

1.   Passing and Conversion in the Contexts of Race and Sex

2.   The Case of Religion

B.   The Unarticulated Benefits of Assimilation

C.   The Problem of Essentialization

 

V.  Conclusion........................................................................... 790

 

I. Introduction

Assimilation is the magic in the American Dream. Just as in our actual dreams, magic permits us to transform into better, more beautiful creatures, so too in the American Dream, assimilation permits us to become not only Americans, but the kind of Americans we seek to be. Justice Scalia recently expressed this pro-assimilation sentiment when he joined a Supreme Court majority to strike down an affirmative action program. Calling for the end of race-consciousness by public actors, Scalia said: “In the eyes of government, we are just one race here. It is American.”_[1] Packed into this statement is the idea that we should set aside the racial identifications that divide us—black, white, Asian, Latino—and embrace the Americanness that unites us all.

This vision of assimilation is profoundly seductive and is, at some level, not just American but human. Surrendering our individuality is what permits us to enter communities larger than the narrow stations of our individual lives. Especially when the traits that divide us are, like race, morally arbitrary, this surrender seems like something to be prized. Indeed, assimilation is not only often beneficial, but sometimes necessary. To speak a language, to wear clothes, to have manners—all are acts of assimilation.

This assimilationist dream has its grip on the law. The American legal antidiscrimination paradigm has been dominated by the cases of race, and, to a lesser extent, sex. The solicitude directed toward racial minorities and women has been justified in part by the fact that they are marked by “immutable” and “visible” characteristics—that is, that such groups cannot assimilate into mainstream society because they are marked as different. The law must step in because these groups are physiologically incapable of blending into the mainstream. In contrast, major strands of American antidiscrimination law direct much less concern toward groups that can assimilate. Such groups, after all, can engage in self-help by assimilating into mainstream society. In law, as in broader culture, assimilation is celebrated as the cure to many social ills. One would have to be antisocial to argue against it.

So it is with great trepidation but greater conviction that I come to do so. For the past few years, I have been working on issues relating to sexual minorities._[2] That work has persuaded me that gays (by which I mean both lesbians and gay men) can proffer a new perspective on the relationship between assimilation and discrimination. I believe that the gay context demonstrates in a particularly trenchant manner that assimilation can be an effect of discrimination as well as an evasion of it. My goal here is to develop this idea in the context of orientation, and then to demonstrate the applicability of this insight to the race- and sex-based contexts.

I believe gays may have theorized some dimensions of the relationship between assimilation and discrimination differently from either racial minorities or women. This is because gays are generally able to assimilate in more ways than either racial minorities or women. In fact or in the imagination of others, gays can assimilate in three ways: conversion, passing, and covering. Conversion means the underlying identity is altered. Conversion occurs when a lesbian changes her orientation to become straight. Passing means the underlying identity is not altered, but hidden. Passing occurs when a lesbian presents herself to the world as straight. Covering means the underlying identity is neither altered nor hidden, but is downplayed. Covering occurs when a lesbian both is, and says she is, a lesbian, but otherwise makes it easy for others to disattend her orientation.

Of these three forms of assimilation, covering will probably be least familiar. The term and concept come from sociologist Erving Goffman’s groundbreaking work on stigma._[3] Goffman observed that even “persons who are ready to admit possession of a stigma . . . may nonetheless make a great effort to keep the stigma from looming large.”_[4] Thus a lesbian might be comfortable being gay and saying she is gay, but might nonetheless modulate her identity to permit others to ignore her orientation. She might, for example, (1) not engage in public displays of same-sex affection; (2) not engage in gender-atypical activity that could code as gay; or (3) not engage in gay activism.

As Goffman realized, these modes of assimilation are not always easily distinguishable from one another. For example, Goffman recognized that the same action could be either passing or covering depending on the knowledge of the audience before whom it was performed._[5] A woman who refrains from holding hands with her same-sex partner may thus pass with respect to those who do not know her orientation but cover with respect to those who do. This does not mean that the modalities of assimilation are indistinguishable. Rather, it means that one must know not only the performance of the actor, but also the literacy of the audience, to make that distinction. The relational aspect of presentations of the self is a preoccupation of this Article, as it was of Goffman’s work._[6] I recur repeatedly to the concept that assimilation is not a simple performance on the part of an agent, but rather a dialectic between an agent and her audiences.

While I have not seen it explicitly theorized, I believe that much of contemporary antidiscrimination discourse operates on a model—which I call the classical model—that incorporates the three assimilationist demands of conversion, passing, and covering. The classical model can be distinguished from others through two assumptions. First, the classical model assumes that the demands operate independently of each other—that is, that one can cover without passing and that one can pass without converting. Second, the classical model assumes that the demands are rigidly ordered in terms of their severity, with conversion always being a more burdensome demand than passing, and passing always being a more burdensome demand than covering. This model of identity can be conceived as a set of concentric circles rippling outward from a core, with having a certain status (failure to convert) at the center, disclosing that status (failure to pass) in the first circle around the core, and signaling that status (failure to cover) in the second circle. The model can be represented as follows:

Figure 1.

 

I later revise this model by noting that some activities denominated as covering are often deeply constitutive of identity. Yet it is heuristically useful to develop the classical model before challenging it in this way.

The classical model of identity is also a model of discrimination. If individuals have multiple ways of modulating their identities, discrimination against them will take multiple forms, including the demands to convert, to pass, and to cover. The form of assimilation required of an identity will often be correlated to the strength of the animus against it. When discriminatory animus against an identity is particularly strong, it may require conversion. When that animus is weaker, it may permit individuals to retain the targeted trait, but require them to pass._[7] When the animus is weaker still, it may permit individuals to retain and disclose their trait, but require them to cover it.

In Part II, I develop the classical model of discrimination in the context of sexual orientation. I retell the history of the gay rights movement as a history of the increasingly attenuated assimilationist demands placed on gays by mainstream society, in both nonlegal and legal contexts. I show that as the gay rights movement has become stronger, the assimilationist demands made on gays have become weaker, shifting in emphasis from conversion, to passing, to covering.

A quick way of demonstrating that shift is to consider the gay-related issues that have figured in the mainstream press over the last decades. In the early 1970s, the press widely discussed the American Psychiatric Association’s (APA’s) deletion of homosexuality from its taxonomy of mental disorders._[8] The controversy over this deletion was a debate about conversion, that is, about whether gays were mentally diseased individuals who needed to change their orientations. In the early 1990s, the press debated the practice of outing_[9]—the revelation of an individual’s homosexuality against her will—and the military’s “don’t ask, don’t tell” policy._[10] These topics pertained not to conversion, but to passing, that is, to whether a gay individual could or should self-identify as straight. Finally, at the turn of the millennium, the press has been devoting much of its gay-related coverage to same-sex marriage._[11] The right of gays to marry is a question of covering, as it pertains not to the ability of gays to be gay or to self-identify as gay, but to their ability to signal that identity beyond the simple act of self-identification. (As I demonstrate below, marriage can also paradoxically be seen as an act of covering, by those who take it to be a form of domesticating gays into straight norms.)_[12]

Again, the demand to cover may be the least intuitive. A recent example may clarify how gays are increasingly encountering covering demands. In 1990, a lesbian lawyer named Robin Shahar was fired from her job at the Georgia Attorney General’s Office._[13] Her employer emphasized that he had not fired Shahar for being a homosexual or for saying she was a homosexual, but for flaunting her homosexuality by engaging in a same-sex commitment ceremony._[14] Thus Shahar was terminated not for failing to convert or to pass, but for failing to cover. As time progresses, I posit that more and more discrimination against gays will take the form of covering demands, rather than taking the historical forms of categorical exclusion or “don’t ask, don’t tell.”

Part II describes this shift in much greater detail in the hope of accomplishing four goals. First, I seek to provide a new way of thinking about gay rights. To my knowledge, no account of gay history has described it as a series of increasingly attenuated demands for assimilation. By providing this account, I aspire to develop a new taxonomy of the forms anti-gay animus can take, not only historically, but in our contemporary moment. I believe that many individuals intuit a difference among current attempts at reparative therapy (the demand to convert), “don’t ask, don’t tell” (the demand to pass), and opposition to same-sex marriage (the demand to cover). I wish to provide some analytic distinctions to rationalize that intuition.

My second reason for spending some time describing this shift is to show its messiness and contingency. Synoptic accounts like the preceding vignettes from the media may give the impression that one assimilationist demand has neatly replaced another in an inexorable teleology in which gays are moving ever closer to full equality. I emphasize my disagreement with that view. No historical moment has existed in which one demand has categorically supplanted another, as suggested by the coexistence of all three demands today. I thus am not arguing that we have definitively moved into a covering phase of anti-gay discrimination in which conversion and passing are no longer at issue. To the contrary, I believe that one of the challenging aspects of being gay today lies in the very multiplicity of the assimilationist demands that gays encounter. Moreover, the attenuation of assimilationist demands made on gays in the past few decades does not mean this shift is inexorable. The ascendance of the demand to convert in medical circles in the years after Freud’s death is but one of many examples of how fragile progress in this area has been. The qualified progress narrative I tell here is not presented as an iron law of history.

My third goal in this Part is to describe some relationships between law and society with regard to conversion, passing, and covering. I have divided the discussion of each demand into its cultural and legal manifestations._[15] In doing so, I consider how the law both converges with and diverges from a broader culture of gay assimilation. In some crude sense, every cultural demand for gay conversion, passing, or covering finds voice in the law. Through translation into legal argot, however, the demands are also transformed. The ways culture gets refracted when articulated as law are not easy to describe systematically, and I have erred here on the side of being a witness rather than a theorist. Yet I pay particular attention to the dimension of time—to how legal discourse sometimes harks back to older cultural discourses and at other times reifies nascent ones.

My fourth and final goal is to use this history to enter an important qualification to the classical model of identity. The classical model assumes that the assimilationist demands (1) are always independent, such that one can cover without passing and pass without converting; and (2) are rigidly ordered in terms of severity, with conversion being the most severe, then passing, then covering. If we adopt this model, gay history can be told as an unqualified progress narrative, as the demands for assimilation have shifted in emphasis from conversion to passing to covering.

Any real engagement with gay history, however, shows that in some instances, the shift from conversion to passing or covering can be experienced by gays as no shift at all. One such shift is the military’s movement from its 1981 policy, which excluded gays on the basis of their homosexual status, to its 1993 “don’t ask, don’t tell” policy (still in effect), which excludes gays on the basis of homosexual self-identification_[16] or homosexual conduct. The 1981 policy was a conversion policy, as it required gays to convert to heterosexuality to serve. The “don’t ask, don’t tell” policy is popularly understood as a passing policy (as its moniker would suggest) and is defended by the military as a covering policy. This shift thus appears to represent progress for gays—no longer will they be excluded for their status, but only for their self-identification or conduct. Yet this shift has not improved the material or dignitary conditions of gays in the military, as homosexual self-identification and homosexual conduct are sufficiently central to gay identity that burdening such acts is tantamount to burdening gay status. Indeed, exclusions under the new policy have skyrocketed,_[17] suggesting that the shift is the reverse of a progress narrative for gays.

Qualifying the progress narrative should lead us to qualify the classical model that frames it. The instance of sodomy subverts both assumptions underlying the classical model. It shows that assimilationist demands are not always independent, in that a demand to cover can be tantamount to a demand to convert. Relatedly, it also demonstrates that we cannot assume that acts of covering are always less severe than acts of conversion. I am therefore led to propose a modification of the classical model that I call the weak performative model. The weak performative model, which draws on the work of Judith Butler, suggests that statuses can be partially constituted by acts. This model suggests that acts denominated as acts of covering might simultaneously be denominated as acts of conversion.

My hope is that Part II provides a free-standing contribution to an understanding of gay rights. As I seek to show in Part III, however, the lessons of this new assimilationist paradigm extend far beyond the context of orientation. Distinguishing among conversion, passing, and covering allows us to speak more precisely about how gays are like and unlike racial minorities and women._[18] In particular, the paradigm suggests that the claims of all three groups converge around covering.

I begin Part III by describing the distinctions that civil rights discourse often draws between gays on the one hand and racial minorities and women on the other. Two posited distinctions are that gays can convert and pass, while racial minorities and women cannot. To a significant extent, the antidiscrimination jurisprudence arising under the equal protection guarantees of the Federal Constitution and Title VII of the Civil Rights Act of 1964 has accepted these distinctions, maintaining that racial minorities and women are more deserving of legal protection in part because they cannot convert or pass. Such jurisprudence embodies an assimilationist bias. It maintains that groups that can assimilate are less worthy of protection than groups that cannot. It further suggests that the only acceptable defense to a demand for assimilation is the inability to accede to it. In doing so, the jurisprudence reflects and reinforces a schism between gays on the one hand and racial minorities and women on the other.

I seek to demonstrate that even if we accept these distinctions for the sake of argument, gays can still find common cause with racial minorities and women. Conversion and passing do not exhaust the forms of assimilation. There is also covering. And while racial minorities and women may be differently situated from gays along the axes of conversion and passing, all three groups are similarly situated along the axis of covering.

Like gays, racial minorities and women cover, and are asked to cover, all the time. The African-American woman who stops wearing cornrows to succeed at work may be covering._[19] The native Hawaiian broadcaster who mutes his accent to retain his broadcasting job may be covering._[20] The Latino venireperson who denies knowledge of Spanish to remain on a jury may be covering._[21] Women also cover. The woman who seeks to downplay her status as a mother or her pregnancy for fear of being penalized as an inauthentic worker may be covering. The female scholar who eschews feminist topics may be covering. The woman who strives to be as aggressive or tearless as the stereotypical man may be covering._[22] In all these instances, the individual is not attempting to change or hide her identity. Nonetheless, she is assimilating by making a disfavored trait easy for others to disattend.

Framing analogies among the covering strategies of these different groups merits some qualification. As an initial matter, these groups are obviously not distinct. When I state that women cover, I am focusing on how they cover as women—I do not foreclose the possibility that they will also cover along other dimensions. Understanding the intersectionality of identity is crucial to comprehending the difficulty of declaring that an individual is covering. For example, if a lesbian wears her hair long and down, is she covering her status as a gay person, refusing to cover as a woman, or exercising a grooming preference that has nothing to do with either axis of identity? If this question seems unanswerable, it is because—like most intersectional analysis—it honors the complexity of the underlying practice. In most of what follows, I disaggregate different axes of identity for the sake of introducing the concept of covering in an accessible manner. I wish clearly to acknowledge, however, that I regard this heuristic overschematization as a necessary evil. I hope further work will be more attentive to how assimilation occurs at the intersections of multiple identities.

Another important qualification, which I have been able more fully to capture in my analysis, is that covering does not manifest itself in identical ways across all contexts. I believe that women are differently situated from both gays and racial minorities in that they are asked more insistently by the dominant group (i.e. men) to “reverse cover” as well as to cover. By this, I mean that women are asked to emphasize their womanhood as well as to deemphasize it. In the landmark case of Price Waterhouse v. Hopkins,_[23] Ann Hopkins was asked to be assertive enough to make partner, but also asked to “walk more femininely, talk more femininely, dress more femininely, wear make-up, have her hair styled, and wear jewelry.”_[24] This was the “intolerable and impermissible catch 22”_[25] that drove the Supreme Court to hold in Hopkins’s favor. While gays and racial minorities certainly experience pressure to reverse cover, I believe they encounter this demand less systematically than women.

This difference, however, should not obscure how covering affects not only gays, but also racial minorities and women. Indeed, I argue that the contemporary forms of discrimination to which racial minorities and women are most vulnerable often take the guise of enforced covering. A member of a racial minority cannot be sanctioned for failing to convert or to pass without having a Title VII employment discrimination claim. But he can be sanctioned for failing to cover—for wearing cornrows,_[26] for lapsing into Spanish,_[27] or for speaking with an accent._[28] Similarly, a woman generally cannot be burdened for failing to convert or to pass. Yet it is still true that for constitutional purposes, state actors can burden pregnancy without triggering a sex discrimination analysis._[29]

This commonality suggests that racial minorities and women have much to gain from a theory of discrimination that focuses on the harms of coerced assimilation. Members of these groups are not as impervious to the assimilationist bias in the current antidiscrimination paradigm as their inability to convert or to pass might suggest. If the only defense against an assimilationist demand is that one cannot accede to it, racial minorities and women are left completely unprotected against covering demands, as anyone is assumed to be able to cover. My model thus shows a ground on which racial minorities, women, and gays can make common cause. That common ground will become more evident as anti-gay claims shift in emphasis away from conversion and passing and toward covering.

The gay context can also demonstrate the seriousness of the harm the covering demand inflicts. In that context, I demonstrate that under a performative conception of gay identity, certain acts denominated as covering, such as abstention from same-sex sodomy, might be constitutive of gay identity. I argue for this reason that one should not dismiss enforced covering as a trivial burden. That insight is applicable to racial minorities and women as well. I maintain that some forms of race-based covering (such as muting linguistic difference) or sex-based covering (such as muting a pregnancy) might also be constitutive of identity in this way.

In Part IV, I set forth three critiques of my model. The first critique suggests that gays do not deserve the primacy I have accorded them. One objection in this vein addresses my claim that gays have greater insight into some dimensions of the relationship between assimilation and discrimination because they can assimilate in more ways than racial minorities or women. The objection points out that one should not lightly assume that racial minorities and women are not capable of conversion or passing. It would be particularly ironic to claim that passing is a special attribute of gays, as passing was first identified as a phenomenon in the racial context. My response concedes all of this, but nonetheless emphasizes the greater relative frequency with which gays are asked to convert or to pass.

Of course, even if one acceded to this point, there would still be other candidates on which to found my model. Prime among these are religious minorities, who have routinely been asked to convert, to pass, and to cover. Moreover, religious minorities have secured constitutional and Title VII antidiscrimination protections despite the fact that they are capable of assimilating in these ways, making them a superficially more prepossessing site on which to build a model of assimilation as discrimination. I nonetheless argue that orientation may provide a better basis for developing this model because religion has been both marginalized and domesticated in the broader discourse of equality jurisprudence.

The second critique contends that my analysis ignores how assimilation can be both beneficial and necessary. My response—which I hope is implicit in much of my analysis—is that I am not arguing that all assimilation is per se bad. Rather, I am arguing against the countervailing assumption—powerful in both law and culture—that all assimilation is per se good. That response, of course, simply begs the question of how I seek to determine which forms of assimilation are malign. Elucidating my answer to that question permits me to articulate the very different standards I would apply in legal and nonlegal contexts. I am relatively conservative in the forms of coerced assimilation I believe should be remedied by the law. In contrast, I am much more willing to espouse a broader anti-assimilationist ethos in nonlegal contexts.

The final critique maintains that my model risks essentializing identities and engaging in the very stereotyping that the antidiscrimination paradigm is meant to retire. I find this critique the most formidable, as it is easy to envision a covering paradigm in which individuals would be accused of covering if they did not conform to stereotypes about their group. Thus a “masculine-identified” woman who experiences herself as “just being herself” might be told that she was not only covering but suffering from false consciousness. My model thus seems to encourage its adherents to think of groups according to the stereotypical behaviors in which they engage.

While this criticism is weighty, I argue that it can be overcome. I contend that it is possible to observe correlations between certain behaviors and certain identities—such as language and ethnic identity, pregnancy and being a woman, sodomy and being gay—without falling into stereotyping. I maintain that the erroneous conflation of such observations with stereotypes is a rhetorical strategy of those who would adhere to a formalistic regime of race-blindness, sex-blindness, and orientation-blindness. As seductive as those blindness regimes may appear to be, I contend that they actually have extremely pernicious implications for the equality claims of racial minorities, women, and sexual minorities. I therefore seek to articulate a regime in which antidiscrimination discourse would recognize correlations between behaviors and statuses without falling into either the essentialism of stereotyping or the formalism of blindness.

 

 

[Part II on Gay Conversion, Passing, and Covering omitted.]

 

III. Convergence

I hope the preceding discussion makes a free-standing contribution to an understanding of the history of the gay rights movement and of anti-gay discrimination. Yet I generated my model not only to reconceptualize anti-gay discrimination, but also to illuminate discrimination encountered by other groups. I therefore now apply my model to racial minorities and women.

I begin that application by describing the rift between gays on the one hand and racial minorities and women on the other. This rift, which I call the “antidiscrimination schism,” can be seen in both politics and law. The schism justifies itself in part on the ground that racial minorities and women cannot convert or pass, while gays can assimilate in both these ways. The schism, then, can easily be represented by my model, as it relates to the differential capacities of these groups to assimilate.

My model demonstrates that alongside this divergence is a critical and undertheorized convergence. According to my model, even if gays are differently situated from racial minorities and women with respect to conversion and passing, gays are similarly situated to these other groups with respect to covering. The recognition of covering as an assimilationist demand is crucial because covering, unlike conversion or passing, can be required of all three groups.

After positing this convergence around covering, I seek to demonstrate it in greater detail in the contexts of race and sex. In the race context, I note that many of the forms of discrimination from which racial minorities remain unprotected today take the form of enforced covering. In the sex context, I make a similar point. I further show, however, that the sex context differs from both the orientation and the race contexts in that women are more often required by the dominant group not only to cover but also to reverse-cover. Yet this takes nothing away from the fact that all three groups are similarly situated vis-à-vis covering demands. Covering thus provides a ground on which the three groups might make common cause.

A.  The Antidiscrimination Schism

In much of contemporary antidiscrimination discourse, one can see a prioritization of race discrimination over sex discrimination, and of sex discrimination over orientation discrimination. There is a gap between the perceived illegitimacy of race discrimination and that of sex discrimination. There is another gap between the perceived illegitimacy of race and sex discrimination on the one hand and that of orientation discrimination on the other. The second gap, however, is much wider than the first.

One can see the existence and the relative size of the gaps by looking at equal protection jurisprudence. That jurisprudence requires that courts give the strictest form of constitutional scrutiny to race-based classifications,_[30] an intermediate form of scrutiny to sex-based classifications,_[31] and the lowest form of scrutiny to orientation-based classifications._[32] The gap between strict scrutiny and intermediate scrutiny is very narrow, as both lead to the presumptive invalidation of legislation relying upon the classification._[33] In sharp contrast, the gap between the two forms of “heightened scrutiny” (strict scrutiny and intermediate scrutiny) on the one hand and the weaker form of scrutiny (rational basis review) on the other is enormous. If a statute is subjected to heightened scrutiny, it is almost invariably struck down; conversely, if a statute is not subjected to heightened scrutiny, it is almost invariably upheld._[34] The gap between heightened scrutiny and rational basis review is effectively the gap between protection and nonprotection. I call this gap the antidiscrimination schism.

What justifies the antidiscrimination schism? There are a number of answers to this question, but a significant one relates to assimilation. Two criteria the courts employ when determining whether a classification merits heightened scrutiny are the immutability and the visibility of the trait on which the classification is based._[35] Traits like race and sex that are viewed to be both immutable and visible are more likely to get judicial protection than traits like orientation that are viewed to be neither immutable nor visible. The rationale appears to be that groups that can assimilate can engage in self-help, and thus do not need to be judicially protected._[36] A group based on a mutable trait can simply convert when faced with discrimination; a group based on an invisible trait can simply pass.

As I have argued elsewhere, this manifests a strong assimilationist bias in equal protection._[37] If a group is not marked by an immutable or visible trait, it is less likely to receive heightened scrutiny. If a group does not receive heightened scrutiny, burdensome legislation against it is almost always upheld. And if burdensome legislation against a group that can convert or pass is upheld, it becomes more likely that members of the group will exercise those powers of conversion or passing to escape the legislation’s effects. Put differently, the descriptive claim that the group can assimilate because of the mutability or invisibility of its defining trait transmutes into the prescriptive claim that the group should assimilate with very little intervening investigation by a court. Because of this, the immutability factor in equal protection analysis effectively translates into a demand that mutable groups convert, and the visibility factor effectively translates into a demand that invisible groups pass.

The antidiscrimination schism and the assimilationist rationale that undergirds it are not a special creation of the judiciary, but can be seen in broader antidiscrimination discourse. The antidiscrimination schism has been particularly strongly articulated in political statements by racial minorities, some of whom feel that gays are making inappropriate analogies between racial civil rights and gay civil rights. In distinguishing between the military’s historical overt discrimination against African Americans and its current overt discrimination against gays, General Colin Powell has stated that “[s]kin color is a benign, nonbehavioral characteristic, while sexual orientation is perhaps the most profound of human behavioral characteristics.”_[38] Similarly, Alveda King, the niece of Martin Luther King, Jr., has argued that any link between black civil rights and gay civil rights is broken by the mutability of orientation._[39] And Reggie White, the Green Bay Packers’ defensive lineman turned popular preacher, has rejected any comparisons between the status of being black and the “sin” of homosexuality by noting that “[h]omosexuality is a decision[,] . . . not a race.”_[40]

Those calling to close the schism often make emotional appeals about the interconnected nature of discrimination._[41] Thus the late Thomas Stoddard opined that “[t]o a large degree we are bound together by our opponents. Those who hate blacks hate gay people, hate Jews and abuse women, and fighting on behalf of any of us will ultimately lead to the liberation of all.”_[42] Earl Ofari Hutchinson has observed that “Black people, and especially Black leaders, need to understand that when you scratch a homophobe, underneath you’ll invariably find someone who will deny you all your civil rights.”_[43] Mari Matsuda puts it most sparely: “[N]o person is free until the last and the least of us is free.”_[44]

All of these statements have a resonance beyond the reach of logic. It may be that such emotional appeals will ultimately be the most telling in closing the antidiscrimination schism. Yet such statements also fail to engage with the assimilationist rationale adduced to support the schism. Without seeking to undermine these statements, I seek to answer that rationale on its own terms.

One could challenge the schism analytically in a number of ways. I have elsewhere assailed it on the ground that the ability to convert or to pass does not redound to the political advantage of a group in any simple way._[45] A rising number of commentators have also begun to challenge the assumption that racial minorities and women cannot convert or pass._[46] In this discussion, however, I seek to make a claim that is different from, although consistent with, these prior ones. I maintain that, even assuming for the sake of argument that only gays can engage in self-help through conversion or passing, there are still grounds for convergence among all three groups. This is because conversion and passing are not the only ways in which groups can be forced to assimilate. Groups can also be forced to cover. And when we turn to covering, it becomes clear that all three groups—racial minorities, women, and gays—are similarly vulnerable. To substantiate that claim, I discuss race-based and sex-based covering in some detail.

B.   Race-Based Covering

In describing race-based covering, I again divide my discussion into cultural and legal contexts. To demonstrate covering in the cultural context, I focus on a nonfiction narrative that recounts how one individual—African-American lawyer Lawrence Mungin—systematically covered his race in pursuit of professional success. I turn to narrative because I believe that listing the axes along which racial minorities cover—many of which are identical to the axes enumerated in the discussion of gay covering—would not add as much to the analysis as a concrete account of the nature and costs of race-based covering. Through narrative, I seek to make the experience of covering more particular, vivid, and tangible.

In shifting to legal contexts, I revert to discussing cases in which individuals subjected to covering demands seek legal redress. I consider the examples of grooming in the Title VII context and language in the equal protection context. In both instances, I demonstrate that racial minorities find themselves largely unprotected from demands to cover. At the same time, I observe traces in these contexts that suggest some potential for an extension of antidiscrimination protections to covering claims. I argue for the exploitation of that potential.

1.   Cultural Contexts

Paul Barrett’s The Good Black_[47] tells the story of Lawrence Mungin, an African-American attorney who brought an unsuccessful race discrimination suit against his law-firm employer, Katten Muchin & Zavis. A graduate of Harvard College and Harvard Law School, Mungin spent all of his life until he decided to file his lawsuit attempting to be “one of the good blacks,” covering to assimilate as much as possible into the white mainstream._[48] As a lawyer at Katten Muchin, Mungin suffered a series of career setbacks that culminated in the firm’s refusal to consider him for a nomination to partnership._[49] Believing that these setbacks occurred because of his race, Mungin filed suit under Title VII of the Civil Rights Act of 1964._[50] A jury consisting of seven African Americans and one white awarded him a verdict of $2.5 million in compensatory and punitive damages._[51] On appeal, a panel of the D.C. Circuit reversed the jury’s verdict as unreasonable._[52] The sole dissenter, Chief Judge Harry Edwards, was also the only racial minority on the panel._[53]

In this discussion, I deploy Mungin’s story less to argue the merits of his lawsuit than to demonstrate the myriad of ways in which he sought to cover prior to filing it. As Barrett notes, Mungin’s decision to litigate was particularly poignant because it cast him in the racialized terms he had sought to eschew for much of his life._[54] Mungin’s story suggests that litigation and covering are two different strategies for addressing racism. In Shelby Steele’s terms, litigation is an instance of “challenging,” and covering an instance of “bargaining.”_[55]

I would therefore posit that a full-bodied account of covering is not possible if one looks only to primary legal materials. Of course, it is generally true that legal materials provide a reductive representation of broader social phenomena. Yet I wish to suggest that such materials may be particularly prone to ignore phenomena such as covering, which present themselves as alternatives to the strategies provided by law.

To say that law cannot adequately represent the harms of covering is not to say that only individual narratives can do so. As I have shown in the gay context, one can engage in what might be loosely described as a sociological account of covering._[56] I turn to narrative here to make up a lack that is present even in those accounts. This is the absence of a sense of how pervasively and deeply such covering demands affect the individuals on whom they are made. Narrative teaches lawyers to unlearn the distinction between social and legal harm that they have internalized as part of their acculturation into the profession. It reminds lawyers that the ability to critique the law correlates with the ability to describe thickly the social harms they seek to redress.

For Mungin, covering began at home. Mungin was raised in poverty in Brooklyn and Queens by his biracial mother, Helen Mungin._[57] Helen, “who considered herself to be black and had mostly black friends, wasn’t ashamed of her racial identity and didn’t cut her children off from theirs.”_[58] Yet while Helen did not seek to convert or to pass, she emphasized the importance of covering to all three of her children. She was fond of telling them: “You are a human being first, . . . an American second, a black third.”_[59] She punished her children when they spoke “street talk, rather than ‘proper’ English.”_[60] She stressed integrationist rather than activist politics, favoring Martin Luther King over Malcolm X, and advising her children that the existing system would treat them fairly if they played by its rules._[61] As Mungin’s sister observed, Helen Mungin “just didn’t make a big thing of race—it was there, but get past it.”_[62]

Mungin set out to do this. In junior high, he skipped lunch rather than be seen by his white peers in the mostly black “free-lunch” line._[63] In high school, he excelled in academics, debate, and student government, becoming the school’s first African-American senior class president._[64] These achievements led Mungin to be heavily recruited by a number of Ivy League colleges._[65] When Mungin visited Princeton, an African-American guide took him to an all-black dormitory and radio station._[66] When he visited Harvard, his tour was not racially oriented, and the alumnus-recruiter said: “Larry[,] . . . if you go to Harvard, you will never have to worry about money again for your whole life.”_[67] The child was the father of the man—Mungin went to Harvard._[68]

At Harvard College, Mungin continued to deemphasize his racial identity. He laughed along with others at racially laden comments._[69] He stopped cooking collard greens, which his mother used to prepare, when his roommates complained of the odor._[70] He avoided African-American campus groups, dormitories known to be dominated by blacks, and the “soul tables” in the dining hall._[71] The strategy continued to work. After taking a hiatus from college to train in the Navy,_[72] Mungin returned to Harvard and successfully applied to Harvard Law School._[73]

When Mungin arrived at Harvard Law School, members of the Black Law Students Association wanted to know why he had failed to join the organization and why he was rooming with a white student._[74] The white roommate was Paul Barrett, who would later become a legal affairs reporter for the Wall Street Journal and the author of The Good Black._[75] Mungin responded that he was attending law school to receive a credential, not to be an activist._[76] He criticized his black peers for “feeling sorry for themselves and haranguing the law school dean over issues he considered marginal, like how many minority professors had gotten tenure.”_[77] Barrett also recalls an incident in which Mungin castigated some boisterous black teenagers who he believed were “confirming every stereotype” about African Americans._[78]

Mungin also covered by keeping silent about his experiences with racism. During their law school years, Barrett heard nothing of the racism that Mungin had encountered in the past._[79] These racist incidents included being asked to leave the house of a white classmate’s parents on explicitly racial grounds,_[80] being called an “arrogant nigger” in the Navy,_[81] and reading the message “[W]hat are you doing with that nigger friend?” written in toothpaste on his white college roommate’s bathroom mirror._[82] To his credit, Barrett recognizes that this omission was intended to increase, and had the effect of increasing, the comfort level of the whites around Mungin. Barrett describes his relief at how, during law school, Mungin “never tried to make [him] feel guilty with talk of ‘systemic racism.’”_[83]

After graduating from Harvard Law School, Mungin worked at three law firms before landing at Katten Muchin & Zavis._[84] At Katten Muchin, Mungin continued to cover, looking and acting the part of the traditional corporate lawyer. Such acts of covering extended to his dress, his speech, his dissociation from other African Americans, and his silence in the face of perceived racial slights.

In Barrett’s chronicle, repeated reference is made to Mungin’s sartorial style—at one point a colleague praises him as the best-dressed person in Katten Muchin’s Washington office._[85] While this might make Mungin seem something of a fop, Barrett shows that the consequences of Mungin’s style were anything but superficial. Barrett describes how Mungin’s grooming practices directly affected perceptions of his race in the middle-class white circles in which he lived and worked. When wearing a suit, Mungin received friendly nods from his neighbors in the suburbs of Alexandria, Virginia._[86] When attired for the gym, he saw the same neighbors “visibly tense up.”_[87] After Mungin left Katten Muchin, he began to work temporary jobs for which a suit was inappropriate._[88] Mungin testified to the difference this made: “No more am I in Georgetown, dressed like a professional and at least getting some respect on the street. . . . I’m out in Chantilly, Virginia, or wherever, and the secretaries are afraid I’m going to attack them as they go to get in their cars.”_[89]

Mungin also covered by underscoring his educational credentials. Mungin emphasized his double-Harvard pedigree, “both because he was proud of it and because he knew it sent another reassuring signal to whites.”_[90] In Barrett’s words, Mungin spoke “with a precision that guaranteed his being described as ‘very articulate,’ a euphemism used by many whites to describe a black person who doesn’t use street vernacular.”_[91] As if he had taken to heart his mother’s injunction not to speak “street talk,” Mungin scrupulously avoided the use of profanity._[92] Indeed, Mungin’s formal personality concerned his African-American lawyers, who thought he identified himself more with his pedigree than with his race._[93]

Mungin further assimilated by dissociating himself from other African Americans, although here the vectors of his racial identification and disidentification were more complex. In a previous law firm, Mungin had taken a younger African-American associate under his wing._[94] The message he sent to that associate was the message that Mungin himself had received—“Don’t use race as an excuse; just get it right.”_[95] When he arrived at Katten Muchin, Mungin was concerned about being typecast as a mentor for other minorities._[96] It seemed that he had some cause for this concern—when the partners hired an African-American student intern, they asked Mungin to play a double role as her work-assignment coordinator and as her mentor._[97] Mungin refused, believing that such segregation would prevent both Mungin and the intern from integrating into the firm._[98] In a seemingly similar spirit, Mungin did not contact any other black lawyer at Katten Muchin until the eve of his lawsuit._[99] This may have paradoxically hurt his ability to negotiate race in the workplace, as it cut him off from a support network versed in that negotiation._[100]

Finally, Mungin covered by not responding to what he believed to be a racialized atmosphere. Just as he had ignored perceived racial slights in the past, so too did he ignore them at Katten Muchin. Mungin’s e-mails responding to a failure to get a raise and to what he experienced as a constructive discharge were exceedingly mild in tone._[101] Yet even these missives were deviations from a history of responding to perceived racism with nonperception or nonacknowledgement._[102]

In short, Mungin invested heavily in a “racial-comfort strategy”_[103] of covering. As Mungin stated his own credo: “I wanted to show that I was like white people: ‘Don’t be afraid. I’m one of the good blacks.’”_[104] In Barrett’s analysis, Mungin strove to join a select group of individuals of African-American ancestry—including Tiger Woods, Colin Powell, and Arthur Ashe—who are seen “‘not as unblack but as not merely, not primarily, black.’”_[105] In other words, Mungin sought to emulate those who had covered successfully.

Mungin’s relentless covering strategy, however, did not succeed. Isolated in the branch office in which he had chosen to work, Mungin gradually realized he had no chance of making partner._[106] Only at this point did Mungin question his strategy. Initially, that questioning led him to consider a career more related to his racial identity. Mungin spoke to Barrett about moving out of corporate litigation into civil rights work, describing his admiration for Thurgood Marshall, who had recently passed away._[107] Mungin then began to speak of suing Katten Muchin._[108] Barrett’s shock was fueled by his long-term vision of Mungin as someone who “wanted to belong to the system, . . . not challenge it.”_[109]

A cynic might read Mungin’s belated embrace of his racial identity as a strategic reaction to nonracialized career adversity. Yet Mungin’s account makes clear that race was always present in his life narrative. Even while he adopted the covering strategy, his systematic resistance to racial stereotyping defined him according to those stereotypes as surely as a photograph is defined by its negative. Mungin did not suddenly start to perform a racial identity when he stopped covering; rather, he started to perform that identity in a different way. Because his lifelong attempt to defuse racism through covering had failed, he shifted to litigation.

In that moment, Mungin was finally able to express what the covering strategy had cost him. He stated:

I was going to have to be more publicly honest about the lie that I was living. It wasn’t that I was around people who were open minded, who thought blacks are terrific. It’s that I was bending over backward all the time to avoid making white people uncomfortable. Like my neighbors [in Alexandria]: Now I’m just tired of making them feel comfortable, I don’t even talk to them. If they say hello, I’ll say hello, but I don’t even bother anymore making them feel comfortable late at night. It’s too much work._[110]

Mungin’s failure to acknowledge the costs of covering before completely abandoning the strategy suggests that part of covering entails repressing the work it requires. At times, covering may be as reflexive as the dilation or contraction of the pupil in response to changing light conditions. At other times, it may be experienced, as Mungin was ultimately to experience it, as an exhausting burden. In either case, however, covering is work.

One of my central claims is that the work of covering, unlike the work of conversion or passing, is imposed on all groups outside the mainstream. If this is true, one might expect to see similarities between the covering practices of racial minorities and those of gays. Mungin’s story vindicates this expectation, as many of the axes along which Mungin chose to cover are familiar from the previous discussion of gay covering. Mungin, for example, prioritized other identities over his minority identity by accepting his mother’s dictum that he was a “human being first, . . . an American second, a black third.”_[111] He accepted a nonactivist identity over an activist identity through his choice not to join other students who were protesting the paucity of minority faculty._[112] He privileged cultures that were stereotypically white over those that were stereotypically black in his dress and his demeanor._[113] He chose to associate with the mainstream and to dissociate himself from other members of his minority group._[114]

It is, of course, important not to extrapolate too quickly from Mungin’s individual narrative to a general claim about the nature of racial covering. The narrative turn in legal scholarship has been repeatedly criticized for its overly quick movement from a single compelling narrative to a normative legal position that impacts so many._[115] I concede that the burden is on the individual adducing such a racial narrative to ensure that it is roughly representative of the stories other racial minorities might tell. Of course, at some level, this burden is impossible for any one narrative to carry. No individual covering narrative can be representative in any strict sense—racial minorities of different backgrounds, professions, and races will cover in different ways.

Yet it bears emphasis that one can easily find examples in other sources of the four kinds of racial covering described above. Thus, in their article Working Identity, Devon Carbado and Mitu Gulati demonstrate how racial minorities drastically privilege their identities as workers over their racial identities._[116] Discussing how racial minorities are taught to eschew race-based activism, law professor Richard Delgado describes how he was advised not to write on “civil rights or other ‘ethnic’ subjects” prior to tenure._[117] Writing on how demeanor can affect perceptions of race, African American Brent Staples details his strategy of whistling Vivaldi while taking evening walks to counter negative visual data with positive aural data._[118] Meditating on racial self-hatred, Paulette Caldwell mourns how African-American children “reject association with black people and black culture in search of a keener nose or bluer eye.”_[119]

These homologies between race-based and orientation-based covering should lead us to inquire whether covering has a performative dimension in the race context as it does in the orientation context. As I show in the legal discussion, covering in the race context, as in the orientation context, is often described as tangential to identity._[120] While conversion and passing demands are deemed to be futile or unreasonable in the race context, covering demands are assumed to be eminently reasonable. In evaluating that assumption, we should ask whether race-based covering can be constitutive of racial identity.

In the orientation context, I showed that covering could be constitutive of identity by demonstrating how often covering rhetorically and conceptually converges with passing and conversion. In the race context, the rhetorical conflation is accomplished by Patricia Williams:

A man with whom I used to work once told me that I made too much of my race. “After all,” he said, “I don’t even think of you as black.” Yet sometime later, when another black woman became engaged in an ultimately unsuccessful tenure battle, he confided to me that he wished the school could find more blacks like me. I felt myself slip in and out of shadow, as I became nonblack for purposes of inclusion and black for purposes of exclusion; I felt the boundaries of my very body manipulated, casually inscribed by definitional demarcations that did not refer to me._[121]

The colleague’s demand is a covering demand—he is not asking Williams to convert to being white, or to pass as white. He is, rather, asking her to perform her racial identity in ways that make it easy for him to ignore her race. Williams is entitled to her race, but not to make “too much” of it—there is some excess of race that she can and should control. Yet Williams’s reaction to this demand suggests that racial covering may not be so distinct from racial passing or conversion. Williams describes her experience of the covering demand in both the rhetoric of passing (“I felt myself slip in and out of shadow”) and the rhetoric of conversion (“I became nonblack for purposes of inclusion and black for purposes of exclusion”). As in the orientation context, covering looks much less reasonable when it can be linked to conversion or to passing.

Social constructionist theories of race conceptually ground Williams’s evocative rhetoric. Neil Gotanda usefully catalogues four definitions of race: (1) status-race, (2) formal-race, (3) historical-race, and (4) culture-race._[122] Status-race defines race as a trait that carries intrinsic social status,_[123] as in a Jim Crow regime in which whites were viewed to be naturally superior to blacks. A much thinner conception of race, formal-race defines race solely in terms of immutable or visible characteristics such as skin color or ancestry._[124] Formal-race views black and white as “neutral, apolitical descriptions” that “are unconnected to social attributes such as culture, education, wealth, or language.”_[125] As such, formal-race is the only racial definition that severs race from social context, seeking to define race biologically rather than culturally._[126] Historical-race defines race as a historical construct. It distinguishes itself from both formal-race and status-race in denying the existence of any transhistorical racial essence, such as the essentialized hierarchy of status-race or the essentialized biology of formal-race. Rather, it seeks to locate the meaning of race in a history of subordination._[127] Finally, culture-race defines race as including “culture, community, and consciousness.”_[128] Culture-race holds that race is permeable to social discourse in a broader sense than simple historical subordination. It believes that race can encompass “broadly shared beliefs and social practices,” “physical and spiritual” community, and racial “traditions of self-awareness” as well as “action based on that self-awareness.”_[129]

For present purposes, the critical distinction is that between formal-race and culture-race. Under formal-race, covering will always be tangential to race. The formal-race concept deems race to be fixed at birth and impermeable to the behavior of its holder. Under such a formulation, behaviors may be correlated with race, but they will never constitute it. In this sense, the formal-race conception tracks a status conception of orientation under which one’s orientation is defined prior to and independently of one’s acts. Both reason from the classical model of identity.

In stark contrast, under culture-race, covering can constitute race. The culture-race conception deems race to be formed at least in part by the racial performances in which one engages. This conception tracks a conduct-based conception of orientation under which one’s orientation is in part defined by one’s acts. Put more broadly, the distinction between formal-race and culture-race is a distinction between a classical and a performative conception of identity.

Adopting a performative conception of identity has more radical consequences in the race context than in the orientation context. This is because there are much stronger norms against discrimination on the basis of status in the context of race discrimination. If covering is adjudged to interfere with that status, it will be much more likely to be legally prohibited.

In the legal realm, one can see this point by considering the federal bans on race discrimination. Unlike orientation discrimination, race discrimination is explicitly prohibited in both the equal protection jurisprudence and the statutory language of Title VII. How legislators and courts define “race” in those bans against “race” discrimination has a profound effect on the viability of covering claims. If race in these bans is defined as “formal-race,” then resistance to covering will be much less likely to fall within the ambit of protection. To the extent that it does, it will only be protected as behavior that is correlated to race (as in Title VII disparate impact analysis) rather than behavior that is constitutive of race._[130] If race in these bans, in contrast, is defined as “culture-race,” then covering demands will be much more likely to fall. The definition of race in the legal context will thus have immediate and immense consequence for racial covering. It is to that project of legal definition that I now turn.

2.   Legal Contexts

In this discussion, I take two case studies—one involving grooming discrimination and one involving language discrimination—to show how antidiscrimination law underprotects mutable race-related traits. I seek to demonstrate that this leaves racial minorities vulnerable to the demand that they assimilate through covering. As I did in the orientation context, I then question whether a more performative conception of race could be adopted by the law. In making the affirmative case, I demonstrate historical and contemporary traces of such a conception in the law.

a.   Grooming

Mungin covered through his grooming practices._[131] When he wore a suit, he evaded some of the stereotypes that attend African-American men._[132] When he failed to cover in this way, he was immediately assaulted with those stereotypes._[133] This suggests that grooming affects the extent to which racial minorities are perceived as such. In Mungin’s case, the refusal to cover never became a legal issue, as Mungin covered so assiduously. But what would have occurred if Mungin had not covered? Would Title VII have protected his race-based flaunting?

In this discussion, I take up that question by examining how racial grooming is analyzed under Title VII, which bars race discrimination in employment. The leading racial grooming case is Rogers v. American Airlines,_[134] decided in 1981. Renee Rogers was an African-American woman who worked for American Airlines as an airport operations agent._[135] This job fell under a grooming policy that prevented employees from wearing an all-braided hairstyle._[136] On its face, the policy was race-neutral and gender-neutral—whites as well as African Americans, and men as well as women, were prohibited from wearing all-braided hairstyles._[137] But the practice of wearing all-braided hair itself was (and is) neither race-neutral nor gender-neutral; this “cornrow” hairstyle is one strongly associated with African-American women._[138] Rogers, who wore cornrows, therefore challenged the policy under Title VII as a form of race and gender discrimination._[139] I focus here on the race discrimination claim.

In rejecting Rogers’s race discrimination claim, the court noted that the grooming policy on its face applied equally to members of all races._[140] While true, this fact in itself was not fatal to Rogers’s claim. Under the analysis of the landmark case of Griggs v. Duke Power Co.,_[141] a Title VII plaintiff can prevail against a facially neutral policy if she can show that it has a disparate impact on a protected group. Thus, Rogers could still have won her suit if she had demonstrated that the policy had a disparate impact on African Americans. Had she made that showing, the employer could only have defended the policy on the ground that it was a business necessity.

In rejecting the disparate impact claim, the court observed that Rogers had not maintained “that an all-braided hair style is worn exclusively or even predominantly by black people.”_[142] It further observed that the defendants had “alleged without contravention” that Rogers only adopted her all-braided hairstyle after it “had been popularized by a white actress [Bo Derek] in the film ‘10.’”_[143] In noting this contention, the court implied that the cornrow hairstyle was not especially associated with African-American culture. This, of course, strains credulity, as Derek’s cornrows were themselves an appropriation of an African-American style._[144]

Nonetheless, the court could simply have stopped at this point—if the grooming policy was not an instance of disparate treatment or disparate impact, it was permissible under Title VII. Yet the court continued its discussion, turning to other grounds on which it could dismiss Rogers’s claims. As if it recognized the strength of Rogers’s claim, the court bolstered its analysis with alternative rationales.

The main alternative rationale was that Rogers’s cornrows were unprotected because they were mutable. In developing this rationale, the court distinguished between the “Afro/Bush” style and cornrows. The court posited that the Afro/Bush style might be protected under Title VII “because banning a natural hairstyle would implicate the policies underlying the prohibition of discrimination on the basis of immutable characteristics.”_[145] The court then maintained that “an all-braided hairstyle is a different matter,” insofar as “[i]t is not the product of natural hair growth but of artifice.”_[146] The court observed that “[a]n all-braided hairstyle is an ‘easily changed characteristic,’ and, even if socioculturally associated with a particular race or nationality, is not an impermissible basis for distinctions in the application of employment practices by an employer.”_[147]

In an analysis echoing equal protection reasoning, the Rogers court thus made immutability a predicate for protection. Afros were protected only insofar as they were immutable. Mutable traits, no matter how race-related, were not protected. To clarify this point, the Rogers court quoted language from Garcia v. Gloor,_[148] a Fifth Circuit case holding that workplace English-only rules did not violate Title VII even if they had a disparate impact on Mexican Americans:

“[Title VII] is directed only at specific impermissible bases of discrimination—race, color, religion, sex, or national origin. National origin must not be confused with ethnic or sociocultural traits. . . . Save for religion, the discriminations on which the Act focuses its laser of prohibition are those that are either beyond the victim’s power to alter, or that impose a burden on an employee on one of the prohibited bases. . . . ‘[A] hiring policy that distinguishes on some other ground, such as grooming codes or length of hair, is related more closely to the employer’s choice of how to run his business than to equality of employment opportunity.’”_[149]

In this analysis, the Gloor court used immutability as a gatekeeping mechanism in two related senses. First, it observed that only classifications based on immutable attributes were protected under Title VII (with the exception of religion, of which more later_[150]). Thus race, color, sex, and national origin were protected because they were “beyond the victim’s power to alter.”_[151] Second, the court determined that protections for even these classifications extended only to their immutable aspects. Thus national origin—the status of being Mexican American—was not to be confused with “ethnic or sociocultural traits” such as the use of the Spanish language. These two points get conflated in the court’s analysis, which assumes that all that is meant by national origin is the immutable aspect of national origin.

The Rogers court appropriated this analysis to reason that (1) Rogers’s Title VII claim was one based on race; (2) the definition of race was limited to traits that one could not change, like skin color, bloodlines, or, perhaps, the texture of one’s hair; (3) cornrows did not fall within this definition because they were mutable; and (4) cornrows were therefore not protected by Title VII. The Rogers court thus defined race as formal-race rather than as culture-race. Once it adopted that definition, Rogers was not protected from covering demands.

Note that in this vulnerability to covering demands, racial minorities are no more protected by Title VII than sexual minorities. Rogers was protected from the requirements of conversion and of passing—she did not have to become a white man or appear to be a white man to retain her job. In this way, she was more protected than a homosexual, who could be fired for not converting or passing without Title VII redress._[152] But once the traits for which Rogers sought protection were characterized as mutable, no matter how race-salient, she ran into difficulties. She was not protected from the assimilationist demand to cover—to minimize the race-salient traits that made her different from others.

The problem with this Title VII analysis is that it scants the performative dimension of race. Rogers made precisely this argument to the court, contending that the cornrow style “‘has been and continues to be part of the cultural and historical essence of Black American women.’”_[153] In these words, Rogers urged the court to embrace culture-race, suggesting that grooming practices were partially constitutive of the “essence” of her racial identity.

Somewhat surprisingly, the court was not entirely deaf to these urgings. The court could have simply relied on Garcia for the proposition that mutable traits were not protected. Yet again, rather than relying on this ground alone, the court continued with a legally gratuitous, but perhaps sociologically impelled, analysis. In tacit response to Rogers’s claim that her grooming practice was essential to her racial identity, the court countered that the choice of a hairstyle was of “relatively low importance.”_[154] In other words, the court fleetingly entertained the concept that race could be partially defined by one’s acts, but stressed that the act of grooming was too trivial to count as such an act.

As seen in the orientation context, adopting a performative conception of status inevitably raises the question of which performances fundamentally constitute that status. Even after one adopts a culture-race analysis of Rogers, one must still ask whether covering demands pertaining to grooming are sufficiently constitutive of race to violate bans on race discrimination. For many, other race-related traits—such as language—will appear more deeply constitutive of racial identity. This judgment has much to do with the perceived triviality of appearance.

I could challenge that perception by observing that discrimination on the basis of skin color is itself a form of appearance discrimination. Less tendentiously, I could observe that even sartorial appearance has consequences that are anything but trivial. Depending on whether he was wearing a suit or not, Mungin was treated as a lawyer or a potential mugger._[155] One should therefore inquire whether Renee Rogers’s cornrows had similar effects.

In answering that question, I look again to the interiority of narrative. While Rogers did not supply an individual account of her cornrows, law professor Paulette Caldwell has supplied an eloquent one._[156] In discussing the Rogers case, Caldwell acknowledges the popular intuition that “hair is such a little thing.”_[157] Yet Caldwell’s purpose is to deploy her own experiences as an African-American woman to subvert the intuition that covering is trivial. She begins her essay by rooting her hair in a set of associations:

I want to know my hair again, to own it, to delight in it again, to recall my earliest mirrored reflection when there was no beginning and I first knew that the person who laughed at me and cried with me and stuck out her tongue at me was me. I want to know my hair again, the way I knew it before I knew that my hair is me, before I lost the right to me, before I knew that the burden of beauty—or lack of it—for an entire race of people could be tied up with my hair and me.

I want to know my hair again, the way I knew it before I knew Sambo and Dick, Buckwheat and Jane, Prissy and Miz Scarlett. Before I knew that my hair could be wrong—the wrong color, the wrong texture, the wrong amount of curl or straight. Before hot combs and thick grease and smelly-burning lye, all guaranteed to transform me, to silken the coarse, [resistant] wool that represents me. I want to know once more the time before I denatured, denuded, denigrated, and denied my hair and me, before I knew enough to worry about edges and kitchens and burrows and knots, when I was still a friend of water—the rain’s dancing drops of water, a swimming hole’s splashing water, a hot, muggy day’s misty invisible water, my own salty, sweaty, perspiring water.

When will I cherish my hair again, the way my grandmother cherished it, when fascinated by its beauty, with hands carrying centuries-old secrets of adornment and craftswomanship, she plaited it, twisted it, cornrowed it, finger-curled it, olive-oiled it, on the growing moon cut and shaped it, and wove it like fine strands of gold inlaid with semiprecious stones, coral and ivory, telling with my hair a lost-found story of the people she carried inside her?_[158]

Caldwell’s narrative recuperates the complexity of racial meanings lost in the Rogers case. In it, her hair grows into puzzles, becoming a metaphor for the unruliness of those meanings.

One way of making sense of that complexity is to see that all four of Gotanda’s racial categories are represented in this passage. In the first paragraph, Caldwell yearns to regain a view of her hair as a formal-race trait. In that vision, her hair is simply one trait among many that permits her to identify herself in the mirror. The described moment of self-identification is not ordinary, but foundational—it is the moment in which the infant first realizes that the image in the mirror is a representation of herself because the image’s laughter, tears, gestures, and hair correspond to the self’s._[159] In Lacanian theory, this is the moment that inaugurates self-consciousness, when the child is able to differentiate between self and nonself._[160] For Caldwell, hair in this desired moment is an individual trait like a laugh—it identifies her as a person, not as a member of a racial minority. In this important sense, then, Caldwell and the Rogers court agree: Both feel the allure of formal-race as a prelapsarian concept in which hair has no racial meanings. Where Caldwell diverges from the court is in casting this moment as irretrievable. To find the moment before her hair was freighted with racial signification, Caldwell must return to the very origins of self-consciousness—her “earliest mirrored reflection.”_[161]

When racial consciousness explicitly intrudes in this passage, it intrudes as status-race and historical-race. The images of status-race—“Sambo and Dick, Buckwheat and Jane, Prissy and Miz Scarlett”_[162]—are terms from a child’s lexicon, again suggesting the speed with which children are inducted into racial consciousness. Even if that status hierarchy is offered as a vestige of the past—as in Prissy and Miz Scarlett—memory carries it into the present as historical-race. And whether these hierarchies are experienced as contemporary realities or historical heritage, their tutelary effect is the same: Black children are taught that they are the wrong race. That lesson often attaches to their appearance, as evidenced in sources as disparate as the doll studies in Brown v. Board of Education_[163] and Toni Morrison’s The Bluest Eye._[164] In this passage, it attaches to Caldwell’s hair, which is “the wrong color, the wrong texture, the wrong amount of curl or straight.”_[165]

Against these definitions of race, Caldwell posits culture-race as a salvific alternative. Recall that Gotanda’s culture-race included “broadly shared beliefs and social practices,” “physical and spiritual” community, and racial “traditions of self-awareness.”_[166] In Caldwell’s account, cornrowing is cast as one such shared practice that occurs across generations of women in physical and spiritual community. The cornrows further express racial traditions of self-awareness, in that the grandmother’s braiding expresses “centuries-old secrets” and tells the “story of the people she carried inside her.”_[167] Moreover, unlike formal-race, culture-race seems to be retrievable. Culture-race does not require that one forget the narratives of status-race or historical-race, but rather that one remember the counter-narratives that could imbue the trait in question with different significations.

How trivial, then, is hair? One way to pursue that question is to ask why, if hair is such a trivial matter, American Airlines was so insistent on requiring Rogers to alter her hair, even going so far—with the court’s approval—to ask Rogers literally to cover her hair with a hairpiece._[168] In reading the Rogers case, one can hear American Airlines and the court asking Rogers: “Why is this so important to you?” To which it seems Rogers could fairly have responded: “Why is this so important to you?” The vehemence of American’s objection suggests that while hair might be trivial in some Lacanian moment of preconsciousness, it cannot be so in any moment human agents recognize as their own. In the dialogue between American and Rogers, cornrows become a symbol of resistance to assimilation, and therefore a symbol of insubordination. The individual wearing them is “seen as having the stereotypical characteristics commonly associated with black will and willpower—undisciplined, insubordinate, unwilling to melt.”_[169] Rogers’s hair must thus be understood not as a simple attribute but rather as a site of racial contest.

Indeed, one can imagine that an African-American woman might be indifferent to whether she wore cornrows or not until her hair became identified as such a site of contest. Consider the individual who picks up a button emblazoned with “Black Power” at a political rally who absently affixes it to her office bulletin board. Assume that at this moment the individual does not care whether the button is in her work environment or not. Now suppose that the individual’s supervisor tells her (to her surprise) that she must remove the button. Is it still not completely rational for that individual to insist on retaining the button not in spite of, but because of, the prohibition on it? At the point where the supervisor has insisted on the button’s removal, the button changes in social meaning. It becomes fraught with meanings it did not have before.

b.   Language

Mungin covered by eschewing black vernacular,_[170] suggesting a nexus between language and race. This nexus has been explicitly theorized in the debate surrounding Ebonics, where some commentators have argued that “Black English” is a constitutive element of race._[171] How much weight does the law give the view that language might partially constitute race? How much should it give?

The Supreme Court faced these questions in the 1991 case of Hernandez v. New York._[172] At issue in Hernandez was the prosecutor’s use of peremptory strikes to eliminate two Latino prospective jurors in a criminal case with a Latino defendant._[173] Since the Supreme Court’s decision in Batson v. Kentucky,_[174] striking a potential juror on the basis of her race is a violation of the equal protection guarantees of the Fifth and the Fourteenth Amendments. The prosecutor defended the strikes by contending that they were not based on race, but rather on the potential jurors’ facility in Spanish._[175] The Court had to decide whether such language-based discrimination was constitutionally impermissible race discrimination._[176]

Six Justices upheld the constitutionality of these language-based strikes by reasoning that they were not race-based strikes. There was, however, no majority opinion. It is instructive to compare the four-member plurality written by Justice Kennedy_[177] and the two-member concurrence written by Justice O’Connor,_[178] as the opinions agreed on almost everything except the definition of race. The two opinions agreed that the case turned on the plausibility of the race-neutral reason the prosecutor had adduced for striking the Latino jurors._[179] The opinions also agreed that the race-neutral reason was that the prosecutor questioned the ability of Spanish-speaking jurors to respect the official translation of the Spanish-language testimony anticipated in the case._[180] The opinions finally agreed that this reason was plausible, and that it required a decision in the state’s favor._[181] Yet in evaluating that reason, the two opinions revealed very different conceptions of race, conceptions that could have great significance for future cases.

Without categorically differentiating between language and race, Justice Kennedy’s plurality opinion concluded that the discrimination in this case did not constitute race discrimination. The plurality stressed that the prosecutor was distinguishing not between Spanish speakers and non-Spanish speakers, but rather between individuals who would defer to the court translator and individuals who would not._[182] In other words, the plurality framed this particular case as involving discrimination on the basis of deference to translation rather than discrimination on the basis of language. The plurality thereby cast the strikes at two removes from race discrimination, as lack of deference to a translator was related to facility in Spanish, which was in turn related to race. As the defendant correctly observed to the Court, this reasoning was tenuous, as lack of deference to a translator would always arise from facility in the non-English language that the translator was employed to interpret._[183] The plurality, however, rejected this claim on the ground that even a strong correlation between lack of deference and facility in Spanish did not collapse the distinction between the two for legal purposes._[184]

It appears possible that the plurality emphasized the broken link between language facility and lack of deference to give itself room to sustain a link between language facility and race without deciding the case in Hernandez’s favor. The plurality showed a surprising openness to the defendant’s claim that language and race were so intertwined that language discrimination was race discrimination. The plurality opined:

Just as shared language can serve to foster community, language differences can be a source of division. Language elicits a response from others, ranging from admiration and respect, to distance and alienation, to ridicule and scorn. Reactions of the latter type all too often result from or initiate racial hostility. In holding that a race-neutral reason for a peremptory challenge means a reason other than race, we do not resolve the more difficult question of the breadth with which the concept of race should be defined for equal protection purposes. We would face a quite different case if the prosecutor had justified his peremptory challenges with the explanation that he did not want Spanish-speaking jurors. It may well be, for certain ethnic groups and in some communities, that proficiency in a particular language, like skin color, should be treated as a surrogate for race under an equal protection analysis._[185]

Thus the opinion ruled against Hernandez without foreclosing the possibility that language discrimination might be race discrimination.

Justice Kennedy’s discussion of race was remarkable in entertaining the possibility of a juridical definition of race as culture-race. Rather than assuming that race was an obvious, biologically predetermined concept, the opinion expressed uncertainty about “the breadth with which the concept of race should be defined for equal protection purposes.”_[186] At a general level, this statement was subversive in its simple acknowledgement that the definition of race could differ according to the purpose to which it was put. More specifically, Justice Kennedy observed that for equal protection purposes, it could well be that a definition of race should reflect a sociological nexus between race and language. He articulated that nexus by observing that “ridicule and scorn” of a language “all too often result from or initiate racial hostility.”_[187] This intrication of language and race led him to speculate that “for certain ethnic groups and in some communities, . . . proficiency in a particular language, like skin color, should be treated as a surrogate for race under an equal protection analysis.”_[188] His analysis was thus open to the claim that language is no less (or more) tangential to racial identity than skin color. If that claim were true, language discrimination would be race discrimination. This is why Justice Kennedy posits that the Court “would face a quite different case if the prosecutor had justified his peremptory challenges with the explanation that he did not want Spanish-speaking jurors.”_[189]

Justice O’Connor’s concurrence took a sharply different view of race. The concurrence accepted that a nonracial justification might be a pretext for race discrimination._[190] But it stated that if the Court accepted the nonracial justification as true, it could not analyze the case as involving race discrimination. For Justice O’Connor, language in this case was such a nonracial, nonpretextual justification. Thus even if there were a 100% correlation between speaking Spanish and being of Hispanic descent, O’Connor would not analyze language discrimination as race discrimination: “No matter how closely tied or significantly correlated to race the explanation for a peremptory strike may be, the strike does not implicate the Equal Protection Clause unless it is based on race.”_[191]

Justice O’Connor’s and Justice Kennedy’s opinions did not differ in assuming that burdening a nonracial trait is permissible even if it has a disparate impact on racial minorities. That assumption is the legacy of Washington v. Davis,_[192] and all six Justices in the majority acceded to its legitimacy. The difference between the two opinions lay rather in whether they believed language to be a nonracial trait. While Justice Kennedy was open to the possibility that race might incorporate language, Justice O’Connor was not. Justice O’Connor never explicitly addressed the claim that race might include language, but rather assumed it. For her, the fact that Batson only prohibited jurors from being struck “‘solely on account of their race’”_[193] was dispositive of Hernandez’s claim. Justice Kennedy’s opinion and Justice O’Connor’s opinion demonstrate the difference between culture-race and formal-race. While these differing definitions did not lead the Justices to different results in this particular case, they could be immensely consequential in cases to come.

3.   The Performative Turn in Race

One of the notable aspects of Justice Kennedy’s opinion is that it is a legal opinion that could ultimately protect a covering claim under equal protection analysis. This is startling against the background knowledge that covering claims have not been protected under existing equal protection and Title VII regimes. The surprise occasioned by a court’s embrace of a performative definition of race may also be fueled, however, by the fact that the concept of status performativity is widely associated with postmodern theory. Such an association may lead to the perception that a performative theory of race is too radical or destabilizing to be accepted by a court.

If one accedes to this perception, one will approach the argument that courts should extend existing protections for race to individuals resisting covering demands as purely academic. Yet such an accession does an injustice not only to the covering analysis, but also to the courts. While courts may not frame their conceptualization of race in the postmodern argot of performativity, they are certainly able to conceive of race in a performative way. If one looks back in time, one can see historical instances in which courts not only entertained, but actually embraced, such performative conceptions of race.

I will do no more than touch on two historical instances in which courts adopted performative conceptions of race, relying on the incisive work of Ariela Gross_[194] and Ian Haney López._[195] Gross focuses on the performative aspects of race in nineteenth-century racial determination trials—that is, trials held to determine whether individuals were black or white for the purposes of discerning whether they were slaves._[196] According to Gross, legislatures sought to define race according to a binary system based on the formal-race criterion of ancestry._[197] Yet because that racial essence could not be easily detected, courts relied on external evidence to ascertain it. Specifically, courts often relied on racial performances to determine racial identity. Gross’s study encompasses sixty-eight trials of racial determination appealed to state courts in the South in the nineteenth century—a sample that represents all of the extant records of such cases that she could locate._[198] Gross’s examination of these cases leads her to conclude that “over the course of the antebellum period, law made the ‘performance’ of whiteness increasingly important to the determination of racial status.”_[199]

Thus, “[d]oing the things a white man or woman did became the law’s working definition of what it meant to be white.”_[200] Some of the conduct that the courts found salient has a chilling contemporary resonance. Individuals were adjudged white for their association with and acceptance by whites,_[201] for the gentility of their demeanor,_[202] and for the straightness of their hair._[203] For these individuals seeking to escape slavery, no less than for Mungin or Rogers, covering was rewarded. With this difference—in the race trials, covering was more openly acknowledged to be a form of conversion.

The statement that these performances converted race could be contested on the ground that these acts were believed to be evidence of a preexisting racial identity, the language through which blood would tell. Yet this objection simply restates the constative fallacy—the misperception that actions are describing an identity they are actually creating._[204] Because the nature of an individual’s blood was not evident to juries in these cases, the racial performances were only imaginatively describing that blood, and were actually creating a status that might have had no objective referent. Racial performances were not the evidence of race but rather its elements._[205] It may well have been that the investment in the constative fallacy was such that the acts were never explicitly acknowledged to be constitutive. But as a practical matter, they were constitutive in precisely this way—“To be white was to act white: to associate with whites, to dance gracefully, to vote.”_[206]

Formal-race was to receive an even more direct challenge in the early 1920s, in the racial prerequisite cases described in Haney López’s White by Law. The racial prerequisite cases were another set of racial determination trials that occurred to ascertain whether aliens could become citizens of the United States. From 1790 until 1870, only whites were permitted to naturalize._[207] For much of the period from 1870 to 1952 (the year when racial bars on naturalization were abolished), only whites and blacks were eligible for citizenship._[208]

Like the antebellum race trials, then, these cases had to determine the race of individuals for the purpose of allocating a crucial entitlement. In the racial prerequisite cases, however, the courts were forced to be much more explicit about their definitions of race because they could not delegate that definitional project to juries. When the issue came to the Supreme Court, the Court sought to define race as formal-race, first as a matter of skin color, then as a matter of blood. As Haney López describes, however, these attempts encountered serious obstacles in a pair of cases—United States v. Ozawa_[209] and United States v. Thind_[210]—that came to the Court in quick succession in 1922 and 1923._[211]

Ozawa concerned a Japanese immigrant who sought naturalization on the ground that he was white. Ozawa’s brief, which he wrote himself, argued in part on performative grounds._[212] Ozawa set forth the following facts:

“(1) I did not report my name, my marriage, or the names of my children to the Japanese Consulate in Honolulu; notwithstanding all Japanese subjects are requested to do so. These matters were reported to the American government. (2) I do not have any connection with any Japanese churches or schools, or any Japanese organizations here or elsewhere. (3) I am sending my children to an American church and American school in place of a Japanese one. (4) Most of the time I use the American (English) language at home, so that my children cannot speak the Japanese language. (5) I educated myself in American schools for nearly eleven years by supporting myself. (6) I have lived continuously within the United States for over twenty-eight years. (7) I chose as my wife one educated in American schools . . . instead of one educated in Japan.”_[213]

Ozawa’s application, however, was rejected by the District Attorney for the District of Hawaii on the ground that he was not a “white person.”_[214] Formal-race, in the form of skin color, trumped culture-race.

Yet Ozawa had an answer to this objection. As he was to articulate all the way to the United States Supreme Court, which he reached eight years later, Ozawa believed that his skin was, in literal terms, “white.” In support of his argument, he adduced anthropological testimony. This testimony stated, inter alia, that the typical Japanese “‘are whiter than the average Italian, Spaniard or Portuguese,’”_[215] and that “‘the Japanese are of lighter color than other Eastern Asiatics, not rarely showing the transparent pink tint which whites assume as their own privilege.’”_[216]

The Court, of course, rejected this argument, unanimously denying Ozawa’s application._[217] In doing so, however, the Court was pressured into disclaiming full reliance on the test of skin color. It observed that “to adopt the color test alone would result in a confused overlapping of races and a gradual merging of one into the other, without any practical line of separation.”_[218] This rejection of skin color as the ultimate arbiter of whiteness did not mean that the Court moved away from a formal-race conception. Instead, the Court shifted from one formal-race conception to another, moving from skin color to blood. The Court held that “the words ‘white person’ are synonymous with the words ‘a person of the Caucasian race,’” and that Ozawa was “clearly of a race which is not Caucasian.”_[219] Ancestry traceable to a particular ethnographic group, rather than skin color, became the court’s bulwark against the culture-race conception of race that Ozawa had forwarded.

Unfortunately for the Court, another plaintiff was waiting in the wings to challenge its newly minted definition of whites as individuals with Caucasian ancestry._[220] Only months after it handed down its decision in Ozawa, the Court heard oral arguments in Thind._[221] Bhagat Singh Thind was an immigrant from India._[222] As the Court conceded, Indians had been characterized by “certain scientific authorities” as Caucasians._[223] Thus, under Ozawa’s definition of white persons, Thind was white.

What, then, was the Court to do? Formal-race conceptions rely primarily on skin color or blood. Yet such literal definitions of formal-race would not exclude Ozawa or Thind, individuals who were in common knowledge not white. The Court resolved this dilemma by relying on that common knowledge. It stated that “[w]hat we now hold is that the words ‘free white persons’ are words of common speech, to be interpreted in accordance with the understanding of the common man, synonymous with the word ‘Caucasian’ only as that word is popularly understood.”_[224] This definition, it found, “does not include the body of people to whom the appellee belongs.”_[225] As Haney López observes, “the Supreme Court abandoned scientific explanations of race in favor of those rooted in common knowledge when science failed to reinforce popular beliefs about racial differences.”_[226]

In deferring to common knowledge, the Court both exposed and preserved the fiction of formal-race. It exposed that fiction in revealing the impossibility of articulating a definition of race as a particular skin color or ancestral group. Yet the Court simultaneously preserved that concept by placing it in the care of the community, in the same way that antebellum race trials placed racial definition in the hands of the jury. Thind was a stroke of ingenuity and disingenuity.

Reading the Supreme Court’s opinions in Ozawa and Thind places a certain pressure on contemporary judicial deployments of formal-race. Given that the Court has effectively admitted that it cannot coherently define race, how can Justice O’Connor (for example) be so confident that language is not race? Implicit in her confidence is a definition of race as skin color or ancestry that stands in tension with the lessons of the racial prerequisite cases, to say nothing of the antebellum racial determination cases.

More generally, these instances suggest that performative conceptions of race are not a recent postmodern phenomenon. We could ask why the concept of formal-race has had such a long life in antidiscrimination jurisprudence, given that its coherence has been so trenchantly called into question by these historical episodes.

There are doubtless many answers to this question, but I believe one significant one pertains to context. I distinguish here between what I call “formation” cases and what I call “treatment” cases. Formation cases are cases in which determining the racial identity of the party is the issue before the court, as in the antebellum trials or in the prerequisite cases. In these cases, the question before the court is “What race is this individual?” In contrast, treatment cases are cases in which the topic of dispute is how an individual has been treated on the basis of a race that is already known or stipulated, as in the generic equal protection or Title VII case.

My hypothesis is that courts are much more likely to adopt a performative conception of race (culture-race) in the formation context and a classical conception of race (formal-race) in the treatment context. This is because courts have more difficulty sustaining the illusion that race is a prediscursive concept when forced to confront the issue directly, as both the antebellum cases and the naturalization cases demonstrate. Yet because of the convenience—both moral and administrative—of formal-race, the courts fail to internalize the lessons about race they have learned in the formation context. When the courts turn to the treatment cases, in which the definition of race can be assumed without being articulated, the courts slip, either consciously or not, into a discourse of formal-race.

What should not remain occluded in this account is that treatment cases are, in their own way, also formation cases. If a court holds that (1) race is protected, but (2) language is not protected, then the only logical inference is that (3) language is not race. What is pernicious about these cases is that race is being defined sub silentio by statements about what race is not, without any obligation to define precisely what race is. This is disquieting because the Court has shown itself unable to satisfy that obligation—or, put differently, because the Court’s practice suggests that formal-race cannot be formally defined.

The dispiriting ability of courts to sustain such inconsistent definitions of race should not detract from the more hopeful insight that the courts are not endemically incapable of adopting culture-race. History teaches that culture-race is not a radical concept, at least in the formation context. It is possible that Justice Kennedy’s conception of culture-race may yet be embraced.

C.   Sex-Based Covering

I now turn to sex-based covering as my final case study. As in the race context, I begin with a nonfictional narrative. This narrative is an account written by Ann Hopkins_[227] about her experience as a plaintiff in the landmark sex-discrimination case of Price Waterhouse v. Hopkins._[228] Hopkins’s narrative demonstrates how sex-based covering both converges with and diverges from orientation-based and race-based covering. Like Singer and Mungin, Hopkins was required to cover by making her status as a member of a subordinated group easy to disattend along some axes. Unlike Singer and Mungin, however, Hopkins was also required to “reverse cover,” that is, to signal her outgroup status along other axes. Through this narrative, I demonstrate that women are differently situated from gays or racial minorities in that the dominant group routinely asks them both to cover and to reverse cover.

Turning to legal contexts, I demonstrate this convergence and divergence. I show how along some dimensions, such as pregnancy or motherhood, women are systematically asked to cover in ways that are deeply analogous to the ways in which gays or racial minorities are asked to mute their identities. I then show how along other dimensions, such as grooming or demeanor, women are sent much more conflicting messages—they are asked to occupy a middle ground that is neither too masculine nor too feminine.

1.   Cultural Contexts

Ann Hopkins’s So Ordered_[229] is an autobiographical account of her famously successful suit against Price Waterhouse, a major accounting firm._[230] In 1982, Hopkins was the sole woman among the eighty-eight nominees for partnership at the firm._[231] Although she possessed the best record among those nominees for generating new business, Hopkins was not elected._[232] She brought suit under Title VII, alleging that she had been discriminated against because of her sex. She prevailed in her lawsuit at every stage, including at the Supreme Court._[233]

Like Mungin, Hopkins engaged in much activity that could be characterized as covering. Many of the positive comments she received during her partnership review related to stereotypically male attributes—she was described as “strong,” “decisive[],” “independen[t],” and “demand[ing].”_[234] As Mary Anne Case points out in her incisive analysis of the Hopkins case, all of “these adjectives come straight out of the masculinity scale of one of the most influential psychological inventories in sex-role research.”_[235] One account of the injustice of Hopkins’s partnership denial, then, was that she bore all of the burdens of covering without reaping any of its benefits.

In Hopkins’s case, however, it is harder to ascertain whether Hopkins herself would have characterized her actions as covering. In Barrett’s narrative, there is a moment in which Mungin explicitly acknowledged how hard he was working to fit into the mainstream._[236] Hopkins makes no such admission. To the contrary, the impression that Hopkins conveys in the narrative is that she is, in stereotypical terms, male-identified. Describing her childhood, Hopkins speaks of herself as a tomboy. She recounts that she “had to take home economics” because “only the boys could take shop,” and that she brought home a “D” because she “ironed on the wrong side of the ironing board.”_[237] She describes herself as an outsider who “cried a lot over [her] difficulties coping with sororities and cheerleaders and snobbery and bias.”_[238] Speaking of her education in a single-sex college, Hopkins describes her “unstatistical” belief “that most interesting women attended women’s colleges.”_[239] She maintains that going to such a college had a significant impact on her development: “I learned to depend on myself and on the analytical integrity of an answer to a question or a solution to a problem before I was taught to depend on or defer to members of the opposite sex or their point of view.”_[240]

It is, of course, important not to read the book through the filter of the lore surrounding the case—that Hopkins was viewed to be male-identified. Such a filter might obscure many details in Hopkins’s account that demonstrate stereotypically feminine behavior. Thus Hopkins speaks of her youthful enjoyment of “Victorian formalities,” of how she “wore white gloves and hats when the occasion called for it, learned ballroom dancing, went to debutante balls.”_[241] Yet the dominant impression created by the narrative is that Hopkins does possess many stereotypically masculine traits. Thus Hopkins describes her rejection of her mother’s suggestions to study “literature, history, art, and philosophy,” in favor of an undergraduate major in mathematics._[242] After earning a master’s degree in that field, Hopkins returned to her alma mater to engage in the stereotypically feminine profession of teaching, but then left it a year later because it was not her calling._[243] She instead moved into the stereotypically masculine corporate world, working successively for IBM, the Computer Sciences Corporation, the accounting firm of Deloitte & Touche, and finally Price Waterhouse._[244] Hopkins’s account of her working years describes her tendency to engage in binary thinking,_[245] her inattention to context,_[246] her hatred of therapy or counseling,_[247] her preference for action over introspection,_[248] and her motorcycle, which she began to drive after she crashed her car on a camping trip._[249]

This is not to say that Hopkins was exempted from covering demands. One such covering demand was that she make her pregnancies easy to ignore. At Deloitte & Touche, Hopkins became pregnant for the first time._[250] She perceived the time she would need to give birth to her child as a medical procedure that would keep her “away from work for a couple of weeks.”_[251] Her supervisor apprehended the situation differently, reacting as if she “were planning to quit,” and elevating her “pregnancy to the level of a professional crisis.”_[252] Similarly, Price Waterhouse criticized Hopkins for bringing her children to work on certain occasions._[253]

Nonetheless, it was not Hopkins’s failure to cover that doomed her partnership chances at Price Waterhouse. The negative comments during her review centered not on perceptions of her excessive femininity, but on perceptions of her excessive masculinity. One partner advised her to “[w]alk more femininely, talk more femininely, dress more femininely, wear make-up and jewelry [and] have [her] hair styled.”_[254] Another suggested that Hopkins take “a course at charm school.”_[255] Others described Hopkins as “‘macho,’” or as “‘overcompensat[ing] for being a woman.’”_[256] Still others complained of her use of profanity, although one partner candidly admitted that the perception of her swearing was heightened “[j]ust because it’s a lady using foul language.”_[257] Even those who supported Hopkins often framed their support in gendered terms. One partner observed that Hopkins had “matured from a tough-talking somewhat masculine hard-nosed manager to an authoritative, formidable, but much more appealing lady partner candidate.”_[258] Previous female candidates for partner had been criticized for being too much “like one of the boys,” for being “women’s libber[s],” or for behaving like “Ma Barker.”_[259]

These demands strikingly contrast with the demands described in the orientation and race contexts. The demands that employers made of Singer or Mungin were covering demands—demands that the individual make his difference from the dominant group easy to ignore. Hopkins, in contrast, was not asked in these comments to act more masculine but rather to act more feminine. She was not asked to cover, but rather to reverse cover.

This is not to say that reverse covering demands are not made of gays or racial minorities. Perhaps most evidently, members of the groups themselves make such demands, as when “queers” ask gays to be “more gay”_[260] or when racial minorities exhort fellow members not to be “oreos”_[261] or “bananas.”_[262] Less obviously, the dominant groups within these classifications also make reverse covering demands. Straights can ask gays to perform according to stereotype in certain realms, although the contours of those realms may be strictly delimited. Thus Butler describes the irony of how drag (which often codes as gay) is a form of “high het entertainment” when performed on the stage but not when performed by the individual seated next to one._[263] Subtler reverse covering demands are also more pervasive—as when a gay man is asked for advice about fashion or design._[264] Similarly, ethnic and racial minorities can also be subjected to demands for minstrelization. This can take the form of required deference or submissiveness, as when an African American who behaves in what would be an ordinary manner for a white is perceived to be “uppity.”_[265] It can also take the form of a demand that such minorities be more “ethnic,” as when individuals are asked—as Mungin was_[266]—to be available to mentor a coworker of the same background._[267]

Nonetheless, I believe it is no accident that Hopkins’s account features the reverse covering demand more prominently than either Singer’s or Mungin’s. It is my sense—admittedly impressionistic—that the dominant group more routinely requires reverse covering in the sex/gender context than in the orientation or race contexts. This may be because stereotypically feminine traits are more likely to be valued as appropriate to at least some spheres of life. The stereotypically feminine attributes of nurture, empathy, intuition, and so forth, were and are valued in the domestic sphere. In contrast, there are fewer spheres in which traits stereotypically associated with homosexuals or racial minorities are valued. I take the import of Bowers v. Hardwick,_[268] for example, to be that there is no place where consensual homosexual sodomy could be valued, not even in the private space of one’s own home.

Hopkins thus found herself subjected simultaneously to both the demand to cover and the demand to reverse cover. She was asked to strike an Archimedian mean between the poles of being too masculine and being too feminine. The nature of that mean is perhaps best reflected in the partner’s comment that Hopkins had “matured from a tough-talking somewhat masculine and hard-nosed manager to an authoritative, formidable, but much more appealing lady partner candidate.”_[269] This comment reveals that while being too masculine is not valued, being too feminine is not valued either. To succeed as a woman, one must have the correctly titrated balance of masculine and feminine traits. One must be “authoritative” and “formidable,” but remain an “appealing lady.” The necessity of striking such a balance is also evident in another partner’s comment that Hopkins had “overcompensated for being a woman.”_[270] Again, this language suggests that some compensation is appropriate, but that “overcompensation” is not. If a woman covers too much, then the reverse covering demand will be made to bring her back into the zone of appropriate behavior.

The opposition between the two demands should not obscure what they have in common—both relate to performative and mutable aspects of Hopkins’s identity. When Hopkins first perceived that her career at Price Waterhouse was endangered, she expressed the desperate hope that the problem was something “other than sex, that [she] could start to work on, especially given that changing sex was not an option.”_[271] This comment reflects her naiveté about how sex can be “worked on” without having a sex change. In my terms, Hopkins associated assimilation with conversion, failing to discern that Price Waterhouse was asking her to cover and reverse cover. For while Price Waterhouse did not wish Hopkins to change her sex, it certainly wished her to work on her sex-based performances. Only one partner at Price Waterhouse explicitly articulated a categorical animus toward women, and the other partners assiduously disavowed any reliance on this position._[272] Yet many partners unwilling to articulate categorical animus toward women were quite comfortable voicing objections to a certain kind of woman. This was the woman who did not perform her gender in the middle band between hypermasculinity and hyperfemininity. Thus when Hopkins’s friend and colleague Sandy Kinsey was asked in court whether Price Waterhouse treated women fairly, she perceptively responded that it had treated her, as an individual woman, fairly._[273] Because she did not fall out of the median band of acceptable gender performance, Kinsey had not felt the policing effects of either the covering or reverse covering demands.

Hopkins’s “sex change” comment is not an outlier. Hopkins’s narrative gives the impression that she was generally oblivious to gender dynamics. Hopkins did not think of her gender as a potential ground for discrimination until after Price Waterhouse failed to promote her. When asked by a friend whether she believed gender had been a factor, her initial reaction was disbelief. Both she and the friends present at that conversation “found unimaginable any notion that gender influenced business decisions.”_[274] Similarly, when asked by her attorney in an early consultation whether she had encountered sexist remarks at work, she responded that there were “none that [she] could recall,” but that she “infrequently recognized sexist comments.”_[275] This is a remarkable statement in light of the sheer volume of sexist remarks that had been made directly to Hopkins by that point.

A comparison with Mungin’s narrative should lead us not to treat this obliviousness as simple naiveté. Mungin remained in denial about the effects of racism in his life until that denial became intolerable. It may be that part of the process of fitting in at Price Waterhouse for Hopkins entailed repressing her experiences with sexism. Perhaps Hopkins “infrequently recognized sexist comments” because that was a strategy of corporate survival.

In any event, the litigation proved to be such an education in gender consciousness that Hopkins was soon unable to see the case in nongendered terms. One of the fascinating aspects of Hopkins’s account is how much the litigation ultimately framed her final understanding of what had occurred at Price Waterhouse. During litigation, one of her attorneys sent her a poem consisting of five limericks, the first of which read: “There once was a woman named Ann / who was told to act less like a man / told to be very sweet / and to dress oh so neat / and to walk with a shake of her can.”_[276] This is how Hopkins ultimately understood her own case, and how tens of thousands of law students have been taught it. Yet this characterization is a far cry from her initial belief that no sexist comments had been made to her at Price Waterhouse.

Hopkins’s narrative suggests that the gender consciousness the trial instilled in her led her to negotiate gender more carefully. Hopkins soon realized that the press coverage of her litigation was itself a trial of whether she could perform her gender in such a way as to be a sympathetic plaintiff. In the face of this judgment, she became much more receptive to the reverse covering demand. Hopkins describes her own sensitivity—described by her friends as hypersensitivity—to having her “brown Coach bag” described as a “‘beat up brown’ briefcase”_[277] or to having reporters judge her house or her attire when they came to interview her._[278] Even more tellingly, Hopkins tells the following anecdote about preparing for a television appearance with her daughter Tela:

On one occasion, when a producer asked me to step down the hall for makeup, a startled Tela asked, “You’re not going to do that are you, Mom?” She seemed even more startled at my reply: “Tela Margaret, that woman is in charge and I’m going to do what she says.”_[279]

When asked to wear more makeup by someone in charge, Hopkins decided to do so this time._[280] Her accession to such reverse covering demands can also be seen in the photograph on the jacket of her book, where she is pictured with makeup, jewelry, and styled hair.

Hopkins’s narrative demonstrates how she was pervasively subjected to both covering and reverse covering demands throughout her life. The fact that the latter were more strongly evident than the former suggests simply that she was generally more on the masculine side of the spectrum of gender performance. It also suggests that all professional women may be pushed toward the middle of the spectrum through the policing effects of the covering and reverse covering demands.

2.   Legal Contexts

I will return to Ann Hopkins’s Supreme Court victory for the purposes of describing how courts have treated the distinctive double bind in which women find themselves. Before doing so, however, I wish to consider the ways in which courts have treated a simple covering demand, taking up the instance of pregnancy.

a.   Pregnancy

One of the ways in which Hopkins was pressured to cover was around issues of pregnancy and motherhood. When she announced her pregnancy at her first accounting firm, it was taken as an announcement of her resignation._[281] Similarly, Hopkins was criticized by Price Waterhouse for bringing her children to the workplace._[282] To be recognized as authentic workers, then, women must deemphasize their roles as potential or actual mothers._[283] To what extent is discrimination against those who refuse to cover in this way discrimination on the basis of sex?

The Supreme Court delivered a startling answer to this question in 1974 in the case of Geduldig v. Aiello,_[284] where it held that pregnancy discrimination was not sex discrimination for the purposes of the equal protection guarantee. At issue in Geduldig was California’s disability insurance program, which did not cover work loss that resulted from pregnancy._[285] A group of female California employees brought a class action suit claiming that this program violated their rights under the Equal Protection Clause. They contended that such pregnancy discrimination was sex discrimination that triggered intermediate scrutiny._[286] In rejecting this claim in a footnote, the Court observed:

[T]his case is thus a far cry from cases like Reed v. Reed and Frontiero v. Richardson, involving discrimination based upon gender as such. The California insurance program does not exclude anyone from benefit eligibility because of gender but merely removes one physical condition—pregnancy—from the list of compensable disabilities. While it is true that only women can become pregnant, it does not follow that every legislative classification concerning pregnancy is a sex-based classification like those considered in Reed and Frontiero. Normal pregnancy is an objectively identifiable physical condition with unique characteristics. Absent a showing that distinctions involving pregnancy are mere pretexts designed to effect an invidious discrimination against the members of one sex or the other, lawmakers are constitutionally free to include or exclude pregnancy from the coverage of legislation such as this on any reasonable basis, just as with respect to any other physical condition.

The lack of identity between the excluded disability and gender as such under this insurance program becomes clear upon the most cursory analysis. The program divides potential recipients into two groups—pregnant women and nonpregnant persons. While the first group is exclusively female, the second includes members of both sexes. The fiscal and actuarial benefits of the program thus accrue to members of both sexes._[287]

Pregnancy is here dissociated from sex (“gender as such”) through the observation that while all pregnant persons are women, not all nonpregnant persons are men. Because pregnancy is not perfectly correlated to being a woman, pregnancy discrimination is not sex discrimination.

Knowing the Court’s antidiscrimination logic, one finds it hard not to look for the immutable/mutable distinction here. While the language of mutability is sub rosa in the Geduldig opinion, Dan Danielsen has hypothesized that what broke the link between pregnancy and sex for the Court was that pregnancy, unlike sex, can be chosen._[288] The volitional aspect of pregnancy may have been particularly visible to the Court as it had decided Roe v. Wade_[289] in the preceding Term._[290]

The simple statement that the Court does not recognize pregnancy discrimination as sex discrimination should be sufficient to demonstrate the troubling narrowness of the Court’s definition of sex for purposes of equal protection. Closer examination of the sex discrimination jurisprudence, however, reveals an even more disturbing tension in the Court’s analysis. The Court’s grant of intermediate, rather than strict, scrutiny to sex is sometimes justified on the ground that there are “real differences” between the sexes._[291] Thus, unlike any two racial groups, men and women are deemed to be biologically different in ways that could justify their differential treatment._[292] One of the fundamental differences that the Court has found between men and women is their differential ability to become pregnant._[293] Thus pregnancy is simultaneously not sex and the fundamental difference between the sexes._[294]

In another surprisingly performative turn, Congress rejected the Geduldig Court’s conception of sex. Two years after Geduldig, the Supreme Court extended its holding that pregnancy discrimination was not sex discrimination to the context of Title VII in General Electric Co. v. Gilbert._[295] In Gilbert, the Court rejected a Title VII challenge to a plan very similar to that in Geduldig._[296] The Court again observed that the nexus between pregnancy and sex was insufficient to warrant protection of pregnancy under Title VII’s bar on sex discrimination. This time, however, the Court’s decision was susceptible to legislative override. Congress availed itself of that opportunity by enacting the Pregnancy Discrimination Act of 1978._[297] The Act framed itself as a clarification of the meaning of the existing Title VII provision barring sex discrimination, explicitly defining sex discrimination to encompass discrimination “because of or on the basis of pregnancy, childbirth, or related medical conditions.”_[298] Rather than adding pregnancy as an additional ground to sex, the Act thus articulated a more encompassing definition of sex that included pregnancy. In this sense, the Pregnancy Discrimination Act took a performative view of sex, noting that an individual’s sex incorporated the acts in which she engaged.

In the equal protection domain, however, the Court has not retreated from its view that pregnancy discrimination is not sex discrimination. In 1993, the Court decided Bray v. Alexandria Women’s Health Clinic,_[299] a case turning on the question of whether antiabortion activism represented animus against women. In answering in the negative, an opinion for the Court by Justice Scalia cited Geduldig for the proposition that pregnancy discrimination was not sex discrimination. To demonstrate the “continuing vitality of Geduldig,”_[300] the Court observed that in a pair of earlier cases concerning funding for abortions, it had determined that “the constitutional test applicable to government abortion-funding restrictions is not the heightened-scrutiny standard that our cases demand for sex-based discrimination.”_[301] Far from limiting Geduldig’s reach, the Court has thus affirmed its premise in the abortion context. In doing so, it adheres to a deeply nonperformative conception of sex as constituting only biological genotype or phenotype.

b.   Demeanor and Grooming—The Double Bind Revisited

The judicial treatment of pregnancy is very similar to the judicial treatment of other covering activities such as homosexual sodomy or grooming. This demonstrates that women are in some ways similarly situated to gays or racial minorities. Yet as Hopkins’s narrative demonstrates, women are also differently situated from these other groups because they are caught in a particularly severe double bind. I now consider how the courts have dealt with this distinctive aspect of sex-based covering.

The evidence that Hopkins was explicitly subjected to both covering and reverse covering demands was important to her legal victory. An expert witness for Hopkins—psychologist Susan Fiske—offered a crucial characterization of Hopkins’s predicament as a “double bind.”_[302] Fiske first explained stereotyping, testifying that “[t]he overall stereotype for feminine behavior is to be socially concerned and understanding, soft and tender, and the overall stereotype for a man, all other things being equal, is that [he] will be competitive, ambitious, aggressive, independent, and active.”_[303] Fiske maintained that because stereotypically male traits are valued in many work environments, women who seek to succeed in such environments are placed in a “double bind.”_[304] This bind is a “conflict between the assertiveness and aggressiveness required to get the job done and the image required to fit the female stereotype.”_[305]

Fiske’s concept played a major role in the Supreme Court’s analysis. The plurality ruling in Hopkins’s favor stated that Title VII prohibits “sex stereotyping,” observing that “we are beyond the day when an employer could evaluate employees by assuming or insisting that they matched the stereotype associated with their group.”_[306] This language would appear to protect women from both covering and reverse covering demands. Yet the plurality also opined that Price Waterhouse had placed women in an “intolerable and impermissible catch 22: out of a job if they behave aggressively and out of a job if they do not.”_[307] The plurality thus provided two theories on which Price Waterhouse’s treatment of Hopkins was illegal: (1) It involved sex stereotyping, and (2) it placed Hopkins in a Catch-22.

Under the first theory, Title VII’s conception of sex expands to encompass a performative dimension. A prohibition on sex stereotyping means that one cannot penalize a woman for behaving in ways that corroborate or violate one’s stereotypical conception of a woman, thereby extending protections of “sex” to protections of “gender.” In such a regime, a demand for either covering or reverse covering is cognizable as a Title VII violation. Under the second theory, women can be forced to modify their activity so long as the demands that are made upon them are not inconsistent. In this regime, a demand for either covering or reverse covering is permitted, but demands for both together are not. The second theory obviously affords much narrower protections than the first.

I believe the plurality opinion clearly espouses the first theory._[308] But it is no surprise to me that subsequent courts have notoriously failed to apply that theory, as the theory has a radical breadth._[309] Most of the courts that seek to cabin Title VII’s protections to a nonperformative conception of sex simply ignore the “sex stereotyping” language._[310] Yet at least one court has explicitly used the Catch-22 theory to restrict the reach of the sex-stereotyping theory. In Dillon v. Frank,_[311] the Sixth Circuit considered the case of a man who had been taunted and harassed for being perceived to be a homosexual._[312] Dillon sought to bring his case within the ambit of Hopkins protection by contending “that he was subjected to such stereotyping in that he was not deemed ‘macho’ enough by his co-workers for a man, and that the verbal abuse resulted from this stereotyping.”_[313] One of the grounds on which the court rejected this claim pertained to the Catch-22. The court observed that the Hopkins Court had

emphasized the “intolerable and impermissible Catch-22” in the stereotyping in that case. A desirable trait (aggressiveness) was believed to be peculiar to males. If Hopkins lacked it, she would not be promoted; if she displayed it, it would not be acceptable. In our case, Dillon’s supposed activities or characteristics simply had no relevance to the workplace, and did not place him in a “Catch-22.”_[314]

In other words, Dillon was not in a Catch-22 because no one in the workplace wished him to be more effeminate. Because only the reverse covering demand was made of him, he was not placed in a double bind. This reading of Hopkins thus makes it a case that prohibits only an impermissible double bind, not sex stereotyping alone. This interpretation severely limits the applicability of Hopkins, as can be seen in its premise that Hopkins can never be applied to protect effeminate men.

It is easy to see why courts have sought to limit the reach of the sex-stereotyping theory. Such a theory would make a myriad of covering and reverse covering claims legally cognizable. Yet it is unclear that a “double bind” theory provides a particularly good gatekeeping mechanism, at least insofar as women are concerned. The implication underlying the phrase “intolerable and impermissible Catch-22” is that the covering and reverse covering demands cannot be simultaneously fulfilled. But how much of a Catch-22 is the conjunction of covering and reverse covering demands? It is hard to imagine that most professional women are not subjected to both the demand to cover and the demand to reverse cover. Many of them, like Sandy Kinsey in Hopkins’s account, appear to succeed in balancing these demands.

The Supreme Court’s emphasis on the Catch-22, then, is simultaneously both sophisticated and naive. It is sophisticated in that it recognizes that women may be differently situated from other groups in having the dominant group consistently impose seemingly contradictory demands upon them. It is naive in its belief that these demands cannot be reconciled. This naiveté is troubling in that it obscures the more fundamental problem in these work situations. What is “intolerable” in these cases is not that the demands are contradictory, but rather that either demand is made at all. To the extent that the Catch-22 rationale is deployed, it will still mean that women who can assimilate must do so.

In this sense, the Catch-22 rationale is yet another instantiation of immutability as a gatekeeping concept in antidiscrimination jurisprudence. The limitation of protection to sex-based Catch-22s is a limitation to situations that individuals cannot control. It is because individuals cannot lift themselves out of the double bind that Title VII does so. Implicit within the limitation is the belief that individuals who can control their gender-based performances must do so.

What we see in the sex-stereotyping reading of the Hopkins plurality opinion—which I take to be the most faithful reading of that opinion—is a remarkable yet unfulfilled promise. The sex-stereotyping theory defines “sex” under Title VII in a more expansive way, much in the way that Justice Kennedy’s opinion in Hernandez suggested that race might be defined. Under this more expansive definition, sex includes behavioral attributes that would be favored in men but disfavored in women. As such, sex extends beyond mere biological facts and into cultural ones. That extension blurs the distinction between sex and gender.

The blurring of the distinction between sex discrimination and gender discrimination, however, will obviously be resisted. As noted earlier, there is a conventional wisdom that distinguishes between sex as a biological, prediscursive fact and gender as a cultural, discursive fact._[315] That wisdom has been internalized by the law. As Justice Scalia has noted, “The word ‘gender’ has acquired the new and useful connotation of cultural or attitudinal characteristics (as opposed to physical characteristics) distinctive to the sexes. That is to say, gender is to sex as feminine is to female and masculine is to male.”_[316] Scalia’s words come from an equal protection case, but they diagnose the resistance to the sex stereotyping theory in the Title VII context as well. If “sex” in Title VII were to incorporate Scalia’s “gender,” then cases involving effeminate men or masculine women would fall within the ambit of sex discrimination.

3.   The Performative Turn in Sex

In discussing race, I noted that the best evidence for the proposition that the judiciary can adopt a performative conception of race is that it has already done so. I hypothesized that the courts were particularly likely to do so in formation cases like the antebellum race trials or the racial prerequisite cases._[317] I now seek to demonstrate that the same premise obtains in the sex context.

In the sex context, such formation cases include cases concerning the sex of transsexuals. In these cases, courts have to decide whether transsexuals are men or women for the purposes of determining whether they can change their birth certificates,_[318] whether they can play in women’s sports leagues,_[319] or whether they can have valid marriages to individuals who presumably share their chromosomal sex._[320]

I found fifteen such reported cases. While the majority of these cases adopt a nonperformative conception of sex (i.e., the transsexual is always the sex she was born), a significant minority seemed to adopt a performative conception of sex (i.e., the transsexual can become her postoperative sex). This suggests judicial receptivity to the concept of sex as a performative category that might inspire some optimism about a broader judicial embrace of such a conception. As in the racial prerequisite cases, however, this receptivity is deeply qualified.

To begin with the easy point, courts that assigned a transsexual her pre-operative identity almost uniformly embraced a nonperformative conception of sex. These courts most often relied on the individual’s chromosomes. As one New York court succinctly stated, “‘male-to-female transsexuals are still chromosomally males while ostensibly females.’”_[321] By emphasizing the immutable nature of the chromosomes, courts underscored the futility of human agency to alter an individual’s sex. As one court observed in 1973, “surgery for the transsexual is an experimental form of psychotherapy by which mutilating surgery is conducted on a person with the intent of setting his mind at ease, and that nonetheless, does not change the body cells governing sexuality.”_[322] Next to chromosomes, genitals were the prime nonperformative signifier adduced by the courts. Such courts focused on the fact that no amount of human ingenuity could create a natural penis or a vagina. In ruling that a female-to-male transsexual’s postoperative penis was not a penis, the court noted that hormone treatments and surgery had not produced a penis capable of “assuming male duties and obligations inherent in the marriage relationship.”_[323] In other words, one uncopyable mark of the natural penis was its ability to become erect. These courts seemed to manifest a grim satisfaction in the futility of transsexual performance. If the performative conception holds that iterated acts and citations accrete to form an identity,_[324] the classical conception holds that these acts and citations, no matter how often repeated, do not cumulate and do not change the prediscursive identity.

In contrast, courts that ruled that transsexuals were their postoperative sex at least in part adopted a performative conception of identity. Two kinds of performances—sexual and gender performances—seemed to be particularly salient. Courts strove mightily to read the transsexual’s sex back into a heterosexual matrix, noting, for example, that a male-to-female (MTF) transsexual had to be a woman because she engaged in “normal” heterosexual sex with her male spouse._[325] Other courts emphasized gender performance. As a New York state court noted,

This individual dresses, acts, and comports himself as a member of the opposite sex. The applicant appeared before this court and, were it not for the fact that petitioner’s background was known to the court, the court would have found it impossible to distinguish this person from any other female._[326]

Note that a linguistic sex change occurs over these two sentences, from the individual comporting “himself” in the first one to the person being found indistinguishable from any “other” woman in the second.

These courts in the 1960s and 1970s demonstrated a striking willingness to accede to performative conceptions of sex that are now marked as postmodern. The conventional wisdom—then as now—is that sex is a stable prediscursive substrate on which orientation and gender are built. When Butler suggests that the arrow of causation runs the other way—positing that sex is actually a back formation from orientation and gender—her ideas are labeled radical. Yet here one observes courts some three decades ago quite placidly noting that an individual’s orientation and gender performances are the causes rather than the effects of his or her sex. If the MTF transsexual sleeps with men then she must be a woman—orientation here determines sex, rather than vice versa. If an MTF dresses like a woman, then, for juridical purposes, she is a woman—gender here determines sex, rather than vice versa.

Yet while the courts reifying the postoperative identity of the transsexual seem to embrace a performative conception of that identity, that embrace is always partial. The facially performative discourse of these courts keeps falling back into a kind of essentialization of sex, much as in the racial prerequisite cases.

While this occurs in more than one way, I wish to focus on the essentialism embodied in the concept of the gendered “soul.” I take this term from expert testimony in one case, where a physician stated that “‘Elaine Francis Ladrach has undergone a gender transformation from male to female establishing a somatic gender to match that of her soul.’”_[327] The courts themselves use less religious language, relying on concepts like “psychological sex” rather than the concept of the gendered soul._[328] But the idea is the same—like nested Russian dolls, the female soul is nested inside a male body which is nested inside a female performance. To frame it in the rhetoric of fraud that is so pervasive in these opinions,_[329] it is as if the court were to say that one could remove Elaine Ladrach’s clothes and postoperative body to expose her male chromosomes, but why stop there? Why not remove the chromosomes and expose the underlying female soul? If the two guarantors of essentialist reasoning are Nature and God, here the latter steps into the void left by the former.

The virtue of the soul is that it cannot be known except through behavioral signifiers—the indicia of the soul are thus performative indicia. The soul thus effectively reifies the performative acts in which the individual engages, but disguises that project of invention as a project of detection. Yet this, again, is the constative fallacy against which Austin inveighed._[330] It reads the performative formation of gender back into a constative enterprise. In this constative paradigm, the essentialist substrate has not been rejected, but rather shifted from body to soul. Thus, the courts fall back into the constative fallacy even as they reject the concept of biological essentialism.

Thus, like the racial prerequisite cases, the transsexual cases that reify postoperative identity both expose and preserve the classical conception of identity. The cases expose the fiction that sex is a biological phenomenon by observing that gender performances, rather than chromosomes, determine one’s sex. Yet at the same time, they preserve the fiction that sex is a prediscursive phenomenon (even if not a biological phenomenon) by observing that these performances are evidence of an underlying gendered soul. These acts are the evidence, not the elements, of identity.

This insight should again inspire both optimism and pessimism for the judiciary’s ability to fashion a jurisprudence that would protect gender under protections for sex. On the one hand, it demonstrates that courts are in some ways the ultimate institutions for reifying performative identities—for transforming “acting as if” into “is” through their linguistic performativity. Yet at the same time, it suggests the reluctance of the courts to relinquish a classical model of identity in which they are detecting, rather than in any sense inventing, those identities.

D.  Synthesis

It is time to pull some information together. The previous discussion of discrimination can be diagrammatically expressed as follows:

Figure 2.

	Conversion	Passing	Covering
Race	Blood       (formal-race)	Skin Color       (formal-race)	Race-Salient Behavior (e.g., language)   (culture-race)
Sex	Chromosomes       (sex)	Secondary Sex Characteristics     (sex)	Sex-Salient Behavior (e.g., pregnancy)   (gender)
Orientation	Orientation “Gene”     (status)	Orientation Self- Identification     (status)	Orientation-Salient Behavior (e.g., sodomy)   (conduct)

 

For the three classifications (race, sex, and orientation), I have identified the traits to which the three assimilationist demands (conversion, passing, and covering) apply. Thus, for race, the conversion demand applies to blood or ancestry, the passing demand to skin color, and the covering demand to race-salient behavior such as language. For sex, conversion applies to chromosomes, passing applies to visible secondary sex characteristics, and covering applies to sex-salient behaviors, such as pregnancy. For orientation, conversion applies to the (hypothetical) orientation gene, passing applies to self-identifying speech about orientation, and covering applies to orientation-salient conduct such as sodomy.

This schematization illuminates both divergence and convergence among the social situations of racial minorities, women, and gays with regard to assimilation. The divergence is represented by the bold horizontal line, which represents the antidiscrimination schism. This line reflects the perception that individuals cannot change or hide their formal-race attributes or their sex attributes, but that individuals can change or hide their orientation-status attributes. This perception can be used to justify protecting gays less than racial minorities or women.

I have argued that focusing on this divergence has obscured an equally important convergence among the social situations of the three groups. The bold vertical line distinguishes between immutable and mutable aspects of each classification. It demonstrates that if immutability is the only ground on which a group can defend against assimilation, then all three groups are vulnerable to the demand to cover their race-salient, sex-salient, or orientation-salient attributes. It thus suggests a ground of coalition among the interests of all three groups.

Both the divergence and the convergence have been reified by antidiscrimination law’s emphasis on immutability. Major strands of Title VII and equal protection jurisprudence have used immutability to justify protecting race and sex but not orientation (the horizontal line). Major strands of Title VII and equal protection jurisprudence have also used immutability to justify protecting only the immutable aspects of race or sex (the vertical line). The result of these two moves has been to restrict much of federal antidiscrimination protection to the shaded area bounded by the two lines.

 

 

[Part IV on Critiques of the Covering Model omitted.]

 

V. Conclusion

I conclude with my aspiration for the model of discrimination I have been propounding. My critique of the current paradigm of discrimination is that it fails to focus on the question of assimilation, or, more broadly, on the question of change. This failure is somewhat puzzling. Civil rights practice, after all, is fundamentally about who has to change: The homosexual or the homophobe? The woman or the sexist? The racial minority or the racist? Yet the current paradigm errs prescriptively in extending greater protections to those who cannot change, and errs descriptively in characterizing identities like race and sex as being incapable of any kind of change. I believe that it could not err in either of these ways if it more closely examined the ubiquity of assimilation. When we see how much we all can and do change along every axis of our identity, we should apprehend that any account of discrimination that does not take assimilation into account is fundamentally impoverished. Just thinking of such change, I hope, will change us.