The Yale Law Journal

November 2017

Title VII’s Statutory History and the Sex Discrimination Argument for LGBT Workplace Protections

Gender and Sexual OrientationStatutory InterpretationCivil-Rights Law

abstract. The Equal Employment Opportunity Commission (EEOC) and the Seventh Circuit have taken the position that Title VII’s bar to employment discrimination “because of . . . sex” applies to discrimination against lesbian, gay, bisexual, and transgender (LGBT) persons. This interpretation follows from the ordinary meaning of the statute, read as a whole and in light of its purpose. If an employer fires a woman because she is married to another woman, rather than a man, the employer has, literally, acted “because of” her sex (if she had been a man, marriage to a woman would have been fine) and because of the sex of her partner. It is difficult to deny that “sex” is not at least one “motivating factor” in the employment decision, which is all that the current version of Title VII requires for liability. Moreover, this reading of Title VII accords with its purpose, which is to entrench a merit-based workplace where specified traits or status-based criteria (race, color, national origin, religion, and sex) are supposed to be irrelevant to a person’s job opportunities.

Treating antigay discrimination as a form of sex discrimination is not a new idea, but for several decades most federal judges have rejected it, and most members of Congress have ignored it. This, however, is an idea that has ripened over time. New circumstances have rendered the argument not only plausible but also compelling. The most significant new development has been evolving social facts and assumptions about sex minorities: in 1964, employees thought to be “homosexuals” were outside the scope of the merit-based workplace, because Americans believed them to be mentally ill, psychopathic, and predatory. Today, those views have been discredited. This shift in thought connects with a second new circumstance: a radically different constitutional baseline. As late as 2003, “homosexuals” could constitutionally be considered presumptive criminals, but the Supreme Court has for twenty years been developing a constitutional norm that gay people cannot be singled out for special legal exclusions without a rational public justification. Indeed, the Court has ruled that the constitutional right to marry applies to same-sex (i.e., “homosexual”) couples. It is constitutionally jarring to know that, in most states, a lesbian couple can get married on Saturday and be fired from their jobs on Monday, without legal redress.

A third new development has been the formal evolution of Title VII itself. Judges as well as commentators have largely ignored the “statutory history” of Title VII—its formal evolution through a process of amendment by Congress and authoritative interpretation by the Supreme Court. The Trump Administration and other skeptics of a broad reading of sex discrimination maintain that Title VII divides the world into males and females, and does nothing more than require employers to apply the same rules to both sexes. According to this view, antihomosexual workplace exclusions or harassment operates equally on both sexes (i.e., both lesbians and gay men are harmed). But the Supreme Court has authoritatively interpreted Title VII to bar gender stereotyping, which also operates to protect both male and female employees alike. Congress ratified and expanded upon that interpretation in its 1991 Amendments to Title VII, which also reaffirmed its statutory mission to ensure a merit-based workplace free from sex-based decision making, even when sex is but one “motivating factor” in the discrimination. Because LGBT persons are gender minorities and because anti-LGBT discrimination is rooted in rigid gender roles, Title VII today bars discrimination because of the sex of the employee’s partner/spouse, just as it bars discrimination because of the race or religion of his or her partner/spouse.

author. John A. Garver Professor of Jurisprudence, Yale Law School. I deeply appreciate excellent research assistance and editorial guidance from Eric Baudry, Yena Lee, Charlotte Schwartz, Daniel Strunk, and Todd Venook (all Yale Law School Class of 2019). I am also grateful for valuable comments from Martha Ertman, Steve Calabresi, Guido Calabresi, Sharra Greer, Heather Gerken, Abbe Gluck, Pat McGlone, John Davis Malloy, Judith Sandalow, Vicki Schultz, Amanda Shanor, Reva Siegel, Susan Silber, and John Witt, as well as participants at a Yale Law School faculty workshop.


Title VII of the Civil Rights Act of 1964 bars employment discrimination “because of . . . sex.”1 A successful part-time teacher at Ivy Tech Community College, Kimberly Hively, complained that the college refused to consider her for a permanent job because she is a lesbian. If true, does that refusal constitute discrimination because of sex? Because Title VII does not bar discrimination “because of . . . sexual orientation,” federal appeals courts have uniformly said “no” to this question—until the Seventh Circuit, sitting en banc, reconsidered the issue.2

Writing for the en banc, eight-judge majority in Hively v. Ivy Tech Community College of Indiana,3 Chief Judge Diane Wood offered two interconnected arguments to support the holding that an employer’s refusal to hire a lesbian constitutes discrimination because of sex. Reading the statute literally, judges in sex discrimination cases often ask whether the plaintiff has shown that the employer would have treated a similarly situated “comparator” (a person of the opposite sex) more favorably. If Hively had been a man, sexually cohabiting with or married to a woman, the college would have considered him for permanent employment on his merits. Ivy Tech allegedly rejected Hively out of hand because she was a woman partnered with another woman. Hence, she was allegedly denied the job “because of . . . [her] sex” as a woman rather than a man. This line of reasoning is often called the “comparator argument.”4

In the alternative, Chief Judge Wood reasoned that Hively was discriminated against because of the sex of her intimate associate (her partner).5 A precedential basis for this “associational discrimination argument” is Loving v. Virginia,6 where the Supreme Court ruled that state discrimination against interracial couples constitutes discrimination because of race. For the same reason that state discrimination against a black woman cohabiting with or married to a white man is race discrimination, Ivy Tech’s discrimination against a woman cohabiting with or married to another woman is sex discrimination. In the first case, the regulatory variable—the factor that changes the legal treatment—is the race of the associated person; in the second case, the regulatory variable is the sex of the associated person.

Writing for three dissenting judges, Judge Diane Sykes criticized these formal arguments as an excessively dynamic, “judge-empowering” interpretation of the statutory language.7 Focusing on the original meaning of the statutory text, Judge Sykes argued that “sex” in 1964 only meant one thing—the two biological sexes (male and female)—and could not have meant “sexual orientation,” a term not widely used in 1964.8 In contrast, Congress has in subsequent statutes, such as the Matthew Shepard and James Byrd, Jr. Hate Crimes Prevention Act of 2009,9 specifically prohibited discrimination because of “sexual orientation” as well as because of “sex,” indicating that the two terms have different meanings.10 In ordinary parlance today, Judge Sykes argued, no one would say that workplace gay-bashing constitutes “discrimination because of sex”; almost everyone would say that it is “discrimination because of sexual orientation.”11 Not only does “sexual-orientation discrimination spring[] from a wholly different kind of bias than sex discrimination,”12 but the only legitimate means to update the statute in this way is through the legislative process.13

Concurring in most of Chief Judge Wood’s opinion, Judge Joel Flaum (joined by Judge Kenneth Ripple) maintained that the text, read in light of the whole Act, supported Hively’s Title VII claim. Given the allegations in the complaint, Judge Flaum found it hard to deny that Ivy Tech excluded her at least in part either because of her sex (the comparator argument) or because of the sex of her romantic partner (the associational discrimination argument).14 In response to Judge Sykes’s point that this was just a case of sexual orientation discrimination, Judge Flaum observed that it had elements of both sex-based and sexual orientation-based discrimination.15 Because Title VII was amended in 1991 to bar sex discrimination even when sex is only one “motivating factor,” today’s statute supported Hively.16

In contrast to other judges in the majority, concurring Judge Richard Posner suggested that Chief Judge Wood’s opinion was not dynamic enough. He joined her majority opinion but also maintained, contrary to Judge Sykes, that judges should “update” statutes based upon current social norms and their understanding of the best workplace policy.17 Homosexuality has become sufficiently normalized in society that judges should dynamically interpret Title VII to reflect its most efficient deployment.

The same issue also recently divided the Eleventh Circuit. In Evans v. Georgia Regional Hospital,18 the panel, in an opinion by District Judge Jose Martinez, held that a lesbian might have a Title VII claim for sex discrimination if the employer denied her opportunities because of gender stereotyping but has no Title VII claim for simple sex discrimination, even if she was denied a job because she is romantically attracted only to women.19 Dissenting on the latter point, Judge Robin Rosenbaum argued that the gender stereotyping argument went further than the majority recognized and instead justified Title VII protection for female employees who depart from the deep stereotype that women should find romantic love with the right man, not the right woman.20 Concurring in Judge Martinez’s opinion, Judge William Pryor Jr. argued that Judge Rosenbaum’s interpretation would amount to an amendment, rather than an interpretation, of the statute.21 Counsel for Jameka Evans plans to petition the Supreme Court for review of the Eleventh Circuit’s decision.22

Chief Judge Robert Katzmann of the Second Circuit has also recently urged a reconsideration of his court’s precedent declining to apply Title VII to bar sexual orientation discrimination.23 Like Chief Judge Wood, Chief Judge Katzmann found persuasive both the comparator and the associational discrimination arguments;24 like Judge Rosenbaum, he opined that anti-LGBT discrimination involves gender stereotypes that the Supreme Court has ruled cannot be the basis for employment decisions in the merit-based workplace.25 Chief Judge Katzmann also addressed the argument, previously accepted by the Second Circuit, that Congress had “ratified” the previous court of appeals decisions because it did not override them in its 1991 Amendments and because it declined to enact any one of several dozen bills specifically seeking to bar sexual orientation discrimination. Following the Supreme Court, he cautioned against drawing legal meaning from “a proposal that does not become law. Congressional inaction lacks persuasive significance because several equally tenable inferences may be drawn from such inaction, including the inference that the existing legislation already incorporated the offered change.”26

Following Chief Judge Katzmann’s suggestion, the Second Circuit has granted en banc review of this issue in Zarda v. Altitude Express, Inc.27The EEOC filed an amicus brief in that case, defending its view that LGBT employees like Donald Zarda are protected by Title VII.28 The Trump Administration’s Department of Justice, however, subsequently filed an amicus brief supporting the opposite interpretation on precisely the grounds rejected by Chief Judge Katzmann, namely, that Congress ratified the older court of appeals cases when it amended Title VII in 1991 and when it rejected bills that would have amended Title VII or created a new statute to protect against sexual orientation discrimination specifically.29 In an appendix, the Justice Department’s brief listed sixty-two bills introduced in Congress between 1974 and 2017 that would have addressed, in a variety of ways, the treatment of sexual and gender minorities in the workplace. Like most proposed legislation, a large majority of the bills died in committee, without hearings or any kind of vote. Over the last forty years, congressional committees held hearings on ten of these bills, and three were voted upon by one chamber (failing 49-50 in the first case, passing one chamber in the other two cases).30

With such a dramatic split in the circuits and even within the executive branch, the issue should soon reach the Supreme Court.31 Ironically, the formalism of Chief Judge Wood’s and Judge Flaum’s approach would be attractive to the Supreme Court Justices (such as Justice Thomas) least inclined, ideologically, to read Title VII to protect LGBT employees and would be an incomplete analysis to some of the Justices (such as Justice Breyer) most likely to read Title VII more broadly. The whole act analysis suggested by Judge Flaum ought to be appealing to Chief Justice Roberts,32 but would the Chief be willing to interpret a civil rights law expansively? Will the Supreme Court divide along predictably ideological and political lines—or might the legal arguments provide a canvas to debate the issue in the relatively nonideological manner the Seventh Circuit did?33 A related issue is whether Title VII’s sex discrimination bar protects transgender employees. For reasons I find persuasive, Judge Pryor has joined Judge Rosemary Barkett’s panel opinion ruling that it does, a judgment followed by most other federal judges, including a recent Seventh Circuit panel.34 Note that Judge Pryor liberally interprets discrimination because of sex to include discrimination because of one’s self-reported sex and discrimination because of gender stereotyping, but is not willing to include discrimination because of the sex of one’s partner.

Hively, Evans, and Zarda not only tee up important substantive issues of job discrimination, but also methodological issues regarding the proper approach to statutory interpretation in general, and dynamic statutory interpretation in particular. Broadly speaking, the Seventh Circuit judges followed three different methodological approaches. Judge Sykes applied what she considered to be Title VII’s original meaning: the objective meaning (to a reasonable speaker) entailed by the statutory text, “discrimination because of . . . sex.”35 Judge Posner called for judicial dynamic interpretation of Title VII, whose original meaning, in his view, has been rendered obsolete by changed social and workplace norms.36 Chief Judge Wood followed a pragmatic approach that considers statutory text, purpose, and precedents, as well as relevant constitutional norms and directives.37 Chief Judge Katzmann is also an advocate of a pragmatic approach that considers various sources for understanding statutory meaning.38

A pragmatic approach is most consistent with the approach to statutory interpretation long followed by the U.S. Supreme Court. Part I of this Feature will suggest that the seemingly simpler approaches followed by Judges Posner and Sykes require consideration of the wider array of sources explored by Chief Judge Wood and by this Feature. Generally, Part I maintains that Title VII’s statutory plan—entrenchment of a merit-based workplace as regards the criteria listed in the law—must inform the analysis of any judge, whether she be an original meaning textualist, a dynamic interpreter, or a pragmatist. Original meaning analysis, properly set forth, produces a much more complicated inquiry than the arbitrary choices made by Judge Sykes. Dynamic interpretation requires better grounding in traditional legal materials than Judge Posner attempts. And a legal pragmatic analysis needs to explore more materials than Chief Judge Wood analyzes in her opinion for the court in Hively.

The main theme of this Feature is that any judicial interpretation of Title VII’s text, purpose, or precedents is incomplete, from any methodological point of view, without an exploration and understanding of the statutory history (congressional amendments and authoritative interpretations) as well as the legislative history of Title VII and its amendments. Part II will offer such an exegesis of Title VII’s statutory history, namely, its formal evolution through authoritative interpretation and congressional amendments. Part II will contrast statutory history, which has never been a controversial source of guidance in statutory interpretation, with legislative history, the internal congressional materials generated in the process of statutory deliberation and enactment, which has been controversial within the Court. Notwithstanding criticism, the Court still relies on legislative history tied to statutory text created by Congress. Indeed, Justice Scalia, the father of the strict new textualism, has said that judges should consult a law’s legislative history for the same reason he and other textualists consult the Federalist Papers in constitutional cases—namely, to help the judge understand the meaning of words, phrases, and structures Congress has enacted.39 Accordingly, Part II will examine the legislative history of Title VII and its amendments only insofar as they help us understand the meaning of the words, phrases, and structures actually enacted into law. Contrast this approach with the Department of Justice’s Zarda amicus brief, which has the vice of not tying its legislative inaction arguments to text adopted by Congress. Instead, the Department gestures toward a link to the 1991 Amendments without actually citing any relevant new statutory text.

Title VII’s statutory history establishes the following points of law that are relevant to Hively, Evans, and Zarda. First, Title VII is equally committed to purging the workplace of arbitrary sex-based, race-based, and religion-based criteria. Congress and, eventually, the Supreme Court have rejected the approach initially followed by the EEOC, whereby enforcement was focused on race discrimination, with discrimination because of sex and religion assuming a subordinate position. As the plain language suggests, the statute presumptively follows the same precepts for race-, sex-, and religion-based discrimination—except where the text explicitly creates a distinct regime, as it does for bona fide occupational qualifications, which are not available to justify discrimination because of race.40 Even employment decisions motivated only in part by a disapproved criterion are now questionable under the statute as amended in 1991.

Second, Title VII not only bars employment practices that treat all women differently from all men, but also bars practices that treat some men or some women differently because of their sex or because of the sex of their spouses or partners. Moreover, binding Supreme Court precedent, ratified and expanded by Congress in 1991, commits Title VII to the broader principle that employers cannot prescribe non-merit-based gender roles on both men and women equally. The same idea has been uncontroversial with regard to race among lower court judges.41

Third, at the same time Title VII was formally evolving, the Constitution’s treatment of LGBT persons has formally evolved. Between 1967 and 1996, authoritative Supreme Court decisions advised Congress and state legislatures that homosexual or bisexual persons could be treated as per se “afflicted with psychopathic personality” and that private nonprocreative sexual acts between consenting same-sex couples could be criminalized as felonies.42 Psychopaths and presumptive felons do not enjoy all the protections of a merit-based workplace, an important reason for Congress’s longstanding lack of interest in deliberating about any workplace rights for LGBT people. But, reflecting changes in social attitudes and in medical and psychiatric views, the Court changed the constitutional baseline in a series of landmark Fourteenth Amendment decisions handed down between 1996 and 2015.43 This dramatically evolving constitutional context upends the assumptions of the leading circuit court precedents excluding “homosexuals” and “transsexuals” from the protections of Title VII and further undermines the congressional acquiescence arguments made by the Trump Administration’s Justice Department in Zarda. That is, judicial precedents premised on the assumption that Congress and employers can discriminate against gay and lesbian employees because they are presumptive criminals or psychopaths not only can be but must be revisited once that presumption has not only been revoked but reversed.44

Part III pauses to summarize the lessons of our extensive investigation of Title VII’s statutory history. Looking forward, this Feature then addresses the possibility that an ideologically driven but textualist Supreme Court no longer effectively monitored by a gridlocked Congress will be tempted to impose an anti-LGBT reading on a statutory text, structure, and history that strongly support the plaintiffs’ claims in Hively, Evans, and Zarda. The web of statutory text, structure, precedent, practice, and constitutional background norms is so tightly interconnected and strongly hostile to reading LGBT employees out of the protections of Title VII’s merit-based workplace that an effort by the Supreme Court to turn back the clock through a stingy reading of the statutory text would probably be an embarrassment to original meaning jurisprudence. Indeed, a poorly researched textual analysis, divorced from statutory history, would amount to an assault on the rule of law itself. The predictability promised by the rule of law requires even the Supreme Court to respect the web of interconnected rules and neutral principles that administrators have created, judges have ratified or altered, and legislators have relied on when they revisit the statute through amendments and overrides. As the Seventh Circuit recognized in Hively, that web of rules and neutral principles supports a simple directive that Title VII’s merit-based workplace protects LGBT individuals.

I. the merit-based workplace and the meaning of title vii

A. The Merit-Based Workplace

Start with the normative ideal of a merit-based workplace, where all people would have and retain jobs based upon their ability to perform and would not be excluded from jobs or harassed at work because of personal characteristics irrelevant to their capabilities. This ideal is inspired by sociologist Max Weber’s contrast between a patriarchal culture, wherein economic and social structure was organized around status and personalized around the father or master, and modern culture, which seeks to organize production around objective, meritocratic rules, rather than subjective, caste-based practices.45 Under Weber’s framework, a merit-based workplace is axiomatic to the assumptions and practices of modern culture—and hence a highly desirable norm.

A wide range of modern normative philosophies would view the merit-based workplace as a worthy ideal. For example, most utilitarian thinkers, seeking the greatest good for the greatest number, would value such a workplace because it allocates resources much more efficiently and promises to reduce disruptive workplace practices (like hazing and harassment).46 Many philosophers of justice would endorse a merit-based workplace because it focuses evaluation on the individual and establishes baselines that are fair and just, the kind of rule that most of us would choose behind the veil of ignorance (i.e., we do not know what entitlements we would have in a world following the rule).47

Finally, many civil rights advocates and theorists, including feminists, would maintain that a genuine merit-based system would undermine the operation of prejudices and stereotypes that hold back women and minorities from equal opportunities in the workplace.48 Their vision of a pluralistic workplace is sometimes in tension with the utilitarian or economic vision, but as to many issues they press in the same direction. I should note here that the merit-based workplace is not a panacea, especially for inequities or discriminations faced by persons who are poor or come from disadvantaged backgrounds. My only point here is that such a norm enjoys support from a variety of perspectives and fits the aspirations of our society.

The merit-based workplace norm is dynamic. A 1964 statute barring job discrimination for reasons unrelated to people’s capabilities would not immediately have been applied to protect “homosexual” employees, as there was a social consensus that “homosexuals” were not capable of doing most jobs due to their inherent “psychopathic personality” and other mental disorders.49 In addition, every state but Illinois at that time considered “homosexuals” to be criminals by reason of their characteristic sexual activity (consensual sodomy).50 The same statutory language, read today, could protect gay people against discrimination, because both social facts and the law have changed: it can no longer seriously be maintained that homosexuality is evidence of psychopathy or mental illness, and consensual sodomy can no longer be a crime after Lawrence v. Texas.51 Indeed, a job discrimination law that authorized employers to discriminate against LGBT workers because of traditional stereotypes would be constitutionally suspect under Romer v. Evans,52 in which the Supreme Court ruled that laws excluding gay, lesbian, and bisexual persons from general legal protections without plausible justification (or because of antigay animus) violate the Equal Protection Clause.

Enacted as part of the Civil Rights Act of 1964, Title VII, on its face, legislates the merit-based workplace as regards characteristics associated with race, color, national origin, religion, and sex—precisely the modern Weberian approach, as limited to the enumerated traits, whose consideration is presumptively not consistent with workplace meritocracy. While some utilitarian thinkers (including Judge Posner) argue that Title VII does not efficiently promote this meritocratic ideal, others have endorsed its mission.53 (And of course Judge Posner enthusiastically advanced Title VII’s mission by his concurring opinion in Hively.54) From most feminist perspectives, Title VII properly “seeks to . . . identify, and change, discriminatory employment practices that reinforce negative stereotypes and foster unnecessary difference and division, and . . . [to] endorse practices that encourage people to relate to each other as equals across boundaries of sex and race.”55 As Part II will demonstrate, the statutory and legislative history of Title VII not only makes clear that the merit-based workplace represented Congress’s plan for the statute, but also elaborates on the threats to that plan posed both by sexuality-based workplace discrimination and by gender stereotypes regarding the role of men and women in the family, the workplace, and society.

Like the generic statute described at the beginning of this part, Title VII in 1964 would not have been applied to ensure a liberal workplace for “homosexuals” or “transsexuals,” for those Americans were, literally, considered psychopaths, criminals, and enemies of the people—propositions that no longer enjoy respectable support. The question in Hively and the other recent cases is whether judges should interpret the current version of Title VII to include gay and lesbian employees within the merit-based workplace. To answer that statutory interpretation question, judges are supposed to consider (1) the text of the statute, including the whole act and the statute’s evolution through legislative amendments and repeals, (2) the legislative purpose, (3) Supreme Court precedents authoritatively interpreting the statute, (4) the legislative deliberations preceding and accompanying the enactment of the statute and its amendments, (5) the regulatory history of the statute and how agency interpretations interact with the legislative deliberations, and (6) larger constitutional, institutional, and social norms.56

These standard legal sources ought to be applied with an eye to the purposes of statutory interpretation as an operation of the government. Specifically, the purposes of statutory interpretation are (1) ensuring the rule of law (predictable and consistent application of concrete legal authorities using an objective, transparent method of analysis), (2) in a manner that respects the legitimate deliberation and actions of our democratically elected lawmakers, and (3) that reasonably adapts statutes to the evolving governance needs of society.57 In Hively, Judge Posner lionized the third element, dominating the other two,58 but I (like Chief Judge Wood and Judge Sykes) maintain that dynamic judges ought to pay close attention both to the operation of the democratic process and the most concrete sources of statutory meaning, namely, statutory text, structure, and precedent.

A central contention of this Feature is that the rule-of-law norms emphasized by Chief Judge Wood and Judges Flaum and Sykes complement rather than conflict with the updating needs emphasized by Judge Posner. To accomplish this synthesis, I shall examine standard legal sources that were not explored in Hively or Evans—namely, those reflecting authoritative interpretation and democratic deliberation. Subsequent congressional amendments, their legislative purpose and history, and administrative and judicial precedents provide democratic support for a synthesis of the approaches taken by Chief Judge Wood and Judge Rosenbaum (dissenting in Evans).

B. Original Meaning of Title VII

In Hively, Judge Sykes relied upon a simpler approach to statutory interpretation that has enjoyed enthusiastic support from many judges and some scholars. Following Justice Scalia, Judge Sykes invoked the original meaning of the statutory text, which bars employer discrimination “because of sex.”59 Is she right to say that the original meaning of this language requires faithful textualist judges to dismiss Hively’s claim? Several strict adherents of original meaning jurisprudence—notably Judges Frank Easterbrook and Kenneth Ripple—joined Chief Judge Wood’s Hively opinion, which suggests that the original meaning inquiry might not be so straightforward.

As enacted in 1964, Title VII did not define “sex.” Federal judges often start with standard dictionaries of the era to think about statutory meaning of undefined terms.60 Judge Sykes found that “sex” in 1964 had only one meaning—the division of humanity into biological males and females.61 Such a simple understanding is incomplete, at best. The unabridged 1961 printing of Webster’s (the most-cited dictionary in Supreme Court opinions) defined the word “sex” to mean three different things:

  • “[o]ne of the two divisions of organisms formed on the distinction of male and female,” or sex as biology;
  • “[t]he sphere of behavior dominated by the relations between male and female,” or sex as gender (man=masculine, woman=feminine);
  • “the whole sphere of behavior related even indirectly to the sexual functions and embracing all affectionate and pleasure-seeking conduct,” or sex as sexuality.62

Webster’s was not alone in this broader array of ordinary meanings that were, in 1964, associated with the term “sex.”63 Can it really be maintained then that “sex” had but one meaning in 1964?

Consider a concrete example. In the 1960s, an increasing number of schools had “sex education” programs. What topics might have been covered in programs teaching about “sex”? According to Judge Sykes’s definition, “sex education” would cover only sex as biology, teaching kids the morphological differences between men and women. But might a sex education course also teach about gender roles, either descriptively or prescriptively? And surely a sex education course might teach about human sexuality—including variation in sexual practices as well as the mechanics and norms relevant to reproductive sex.

In fact, sex education courses of the 1960s offered much more discussion about gender roles and human sexuality than they did about the biological differences between the sexes.64 For example, the sex education curriculum followed by many secondary schools in California asked students to engage in role-playing exercises contrasting authority relations between husbands and wives; high school students tackled “changing male and female roles” by discussing differing interpretations of masculinity, femininity, and differentiated sex roles.65 As part of a ninth-grade unit on “sexual deviations,” California’s sex education students learned that homosexuality “has been known throughout human history and occurs in many societies.”66

The foregoing analysis suggests that the meaning of “sex” in 1964 was not as one-dimensional as Judge Sykes asserted. Judge Sykes also maintained that discrimination “because of sex” must mean something different than discrimination “because of sexual orientation.”67 That is a better argument, but has Judge Sykes posed this question in a neutral manner? Chief Judge Wood posed the textual question another way: What is the relationship of “discriminate because of sex” and “discriminate because of the sex of your spouse/partner”? Is original meaning textualism nothing more than a clever shell game, where each side can secure the “original meaning” it prefers by the way it poses the inquiry?

Stick with Judge Sykes’s manner of posing the question. The honest textualist still needs to explore the exact relationship of “discriminate because of sex” and “discriminate because of sexual orientation.” Do the two types of discrimination have no overlapping application, as Judge Sykes assumed? Might they overlap (Judge Flaum’s view)? Or might one kind of discrimination be a subset of the other (Chief Judge Wood’s perspective)? Overlapping or subsumed terms are common in antidiscrimination law. Title VII bars discrimination because of skin color, nationality, and race—terms that overlap considerably. Many state and municipal laws prohibit discrimination both because of sex and because of pregnancy, even though the latter is widely accepted as a subset of the former.68 Other state statutes, as well as Title VII today, explicitly define “discriminate because of pregnancy” as a subset of “discriminate because of sex.”69

As Judge Flaum argued, it is difficult to believe that an employer who objects to female employees because of the biological sex of their romantic partners is not discriminating at least in part “because of sex” even if the statutory term is understood only to entail sex as biology.70 As amended in 1991, Title VII provides that an employer can violate the law in mixed-motive cases, so long as one significant “motivating factor” is sex, even if “other factors also motivated the practice.”71 This is an example of how statutory interpretation can be highly dynamic. Because of this and other amendments, Title VII means something different today than it did when it was enacted in 1964—based upon new statutory texts adopted by our democratically elected representatives. Thus, even if Judge Sykes were right about the original 1964 meaning of “sex” and its relationship to “sexual orientation,” the 1991 amendment requires her to demonstrate that “sex” is not even a motivating factor when an employer discriminates against an employee because of the sex of her partner (or because of her sex, in light of the partner’s sex). For this reason, Judges Flaum and Ripple emerge as the most faithful agents of a strictly textual meaning for Title VII.

C. Dynamic Title VII

Because she ignored most of the relevant statutory text and imposed arbitrary, unhistorical choices on the text she did analyze, the application of Judge Sykes’s approach is just as “judge-empowering” as Judge Posner’s openly dynamic approach. The challenge for both judges would be to demonstrate that their creative readings of Title VII rested upon legitimate reasoning from legal authority and relevant facts about the world. Setting aside his earlier criticisms of Title VII, Judge Posner argued that judges ought to expand the merit-based workplace to include lesbian and gay employees, based upon current knowledge about these employees.72 Judge Sykes, in contrast, was reluctant to apply the statute to create new liabilities for employers without more explicit guidance from Congress, which has repeatedly failed to act upon bills to bar workplace discrimination based on sexual orientation (and, recently, gender identity).73 Both Judge Posner and Judge Sykes would have written more persuasive legal opinions if they had been attentive to what Congress actually did when it enacted the law in 1964 and when it subsequently amended the law or adopted related statutes. In my view, what Congress actually does—the point of statutory history—is more important than what Congress fails to do—the point of Judge Sykes’s and the Justice Department’s neglected proposals argument. (Indeed, strict textualists are committed to following the original meaning of the words Congress actually enacted—not words that it later failed to enact—and have rejected the idea that a subsequent Congress can tell us the original meaning of a previously enacted statute.74)

Authoritative judicial interpretations of the governing statute are another inevitably dynamic source in statutory interpretation; those interpretations will affect the path taken by the statute. In Hively, neither Judge Posner nor Judge Sykes applied Supreme Court Title VII precedents with notable enthusiasm, an attitude at odds with the hierarchical structure of the federal judiciary and, often, with the rule of law itself.75 Lower court judges are expected to treat Supreme Court precedents seriously—and should typically follow them faithfully and not begrudgingly. While ambiguous Supreme Court precedents will inevitably be applied dynamically by lower court judges,76 these lower court judges have an obligation to justify their views that Supreme Court precedents offer them no clear guidance. In my view, such a justification requires judicial attention to legislative deliberations about the statutory purpose(s).

In the end, I believe that the methodological debate between original meaning (Judge Sykes) and dynamic interpretation (Judge Posner) has little to do with the proper interpretation of Title VII. A statute—like Title VII—that has been authoritatively interpreted, amended by Congress on several occasions, and then reinterpreted is a statute where original meaning itself is a dynamic process and involves updating. The clashing interpretations in Hively cannot be legally evaluated without understanding the formal evolution of Title VII—from its enactment in 1964, through its amendment in 1972, continuing with the congressional override of the Supreme Court’s pregnancy discrimination decision in 1978, referencing the EEOC’s application of Title VII to sexual harassment and the Supreme Court’s and Congress’s ratification of that application, emphasizing the Supreme Court’s landmark decision holding that workplace exclusions based on gender stereotypes can violate the statute, exploring Congress’s ratification and expansion of this norm in the 1991 Amendments, and finally, including the Family and Medical Leave Act in 1993. To that evolution I now turn.

II. the three faces of discrimination “because of sex”: ensuring a merit-based workplace

In surveying the statutory history of Title VII, I shall apply each of the three meanings that “sex” had in 1964 and, in the process, more deeply explore how those meanings are interrelated. The statutory history establishes, as a matter of relatively settled law, that Title VII guarantees individual employees a merit-based workplace where their opportunities will not be impeded by their biological sex (or that of their intimate associates), descriptive or prescriptive gender stereotyping, or sexualized harassment.

The statutory history is highly relevant to all the judges considering the Hively/Evans/Zarda issue. Contrary to Judge Sykes and the Trump Administration, Title VII is not simply class-based legislation, aimed only at employer policies or workplace conditions that disfavor women and favor men, or disfavor blacks and favor whites, or disfavor Catholics and favor Protestants. Instead, as stated by its text and entrenched by its statutory history, Title VII operates as classification-based legislation, aimed at employer policies or workplace conditions that disadvantage any employee because of her or his race, sex, or religion—including the race, sex, or religion of her or his intimate associates. (These classifications have traditionally burdened blacks, women, and Catholics more than whites, men, and Protestants, but the statutory command also protects whites, men, and Protestants against improper discrimination.) The statutory history also makes clear that “because of sex” has a broad meaning that includes gender and sexuality as well as biological sex. Finally, the history supports my intuition that you cannot linguistically or conceptually separate biology, gender, and sexuality when talking about “sex.” Workplace rules that arbitrarily exclude or disable employees because of their race, sex, or religion are suspect under Title VII, even when they apply to whites as well as blacks, men as well as women, Catholics as well as Protestants, and gays as well as straights.

Title VII’s statutory history has created a web of interconnected rules and principles that have been responsive to emerging facts about what undermines the possibility of a merit-based workplace for all employees—the pervasiveness and toxicity of sexual harassment is one example, and the falsity of antigay stereotypes is another. As administrators, judges, and legislators have responded to our evolving understanding of the workplace, they have crafted a series of legal rules and precedents that render the exclusion of LGBT employees from Title VII increasingly anomalous and profoundly unworkable. In light of the Supreme Court’s recent Fourteenth Amendment precedents, the blanket exclusion is constitutionally problematic as well.

The materials that follow provide some interesting arguments that might enrich the approaches of Judges Sykes and Posner. As you read the statutory history, consider the important role Congress has played in updating the statute (an idea that gives Judge Sykes some ammunition), but also the ways in which Title VII jurisprudence has worked itself into a hopeless muddle if interpreted to exclude sexual orientation claims (a practicality that supports Judge Posner). On the whole, Title VII’s statutory history suggests that the opinions of Chief Judge Wood and Judge Rosenbaum, read together, provide a way forward consistent with the rule of law.

A. Sex as Biology: Relational Discrimination

Chief Judge Wood’s formal argumentation assumed that women and men are separate biological sexes and reasoned that discriminating against a woman (like Hively) because she is romantically involved with another woman is discrimination “because of . . . sex.” Recall that two different kinds of arguments flow from this proposition. The first is the comparator argument: if Hively had been a man (the sex-based comparator), she would not have been discriminated against, because Ivy Tech would not have excluded men romantically involved with women. “But for” Professor Hively’s sex, she would not have been excluded from consideration by the college.77

The second kind of argument, which Chief Judge Wood also offered, points to associational discrimination based on Loving v. Virginia78 and, implicitly, McLaughlin v. Florida.79 In the South as late as the 1960s, different-race couples were subject to special penalties for marriage (Loving) and sexual cohabitation (McLaughlin) not applicable to same-race couples. In both cases, the Supreme Court treated the state rules as discrimination because of race.80 Chief Judge Wood read Loving (and I read McLaughlin) for the idea that discrimination because of one’s association with a person of a different race is a cognizable form of discrimination. By analogy to Loving, Hively was discriminated against because of her romantic association with someone of the same sex. As Andrew Koppelman has long argued, “If a business fires Ricky, or if the state prosecutes him, because of his sexual activities with Fred, while these actions would not be taken against Lucy if she did exactly the same things with Fred, then Ricky is being discriminated against because of his sex.”81

Notice that both the comparator point and the analogy to Loving and McLaughlin are standard original meaning arguments: they are both linked to the original text of Title VII as enacted and amended by Congress. Handed down shortly after Congress adopted Title VII, McLaughlin was the leading Supreme Court case on this issue in 1964. Loving was the leading case when Congress expanded Title VII in the 1972 Amendments. (Unlike the Department of Justice and Judge Sykes, I am not relying on circuit court decisions; Supreme Court decisions are the authoritative final interpretations of the Constitution and federal statutes, of which Congress is presumed to be aware and in fact does follow fairly carefully.82) To say that members of the 1964 Congress themselves would not have drawn the conclusions Chief Judge Wood draws from this text is the sort of counterfactual speculation that discredited the “original intent” arguments made against Loving.83 As a matter of original legal meaning, firing a white employee dating a black woman because of prejudice against “racial mixing” would have been discrimination “because of race” in 1964 and in 1972. In parallel fashion, why then would it not be discrimination “because of sex” to fire a female employee because she is dating a woman?

Judge Flaum’s concurring opinion treats the comparator argument and the associational discrimination argument as more or less the same claim.84 I find Chief Judge Wood’s separate analysis useful, but—inspired by Judge Flaum—I should like to link both of her arguments with the concept of relational discrimination. Relational discrimination refers to adverse treatment of an individual because of her relationship to others; relational discrimination because of a regulated trait (race, ethnicity, religion, sex) refers to adverse treatment where that relational trait is the variable that determines who is discriminated against.85 An employer’s discrimination against a white woman married to a black man is relational discrimination because of race—either the race of the woman (white) or of her spouse (black) or, best conceived, the relationship or interaction between the two. Likewise, discrimination against Kimberly Hively because she was romantically attracted to women rather than men was relational discrimination because of sex—Hively’s sex or her partner’s sex or, best, the relationship or interaction between the two.

Notice that relational discrimination disrupts the narrative that Judge Sykes and the Trump Administration’s Department of Justice offer for their view that “because of sex” cannot be read literally (Hively would have been treated differently if she had been a man who loved women) or cannot be read to mean “because of the sex of her partner.” By their account, antigay employment policies or practices do not violate Title VII because they involve “differential treatment of gay and straight employees for men and women alike.”86 Their assumption is that discrimination because of sex must affect men or women in different ways. The text of Title VII, however, is inconsistent with that view; Section 703 does not announce protected “classes” of employees but, instead, sets forth status-based “classifications” (race, sex, etc.) that cannot be the basis for workplace discrimination. A classification-based approach can apply equally to all classes, in contrast to the class-based approach hypothesized by Judge Sykes and the Trump Administration—without citing a single provision of Title VII. For example, it would violate Title VII for a secular business to hire only Roman Catholics, but it would also be a violation for the secular business to refuse to hire employees who say they are religious but who do not attend religious services every week. The latter policy affects Catholics and Protestants the same way, but that does not protect it from the merit-based rules of Title VII. The same point, of course, can be made about Loving and McLaughlin: the state policies affected white people and people of color the same way, but they still represented unconstitutional discrimination because of race.

Consistent with this text and with the analogy to Loving, the original meaning of Title VII is that antimiscegenation employment policies violate Title VII, even though they do not involve “differential treatment” for black and white employees (who are treated alike by the employer who does not tolerate interracial intimacy). The Justice Department and Judge Sykes can escape the Loving analogy only by claiming that discrimination because of sex is less important than or has a different structure than discrimination because of race or discrimination because of religion for that matter. Like their other arguments, they cite no statutory text for this proposition. What is the point of original meaning textualism when its adherents ignore statutory text that cuts against their desired result or argumentation?

Moreover, the Trump Administration and Judge Sykes’s special treatment of sex discrimination is inconsistent with the statutory history of Title VII. After public debate and statutory amendments, Title VII targets sex discrimination with as much force and in the same way it targets race discrimination. That includes relational discrimination because of either sex or race (or religion).

1. The Early History of Title VII and the 1972 Amendments

Administrators and judges were slow to apply Title VII’s sex discrimination bar vigorously, in part because they believed the primary statutory mission to be eradication of race-based discrimination and did not think Congress expected them to dislodge traditional gender roles or, perhaps, to do much about workplace sex discrimination at all.87 Indeed, many considered the sex discrimination amendment to Title VII to have been either a joke or a subterfuge plotted by its sponsor, Representative Howard Smith, a well-known foe of racial integration but in fact also a close ally of the women’s rights movement.88 As scholars have documented, the joke or subterfuge reading of the sex discrimination amendment is greatly exaggerated, at the very least.89 Nonetheless, for some years it had traction as a reason or an excuse for decisionmakers to read the sex discrimination bar to be much narrower than the race discrimination bar. Under such circumstances, the Loving analogy would have been a much less persuasive argument.

After the EEOC initially took the absurd position that employers could openly advertise jobs in a sex-segregated manner, female activists and lawyers from various perspectives created an umbrella group, the National Organization for Women (NOW), specifically to resist the EEOC’s narrow view of sex discrimination.90 More faithful text-based decisions from agencies as well as courts followed that feminist educational and political effort. In its first Title VII sex-discrimination case, Phillips v. Martin Marietta Corporation, decided in 1971, the Supreme Court unanimously overturned an employer policy refusing to employ women (but not men) with pre-school-age children.91 The unanimous Court reasoned that Title VII enforced a merit-based workplace by reference to classifications, not classes. “Section 703 (a) of the Civil Rights Act of 1964 requires that persons of like qualifications be given employment opportunities irrespective of their sex.”92 Moreover, it is apparent that the employer was not discriminating exclusively based upon sex as biology, as 75-80% of those hired were women. Instead, the employer was primarily discriminating because of sex as gender role (women like Ida Phillips should be at home tending the kids) and even sex as sexuality (women with children were presumably sexually active). Thus, from the very first Supreme Court interpretation, discrimination because of sex had a broader meaning than Judge Sykes supposed.

After extensive hearings and floor debate, Congress in the Equal Employment Opportunity Act of 1972 extended Title VII to apply to federal, state, and local government employees and expanded the EEOC’s enforcement authority.93 Because Congress extended Title VII’s application to state employees under the authority of the Fourteenth Amendment, the Supreme Court has ruled that Congress acted to rectify unconstitutional state sex discrimination, such as longstanding rules barring female employees from job opportunities, and surely including the gender-stereotyping rule struck down the year before in Martin Marietta.94 Such an assumption is amply supported by the congressional deliberations, which show that the drafters of the Act agreed with NOW and other groups that the workplace needed to be open to all workers, regardless of sex, and that state and local governments were pervasively violating that norm. The committee reports endorsed the approach to Title VII advanced by NOW and (by 1972) the EEOC: sex discrimination “is no less serious than other prohibited forms of discrimination,” including race discrimination.95 The reports further stated that the EEOC and the courts should make every effort to enforce the sex discrimination bar vigorously.96 Support for progressive views on the statutory sex discrimination ban and a stronger EEOC found voice in both contemporary feminist scholars97 and members of Congress in public deliberation.98 After the 1972 Amendments, which repeatedly equated the evils of sex discrimination with those of race discrimination, the Loving analogy becomes much more cogent. Indeed, 1972 is a key moment for that argument in another way.

The same year that Congress invoked its authority to enforce constitutional equality guarantees in the 1972 Amendments to Title VII, it passed the Equal Rights Amendment (ERA) by overwhelming majorities and sent it to the states for ratification. The ERA would have amended the Constitution to prohibit government discrimination “on account of sex,” similar to the Title VII language. A major concern with the ERA was the associational discrimination point, raised in 1970 congressional testimony by Paul Freund of Harvard Law School. Applying the logic of Loving, Professor Freund argued that for the same reason that denying different-race couples marriage licenses constitutes discrimination because of race (and is unconstitutional under the Fourteenth Amendment), denying same-sex couples marriage licenses would be discrimination because of sex (and thus probably unconstitutional under the ERA).99

Citing Freund’s testimony, ERA opponent Senator Sam Ervin introduced an amendment to the ERA: “[The ERA] shall not apply to any law prohibiting sexual activity between persons of the same sex or the marriage of persons of the same sex.”100 Senator Birch Bayh, the ERA’s floor manager, opposed the Ervin Amendment because “the concern legitimate at first blush dissipates and indeed disappears in toto.”101 The Senate rejected the Ervin Amendment. We cannot know the precise reasons various senators had in mind when they voted against the Ervin Amendment, but the public record does demonstrate that Congress was on notice that the nation’s leading authority on the Constitution (Professor Freund) believed that the legal meaning of discrimination “on account of sex” carried with it the relational discrimination meaning established by Loving. And at the same time Congress was amending Title VII in 1972, it was aware of precise language that could be used to head off the application of Loving to protect same-sex couples and “homosexual” employees.

Like Senator Ervin, ERA ratification opponents, ambivalent state legislators, and many engaged voters agreed with Professor Freund’s argument. Phyllis Schlafly, the founder of STOP ERA, made this argument a key part of her successful campaign to slow down and ultimately prevent the ERA from securing the needed ratification from 38 state legislatures.102 Consider the STOP ERA cartoon below. As you study the cartoon, notice how Mrs. Schlafly understood “on account of sex” mainly in terms of sex as gender and sex as sexuality. Implicitly, the cartoon’s baseline is that sex as biology, sex as gender, and sex as sexuality are normatively interrelated. Fated by biology, women’s role is to bear and rear the marital children, whose well-being is thwarted by a strict rule against sex discrimination, for such a rule would generate “homosexual marriages and adoption” and a culture of “abortion on demand.” The cartoon is roughly contemporaneous with the 1972 Amendments to the 1964 Act. Is it completely clear, as Judge Sykes maintained, that no one would have understood a rule against discrimination “because of sex” to secure rights for gay people?

Due in part to Mrs. Schlafly’s efforts and popular receptiveness to her arguments, state ratifications slowed to a trickle after 1973—and a new ally ensured that STOP ERA would not prevail in the states that had not yet ratified. The Church of Jesus Christ of Latter-day Saints (LDS) came out against the ERA in January 1975, on the eve of a vote in the Utah Legislature (which swiftly rejected the ERA). Like STOP ERA, the LDS Church found the Loving analogy a persuasive reason to reject the ERA.

figure 1.

stop era’s interpretation of discrimination “on account of sex”103


Thus, the LDS leadership expressed grave concern that “passage of the ERA could extend legal protection to same-sex lesbian and homosexual marriages, giving legal sanction to the rearing of children in such homes.”104 Because the ERA barred any state discrimination “on account of sex,” it might bar a state from denying marriage licenses to all-male or all-female couples. LDS opposition through massive grass-roots activism by Mormon congregants was critical to defeating the ERA in the period after January 1975.105

In short, the Loving analogy accepted by the Seventh Circuit in Hively has longstanding historical roots that might have been known to the Congress that adopted the 1964 Act (McLaughlin would be decided later that year) and would surely have been known to the Congress that passed the 1972 Amendments (the same Congress that passed the ERA). In any event, as a matter of statutory interpretation doctrine, the Supreme Court routinely “attributes” to Congress “knowledge” of widely recognized legal parallels and terms of art.106

To be sure, the Supreme Court in the 1970s would not have ruled that Title VII protected gay employees against discharge—not because the Loving analogy was illogical, but because so few LGBT employees were out of the closet at work and the employees who were “out” (or whose gender-bending appearance and behavior made them “stand out”) were widely considered sick, depraved, conspiratorial, disturbing, and even criminal. Thus, the Court could easily have agreed with discriminatory employers that being straight, or at least hiding in the closet (don’t ask, don’t tell), was a bona fide occupational qualification under Title VII.107 At the very least, the issue was not ripe for Supreme Court resolution. Indeed, right before the states started to debate the ERA, the Supreme Court summarily refused to consider the Loving analogy as a constitutional matter in its first same-sex marriage case.108

During the ERA debate of the 1970s, a few lower courts went further. In Smith v. Liberty Mutual Insurance Co., the Fifth Circuit ruled that an employer could discriminate against a man because he was “effeminate.”109 The court rested its reasoning on the premise that Congress only intended to guarantee equal job opportunities to men and women and so did not intend to protect employees whose gender traits did not perfectly match their employers’ expectations.110 In Blum v. Gulf Oil Corp., the Fifth Circuit, in dicta, cited Smith to opine that “discharge for homosexuality is not prohibited by Title VII.”111 A handful of lower court opinions in the 1970s also held that Title VII provided no relief for transgender employees discriminated against because of their asserted sex.112

2. The Pregnancy Discrimination Act of 1978

Another reason the Loving analogy would not have succeeded in the 1970s is that the Supreme Court remained unwilling to apply Title VII’s sex discrimination bar as seriously as it applied the race discrimination bar. In General Electric Co. v. Gilbert,113 the Court interpreted Title VII to allow employers to exclude pregnancy benefits from their health care and disability insurance. Even though the pregnancy exclusion only affected female employees, the Court held that the policy was not disparate treatment because of sex. Central to the Court’s reasoning was its earlier interpretation of the Equal Protection Clause to allow states to exclude pregnancy benefits from state disability plans; in both Gilbert and the earlier constitutional precedent, the Court held that it is not discrimination because of sex unless the policy treats women differently from similarly situated men.114 The Court seemed to believe that there were no male comparators to pregnant women—and the Court explicitly stated that denying pregnancy benefits is not discrimination because of sex, as “traditionally” understood (and as the Court had previously ruled in the constitutional case).115 Because the statute put men and nonpregnant women in one category, and pregnant women in the other, the exclusion was not, strictly speaking, “because of sex.”116 Denying women Title VII protection because only a subgroup of women were disadvantaged, the Court thus took a narrow view of the statutory purpose. The Court also held that the policy was not unlawful disparate impact discrimination, even though any policy with such a strong and exclusive impact on employees of color would surely have been invalidated.117

The Gilbert approach might offer a counter to Chief Judge Wood’s formalist argumentation: Ivy Tech was not discriminating against women as a class, and many people in the 1970s believed that homosexuality was the result of chosen behavior rather than an immutable characteristic of one’s biological sex. There is an echo of this approach in Judge Sykes’s Hively dissent118 and in Judge Pryor’s concurring opinion in Evans119—but this approach is inconsistent with the statute as written in 1964 and as amended in 1972 and 1978.

Responding strongly to Gilbert’s result and reasoning, women’s groups took their case to Congress—which thoroughly repudiated Gilbert in the Pregnancy Discrimination Act of 1978 (PDA).120 The congressional bipartisan super-coalition that supported the PDA repeatedly expressed the view that Gilbert was not only wrongly decided, but also completely misguided in its reasoning and approach to sex discrimination. “By concluding that pregnancy discrimination is not sex discrimination within the meaning of title VII, the Supreme Court disregarded the intent of Congress in enacting title VII. That intent was to protect all individuals from unjust employment discrimination, including pregnant women.”121 The PDA added pregnancy and related medical conditions to Title VII’s definition of “sex” and further provided: “women affected by pregnancy, childbirth, or related medical conditions shall be treated the same for all employment-related purposes, including receipt of benefits under fringe benefit programs, as other persons not so affected but similar in their ability or inability to work.122

The italicized text embodies the regulatory philosophy animating the 1964 Act and the 1972 Amendments, as well as the angry 1978 override. The governing principle is the merit-based workplace, where “ability or inability to work” is the appropriate criterion, and where sex-based exclusions are disallowed or must be justified by the legitimate needs of the business. Reflecting the ability-to-work approach to sex discrimination evident in Congress’s deliberations in 1977-78, Supreme Court decisions since the PDA have generally interpreted Title VII (as amended) to cover any kind of sex-based classification, including those affecting only a small percentage of women or men in the workplace.123

Enactment of the PDA provided the occasion for the Supreme Court to apply Title VII to a matter of relational discrimination. In guidance to employers soon after the PDA took effect, the EEOC opined that health and medical insurance policies could no longer deny pregnancy benefits to spouses of employees, as well as to employees themselves.124 This guidance targeting relational discrimination could have been justified by the same kinds of associational discrimination or comparator arguments that Chief Judge Wood invoked in Hively. Female employees married to men received pregnancy benefits as part of their family health insurance, but male employees married to women did not. As a formal matter, the sex of the employee (the comparator argument) or of the spouse (the associational discrimination argument) is what triggered different treatment. Employers objected that relational discrimination such as this was beyond the coverage of Title VII, even after the PDA.

In Newport News Shipbuilding & Dry Dock Co. v. EEOC,125 the Supreme Court agreed with the EEOC. The Court announced that the PDA had not only overridden the Gilbert result, but had also renounced its reasoning. Ruling that the company’s relational discrimination was “because of sex,” Newport News relied upon the same kind of comparator argument later deployed in Hively. The majority opinion held that the original Title VII (before the PDA), properly interpreted, barred employment practices “treat[ing] a male employee with dependents ‘in a manner which but for that person’s sex would be different.’”126 In other words, Title VII from the beginning had been a legislative endorsement of the merit-based workplace, for the benefit of male as well as female employees. To the extent Gilbert was decided under different premises, it had been repudiated by the PDA. Newport News also conclusively abrogated the Gilbert reasoning that discrimination affecting only a sex-based subgroup is not sex discrimination. As a matter of Title VII doctrine, Newport News reflects the Court’s application of Title VII to relational discrimination and confirms the comparator argument as a valid form of reasoning about whether there is discrimination because of sex.127

Newport News did not explicitly rely on the associational discrimination argument, but lower court decisions since 1975 have all but uniformly interpreted Title VII to regulate discrimination because of the race or ethnicity of one’s intimate associates.128 The Supreme Court, in another statutory context, has also accepted the associational discrimination argument and the Loving principle that race-based discrimination is presumptively illegal even when it equally affects all races (that is, when people with Caucasian as well as African or Asian or Latino backgrounds are all similarly formally affected). Interpreting the Internal Revenue Code, the Supreme Court in Bob Jones University v. United States unanimously treated the university’s bar to different-race marriage and dating to be simple race discrimination.129

Loving’s reasoning ought not be limited to race. If an employer is willing to hire Catholics but not Catholics who marry outside their faith, surely the employer has engaged in discrimination “because of religion,” also generally prohibited by Title VII. The EEOC Compliance Manual notes that “Title VII prohibits discrimination against an individual because s/he is associated with another person of a particular religion. For example, it would be unlawful to discriminate against a Christian because s/he is married to a Muslim.”130 A New York appellate court found an associational discrimination claim based on religion cognizable under the New York State Human Rights Law, under which claims are “analytically identical to claims brought under Title VII.”131 Admittedly, this area of law is undeveloped, perhaps because few employers openly discriminate in this way.132

As consistently applied by the EEOC and lower courts, Title VII applies to race-based and religion-based employer rules or practices that affect all races and all religions equally. Unless Title VII’s bar to sex-based discrimination is analytically different than its bars to race-based and religion-based discrimination, Title VII applies to sex-based rules or practices that affect men and women equally—contrary to the Hively dissent, the Evans majority, and the Justice Department’s position in Zarda. Interestingly, there is a great deal more to be said about this stance, which receives support in some of the state marriage equality cases. Those cases suggest a deeper understanding of judges’ resistance to the sex discrimination argument for gay rights.

3. Constitutional Challenges to Same-Sex Sodomy Laws and Same-Sex Marriage Bars

There is a broader point to be made about the associational discrimination argument. As previous scholars have argued, the Loving analogy has special bite for lesbian, gay, and bisexual employees, because sexual orientation itself is intrinsically relational.133 For gay men, lesbians, straight people, and bisexuals, sexual orientation is relational to another person’s sex—whether oriented toward the same biological sex (gay), the “opposite” biological sex (straight), or both (bisexual). Typically, a lesbian is subject to discrimination not because of her sexual activities (such as oral sex, an activity most Americans enjoy), but because of the biological sex of her partner. Constitutional challenges to laws discriminating against lesbian and gay persons illustrate this point, as well as its complexities.

This phenomenon is apparent in Lawrence v. Texas,134 in which the Supreme Court constitutionally protected two men engaged in private consensual sex from being penalized by the Texas “homosexual conduct law.” The law’s command was simple: “A person commits an offense if he engages in deviate sexual intercourse with another individual of the same sex.”135 Deviate sexual intercourse was defined to include anal sex, oral sex, and sex toys.136 Two persons of different sexes could engage in oral sex without a legal problem; indeed, a gay man and a lesbian (i.e., “homosexuals”) could lawfully engage in consensual oral sex with one another, consistent with the Texas statute.137 Even though the statutory crime was only defined by reference to “sex” and was completely relational (two or more people defined by “sex” have to be acting in concert), the Supreme Court majority opinion assumed that the objects of the statute were “homosexual persons”138 or “homosexuals, lesbians, [and] bisexual[s],” as the majority opinion put it in one passage.139 Justice O’Connor’s concurring opinion said this: “The Texas statute makes homosexuals unequal in the eyes of the law by making particular conduct [i.e., deviate sexual intercourse between persons of the ‘same sex’]—and only that conduct—subject to criminal sanction.”140 That the Justices in the majority reflexively moved back and forth between the sex-based classification and the affected class of “homosexuals” without any explanation suggests how pervasively our society, our language, and our constitutional culture assume that homosexuality is relational and is relational because of sex.

Exactly ten years after Lawrence, the Court struck down Section 3 of the Defense of Marriage Act (DOMA) in United States v. Windsor.141 The challenged provision provided that “the word ‘marriage’ means only a legal union between one man and one woman as husband and wife, and the word ‘spouse’ refers only to a person of the opposite sex who is a husband or a wife.”142 Nowhere does the DOMA text mention discrimination or exclusion because of sexual orientation; the DOMA discriminations and exclusions are solely because of sex (or, more precisely, the sex of the two putative spouses). Yet the Court’s analysis began, and pretty much ended, with the finding that Congress passed the law to express “both moral disapproval of homosexuality, and a moral conviction that heterosexuality better comports with traditional (especially Judeo–Christian) morality.”143 As before, the Court did not require any explanation as to why a sex-based bar should be treated as a law discriminating on the basis of sexual orientation. Even the no-holds-barred dissenting opinion did not challenge this obvious point.

State marriage equality litigation addressed the sex discrimination argument for gay rights more directly. Almost half of states have constitutional ERAs.144 The Supreme Court of Hawaii in Baehr v. Lewin145 famously ruled that the state constitutional bar to discrimination because of sex required the state to demonstrate a compelling justification for its same-sex marriage bar. Hawaii’s Constitutional Convention, convened in 1978, understood that a bar to sex discrimination not only protects women but also protects lesbians and gay men.146 Hawaii’s constitutional framers as well as its justices understood antigay discrimination itself as relational: when an institution discriminates against a lesbian, it is doing so because of the sex of her preferred romantic partner, the object of her desire.

Baehr has been criticized as well as praised, and most judges—including those who have ruled in favor of marriage equality—have been reluctant to address the Loving analogy for gay marriage, even as others have embraced it.147 Some judges have explicitly rejected the argument. In a leading case, California Chief Justice Ron George interpreted the state constitution to invalidate the same-sex marriage exclusion. His opinion for the court applied heightened scrutiny for two reasons: the exclusion of same-sex couples was both a denial of the fundamental right to marry and an exclusion from marital benefits and duties because of a suspect classification, namely, sexual orientation. Although three of the seven justices found the Loving analogy persuasive as applied to sex discrimination as well, the chief justice (and the controlling vote) rejected it explicitly.148 Why should gay rights and marriage equality piggyback on the advances made by women for sex equality?

Chief Justice George’s opinion dovetails with Judge Sykes’s objection to the Loving analogy. The table below explains why lawyers and judges have traditionally not accepted the sex discrimination argument for gay rights: there is a lack of symmetry among the classification, the class that is harmed, and the harmful ideology that the antidiscrimination rule would reject.

table 1.

why some judges do not accept the loving analogy for lgbt rights


Lawyers love symmetry. The foregoing table disturbs them. Nonlawyers like simple, intuitive rules. Occam’s razor cuts against the Loving analogy for many intelligent Americans.

The precepts of symmetry and simplicity underwrite Judge Sykes’s Hively dissent, the Justice Department’s Zarda brief, and, most likely, the Eleventh Circuit’s majority opinion in Evans. The lack of symmetry among the classification, class, and ideology targeted for regulation might explain why a number of bills and statutes have prohibited both sex and sexual orientation (as well as gender identity) discrimination: they are different categories, protect different classes of Americans, and regulate different ideologies (sexism versus homophobia). And, as Judges Sykes and Pryor argue, this also explains why in ordinary parlance sex discrimination is not immediately mentioned when a gay person is being excluded.

Is this a decisive objection to the relational discrimination arguments—and hence to the opinions of Chief Judge Wood and Judge Flaum? Again, the legislative record is instructive. As congressional committees noted in their PDA reports, some states had bars to both sex and pregnancy discrimination.149 Perhaps Gilbert could have been justified, as a matter of text-based interpretation, by this contrast. It is clear from the reports, however, that the PDA Congress did not think this was a valid justification.150

Judges Frank Easterbrook and Kenneth Ripple, no easy votes for a job discrimination plaintiff, joined the Hively majority opinion. So, judges strongly committed to text-based, rule-of-law values can agree with Chief Judge Wood. In such cases, the statutory interpreter needs to dig more deeply. Like Judge Rosenbaum in the Eleventh Circuit and Judge Rovner, who authored the original panel opinion in Hively,151 I also want to think about the issue from another angle, namely, the purpose of Title VII to free workers of gender stereotypes.

B. Sex as Gender: Homophobia as Prescriptive Sex Stereotyping

Although I am inclined to see the issue as the Seventh Circuit majority did, there is room for debate about the application of the statutory language to the facts of Hively and Evans. In such instances, interpreters routinely consult statutory purpose to derive principles that might resolve—or suggest—a statutory ambiguity.152 As with the text, Title VII’s overall purpose—the merit-based workplace with respect to race, color, national origin, religion, and sex—has filled out over time and enjoyed a complicated interplay with constitutional discourse. Title VII has been authoritatively, and repeatedly, interpreted to regulate employer gender stereotyping. Congress has endorsed, in both statutory text and vast legislative deliberations, the notion that discrimination because of sex includes employer policies that impose gender-based norms onto male and female employees alike. The merit-based workplace entrenched by Title VII reflects a statutory purpose to protect workers against being penalized because of gender-based stereotypes. Sex-stereotyping claims are ones that affect male and female employees equally, and this understanding of the statutory purpose provides a way to understand how classification, class, and harmful ideology reconnect in Title VII cases involving LGBT claimants.

1. The 1964 Act and the 1972 Amendments

Because the congressional discussion of the House amendment that added “sex” to the job discrimination title was brief, most judges and scholars have ignored the expressions of purpose in the deliberations underpinning the 1964 Act—but they were abundantly expressed by House members who supported the addition of “sex” to Title VII. The main argument made by representatives against adding “sex” to the law was that prohibiting this form of discrimination would facilitate women’s employment outside the home and thereby undermine the American family. The main argument made by representatives in favor of the addition was that women, in particular, are unfairly denied equal economic opportunities because of widely held stereotypes about women and their capabilities.153 In Martin Marietta, the first Supreme Court decision on Title VII sex discrimination, Justice Marshall’s concurring opinion grounded Title VII in a purpose to outlaw job decisions based on “stereotyped characterizations of the sexes.”154 This is a broader purpose than outlawing job decisions that treat all men and all women differently, for stereotypes will often affect only some and not all women and will often affect both men and women in parallel ways.

Previous scholars have made a useful distinction between descriptive and prescriptive sex-based stereotypes.155 Descriptive stereotypes are assumptions about the different capabilities and limitations supposedly linked with one’s biological sex. If an employer believes most women are emotional and passive and do not possess the same work ethic as men, this descriptive gender stereotype will disadvantage qualified women applying for supervisory positions requiring initiative and long hours.156 Prescriptive stereotypes are preferences about what roles and attitudes ought to be associated with each biological sex. If an employer believes that women ought to be passive and compliant with men’s initiatives, this prescriptive gender stereotype will disadvantage women applying for supervisory positions.157 If Title VII were understood to police prescriptive stereotypes, that would entail a broad understanding of the statutory purpose, well beyond policing policies that treated all men and all women differently.

The original justifications for the sex discrimination provision in Title VII addressed both kinds of stereotyping. Women were disadvantaged in the workplace because employers believed that women’s natural inclination was to be mothers and housekeepers (descriptive stereotyping) and that this was what a woman should be doing (prescriptive stereotyping). For some “helper” occupations such as nursing, men were disadvantaged by stereotypes as well. By its broad text and apparent legislative purpose, Title VII was supposed to make stereotype-based discrimination illegal.158 Qualified women ought to be doctors, and qualified men ought to be nurses. Because stereotyping can affect both sexes, Title VII from the beginning was not just focused on rules that only disadvantaged one sex or the other.

During its deliberations in 1972, Congress’s committees agreed with and expanded upon this broad understanding of the statutory purpose. Arguing for an expansion of Title VII to include local, state, and federal government employers and to empower the EEOC, committee reports understood that sex discrimination remained pervasive in the workplace, in large part because of the traditional stereotype that a woman’s proper role in life was as a wife and mother in the home.159 In other words, a broad anti-stereotype understanding of Title VII was a building block upon which Congress constructed the 1972 Amendments,160 and it was a central justification for Congress exercising its Fourteenth Amendment power to abrogate state immunity from lawsuits for sex discrimination in government workplaces.161

2. The Pregnancy Discrimination Act of 1978 and the Family and Medical Leave Act of 1993

Congress in 1972 emphatically rejected the acceptability of sex-based stereotyping in a meritocratic society—and the Supreme Court followed a similar approach in its equal protection jurisprudence. Notwithstanding the declining fortunes of the ERA, the Court interpreted the Equal Protection Clause to afford heightened scrutiny to sex-based classifications—scrutiny that was generally fatal when the state’s justifications were grounded in traditional stereotypes, even when those stereotypes harmed men. In Craig v. Boren, the leading case, the plaintiff was a man who was disadvantaged by stereotypes of young women as more responsible and less wild than young men.162 The Court ruled that a law allowing eighteen- to twenty-one-year-old women but not eighteen- to twenty-one-year-old men to buy 3.2% beer impermissibly relied upon “outdated misconceptions concerning the role of females in the home rather than in the ‘marketplace and world of ideas.’”163

Craig invalidated a law resting upon descriptive stereotypes that disadvantaged young men (as immature decisionmakers), but Gilbert upheld an employment practice that disadvantaged women based upon both prescriptive and descriptive stereotypes. That is, General Electric and other employers refused to spend more money to cover pregnancies as either insurable events or as disabilities in part because they believed that women were not likely to return to work after they had a child (descriptive), and in part because they believed that the new mother ought to stay at home with her baby (prescriptive). The immediate and powerful feminist indictment of Gilbert rested on the view that descriptive and prescriptive stereotyping based upon pregnancy were self-fulfilling prophecies that formed a deep foundation for workplace inequality because of sex.

Representing the coalition of feminist, religious, civil libertarian, and other groups supporting pregnancy discrimination legislation, Professor Wendy Webster Williams was the lead witness introducing the Gilbert override bill in both House and Senate hearings. She testified that barring pregnancy discrimination was central to the statutory goal of barring descriptive and prescriptive stereotypes from the merit-based workplace. Accordingly, the assumptions “that women would and, in fact, should get married and have children and leave the work force have led to the view that women are marginal workers not deserving of the emoluments of the ‘real’ workers in the work force.”164 In support of the pregnancy legislation, the ABA stated:

Employer pregnancy rules which treat pregnancy differently from other disabilities reflect the generally held stereotype that women will and should marry [men], become pregnant and leave the work force to raise their children. This stereotype has had a serious inhibiting effect on the employment opportunities of women and is not supported by the facts.165

Congress passed the PDA by overwhelming margins, and the committee reports and sponsors’ explanations closely followed Professor Williams’s articulation. For example, the Senate report assailed Gilbert as undermining the “central purpose of the sex discrimination prohibitions of Title VII. As the testimony received by this committee demonstrates, the assumption that women will become pregnant and leave the labor market is at the core of the sex stereotyping resulting in unfavorable disparate treatment of women in the workplace.”166

As a formal matter, the PDA protected pregnant women against various forms of employment discrimination, but Congress’s purpose in enacting the PDA was also to repudiate the form of reasoning the Court followed in Gilbert and to reaffirm and entrench as the central purpose of Title VII the notion that no one should be denied employment opportunities based upon descriptive or prescriptive stereotypes about the capabilities of men and women.167 Like all employees, “pregnant workers [must] be treated the same as other employees on the basis of their ability or inability to work.”168

The PDA was a major congressional rebuke to narrow Title VII constructions based upon judicial speculation about “how far” Congress intended to take antidiscrimination norms regarding sex. Congress’s answer was that the text of the 1964 Act took the country fairly far, and the PDA was merely reaffirming the original meaning of the statutory text, read in light of the anti-stereotyping purpose of the 1964 Congress. As labor economists demonstrated after 1978, however, the PDA was a limited remedy for the pervasive problem of descriptive and prescriptive stereotyping. A common stereotype that held women back in the workplace was the employer view that men are, and should be, more focused on work than on family. The PDA provided no opportunity for new fathers to take care of their infants or otherwise to show a dedication to family traditionally attributed just to mothers.169

Starting in 1985, congressional hearings explored the ways in which men’s inability to take family leave reinforced both descriptive and prescriptive stereotypes of male employees as work-oriented and female employees as family-oriented.170 From the beginning, legislators heard testimony from a variety of sources about the importance of having gender-neutral family leave policies in order to counteract these stereotypes. Law professors and women’s rights advocates emphasized that the lack of family leave standards reflected the idealization of “a male work force with wives performing the traditional and necessary functions in the home,”171 and the “pervasive presumption that women are mothers first, and workers second.”172 Congress also learned that among the employers who did provide some form of family leave, most of them offered this leave only to women, in my view because employers assumed that men would not and should not want to take family leave.173

In 1993, Congress passed and the President signed the Family and Medical Leave Act of 1993 (FMLA),174 which requires employers to give unpaid leave to both male and female employees for family care (and other) purposes. The goal of the statute was to encourage men as well as women to break out of traditional stereotypes. The best strategy for a sex-integrated workforce, at the top end as well as lower down, was to encourage women who wanted or needed to work outside the home to pursue their vocations and to encourage working men to engage in child and elder care inside their families.175

Like Title VII, as amended, the purpose of the FMLA is to undermine disabling stereotypes about men as well as women in the workforce. The 1993 law is remarkable in the salience that it gives to prescriptive stereotypes, based upon employer preferences as well as assumptions about the ways employee family choices played out in the workplace. Some employers might support the norm that working women ought to marry working men, bear their children, and put the family first—but the FMLA makes it illegal for the employer to institutionalize this norm as workplace policy.

3. Hopkins and the 1991 Amendments

Prior to the PDA, the typical Title VII case emphasized descriptive stereotypes that were overbroad (i.e., they applied to many women or many men but not all women or all men). Many of the post-PDA sex discrimination cases reflected stereotypes that were more prescriptive. Indeed, some were entirely prescriptive. For example, the Court held in International Union v. Johnson Controls, Inc.176 that a paternalistic policy barring only female employees of childbearing age from positions where they would be exposed to lead, which affects both the male and female reproductive systems, constituted sex discrimination in violation of Title VII. The Court rejected the Seventh Circuit’s ruling that the employer was responding to legitimate fetal health concerns and not stereotypes, and further held that the discriminatory policy was not a bona fide occupational qualification.177 It was grounded in prescriptive stereotypes and was not merit-based.

The leading case on prescriptive stereotyping is Price Waterhouse v. Hopkins.178Ann Hopkins was allegedly denied partnership not because the employer (the accounting firm Price Waterhouse) did not think women could do a good job, and not because she was unlikely to enrich the firm, but instead because she did not fit the firm’s image of a proper woman. According to some partners, she was too pushy, too masculine, not feminine enough to meet their prescription for a proper female executive.179

In Hopkins, the employer assertedly denied a position based on prescriptive stereotypes alone, rather than descriptive stereotypes or a mix of descriptive and prescriptive. The Supreme Court majority held that an employer decision grounded in prescriptive stereotypes about women (i.e., this is how women should be) constituted discrimination because of sex.180 The plurality opinion responded to the employer’s argument that the evidence of prescriptive sex stereotyping ought not be dispositive:

As for the legal relevance of sex stereotyping, we are beyond the day when an employer could evaluate employees by assuming or insisting that they matched the stereotype associated with their group, for “[i]n forbidding employers to discriminate against individuals because of their sex, Congress intended to strike at the entire spectrum of disparate treatment of men and women resulting from sex stereotypes.”181

If a woman has a claim for prescriptive stereotyping because the employer considers her too masculine, a man ought to have such a claim against an employer who considers him too feminine.182

In the wake of Hopkins, most of the federal courts of appeals have ruled that LGBT employees have a Title VII claim if they adequately allege they were discriminated against because they do not conform to traditional gender roles and stereotypes.183Indeed, the Eleventh Circuit unanimously ruled in Evans that the lesbian employee’s claim for gender discrimination could go to trial. Judge Rosenbaum maintained in dissent that if Jameka Evans had a claim for discrimination because she did not conform to prescriptive stereotypes about proper gender presentation by a woman, the same analysis supports a claim for discrimination because of her sexual orientation, which violates the prescriptive stereotype that women should only find sexual fulfillment through intercourse with (and perhaps marriage to) a man. The deepest violation of entrenched gender roles is a woman’s romantic partnership or marriage to another woman. It is a blatant violation of the core gender role: the gendered requirement that women are not fulfilled unless they find the right man, marry him, and rear his children in their household.184

At the outset, notice that a gender-stereotyping jurisprudence is inconsistent with an understanding of Title VII that limits its prohibitions to practices that affect only male or only female employees. Prescriptive stereotyping often affects both male and female employees—the same kind of symmetry that is reflected in associational discrimination claims following the Loving analogy. Conversely, this symmetry was rejected by Judge Sykes’s Hively dissent and by the Justice Department’s Zarda brief, both of which explicitly limit Title VII to practices that affect only men or only women. As before, please notice that neither Judge Sykes nor the Trump Administration justified such a class-based focus by analysis of the statutory text, which only regulates by classification (rather than by class).

Judge Rovner’s panel opinion in Hively explored the problem appellate courts now face in gender-stereotyping cases brought by gay and lesbian employees. The courts have sought to reconcile Hopkins gender-stereotyping claims with their circuit precedents denying sexual orientation claims “either by disallowing any claims where sexual orientation and gender non-conformity are intertwined (and, for some courts, by not allowing claims from lesbian, gay, or bisexual employees at all), or by trying to tease apart the two claims and focusing only on the gender stereotype allegations.”185

The first approach, simply denying lesbian and gay employees the gender-stereotyping claims that everyone else can make, ought to be off the table for constitutional reasons. Shortly after Hopkins, the Supreme Court overturned an antigay state constitutional amendment in Romer v. Evans.186 The amendment preempted state and local laws and directives specifically barring sexual orientation discrimination, which the Court found problematic, but another feature deepened the equal protection problem with the amendment.

It is a fair, if not necessary, inference from the broad language of the amendment that it deprives gays and lesbians even of the protection of general laws and policies that prohibit arbitrary discrimination in governmental and private settings. . . . If this consequence follows from Amendment 2, as its broad language suggests, it would compound the constitutional difficulties the law creates.187

Without determining exactly how far the state constitutional amendment swept, the Court struck it down:

Central both to the idea of the rule of law and to our own Constitution’s guarantee of equal protection is the principle that government and each of its parts remain open on impartial terms to all who seek its assistance. “Equal protection of the laws is not achieved through indiscriminate imposition of inequalities.”188

Romer recognizes a constitutional background principle admonishing judges against carving LGBT people out of the normal protections of any statutory scheme, including Title VII. The constitutional questions raised by an explicit LGBT carve-out are sufficiently serious to trigger the constitutional avoidance canon the Supreme Court follows in sensitive areas of law.

Like the Eleventh Circuit in Evans, most circuits have tried to tease apart the two claims. Unfortunately, it is hard to differentiate these two types of claims, as Judge Rovner explained:

Discrimination against gay, lesbian, and bisexual employees comes about because their behavior is seen as failing to comply with the quintessential gender stereotype about what men and women ought to do—for example, that men should have romantic and sexual relationships only with women, and women should have romantic and sexual relationships only with men. In this way, almost all discrimination on the basis of sexual orientation can be traced back to some form of discrimination on the basis of gender nonconformity. Gay men face discrimination if they fail to meet expected gender norms by dressing in a manner considered too effeminate for men, by displaying stereotypical feminine mannerisms and behaviors, by having stereotypically feminine interests, or failing to meet the stereotypes of the rough and tumble man. Co-workers and employers discriminate against lesbian women for displaying the parallel stereotypical male characteristics. But even if those employees display no physical or cosmetic signs of their sexual orientation, lesbian women and gay men nevertheless fail to conform to gender norm expectations in their attractions to partners of the same sex. Lesbian women and gay men upend our gender paradigms by their very status—causing us to question and casting into doubt antiquated and anachronistic ideas about what roles men and women should play in their relationships. Who is dominant and who is submissive? Who is charged with earning a living and who makes a home? Who is a father and who a mother? In this way the roots of sexual orientation discrimination and gender discrimination wrap around each other inextricably.189

Illustrating Judge Rovner’s point, the experience of women in the military during the long period of antigay exclusion was that such policies were bad for all female personnel, with straight as well as bisexual and gay women subject to antilesbian witch hunts.190 This reflects another doctrinal point. Laws barring workplace discrimination because of sexual orientation usually include “perceived” sexual orientation. So a straight woman, like Ann Hopkins, might be discriminated against because some supervisors perceived her to be lesbian—and the evidence they would cite, behind closed doors, would be largely the same evidence before the Supreme Court: that she was too butch, needed to act more feminine (i.e., available to men), and should go to charm school (to please the male supervisors).

The tension reflected by these cases calls forth the virtue of common law judging: apply all relevant precedents to new fact situations, and when inconsistencies between precedents become apparent, the judge needs to reconcile the precedents. But if the reconciliation path is unworkable, as Judge Rovner suggested in Hively, the court of appeals needs to narrow or overrule its own precedents, in deference to those of the Supreme Court. This precept, essential to the judicial contribution to the rule of law, inspired Judge Rovner’s suggestion that the Seventh Circuit revisit its sexual orientation precedents and was an important feature of Chief Judge Wood’s opinion for the en banc Court (an opinion Judge Rovner joined).

Conversely, this precept is at odds with the dissenting opinion by Judge Sykes, who sought to avoid conflict by insisting on stare decisis for the circuit precedents and reading Hopkins as narrowly as possible, but for legally weak reasons. Thus, she chided the Hively majority for invoking the Hopkins plurality opinion that she felt is not formally binding on the judiciary.191 That objection strikes me as incorrect. Five and probably six Justices (the four plurality Justices and two Justices concurring in the judgment) agreed that the prescriptive stereotyping alleged by Hopkins constituted sex discrimination in violation of Title VII if she met her burden of proving she was not promoted “because of” that stereotyping. The plurality and the concurring Justices parted company over how the burden of proving “because of” causation should be allocated.192

As suggested by Judges Flaum and Ripple, the congressional aftermath of Hopkins is also significant. The Supreme Court handed down several restrictive statutory workplace discrimination decisions in the 1988 Term (the same Term as Hopkins)—to withering criticism from legal scholars, workplace experts, civil libertarians, the EEOC, and the Department of Justice (regarding some of the decisions).193 Congressional hearings critically examined all of the Court’s recent Title VII decisions, including Hopkins. The sponsors of proposed amendments to Title VII and the witnesses testifying before congressional committees in 1990 and 1991 accepted or endorsed the substantive holding in Hopkins, but wanted to liberalize the burden of proof in mixed-motive cases.194 Section 107 of the Civil Rights Act of 1991 adopted Justice Brennan’s basic rule: once the plaintiff has shown that sex or race discrimination played a significant role in denying job benefits, the burden shifts to the employer to show that it would have made the same decision on legitimate grounds, without the discriminatory considerations. Liberalizing the Brennan approach, however, Congress provided that the employer could still be liable if the discriminatory factor was “a motivating factor,”195 even if not the necessary one, but elsewhere in Title VII limited the relief that could be granted in such mixed-motive cases.196

In 1990-91, Congress heard testimony that Hopkins was substantively correct “in its acknowledgment that evidence of sex stereotyping is legitimate evidence of gender discrimination.”197 Even some of the corporate interests vigorously opposing the 1991 amendments recognized that Ann Hopkins had been wronged because the prescriptive sex stereotyping she suffered was “very impermissible” under Title VII.198 In committee reports, the sponsors were clear that the section aimed at Hopkins only “overrules one aspect of the decision.”199 One committee report emphasized that these amendments would in no way affect Hopkins’s holding that “evidence of sex stereotyping is sufficient to prove gender discrimination.”200 There was also a congressional hearing devoted entirely to the discrimination that women face in the workforce. The subcommittee chair concluded that while women “have the experience and skills for advancement,” they were being held back by “misconceptions and stereotyping” and in particular “outdated stereotypes about what women want in the workplace.”201 This is similar to the discrimination that Ann Hopkins suffered.

The foregoing legislative deliberations enrich our understanding of the text and structure of the congressional override statute. Section 3(4) of the 1991 Amendments announced Congress’s purpose “to respond to recent Supreme Court decisions by expanding the scope of relevant civil rights statutes in order to provide adequate protection to victims of discrimination,” including Ann Hopkins.202 Consistent with that purpose, Congress examined all the recent Supreme Court Title VII decisions (including Hopkins), overrode most of them (including Hopkins on the burden of proof issue), and left the substantive discrimination holding of Hopkins intact. By that point even many employers had accepted the illegitimacy of sex-based stereotypes, including prescriptive stereotypes as illustrated by Hopkins. This text and history are evidence enough that the 1991 Amendments accepted as a building block for future policy the Hopkins holding that prescriptive stereotyping is actionable under Title VII.203 That the supporters of the amendments repeatedly endorsed the substantive holding is icing on a cake already well-frosted.

The Trump Administration’s Justice Department claims that the 1991 Amendments ratified four circuit court decisions that it says held that lesbian and gay employees had no rights under Title VII.204 The Department cites no statutory text that ratified a narrowing of the statutory sex discrimination bar. Moreover, the Department mischaracterizes two of the four cases. In Blum v. Gulf Oil Corp., the Fifth Circuit held that a “Jewish homosexual” could be discharged for using the employer’s telephone during work hours for his private business and then said, in dicta, that “[d]ischarge for homosexuality is not prohibited by Title VII.”205 In Ulane v. Eastern Airlines, Inc., the Seventh Circuit held that a transgender woman had no claim for relief under Title VII.206 Although the court felt that the failure of Congress to pass bills protecting against sexual orientation discrimination was relevant to the claims of a transgender employee,207 the holding of the court was limited to gender identity under Title VII (a stance that has been widely repudiated in the new millennium). Two pre-1991 circuit court decisions did squarely hold that lesbian and gay employees had no rights under Title VII: the Ninth Circuit’s decision in DeSantis v. Pacific Telephone & Telegraph Co.208 and the Eighth Circuit’s decision in Williamson v. A.G. Edwards & Sons.209

The Department of Justice’s claim is not well-grounded in the rule of law. No Supreme Court decision has held that Congress ratified two (or even four) circuit court precedents when it reenacted or revised a relevant statute.210 More important, there was no discussion of DeSantis or Williamson in the 1990-91 congressional deliberations—in contrast to the extended focus on Hopkins. Most important, the Department of Justice’s claim that the 1991 Amendments slammed the Title VII door on gay people is flatly inconsistent with the statutory text. Not only did the 1991 Amendments clearly focus on Hopkins and corrected only its stingy procedural holding, but section 3(4) stated that the purpose of the 1991 Amendments was “to respond to recent decisions of the Supreme Court [not to isolated court of appeals decisions] by expanding the scope [not narrowing the scope] of relevant civil rights statutes in order to provide adequate protection to victims of discrimination [not to provide adequate protection to antigay employers].”211 It is uncharacteristic for the distinguished lawyers of the Department of Justice to make an argument so poorly researched and so weakly justified by reference to statutory text and structure, Title VII’s purpose, and Supreme Court precedent.

Hopkins is also significant because it provides a doctrinal link to a longstanding feminist understanding that a premise of sexism, the regime under which women are subordinated, is compulsory heterosexuality.212 Poet Adrienne Rich famously argued that women are not always naturally oriented toward heterosexuality, nor do they always choose it freely, but are systematically pressed to enter into unequal relationships with men of all kinds by a patriarchal culture replete with “fairy tales, television, films, advertising, popular songs, [and] wedding pageantry.”213 This culture punishes the lesbian experience as socially deviant and, in the past, as criminal. In this culture, to be a woman is to be heterosexual, a prescriptive stance at odds with the integrity of many women. Rich directly addressed the implications of compulsory heterosexuality for lesbians in a workplace where women must endure sexual harassment as part of their employment: “her job depends on her pretending to be not merely heterosexual but a heterosexual woman, in terms of dressing and playing the feminine, deferential role required of ‘real’ women.”214 Et voilà: Ann Hopkins,215 Kimberly Hively, and Jameka Evans.

Consistent with Chief Judge Wood’s relational discrimination arguments in Hively, the poet Rich’s analysis makes clear that prescriptive sexual stereotypes about women contribute to a sexist system the purpose of which is to entrench patriarchy, a system of beliefs and practices that subordinates women and prevents them from having as many choices as men. “A lesbian is perceived as being outside the acceptable, routinized order of things . . . [and] as a threat to the nuclear family, to male dominance and control, to the very heart of sexism.”216 Recall the STOP ERA cartoon above: a strong rule against discrimination “on account of sex” (the ERA language) would entail a purpose to free men and women from traditional gender stereotypes. That rule and that purpose are deeply and not just formally inconsistent with policies excluding lesbian and gay people from jobs and rights.

Now we are in a position to see how relational discrimination because of sex fits into the central purpose of Title VII. If statutory text, structure, and binding precedent are any guide, the purpose of Title VII is not to favor any single group, but to entrench a merit-based workplace that does not tolerate sex-based decision-making—including employer enforcement of prescribed gender roles (Hopkins) and relational discrimination (Newport News). The beneficiaries of the sex discrimination rule are men (Newport News) as well as women (Evans and the PDA)—and gays (Hively) as well as straights (Hopkins).

This analysis also suggests a deeper point about Loving, which ruled that antimiscegenation laws represented unconstitutional racial discrimination. Virginia’s marriage exclusion discriminated equally against both white people (who married people of color) and people of color (who married white people), so the class directly harmed by the different-race marriage bar was “miscegenosexuals,” not people of color generally.217 Thus, if Judge Sykes really understood “discrimination” the way she articulated it in her Hively dissent, she ought to complain that Loving was not entirely logical, because it was invoking a suspect classification (race) to protect a group defined by their sexuality (miscegenosexuals). But no one is going to reject Loving and McLaughlin—universally admired, landmark constitutional precedents that should be the basis for legal reasoning in related arenas.

Indeed, the apparent lack of symmetry between class and classification in Loving is easily resolved—and its resolution suggests a similar path for thinking about the relational discrimination in Hively and Evans. While the class directly harmed by different-race marriage laws is miscegenosexuals, the class indirectly but deeply harmed is all people of color. What makes the latter conclusion possible is the purpose of these statutes (white supremacy), which is at odds with the purpose of equal protection (no race-based castes). Just as the Equal Protection Clause prohibits (relational) race-based laws forcing prescriptive stereotypes onto partnerships between people of different races, Title VII as interpreted in Hopkins and amended in 1991 prohibits (relational) sex-based employer policies forcing prescriptive stereotypes onto partnerships between people of the same sex. While the class directly harmed by policies excluding persons with same-sex romantic feelings is lesbian and gay people, the class indirectly but deeply harmed is all women, because the discrimination reflects an insistence on rigid gender roles that is foundational to patriarchy or sexism, a philosophy that has been particularly harmful to women’s liberties and equal participation. What makes the latter conclusion possible is the purpose of the employer policy (prescribing rigid gender roles) and its clash with a central purpose of Title VII (no prescriptive sex stereotyping).

table 2.

why judges should accept the loving analogy for lgbt rights


Reflect upon the foregoing table in light of the pre-1991 statutory history of Title VII, recounted in Section II.A. From the very beginning, an important argument in favor of barring workplace sex discrimination is that it would free women from employers’ tendency to view them as best suited for domestic duties, namely, keeping house and raising the children borne of heterosexual marriages to men. This was the point of the Supreme Court’s first decision interpreting Title VII’s sex discrimination provision in Martin Marietta. Strongly endorsing this purpose, Congress in 1972 expanded Title VII’s sex discrimination rules to government employers. And the PDA in 1978 was entirely focused on pressing employers away from the stereotype that female employees were not temporary workers, biding their time until they could leave their jobs to raise a family. The FMLA in 1993 complemented Title VII in this larger project of freeing female as well as male employees from such a stereotype. In this way, Hopkins was a natural conclusion of the conversation by which Congress, the EEOC, and the Supreme Court were thinking about Title VII over the years. The substantive holding of Hopkins, in turn, found specific as well as general affirmation in the congressional override law adopted in 1991.

In short, a legally compelling argument for Kimberly Hively and Jameka Evans (as well as Donald Zarda) is a synthesis of the Wood-Flaum argument with the Rosenbaum-Rovner argument. Excluding lesbians and gay men from employment opportunities is, strictly speaking, a discrimination because of the sex of the employee and/or the sex of the employee’s romantic partner/spouse—and the case for discrimination “because of sex” is much strengthened because such discrimination is fundamentally at odds with the central purpose of Title VII (repeatedly reaffirmed, strengthened, and amplified by Congress), namely, to protect employees against employer insistence upon conforming to old-fashioned, rigid gender roles.

C. Sex as Sexuality: Sexual Harassment and the Merit-Based Workplace

Recall Judge Sykes’s argument that when we talk about sex, we are not talking about sexual orientation or even sexuality. She reads the lesbian discrimination cases to be all about sexuality, which she seems to understand as intrinsically different from biological sex and gender. Yet her understanding is hard to defend as a matter of original 1964 meaning, even harder to defend as a matter of original meaning after the 1991 Amendments, and virtually impossible to defend as a matter of ordinary meaning today.

To the contrary, for more than a generation, “sex discrimination” has been synonymous with “sexual harassment.” Although lacking substantive lawmaking authority under Title VII, the EEOC in 1980 promulgated sexual harassment guidelines to implement the sex discrimination bar.218 The guidelines interpreted the statute to prohibit employer tolerance of quid pro quo harassment, where a supervisor demands sexual favors in return for workplace advancement or maintenance, and hostile work environments, where there is pervasive and unwelcome sexual harassment by coworkers. While often reluctant to defer to the EEOC in statutory cases, a unanimous Supreme Court wrote the agency’s sexual harassment guidelines into law in Meritor Savings Bank v. Vinson.219 Justice Rehnquist’s opinion for the Court drew upon and explicitly endorsed “a substantial body of judicial decisions and EEOC precedent holding that Title VII affords employees the right to work in an environment free from discriminatory intimidation, ridicule, and insult.”220 Grounded in the merit-based workplace norm developed in this Feature, this body of judicial and agency precedent had been developed in the context of race-based discrimination, but the Court without dissent held that the precept of the race discrimination cases carried over without diminution to cases of sex discrimination.

The EEOC’s sexual harassment guidelines are a classic example of successful administrative policy entrepreneurship. Not only did the Supreme Court unanimously adopt their basic structure and prohibitions, but the Court subsequently held that Congress implicitly ratified Meritor in the 1991 Amendments.221 The guidelines were a policy whose time had come. For many employees, the biggest failure of our aspiration for a merit-based workplace has been sexual harassment by supervisors and coworkers. Speaking for millions of women in the workplace, Catherine MacKinnon famously objected that the sexualization of the workplace is a primary reason for women’s continuing inequality.222 Although the Chamber of Commerce had dragged its feet on the PDA (and would do so again on the 1991 Amendments), it was in agreement with progressive feminists about sexual harassment: a sexually harassing workplace is, from a business perspective, a major distraction from a productive workforce.223

Title VII’s sexual harassment jurisprudence represents a major development not only in the statute’s evolution, but also in the nation’s small “c” constitutional culture and even in its vocabulary. Professor MacKinnon’s pathbreaking book, which inspired the EEOC’s guidelines, was entitled Sexual Harassment of Working Women: A Case of Sex Discrimination. Even if sex in 1964 did not entail sexuality in ordinary parlance (a proposition I dispute), surely by the 1980s it was commonplace to understand sex discrimination to include sexual harassment.

The phenomenon of sexual harassment has deeply affected our culture’s understanding of workplace dynamics, by revealing both the pervasiveness of sexuality, its relationship to sex as biology and sex as gender, and the threat that this combination poses to the merit-based ideal. The sexualized workplace is one where some employees will be distracted, immobilized, passed over for promotions, or fired for reasons that have nothing to do with their capabilities. While Professor MacKinnon classically articulated the anti-harassment policy to protect women in the workplace, she has also emphatically argued that the anti-harassment policy extends to protect LGBT and straight male employees as well.224 Unlike Judge Sykes’s views, Professor MacKinnon’s views are consistent with the classification-based text of Title VII.

Flowing from Professor MacKinnon’s vision, some of the most dramatic applications of Title VII in the last generation have revealed the interconnection among biological sex, gender role, and sexuality. Gay men sexually assaulting male employees have been held to create employer liability under Title VII,225 and gay men sexually harassed or even assaulted by male coworkers have often been afforded relief under Title VII as well.226

Shortly after the 1991 Amendments, the Supreme Court took a case in which a straight man claimed he was sexually harassed by his male coworkers. Although a case of male-on-male harassment may have been beyond the imagination of the 1964 Congress, a unanimous Supreme Court ruled that this conduct fell within the broad statutory language in Oncale v. Sundowner Offshore Services, Inc.227 Both parties to the appeal agreed that plaintiff Joseph Oncale sufficiently alleged the following conduct on the part of his immediate supervisor: on one occasion, the supervisor placed his penis on the back of Oncale’s head and threatened to anally rape him, as another employee restrained Oncale; on another occasion, the supervisor placed his penis on Oncale’s arm, as another employee restrained Oncale; in the most shocking incident, the supervisor and another employee jumped into the shower with a naked and unconsenting Oncale, and the supervisor inserted a bar of soap between Oncale’s buttocks and threatened to rape him, as the other employee restrained Oncale; in addition to these incidents, the supervisor at least once repeated his threat to anally rape Oncale.228 Oncale repeatedly complained about this sexual harassment, but quit the job when he found no relief from the company.229

The Fifth Circuit held that same-sex (i.e., homosexual) harassment of this sort fell outside the ambit of Title VII, which is limited to harassment of men against women and, probably, women against men. Oncale’s attorneys on appeal argued that the plain meaning of Title VII suggests no such limitation so long as there is discrimination “because of sex,” and that the latter requirement is met whenever an employee is subjected to unwelcome sexual abuse or threats. Represented on appeal by one of the nation’s preeminent Title VII scholars, Sundowner responded that Oncale’s interpretation “would expand Title VII beyond its language and legislative purpose by conflating sex discrimination with sexual orientation discrimination.”230 Sundowner demonstrated that Congress had repeatedly declined to enact proposals to protect against sexual orientation discrimination and harassment in the workplace; indeed, such a proposal was rejected on the floor of the Senate in 1996.231 Leave such an expansion of Title VII to the political process, urged Sundowner.

Sundowner’s arguments had prevailed in the Fifth Circuit, but a unanimous Supreme Court reversed. The Court held that there is nothing in the text or structure of Title VII that would preclude relief for same-sex workplace harassment. In some respects, the opinion for the Court was unusually coy: it discreetly set forth the facts of the case in strikingly vague terms, never used the word “homosexual” to describe the well-pleaded harassment, and ignored Sundowner’s rejected proposals argument. Unpersuaded by Oncale’s argument that sexual harassment per se satisfied the “because of sex” requirement of Title VII, the Court remanded the case to the lower courts to give the plaintiff an opportunity to satisfy that requirement. In dicta (but dicta joined by all nine Justices), the opinion for the Court provided some evidentiary routes available to Oncale on remand, which also illuminate what discrimination because of sex might mean in Hively and Evans.

First, the Court said that Oncale might offer “direct comparative evidence about how the alleged harasser treated members of both sexes in a mixed-sex workplace.”232 Thus, he could argue that, if he had been a female employee, he would not have been subjected to unconsented sexual contact, threats of rape, and anal assault while he was naked in the shower. More on point (because a Title VII lawsuit is filed against the employer, not against fellow employees), Oncale presumably could have argued that a female employee complaining to the company that she was being sexually hazed, subjected to unconsented sexual contact, threatened with rape, and anally assaulted while she was naked in a shower would have been taken seriously, and the employer would have responded in some way. Oncale alleged that his complaints of male-on-male sexual abuse, threats of rape, and anal assault had not been taken seriously by the employer (“boys will be boys”).233

If Oncale had been able to make this kind of showing, he would have established that the company was more intolerant of, and willing to remedy, “heterosexual” assaults (men on a woman) than “homosexual” assaults (men on a man). Even if neither Oncale nor his harassers were self-identified as “gay,” the touching of a male penis on the head and arm of another man, the threats by one man that he intended to anally rape another man, and the anal abuse of a naked man by another man wielding a bar of soap while the first naked man was restrained by a third man, were certainly “homosexual.” In this respect, therefore, employer tolerance of “homosexual” abuse is actionable under Title VII, as interpreted in Oncale.

Joseph Oncale represented himself as straight, though he may have been perceived or denigrated as gay by the other men. Should his case come out differently if Oncale were gay? Surely not. In light of Romer and Lawrence, this reading of Title VII is supported by the avoidance canon, the principle that statutes should be interpreted to avoid serious constitutional difficulties.234 That is, a statute requiring employers to discourage heterosexual (man-on-woman) rape threats, unwelcome penile contact, mock rapes, and other sexual assaults in the workplace, but allowing them to ignore homosexual (man-on-man) rape threats, unwelcome penile contact, mock rapes, and other sexual assaults against gay or straight men, would surely raise equal protection concerns. As described above, Romer stands for the proposition that the state cannot, without some public-regarding purpose, exclude lesbians, gay men, and bisexuals from the ordinary protections of the law.

Second, and relatedly, Oncale would have established discrimination because of sex if he could have shown that his attackers were themselves “homosexuals” wanting to molest him (and the employer tolerated that conduct). Lower courts have ruled that same-sex harassment is actionable if motivated by homosexual lust,235 and the Supreme Court said in Oncale that same-sex harassment by a “homosexual” supervisor would meet the because-of-sex criterion.236 The logic of these rulings is that “but for” the employee’s sex, the sexual harassment would not occur; the gay supervisor would not make sexual advances to a woman. This reasoning can be extended to cases where the harassers are motivated by homophobia, which is a kind of sexual panic aimed at men or women because of the sex of their preferred partners. Psychology studies suggest that at least some homophobic harassment stems from the harasser’s own (often latent) homosexuality.237 In Oncale, the apparently homophobic harassment involved openly homosexual acts by the hazing supervisor and other crew members. And common sense suggests that homophobia is a fear mobilized by same-sex intimacy and/or by a man’s departure from his traditional gender role as inseminator of women.238

Third, Oncale probably would have succeeded if he could have shown that he was attacked and abused because he was gender deviant. The Court’s Oncale opinion did not discuss this avenue for relief, but almost all of the circuit courts that have ruled on such a case have applied Hopkins to support a Title VII claim for LGBT workers harassed because of prescriptive gender stereotyping.239 Indeed, this was the stance taken by the Eleventh Circuit in Evans: even though Jameka Evans did not have a Title VII claim for discrimination because of sexual orientation (said two of the three judges), she did have a claim for discrimination because of gender deviance (said all three judges).

Oncale neither holds nor denies that Title VII provides a remedy for (homo)sexual harassment against gay or bisexual men, but the door for such claims that was opened wide by Hopkins is ripped off its hinges by Oncale. Put somewhat differently, it is harder to deny the force of Chief Judge Wood’s and Judge Rovner’s opinions in Hively or Judge Rosenbaum’s opinion in Evans after reading the thorough exegesis of Oncale provided above.

To appreciate this last point, consider this remarkable dictum at the end of the Court’s opinion in Oncale: “A professional football player’s working environment is not severely or pervasively abusive, for example, if the coach smacks him on the buttocks as he heads onto the field—even if the same behavior would reasonably be experienced as abusive by the coach’s secretary (male or female) back at the office.”240 Penned by the late Justice Scalia, a strict constructionist for antidiscrimination laws, this guidance suggests a dawning appreciation of the complicated, context-dependent interconnections among sexuality, gender, and the law of Title VII. What I now suggest is that the relationship must be more complicated than Justice Scalia apparently realized.

Start with the coach’s interaction with his football player, which Justice Scalia seemed to think is safe from legal liability: Title VII allows male coaches to butt-smack male players. Based upon his earlier analysis of Oncale’s claim, however, even Justice Scalia would have to concede that a variation of his example might be sexual harassment because of sex, that is, if the coach were gay and “intended” the buttocks-smacking to be “homosexual” (i.e., because of the male sex of the butt of the abuse). Even if the coach were straight, the player may still object because he considers the coach’s particular butt-smacking technique to be generally “queer.” Once the player makes clear to the coach that this is unwelcome and disruptive (homo)sexual touching, does the coach not have a duty to lay off his player’s buttocks?

How about this variation: Assume that the player himself is gay and that a homophobic coach regularly smacks his buttocks while uttering lewd antigay remarks and threats of sexual assault, along the lines of the supervisor in Oncale. Does this sexual harassment violate Title VII? Depending on how pervasive, hostile, and unwelcome the comments are, I do not see why the player would not have a potential Title VII claim. If a gay man’s unwelcome touching of a straight man’s buttocks is a potential violation of Title VII, why shouldn’t a straight man’s unwelcome touching of a gay man’s buttocks not be potentially a violation as well? Romer demands some kind of policy explanation for treating the cases differently. (Even without the benefit of the symmetry of legal rights suggested by Romer, the player ought to have a claim if the coach would not have taunted and sexually touched a female because she dated men.)

Justice Scalia intended the example of a coach smacking a jock’s butt to be an “easy case,” where there is no Title VII liability—but even the most elementary understanding of sexuality and gender in the workplace (or on the football field) reveals this to be a more involved scenario.

Now consider the coach’s interaction with his secretary, whom Justice Scalia assumed would have a potential Title VII claim for buttock-smacking. Note that Justice Scalia, writing for the Court, included male as well as female secretaries. If the coach smacks the buttocks of his male secretary and thereby creates a hostile environment for him, I do not see why the secretary’s or the coach’s sexual orientation should make a difference (and Justice Scalia’s opinion did not suggest that it would). Thus, if the coach knows that he is not supposed to smack the buttocks of his female secretary but feels free to manhandle the rear end of his male secretary, this conduct might well be a Title VII violation—regardless of the coach’s or the secretary’s sexual orientation. The key inquiry is whether the sexual contact upends the merit-based workplace because it is unwelcome and disruptive. The male secretary may consider the buttocks-smacking unwelcome and disruptive because it is demeaning or effeminizing, because it is homosexually aggressive, or because it is homophobic—or some combination of all three perceptions.

The Supreme Court’s decision in Oncale also sheds light on the Trump Administration’s interpretation of Title VII. Recall the central argument made in the brief for Sundowner, the defendant employer: Title VII provides no remedy for homosexual harassment, because Title VII prohibits only sex discrimination and not sexual orientation discrimination. Sundowner’s main evidence was that lower courts had so held and that Congress had rejected bills that would have provided relief to gay employees.241 Notice that this is exactly the same argument briefly noted by Judge Sykes in Hively and made in detail by the Department of Justice in Zarda. The Supreme Court’s opinion in Oncale rejected that argument and implicitly but clearly endorsed the application of Title VII to provide a remedy for a great deal of homosexual harassment. Indeed, employer policies that tolerate harassment of gay or bisexual men but provide remedies for harassment of women are suspect under virtually any reading of Oncale.242

When the Supreme Court handed down Oncale, Congress had declined to act on thirty-seven bills, introduced between 1974 and 1996, that would have provided relief under some circumstances for gay or lesbian employees fired or harassed because of their sexuality.243 Even though gender and sexual policing were at the heart of Oncale, neither Justice Scalia, for the Court, nor any other Justice found Sundowner’s argument even worth mentioning. (And recall that this was the main argument made by Sundowner.) The reason is apparent. Although the Warren and Burger Courts sometimes found meaning in congressional refusal to adopt legislation or in rejection of such proposals on the floor of the House or Senate, the Justices were cautious about such evidence, because it is hard to say why Congress does nothing (the usual explanation is inertia).244 Inspired by Justice Scalia’s legisprudence, the more textualist Rehnquist and Roberts Courts have been even more leery of such evidence. For the current Court, rejected proposals have only been considered relevant when Congress enacts a law, considers alternatives, and explicitly rejects those alternatives.245 Under Roberts Court practice, Congress’s explicit rejection of the Ervin Amendment in 1972 (a rejection that supports the Hively rule) is more relevant than Congress’s failure to move forward with proposals to bar sexual orientation discrimination, with a variety of special rules and limits.

The Roberts Court’s caution about legislative inaction is well-illustrated in the context of the Hively issue. Before Oncale, proposals to remedy workplace discrimination or harassment because of “affectional orientation” or “sexual orientation” or “nonmarital status” suffered from an almost complete lack of congressional attention. As Chief Judge Katzmann observed in Christiansen, Congress usually fails to vote on bills or even hold hearings because its agenda is filled with other proposals its members or the relevant committee chairs consider more urgent. Another reason Congress neglects proposals is that they raise a host of collateral issues, as was the case with all of the early bills, including the proposed Civil Rights Amendments Act of 1979, which would have added “affectional or sexual orientation” as prohibited characteristics to federal law barring discrimination in public accommodations, federal programs, employment, housing, and crimes of intimidation.246 (In the 1964 Act, only the employment title barred sex discrimination.)

The 1979 bill was the first to generate congressional hearings. At the beginning of the House subcommittee hearing, the minority counsel noted that the still-pending ERA might protect “homosexuals,” and that the current EEOC Chair Eleanor Holmes Norton viewed Professor Freund’s argument for gay marriage and perhaps gay rights under Title VII as quite plausible.247 Counsel worried that the proposed legislation would lend support to the ERA argument for gay marriage, and the witness (San Francisco politician and later Mayor Art Agnos) very much agreed, to the chagrin of supporters of the legislation (who probably considered the issue a diversion). Critics of the bill expressed concern that “affectional or sexual orientation” was too broad a term (they worried that it would include pedophiles), that homosexual behavior is immoral and should not be encouraged, and that supporters had not demonstrated that the generally wealthy and privileged (white) homosexual population actually suffered discrimination.248 Similar problems bedeviled the next bill, which endured a contentious House subcommittee hearing in 1982.249 The AIDS epidemic effectively killed congressional interest in this legislation for the next dozen years.

In 1994, proponents abandoned the strategy of amending the entire Civil Rights Act of 1964 and focused on creating a new statutory scheme just for employment discrimination: the proposed Employment Non-Discrimination Act (ENDA).250 For the first time, a Senate committee held hearings on the issue in 1994, followed by ENDA hearings in a House committee in 1996 and by a Senate committee in 1997 and 2002.251 In the first full-blown committee report, a Senate committee in 2002 reported ENDA favorably, but minority senators objected that the bill was not ready for a vote because its text was “overly-broad and unclear in many respects.”252

An even more detailed report, following subcommittee hearings, came in 2007, when the House Education and Labor Committee reported ENDA favorably (the House passed the bill, but it was not taken up by the Senate).253 The report noted that by 2007 almost all of the federal circuit courts had interpreted Title VII to reject claims by lesbian and gay employees—but concluded that those lower court decisions should not be treated as authoritative, because they were inconsistent with Hopkins and Oncale.254 Dissenting from the report, ten committee members said that they “have consistently stated their opposition to intentional workplace discrimination. However, H.R. 3685 as reported out of Committee raises many legitimate concerns that remain unresolved.”255 Among their complaints about the draft legislation were the following: its religious exemption was too narrow, the introduction of a claim for discrimination based on “perceived” sexual orientation was confusing and ill-defined, the ERISA exemption provision was controversial, and there was confusion about whether gender identity should be added as an additional grounds for nondiscrimination.256 Additionally, “the bill fails to provide a proper balance with respect to retaliation, unfairly according protections to one class of employees but not others.”257 Both the majority and the minority on the committee agreed that Title VII protects all employees against status-based discrimination and that intentional discrimination against gay employees is not defensible.

At no point did the dissenting members disagree with the Committee’s view that Hopkins and Oncale overrode lower court decisions barring gay and lesbian employees from Title VII relief. Also, notably absent from the minority views were the arguments made in the 1980s that homosexual conduct is immoral and ought not to be encouraged, that homosexuality is a mental defect, or that gay workers are predatory and disruptive; not a single member of the committee claimed that lesbians, gay men, and bisexuals do not belong in the merit-based workplace on the same terms as straight employees. And, as quoted above, all the dissenters said they “have consistently stated their opposition to intentional workplace discrimination,”258 presumably including discrimination against lesbian and gay employees who are capable of doing their jobs.

The most interesting punch line of the foregoing discussion is that the relevant congressional committees were generally aware that an increasing number of lower courts were closing off Title VII to lesbian and gay claimants—at the same time that the Supreme Court was opening up Title VII to claims of gender stereotyping and (homo)sexual harassment. Ironically, Congress publicly acknowledged the tension between the Supreme Court’s precedents and the circuit court precedents before any of the circuit courts did.259

The congressional deliberations and case law teach that Title VII’s meta-purpose is to ensure a merit-based workplace, namely, a productive space free of sexual or gender harassment. Now apply Title VII’s sexual harassment jurisprudence to Hively. If Ivy Tech’s football coach had publicly humiliated Hively by smacking her buttocks and denigrating her for being a lesbian, she probably would have a Title VII sex discrimination claim. So why should Hively not have a Title VII claim if Ivy Tech’s administration refuses to hire her because she is a woman who loves other women? The Loving analogy suggests she has an associational discrimination claim and Hopkins suggests she has a claim grounded in gender stereotypes.

III. title vii’s merit-based workplace under a textualist court and a gridlocked congress

As a matter of the nation’s antidiscrimination policy, our deep dive into Hively, Evans, and Zarda reveals several important conceptual themes. First, Congress, the EEOC, and the Supreme Court have committed the nation to a meritocratic norm of employment evaluation that disfavors the listed personal characteristics, because they are unrelated to merit, and favors the integration, under conditions of equality, of women, racial minorities, religious minorities, and sexual and gender minorities long excluded from and harassed within the workplace. This commitment has long entailed skepticism about employer policies that exclude or penalize employees because of the race or sex of their spouses, romantic partners, or families. This concept of relational discrimination is at the heart of Title VII, and if McLaughlin and Loving are any guide, relational discrimination has been an important target since the enactment (1964) and early expansion (1972) of Title VII. Because sexual orientation is a classic example of relational identity, gay or lesbian employees like Kimberly Hively, Jameka Evans, and Donald Zarda are exemplary of this concept.

Second, cases like these illustrate a distinction that has long been integral to Title VII but was controversial until the 1980s. From the beginning, and deepening over time, the purpose of Title VII has been to encourage employers to make merit-based decisions and to discourage employers from making decisions based upon stereotypes about men and women, as well as workers of different races, ethnicities, or religions. Merit is the antithesis of stereotyping. Often, these stereotypes are descriptive—women are too emotional or weak to do this kind of job—and most sex discrimination lawsuits have attacked generalized views that are often patently false as a general matter or unjustified in particular cases. Increasingly apparent are stereotypes that are prescriptive—women ought to be feminine, cooperative, or married/partnered with men—and many classic sex discrimination cases (such as the pregnancy cases) involve prescriptive stereotypes as a suppressed feature. Hopkins and Johnson Controls illustrate the centrality of prescriptive stereotypes as an object of Title VII. Hively and Evans are recent examples.

Third, some judges say they are reluctant to sexualize Title VII—but that ship sailed decades ago. Judges cannot blink the reality that the workplace is sexualized, perhaps more so than ever before, as it has become more integrated with female workers, employees of color, and openly LGBT employees. The norm against sexual harassment in the workplace—propounded by feminist thinkers, adopted by the EEOC, assimilated into governing precedent by the Supreme Court, and implicitly ratified by Congress—has entrenched sexuality as a core project of Title VII. It has long been clear in American public law (as well as Title VII jurisprudence) that sexuality and gender are inseparable. Hively and Evans illustrate how deeply the fates of women, gay men and lesbians, and transgender people are intertwined in American public law—an underappreciated theme that animates the marriage equality constitutional norm as well.

The foregoing analysis has been backward-looking. Now look forward. What does the future look like for Title VII?

The next half century of Title VII will surely have some important discontinuities with its first fifty years. Start with the next three years, during which the Supreme Court will most likely address this issue: does employer discrimination against gay, lesbian, and bisexual workers amount to discrimination because of sex?260 The Court that will decide the issue will be more textualist and more ideologically splintered than earlier Courts. Congress may or may not remain strongly conservative, but it will probably remain gridlocked and, therefore, incapable of overriding the Court as it frequently did during the first fifty years of civil rights law. The executive branch will be less supportive of the rights of sexual and gender minorities, as illustrated by the Trump Administration’s astonishing amicus brief in Zarda.261

What might the five most textualist Justices (all older men appointed by Republican Presidents), nudged to the right by the Department of Justice and unchecked by Congress, do with Title VII? One path would be that suggested by Judges Sykes and Pryor. Applying “original meaning,” a textualist Court could read the statutory language narrowly, that is, to deny rights to a small group of sexualized and gender-bending employees. This Feature has suggested the conceptual moves such a Court might make: Insist that the ordinary meaning of discrimination because of sex is biology and not gender or sexuality. Articulate Title VII’s purpose as equalizing opportunities for women and men in the workplace, and nothing more. Marginalize relational discrimination claims as exceptional and confine them to the race arena. Redirect the focus away from the 1991 Amendments, object to any reference to the legislative history of those Amendments, and refocus on the whole code, where Congress sometimes bars discrimination because of both sex and sexual orientation. Leave the issue to the legislative process and eschew any judicial authority to radically expand Title VII. To accomplish this analytical coup, the Court would have to ignore a great deal of the statutory history; explain away the broad Title VII text (including text added in 1991); exclude in-depth congressional deliberations; and narrow, reconceptualize, or overrule some of its precedents, such as Newport News, Oncale, and, especially, Hopkins. That Hopkins has been underwritten by Congress renders an attack on that precedent especially questionable as a matter of law and democracy.

To be sure, the Roberts Court has shown itself willing to interpret Title VII stingily, as illustrated by its decisions following and expanding upon Ledbetter v. Goodyear Tire.262 That decision ruled that women denied equal pay are barred from Title VII lawsuits because of the short statutory limitations period, even though the female employees would not reasonably have known about the pay disparity within the limitations period. Although Congress overrode Ledbetter in 2009, the Court has continued to treat it as a “shadow precedent” (Deborah Widiss’s felicitous term), and has essentially confined the override to a narrow space, notwithstanding the congressional deliberations.263 In University of Texas Southwestern Medical Center v. Nassar,264 the Court imposed on plaintiffs in Title VII retaliation cases a high burden of showing causation. The 5-4 majority not only found inapplicable the Civil Rights Act of 1991’s partial override of Hopkins (which did not revise Title VII’s retaliation provision), but invoked, as precedent for a higher burden on plaintiffs, the Hopkins dissenting opinion, which the Court had recently followed when interpreting the age discrimination law.265

Ledbetter and Nassar suggest that as many as five Justices are sometimes willing to read the text of Title VII narrowly and to interpret statutory precedents and even overrides narrowly. It would be a much bigger move, however, for those Justices to attack Hopkins—either by overruling or (more likely) narrowly interpreting that super-precedent. An even more daring, but no longer inconceivable, strategy would be to reconceive antigay employment rules as policies involving disruptive expression or disapproved conduct or both. This argument works best in cases where the employer can conceptualize its discrimination as a matter of its own religious expression. Such Justices can then argue that the employer is certainly not discriminating “because of sex” or even “because of sexual orientation”; at most, the employer is discriminating because of the alarming messages conveyed by an employee’s unfortunate presumed behavior. The Alliance Defending Freedom (ADF) has been running this kind of argument in antigay discrimination cases,266 and Judge Pryor’s concurring opinion in Evans gestures toward this argument as well.267 Note that this conflation of status/identity and choice/behavior finds a direct parallel in segregationists’ insistence two generations ago that God required separation of the races, especially their sexual separation.268 Indeed, the trial court in Loving made pretty much the same argument: Virginia’s antimiscegenation law was not at all discriminatory and was not a status-based law, for it merely implemented God’s Word and regulated immoral behavior.269 (Obviously, the Supreme Court rejected that argument in Loving.) The parallels between antigay tropes and argumentation and long-renounced racist tropes and argumentation ought to give judges pause before they accept them.

What a conservative textualist majority has the power to do is not necessarily what a conservative textualist majority ought to do in a case raising the Hively issue, however. To begin with, many conservatives and most textualists believe in the rule-of-law virtues of predictability, objectivity, and consistency in statutory interpretation—and I have yet to see a serious rule-of-law engagement with the Hively issue that refutes the conclusion reached by Chief Judge Wood. In my opinion, her interpretation is supported by original meaning, the often-reiterated statutory purpose of the merit-based workplace (also explicitly written into the statutory text by the PDA), Supreme Court statutory precedents ratified by Congress, congressional deliberations in 1964 and over time as the law has been amended, practical considerations articulated by the implementing agency and by judges, and the constitutional equality norms announced in Romer and Lawrence.

Any poorly researched effort to reject Chief Judge Wood’s or Judge Flaum’s interpretation based upon original meaning risks exposing original meaning jurisprudence as entirely unconstraining or result-oriented. Scholars have already lampooned original meaning jurisprudence as nothing more than looking out over the crowd and picking out your friends.270 Is this a jurisprudence of judicial restraint? Conversely, the opinions of Chief Judge Wood and Judge Flaum are better exemplars of conservative textualist analysis, a point as to which strong conservative textualist Judges Easterbrook and Ripple seem to agree. If law is to be neutral, and if judges see themselves as following the logic of preexisting law wherever it leads them, Hively is an excellent case for conservative textualist judges to offer some proof that their method is not ideologically slanted.

Consider this thought experiment. Assume a judge who is politically conservative but determined to read statutes to be consistent with the rule of law and the normal operation of the democratic lawmaking process. Call her a “rule-of-law judge.” Such a judge might not be delighted that Title VII protects lesbian employees, would oppose such a policy were she a legislator or an agency head, and would not strain the statute to do so—yet there are strong arguments in support of the Hively interpretation that such a principled jurist ought to find persuasive.

To begin with, a rule-of-law judge would want to ground her interpretation in statutory text. As demonstrated above, the plain meaning of Section 703(a)(1) supports relief. If the allegations of the complaint are true, Ivy Tech discriminated “because of . . . sex” when it refused to consider Hively for a permanent position because of her sex (as a woman attracted to other women) or because of the sex of her partner (female). Would it not be discrimination because of race if Ivy Tech had excluded her because she was a white woman married to a black man? Or a Catholic married to a Muslim? Nothing in the text of Section 703(a)(1) suggests that the race case, the religion case, and the sex case ought to be decided differently. If one wants to be a neutral textualist, one needs to apply plain meanings to reach some results on grounds of consistency that one would not vote for as a legislator.

Second, rule-of-law judges (especially the new textualists) maintain that statutory interpretation is a holistic endeavor: the interpreter needs to examine the text in light of the whole act.271 The Title VII we interpret today is not the 1964 statute, but a statute that has been created over time. The whole act supports relief in a variety of important ways: Title VII presumptively treats race and sex discrimination the same (thereby strengthening the Loving analogy), and when Congress wanted race and sex to be treated differently, the statute says so explicitly in the provision creating a defense for sex (but not race) discrimination if sex is a bona fide occupational qualification, namely, the Weberian, merit-based demands of the job require it.272 The PDA amendment to Title VII explicitly commits the statute to the merit-based workplace norm and implicitly commits the statute to extensive regulation of purely prescriptive stereotypes (a project further advanced by the FMLA). And the 1991 Amendments both confirm Hopkins and expand its allowance of Title VII lawsuits where sex is one “motivating factor” (the point emphasized by Judge Flaum’s Hively opinion). By no stretch do the 1991 Amendments ratify the isolated court of appeals decisions that had read lesbian and gay employees out of Title VII.

Third, and relatedly, a rule-of-law judge respects the operation of the legislative process as a means for our polity to tackle significant problems in a democratically accountable way. For this reason, such a judge ought to take seriously the statutory plan.273 In my view, this is the key area of contest. What exactly is Title VII’s plan? A reading of the statute that limits its aim to employer policies that apply different rules to men and women, without more, is not a plausible understanding of Title VII. It does not track the text of the statute (which focuses on illegal classifications, not treatment of classes of employees), is inconsistent with Title VII’s treatment of race discrimination, and is at odds with Section 703(m). A better reading of the statutory plan links it to the merit-based workplace, freed of sex-based criteria. As Section 701(k) says, pregnant women—like all other women and men—must be treated the same as other employees “similar in their ability or inability to work.”274

While some rule-of-law judges are famously suspicious of legislative history, the most famous textualist in American history (i.e., Justice Scalia) was willing to consult it, for the same reasons he freely consulted The Federalist Papers, to understand how text was understood by contemporaries and to discern the statutory plan.275 It is hard to read the legislative history of the 1964 Act, the 1972 Amendments, the 1978 PDA, and the 1991 Amendments without appreciating the overwhelming congressional consensus that Congress meant what it said in Section 703: Title VII has a broad, ambitious plan, and that plan is to entrench a merit-based workplace that is free of stereotype-laden employer policies, harassing conduct, and non-meritocratic exclusions. It would require a relatively headstrong judge, determined to read her own views into the statute, to ignore this overwhelming history and the democratic legitimation it bestows on such a reading of Title VII’s plan.

Fourth, a rule-of-law judge ought to consider binding statutory precedents and ought to give them a reading consistent with the statutory plan.276 As suggested in this Feature, the Supreme Court’s precedents in Newport News, Hopkins, and Oncale strongly support Hively’s position. It is possible to read these decisions more narrowly than I have read them, but few lower court judges have done so, and a narrow reading would be strongly inconsistent with the statutory history of Title VII. Indeed, Congress itself closely examined Hopkins, heard overwhelming support for the substantive, prescriptive stereotype portion of that precedent, and overrode the burden of proof portion of that precedent. Textualists and pragmatists alike have said that the Supreme Court should follow its own statutory precedents, and have an especially keen duty to apply them faithfully when Congress has deliberated about them and implicitly ratified them.277

Fifth, a rule-of-law judge ought to consider the regulatory history of a statute, namely, its application by administrators over time.278 That the EEOC has championed the broad statutory plan (i.e., the merit-based workplace) in its pregnancy guidance, its sexual harassment guidelines, and its strong and sustained opposition to prescriptive as well as descriptive stereotypes in the workplace is just as significant as its relatively recent view that discrimination against LGBT employees is discrimination because of sex. This regulatory history is significant because it reflects a longstanding body of experience with Title VII that employers have accepted and relied on to shape their policies. Industry reliance or acquiescence in a longstanding regulatory history is not the equivalent to statutory text and precedent in the rule-of-law universe, but a rule-of-law judge ought to be fairly certain the text is clear before she ignores regulatory history.

Sixth, a rule-of-law judge ought to interpret statutes to avoid conflict with the Constitution.279 The original meaning of the Equal Protection Clause was to abolish “class legislation,” laws privileging the few or excluding particular groups from the protections and benefits of the law.280 As the Cato Institute, Professor Steven Calabresi, and other principled constitutional conservatives have argued, such a constitutional original meaning militates against exclusions of LGBT persons from general protections of the law.281 While this point remains controversial (but in a diminishing number of communities), principled conservatives ought to recognize LGBT persons as equal citizens.

An interpretation of Title VII that protects straight employees who happen to be gender-deviant (Hopkins) or whose buttocks have been smacked by gay supervisors or other employees (dicta in Oncale), while at the same time refusing to protect gay employees who are gender-deviant (Hively) or whose faces have been smacked by homophobic supervisors or other employees (Rene), is a Title VII regime at odds with the original meaning of the Equal Protection Clause. That Romer and Lawrence have applied the Equal Protection Clause to protect against this kind of special exclusion of gay people is a further rule-of-law reason to follow Chief Judge Wood and Judge Flaum.282

Another kind of conservative judicial philosophy (associated with legal process theory) admits that strict legal analysis does not always produce clear answers and that hard cases call forth responses that consider institutionalist values.283 The judiciary is one of the three branches of the federal government—the “least dangerous branch,” according to the Framers of the Constitution.284 If the Hively issue were a close call on the merits, a conservative institutionalist such as the Chief Justice might worry about the effect of different rulings on the judiciary’s political capital—the confidence that We the People have in judges’ ability to deliver a rule of law that makes sense to them.285 Although a hard core continues to favor job discrimination against sexual and gender minorities, at least in some circumstances, overwhelming majorities of Americans believe that job discrimination based on sexual orientation is inconsistent with the merit-based workplace.286 Indeed, in my experience, many Americans are astounded that Title VII does not already protect lesbian, gay, and bisexual employees.

Another conservative institutional norm is what Anita Krishnakumar calls the “anti-messiness principle.”287 When the Supreme Court decides statutory cases, it usually sets forth a rule of law that the lower courts will implement. The anti-messiness principle suggests that the Court will and ought to consider which point of law would be easiest for lower courts to administer in a predictable and consistent manner. Judge Rovner’s panel opinion in Hively (extensively quoted above) lays out a persuasive administrability case for the Hively en banc point of law.288 The current regime, where most circuits say that gay and lesbian employees have no claim for relief under Title VII as “homosexuals” but have a potential claim as Hopkins-like gender-benders, is very hard to administer, for the reasons laid out by Judge Rovner. Unless the Supreme Court completely overrules Hopkins, the lower courts will continue to struggle with the vexing issue of figuring out when a gay employee’s case includes enough gender stereotyping to fit under Hopkins or whether it really just involves sexual orientation discrimination, and so falls under a Title VII embargo (if the Supreme Court rejects Hively). The simplest approach, consistent with the rule of law and precedent, for the Supreme Court to take would be to include relational discrimination within Title VII’s bar to discrimination “because of sex.” To be sure, a conservative legal-process judge such as the Chief Justice would not find administrability decisive if the statutory text, plan, and precedents cut only one way—but in close cases or cases where the Justices struggle to articulate the precise rule of law, administrability should be and often is an important consideration.289

Consider a final way some conservative judges might think about these issues: Judge Posner’s concurring opinion. Shorn of his impatience with close textual analysis or governing precedent and stripped of his provocative “judicial updating” rhetoric, Judge Posner’s concurring opinion is best read as an ongoing implementation of the merit-based workplace norm protected by Title VII. Just as the Sherman Act’s open-textured prohibition of restraints of trade and monopolization (terms left undefined) authorizes judges to apply the statute to carry forth its liberal purpose (consumer welfare), so Judge Posner argues that Title VII’s open-textured prohibition of discrimination because of sex (terms left largely undefined) authorizes judges to apply the statute to carry forth its liberal purpose, namely, a Weberian workplace where merit-based performance and not status-based characteristics (race, sex, religion) are determinative.290

Judge Sykes urges caution before treating Title VII as a “common law statute” just like the Sherman Act; I agree with her that unguided judicial updating is not the norm, especially when Congress has revisited the statute as often it did Title VII between 1964 and 1991. But the analysis in this Feature suggests that a faithful attention to the textual and legislative evolution of the law creates a more legitimate foundation for statutory updating. Like most civil rights groups, many and perhaps most employers will agree with Judge Posner when this issue is considered by the Supreme Court: workplace efficiency as well as fairness are undermined by tolerating gendered stereotypes, whether they hold back all women or men, or just some men or women, or lesbian and gay employees, or transgender workers.

Most appealing for Judge Posner’s perspective is that statutes have got to evolve to meet the social or economic necessities that called them forth in the first place.291 For most of its history, Title VII has not really been a common law statute, as Congress and the EEOC have aggressively participated in its evolution. Since 1991, Congress has amended Title VII in one relatively minor way (the Ledbetter override), and the EEOC has more often seen its suggestions dismissed by the Solicitor General and by the Supreme Court.292 In that sense, Judge Posner is right: since 1991, Title VII has become a common law statute like the Sherman Act (which has seen no significant congressional amendment for a longer time). Like the Sherman Act, Title VII has generated thoughtful Supreme Court precedents that have applied the statute and its underlying purpose to address new social and workplace issues that were not salient in 1964 or 1972. From a Posnerian perspective—which is a libertarian conservative one—the simple efficiency aspiration of the Sherman Act (consumer welfare) finds a parallel in the simple efficiency aspiration of Title VII (the merit-based workplace).

The merit-based workplace, like Title VII, is a work in progress. The debate among court of appeals judges in Hively and Evans has enriched this work. My hope is that a deeper examination of the legal and constitutional sources relevant to the issue of lesbian, gay, and bisexual persons in the workplace will contribute to the Supreme Court’s examination as well.