Notre Dame Law Review

View PDF

Andrew Koppelman*

In Lawrence v. Texas,¹  the Supreme Court invalidated a Texas statute criminalizing homosexual sex.²   As it happened, there was no litigated dispute about the underlying facts in that case, which allegedly involved sexual conduct between two men.³   But suppose that one of the parties had demanded a trial.  One of the elements of the crime that the prosecution would have been obligated to prove was the sex of the defendant.  He could only be guilty if he was male.

Suppose, further, that Lawrence challenged the statute on the grounds that it discriminated on the basis of sex in violation of the Fourteenth Amendment.⁴   The legal authority for such a challenge was clear.  The Supreme Court had said decades earlier that “the party seeking to uphold a statute that classifies individuals on the basis of their gender must carry the burden of showing an ‘exceedingly persuasive justification’ for the classification.”⁵   “The burden of justification is demanding and it rests entirely on the State.”⁶ 

But does the statute classify individuals on the basis of their sex?  The reasoning proposed by Mitchell Berman and Guha Krishnamurthi entails that there was no sex-based classification in the Texas statute, which classifies on the basis of homosexuality, not sex.  Classifications on the basis of homosexuality treat men and women equally.  (Their argument pertains to classifications by discriminatory employers, but the logic must be the same when assessing criminal statutes.)  A woman could be in prison because she had done something that only men are allowed to do, and she still would not be punished because of her sex.

The confusion that leads them to this weird position is complex.  Here I will seek to unpack it, and incidentally to show (again)⁷  the correctness of Bostock v. Clayton County,⁸  in which the Supreme Court held that Title VII of the Civil Rights Act of 1964 prohibits employment discrimination against lesbian, gay, bisexual, and transgender (LGBT) people.

In Bostock Was Bogus: Textualism, Pluralism, and Title VII,⁹  Berman and Krishnamurthi argue that the ordinary meaning of Title VII’s ban on discrimination “because of” an employee’s “sex” does not cover sexual-orientation discrimination.  In particular, and most originally, they claim that the Court’s reasoning “depends upon a fatally flawed application of the ‘but-for’ test for causation, one that flouts bedrock principles of counterfactual reasoning.”¹⁰   The appropriate approach, they claim, is their Principle of Conservation in Motivational Analysis, which would disregard facts that are not likely, or less likely, to have been among the actor’s motivating reasons.¹¹   The Principle is relevant, they think, because a discriminatory employer is not motivated by the employee’s sex, but rather by the employee’s sexual orientation.¹² 

The Bostock Court adopted an argument I’ve been making for years,¹³  and that I pressed upon it in an amicus brief: that discrimination against gay people is necessarily sex discrimination.¹⁴   I defended Justice Neil Gorsuch’s opinion for the Court in my article, Bostock, LGBT Discrimination, and the Subtractive Moves, which catalogues various common but unsuccessful strategies for evading the force of the sex discrimination argument.¹⁵   That piece, originally drafted before the Supreme Court’s decision as a critique of arguments by Court of Appeals judges,¹⁶  was easy to revise and update.  The dissenters, Justices Samuel Alito (joined by Clarence Thomas) and Brett Kavanaugh, mostly repeated arguments that I had addressed in my earlier version.¹⁷   All I needed to do was note that they made the same mistakes as the lower court judges.

Berman and Krishnamurthi respond to me as well as to Gorsuch.  They, too, make errors that I have already catalogued, though they combine them in novel ways.¹⁸   Both are major scholars.  Berman is one of the smartest constitutional theorists writing today, and a skillful deflater of bad arguments.  To take only one example, his article, Originalism Is Bunk,¹⁹  remains the single most devastating critique of that school of thought.  Until now, I cannot recall ever disagreeing with him about anything.  Guha Krishnamurthi’s earlier coauthored article, Bostock and Conceptual Causation,²⁰  is an insightful analysis of the but-for causation issue that is key to that decision.²¹   Since they are unpersuaded, I must not have been clear, so I will use this occasion to restate the argument.²² 

Part I of this response explains why their clarification of the causation question casts no doubt on Bostock.  Part II exposes the error of taking the linguistic happenstance of a separate term for gender-atypical behavior—here, “homosexuality”—to subtract those whom the term describes from the statute’s protection.  Part III takes up the fallacy that parallel conjunctions of discriminations balance out, so that there is no violation if male and female homosexuals are both discriminated against.  The fallacy is an old one: it was deployed in 1883 to uphold prohibitions of interracial sex, in a decision that was overruled in 1964, the same year the Civil Rights Act was passed.  Part IV critiques Berman and Krishnamurthi’s deployment of their proposed Principle of Conservation in Motivational Analysis.  The Principle is no objection to Bostock, because in LGBT discrimination, the victim’s sex is always among the actor’s motivating reasons (or reasons delegated to a subordinate, for which the actor is responsible).  Part V addresses their claim that discrimination law should target practices that reinforce racial and gender hierarchy.  I agree with that claim, and with their rejection of textualism.  But given the Court’s commitment to textualism, Bostock is correctly reasoned.

I.      CAUSES AND MOTIVATIONS

Berman and Krishnamurthi challenge as “eccentric” my claim that the plain language of Title VII covers sexual orientation.²³   They do not however refute my most straightforward argument for that conclusion, which appears in the third sentence of my article: “It is not possible to discriminate on these bases without treating a person worse because of their sex.”²⁴   I state the argument in simple form, thus:

An actor discriminates on the basis of trait T if its decision depends on its determination in specific cases whether T is present.  Consequences turn on the presence or absence of T.  That is what it means to classify.  And if bad consequences turn on the presence or absence of T, if you treat someone worse than you would otherwise because they have trait T, then you discriminate against them on the basis of T.²⁵ 

Note that these terms are normatively inert.  They do not specifically concern wrongful discrimination.  They do not depend on the specification of T: the logic is the same whether T is race or sex.

Gorsuch, who wrote the Court’s opinion, and the dissenters, Kavanaugh and Alito, are all adherents of the “New Textualism,” the theory that laws should be interpreted only on the basis of a statute’s text and not extratextually derived purposes.²⁶   New Textualists aim, as Justice Antonin Scalia explained, to derive interpretation of statutes from their words alone, and to ignore unenacted context such as legislative history: “it is ultimately the provisions of our laws rather than the principal concerns of our legislators by which we are governed.”²⁷   Gorsuch offers a straightforward New Textualist (hereinafter I will just call it textualist) reading of the statute.  Its words prohibit discrimination because of sex, and discrimination against LGBT people always involves that kind of discrimination.

I have doubts about textualism, and so do Berman and Krishnamurthi.²⁸   However, in our dispute about Bostock, we all stipulate the textualist method.  We disagree about its implications in this case.

So here is the text in question.  Title VII of the Civil Rights Act of 1964 provides, in relevant part:

It shall be an unlawful employment practice for an employer . . . to fail or refuse to hire or to discharge . . . or otherwise to discriminate against any individual with respect to his [or her] compensation, terms, conditions, or privileges of employment, because of such individual’s race, color, religion, sex, or national origin . . . .²⁹ 

What does it mean to discriminate “because of . . . sex”?  The statute explains that an employer has engaged in “[i]mpermissible consideration of . . . sex . . . in employment practices” when “sex . . . was a motivating factor for any employment practice,” irrespective of whether the employer was also motivated by “other factors.”³⁰   The statute is concerned with motives, not causes.  So it is motivation that a court is looking for when it asks “whether the evidence shows ‘treatment of a person in a manner which but for that person’s sex would be different.’”³¹ 

Justice Gorsuch understands that the statute is concerned with motivation.  He writes that “the ordinary meaning of ‘because of’ is ‘by reason of’ or ‘on account of.’”³²   Berman and Krishnamurthi think however that Justice Gorsuch invites confusion when he writes: “In the language of law, this means that Title VII’s ‘because of’ test incorporates the ‘simple’ and ‘traditional’ standard of but-for causation.  That form of causation is established whenever a particular outcome would not have happened ‘but for’ the purported cause.”³³ 

Berman and Krishnamurthi accurately observe that many facts can be among the causes of an action without being among the agent’s reasons for action: “a fact or event can be a ‘but-for cause’ of some agent’s doing something without it being the case that the agent did that thing ‘because of’ that fact or event.”³⁴   Causation is a relation between physical objects in time.  The statute, however, is concerned not with causation, but the motives of intentional actors.  Actions have many causes that are not part of the actor’s motive.  The fact that you were born is a but-for cause of, but is not your reason for, reading this.

The potential confusion is however harmless, because Justice Gorsuch is right that sex is a necessary part of the motivation for, not merely a but-for cause of, any possible LGBT discrimination.  The discriminator must act on account of the sex of the person discriminated against.

An employee who dates women is “homosexual” only if that employee is female.³⁵   Discrimination against “homosexuals” must intentionally target individuals on the basis of their sex.  Justice Gorsuch thus properly concluded: “An employer who fires an individual for being homosexual or transgender fires that person for traits or actions it would not have questioned in members of a different sex.  Sex plays a necessary and undisguisable role in the decision, exactly what Title VII forbids.”³⁶   Justice Gorsuch is not saying merely that sex is a cause of the decision to discriminate.  He is saying that sex is a reason for the decision.

Berman and Krishnamurthi distinguish reasons from causes because, as we will now consider, they do not think that sex is a motivating factor in acts of discrimination against gay people.  They think that it is merely an antecedent cause of such acts of discrimination.  Since they are mistaken about that, the distinction between reasons and causes is irrelevant to the assessment of Bostock.

II.      LINGUISTIC HAPPENSTANCE

Berman and Krishnamurthi write that sex discrimination “does not cover discrimination taken by reason of a person’s sexual orientation as a matter of ordinary meaning or common parlance.”³⁷   They quote with approval Judge Sykes: “Classifying people by sexual orientation is different than classifying them by sex.  The two traits are categorically distinct and widely recognized as such.”³⁸ 

But I don’t dispute this.  This is one of the subtractive moves that I critique in my earlier article.

This subtractive move . . . is available in any novel sex discrimination case: one could make it about “persons sexually harassed at work” or “persons discriminated against based on gender stereotypes.”  The difference is that these do not have common colloquial terms that refer to them, while “homosexuals” do.  But the linguistic happenstance that such a term exists, that there are “other social categories,” does not mean that “homosexuals” are excluded from the statute’s coverage, or that discrimination against them is not sex discrimination. . . .  In fact, even when there is such a familiar term, it is well settled that Title VII nonetheless applies, for example to discrimination against “mothers.”³⁹ 

Berman and Krishnamurthi’s invocation of “common parlance” turns on this happenstance.⁴⁰   But, although they make this point, it turns out not to be a weight-bearing pillar in the overall structure of their argument, as I’ll shortly explain.  It does, however, play a role in leading them to endorse the parallel discriminations move that I discuss in Part III.

The linguistic happenstance in question is no happenstance at all.  It is a common and predictable consequence of the precise ascriptive hierarchies, deeply embedded in the culture, that antidiscrimination law aims to disrupt.⁴¹   When there is such a hierarchy, there will be people on the bottom who resist their place there.  There will be labels for such people.  Identifying and punishing such people is one of the prime mechanisms by which the hierarchies reproduce themselves.  If the availability of such labels could defeat a discrimination claim, then it is not clear what would be left of Title VII.

Suppose an employer declares that he has no objection to hiring African Americans.  He just doesn’t want employees who are “uppity.”⁴²   The quality of being uppity turns out, upon analysis, to pertain to any African American who aspires to a nonsubordinate position.  The adjective is commonly conjoined with a vile racial slur, but it retains its racist connotation standing alone.⁴³   Deference to an employer’s discrimination against the “uppity” would generate a workforce stratified by race, with African Americans at the bottom.

In Price Waterhouse v. Hopkins, the plaintiff was denied a partnership because her hard-charging demeanor, which was valued and rewarded in male employees, made her male colleagues uncomfortable.⁴⁴   She did not act as a woman should.  The Supreme Court held the company liable: “an employer who acts on the basis of a belief that a woman cannot be aggressive, or that she must not be, has acted on the basis of gender.”⁴⁵   But the defendant could say that it has no objection to employing women.  It simply does not want an employee, gender neutral, who happens to be a virago, harridan, termagant, harpy, or bitch.

Berman and Krishnamurthi address this subtractive move when they take up the question of discrimination against “mothers.”  The Court held in 1971, in Phillips v. Martin Marietta, that this was sex discrimination.⁴⁶   Gorsuch thought the case was analogous.  They respond:

Phillips involves employment discrimination on a basis—being a mother—that is a true subset of one sex (or gender).  It raises the question whether adverse treatment of an individual on account of a property that only women possess counts as discrimination “because of” that individual’s sex when it is not a property that all women possess.  The Court answered that it does.  We think that answer probably correct, even as a textualist matter. (And if it wasn’t, then so much the worse for textualism.)  But the property of being gay (or of being straight) is not similarly a property that only, though not all, women (or men) possess.⁴⁷ 

This (curiously tentative) move depends on the happenstance that the employer is relying on a gender-specific label to implement its policy.  Suppose that the employer refuses to employ “nontraditional parents of young children.”  That would in practice be the functional equivalent of the policy in Phillips, but this category is not a property that only, though not all, women (or men) possess.  This same shift in labels would be available to defeat any discrimination claim.  Instead of “uppity,” the employer could discriminate against “employees who do not follow the patterns of authority traditionally associated with their race.”⁴⁸   As it happens, there is no single word available to encapsulate that idea, but is a valid antidiscrimination claim contingent upon that?

In 1964, the notion that discrimination against mothers was not sex discrimination “as a matter of ordinary meaning or common parlance”⁴⁹  would have been so familiar to many people as to be unnecessary to spell out.  It would “fully comport[] with our linguistic intuitions.”⁵⁰   Many thought that women as a general matter should have equal opportunities at work, but that of course mothers of young children were different and should not be employed.⁵¹   When the Fifth Circuit considered Phillips, it held that such discrimination was not sex discrimination (the decision the Supreme Court reversed), explaining that the statute must permit “consideration of the differences between the normal relationships of working fathers and working mothers to their pre-school age children,” citing the “common experience . . . of mankind in general.”⁵² 

One way of discerning the public meaning of a statute, pressed at some points in Justice Alito’s dissent in Bostock, is to take as a source of law the entire background culture at the time the law was enacted.⁵³   That is essentially what the Fifth Circuit did.  It aimed to imagine and reconstruct what “a group of average Americans” would think the statute meant.  Any set of claimants might have been denied protection if such average Americans were (according to judicial speculation, since courts cannot compile poll data) uninterested in protecting them.

This approach demands that the interpreter deploy the technique of Method Acting, pioneered by Constantin Stanislavski.⁵⁴   That is a problem for the textualist.  The Stanislavski method demands that the actor supplement any ambiguities in the text of the play (there are always ambiguities), and thereby recreate an entire world in which it makes sense for the character to say and do these things at this point in the story—a task that sometimes involves detailed historical research into the context in which the play’s events occur.  The Stanislavski method requires imagination and creativity, just what textualism is trying to keep out of statutory interpretation.⁵⁵   It does not purport to constrain the interpreter.  On the contrary, it invites the interpreter to invent, but in a way that is consistent with the text.  If textualism means anything, it means reading the text while determinedly ignoring certain aspects of the context in which it was enacted, such as its legislative history.  It would be strange to rule out consideration of the precise problem that the legislature was addressing, while taking into account other cultural facts that are more distant from the law’s enactment.⁵⁶ 

The effect of Justice Alito’s approach, to the extent that it is determinate, would be to make “the societal norms of the day”⁵⁷  constrain the operation of antidiscrimination law.  This is the opposite of textualism: it draws on the entire universe of extratextual sources, and in particular the prejudices that were common at the time of enactment, in order to limit the effect of a statutory command.  Professor William Eskridge and I addressed this issue in the amicus brief we filed in the case: “The statute attacks an injustice that is present in virtually every known civilization.  What would be surprising would be if that broad project did not have surprising implications . . . .”⁵⁸   This particular subtractive move, however, uses the very norms that the statute attacks in order to limit the statute’s application.  It imagines Congress to declare, “Oh, Lord, make America nondiscriminatory, but not yet.”⁵⁹ 

Berman and Krishnamurthi, however, do not finally take Justice Alito’s approach, though they occasionally gesture toward it.  In addressing the question whether sex discrimination covers discrimination against mothers, they ignore cultural context and original expectations.  “[H]eavy reliance on armchair theorizing”⁶⁰  does all the work: “being a mother . . . is a true subset of one sex (or gender).”⁶¹   They thus reject this particular subtractive move, on essentially the same basis that I did.⁶²   But ordinary meaning and common parlance still do a lot of work for them, in a different way.

III.      SHORTHAND OPERATORS

The term “homosexual” is simply a pair of parallel conjunctions: men attracted to men + women attracted to women.  Could such parallel conjunctions defeat the sex discrimination claim, when either of them alone would admittedly be sex discrimination?

The locus classicus of the parallel-conjunctions argument is an 1883 decision, Pace v. Alabama, in which the United States Supreme Court considered for the first time the constitutionality of miscegenation laws.⁶³   The statute in question in Pace prescribed penalties for interracial sex that were more severe than those imposed for adultery or fornication between persons of the same race.⁶⁴   The Court unanimously rejected the equal protection challenge to the statute, denying that it discriminated on the basis of race:

[The section prohibiting interracial sex] prescribes a punishment for an offence which can only be committed where the two sexes are of different races.  There is in neither section any discrimination against either race. . . .  Whatever discrimination is made in the punishment prescribed in the two sections is directed against the offence designated and not against the person of any particular color or race.  The punishment of each offending person, whether white or black, is the same.⁶⁵ 

I noted earlier that, in Lawrence, the sex of the defendant was an element of the crime that the prosecution needed to prove.  The same was true of the criminal prohibition of interracial sex and marriage: “To be a negro is not a crime; to marry a white woman is not a crime; but to be a negro, and being a negro, to marry a white woman is a felony; therefore, it is essential to the crime that the accused shall be a negro—unless he is a negro he is guilty of no offense.”⁶⁶   But the Pace Court explained that there was nonetheless no race discrimination, because “the offence designated” is distinct from any racial classification that operates within its application.

When the sex discrimination argument was made against the kind of law challenged in Lawrence, a law specifically criminalizing homosexual sex, it was rejected with an argument that precisely mimicked that of Pace.  In 1986, the Supreme Court of Missouri in State v. Walsh reversed a lower court’s declaration that a statute prohibiting “deviate sexual intercourse with another person of the same sex”⁶⁷  deprived the defendant of equal protection because “the statute would not be applicable to the defendant if he were a female.”⁶⁸   The state did not dispute that the defendant’s gender was an element of the crime.  But, the court explained, there was still no sex discrimination:

The State concedes that the statute prohibits men from doing what women may do, namely, engage in sexual activity with men.  However, the State argues that it likewise prohibits women from doing something which men can do: engage in sexual activity with women.  We believe it applies equally to men and women because it prohibits both classes from engaging in sexual activity with members of their own sex.  Thus, there is no denial of equal protection on that basis.⁶⁹ 

Once more, note what this concretely means: a woman could be criminally punished for doing something that only guys are allowed to do in Missouri.  According to the court, this would not be discrimination because of her sex.

Justice Alito made the same move when he wrote in Bostock that:

it is quite possible for an employer to discriminate on those grounds without taking the sex of an individual applicant or employee into account.  An employer can have a policy that says: “We do not hire gays, lesbians, or transgender individuals.”  And an employer can implement this policy without paying any attention to or even knowing the biological sex of gay, lesbian, and transgender applicants.⁷⁰ 

The employer can delegate to a subordinate the job of figuring out who is gay, and even instruct that subordinate not to report the specifics of cases in which the policy was implemented.  Similarly, the Missouri legislature did not know, when it enacted the ban on homosexual sex, whether it would be applied to men or to women.

To understand the trick that has been performed here, it may be helpful to compare the concept of “shorthand operators” from computer programming.  A shorthand operator is a way to briefly invoke an operation that is already available.  (“Macros,” in Microsoft Word, are a familiar example.)  It is equivalent to the full operation, triggering exactly the same series of steps, but spares the programmer the trouble of spelling out all of those steps.⁷¹   The linguistic happenstances discussed above, which sometimes include parallel-conjunctions terms, are a kind of shorthand operator.  They encode ascriptive hierarchies without spelling out their racist or sexist character.⁷² 

Berman and Krishnamurthi think I am “really relying on . . . a conflation of (a) facts that must be known for an agent to draw a warranted inference about a fact that is operative in their decision making with (b) the operative fact itself.”⁷³   That is, they accuse me of treating shorthand operators as equivalent to the operations for which they are shorthand, as though I am under the impression that two squared is the same as two times two.  But it is the same.  Guilty as charged.

In the Pace Court’s reasoning, “the offence designated” functions as a legal shorthand operator.  It designates a crime that has a racial element, but since the racial element is encoded within the general description of the crime, the Court concludes that there is no discrimination because of the defendant’s race.

The use of this shorthand operator to justify discrimination was consistent with the legal culture at the time of Pace.  There was no general rule against racial classification.  The closest the Court had come to stating such a rule was its declaration that under the Fourteenth Amendment, “the law in the States shall be the same for the black as for the white.”⁷⁴   The Pace Court construed the statute to be formally consistent with that requirement.⁷⁵   But this kind of move was no longer legitimate by the time of State v. Walsh.

The Court devised its present anticlassification rule precisely for the purpose of overruling Pace.  In McLaughlin v. Florida,⁷⁶  it unanimously invalidated a criminal statute prohibiting an unmarried interracial couple from habitually living in and occupying the same room at night.  “It is readily apparent,” the Court held, that the statute “treats the interracial couple made up of a white person and a Negro differently than it does any other couple.”⁷⁷   Racial classifications, it concluded, can only be sustained by a compelling state interest.  Since the State had failed to establish that the statute served “some overriding statutory purpose requiring the proscription of the specified conduct when engaged in by a white person and a Negro, but not otherwise,”⁷⁸  the statute necessarily fell as “an invidious discrimination forbidden by the Equal Protection Clause.”⁷⁹   Justice Stewart was unimpressed by the claim that the law affected both races equally: “I cannot conceive of a valid legislative purpose under our Constitution for a state law which makes the color of a person’s skin the test of whether his conduct is a criminal
offense.”⁸⁰ 

Berman and Krishnamurthi never mention McLaughlin.⁸¹   They come closest to addressing its reasoning when they engage with Price Waterhouse.  They respond that “to accept that Price Waterhouse was rightly decided is not to grant that it reflects the ordinary meaning of the statutory text at enactment; we can’t simply assume that it was a sound decision on textualist premises.”⁸²   It is possible that, if a person is penalized for failing to conform to gender norms—if, say, Fran is fired for wearing her hair short—“the firing of Fran could fall within the ordinary meaning of ‘sex discrimination,’ but not count as discrimination ‘because of Fran’s sex.’”⁸³   That is because the employer may not believe himself to be discriminating: “the locution ‘A does X because of Y’ tracks, at least in part, the actor’s view of Y, not only the philosophically informed view of Y.”⁸⁴   Even if the plaintiff can “prove that the employer relied upon sex-based considerations in coming to its decision,”⁸⁵  employers who use shorthand operators, and manage not to think about the discriminatory components of those shorthand operators, are in their view not discriminating at all.

If this move is available, the implications are far-reaching.  I wrote:

Suppose an employer decides to demand equally of men and women that they comport themselves in a manner consistent with the traditional understanding of their gender.  As the Court observes, we might hypothesize “an employer eager to revive the workplace gender roles of the 1950s,” who “enforces a policy that he will hire only men as mechanics and only women as secretaries.”⁸⁶ 

Similarly with race.  Racial segregation would not be deemed to classify on the basis of race so long as a parallel-conjunctions move is available.  It is available a lot.  Segregated railroad cars, for example, would not be deemed to classify so long as they were separate but equal.  More on this shortly.

To say it again, the statute declares that an employer has engaged in “[i]mpermissible consideration of . . . sex . . . in employment practices” when “sex . . . was a motivating factor for any employment practice,” irrespective of whether the employer was also motivated by “other factors.”⁸⁷   That precludes any Pace-like, shorthand operator excuse.  Textualists are supposed to read the statute.

IV.      CONSERVATION IN MOTIVATIONAL ANALYSIS

Berman and Krishnamurthi’s most novel argument responds to this difficulty by challenging Bostock’s account of motive.  They begin by scrutinizing the counterfactual claim that Bostock would not have been fired if he were female, keeping all other facts about him constant:

There are three relevant facts about Bostock, not two.  Bostock is (1) a man, (2) gay, and (3) attracted to men.  So, when “changing the employee’s sex,” Justice Gorsuch has not kept everything else the same.  If we change (1) from a man to a woman, then we can’t keep both (2) and (3) constant: by definition, a gay woman is not attracted to men.  So we must keep one and change the other.  The hypothetical woman version of Bostock must be either (a) gay and attracted to women or (b) straight and attracted to men.⁸⁸ 

It is thus a mistake to claim that Bostock would not have been fired if he were female, because

to the employer’s mind, the two employees are not materially identical in all respects except that one is a man and the other a woman; they are also non-identical in the respect that one is gay and the other straight.  And that latter material non-identicality is precisely the one that, by hypothesis, motivated the employer.⁸⁹ 

In order to avoid this error, they propose the single most innovative idea in their article:

The Principle of Conservation in Motivational Analysis (PCM): In performing counterfactual analysis, when changing one fact requires changing other facts too, the analyst must not change facts that are known, confidently believed, or stipulated to have been among the actor’s motivating reasons in favor of facts that are not likely, or less likely, to have been among the actor’s motivating reasons.⁹⁰ 

Since the employer was motivated by Bostock’s homosexuality rather than by Bostock’s sex, they conclude, it is not true that Bostock was fired because of his sex.

There is nothing wrong with the proposed PCM, if it means that when legal consequences turn on the intention with which an act is done, we should focus on an actor’s actual intentions.  Berman and Krishnamurthi however construe it to mean that if discrimination is not explicitly embraced, but the actor uses a shorthand operator that has a discriminatory element, then the discriminatory element is not to be deemed part of the actor’s motivation.

Shorthand operators (to say it again) are equivalent to the operations for which they are shorthand.  Justice Alito invoked a kind of shorthand operator when he observed (as quoted earlier) that an employer could discriminate against gay people without ever knowing the sex of those who are discriminated against.⁹¹   The employer would know that he has commanded that some individuals be treated worse because of their sex, but he is insulated from knowing about the specifics of any episode of discrimination.  The employer might say that he is invoking PCM: his motivating reason is excluding gay people, not discriminating on the basis of sex.  Berman and Krishnamurthi are
offering, in some ways, an extended defense of this claim of Justice Alito’s.

Can it be true that a boss is not responsible for discriminatory decisions he delegates?  Suppose I tell a manager at the business I own that I don’t want him ever to hire any black people.  “But,” I add, “if you do get any black applicants and turn them away, don’t tell me.  I don’t want to hear about it.  It is entirely possible that we won’t happen to get any qualified black applicants.  Make sure that I never know whether we do or we don’t, so that I can honestly say that I don’t know of or specifically intend any discriminatory decisions.”⁹² 

This kind of behavior is familiar in criminal law, which addresses it with the doctrine of “willful blindness.”  Here is the rule, laid down by the Supreme Court (per Justice Alito.): “(1) The defendant must subjectively believe that there is a high probability that a fact exists and (2) the defendant must take deliberate actions to avoid learning of that fact.”⁹³   Justice Alito cited with approval the following: “The willful blindness instruction allows the jury to impute the element of knowledge to the defendant if the evidence indicates that he purposely closed his eyes to avoid knowing what was taking place around him.”⁹⁴ 

This excuse would be even easier if the linguistic happenstance, discussed above, should be present.  That would defeat discrimination claims in the context of sexual harassment, hostile work environment, and even discrimination against mothers.  Or discrimination against uppity people, or viragoes.  Or miscegenators.

To borrow a term from Krishnamurthi’s earlier work, an employer might invent his own linguistic happenstance, deciding that any job inconsistent with the traditional social role of one’s race or sex is not “proper.”⁹⁵   The term “proper” could then function as a shorthand operator for the entire universe of discrimination that Title VII aimed to prohibit.  This is a broader exception than the one that relies on the culture of 1964 to argue that the authors did not intend to protect homosexuals.  It would be available as a defense in any Title VII case, including the paradigmatic cases of discrimination that the law’s authors had in mind.  The statute would be nullified in all its possible applications.⁹⁶ 

Berman and Krishnamurthi offer examples of this pathology.  “Suppose that Employer fires Fran, a straight ciswoman, for wearing her hair short.”⁹⁷   And suppose that “Employer announces and adheres to a sex-neutral policy according to which all employees must abide by gender-appropriate hair-length norms: short for men, long for women.”⁹⁸   In this case, they claim, “the firing of Fran could fall within the ordinary meaning of ‘sex discrimination,’ but not count as discrimination ‘because of Fran’s sex.’”⁹⁹   But if the statute does not prohibit the enforcement of gender-appropriate norms—if employers are still entitled to insist that women must not be supervisors, doctors, lawyers, construction workers, etc.—then it is not clear that the statute has any applications.  If Berman and Krishnamurthi think that textualism entails that Title VII is so easily evaded as to have no effect at all, they ought to tell us so.

The “proper” hypothetical sounds fanciful, but in fact a homosexuality exception can be the functional equivalent.  It is available in any case of discrimination against women, as I noted in my earlier defense of Bostock:

any time a woman occupies a position of authority, a significant strand of popular culture will use that position in order to impute lesbianism, which it deems intolerable.  And if discrimination is permissible whenever the discriminator plausibly recites a purpose of excluding lesbians, then discrimination against women will often be permissible.  More generally, any mistreatment on the basis of imputed homosexuality reinforces gender roles and contributes to the subordination of women.¹⁰⁰ 

One might respond that “proper” is an abuse of the linguistic happenstance, because it lumps together radically heterogenous activities that are really entirely distinct from one another.  “Interracial sex” involves the same physical action regardless of the race of the specific partners.  In this respect, though, “homosexual sex” is more like “proper.”  Male homosexual sex and lesbian sex are different physical activities, involving different body parts.  Lesbians do not engage in fellatio.  Male homosexuals do not engage in cunnilingus.

In a sound textual reading of Title VII, the only mental state of the defendant that matters is an intention to discriminate on a statutorily forbidden basis.  The PCM should not be construed to obscure that intention when it is in fact present and operative.  He who wills the end wills the means.  If the employer delegates a task that necessarily involves sex discrimination, then the employer is responsible for the discrimination even if it is carried out by a subordinate, and even if the employer was only thinking about a shorthand operator such as “homosexual.”

Return to the problem of parallel discriminations.  A law requiring separate but equal segregated railroad cars, for example, could be (and, in fact, was) deemed to have been enacted for race-neutral purposes, “with reference to the established usages, customs and traditions of the people, and with a view to the promotion of their comfort, and the preservation of the public peace and good order.”¹⁰¹ 

Suppose Homer Plessy, prosecuted for riding in the car reserved for white people, claimed that he was discriminated against because he was black.  He would claim that the relevant facts were that he was (1) black, and (2) in possession of a valid ticket.  But the state could retort that there were actually three relevant facts: he is (1) black, (2) in violation of “the established usages, customs and traditions of the people,”¹⁰²  and (3) in possession of a valid ticket.  You can’t change (1) without changing (2).¹⁰³   If whites would likewise be ejected from the cars reserved for blacks because they too are in violation of “the established usages, customs, and traditions of the people,” then Plessy’s discrimination claim, Berman and Krishnamurthi would have to say, “puts the rabbit in the hat.”¹⁰⁴   Removing those established usages, customs, and traditions from the analysis “does great and unnecessary violence to our ability to explain what did in fact happen in the real world.”¹⁰⁵   To say that Plessy is similarly situated to a white passenger

is screamingly false to the facts as stipulated: to the [Louisiana legislature]’s mind, the two [passengers] are not materially identical in all respects except that one is [white] and the other [black]; they are also non-identical in the respect that one is [defying established usages, customs, and traditions] and the other [is not].¹⁰⁶ 

An argument that logically entails that there was no discrimination in Plessy v. Ferguson has (if you will pardon the expression) gone off the rails.

V.      ANTICLASSIFICATION AND ANTISUBORDINATION

I have pressed the analogy with discrimination against interracial couples, which is uncontroversially forbidden by Title VII.  Berman and Krishnamurthi concede that the analogy is “structurally perfect.”¹⁰⁷   But they bite the bullet and say that this kind of discrimination “is not, as a textual matter, discrimination ‘because of the individual’s race.’”¹⁰⁸   They think the correct answer is to leave textualism behind, and claim that “a national commitment to combatting social practices that are rooted in, and further, white supremacy and racial subordination is properly attributed to or located within Title VII.”¹⁰⁹ 

So it turns out that they think Bostock was rightly decided, because “Title VII is rightly understood to target employment practices that arise from and reinforce sex-based hierarchy, in the same way that Title VII attacks racial subordination.”¹¹⁰   The proper basis for the Court’s decision, they write, is the fact that “[a]nti-gay and anti-transgender prejudice arise from the same soil as does prototypical sexism and serve the same structures of power and privilege.”¹¹¹ 

I wholeheartedly agree.  It’s what I’ve been saying for decades.¹¹²   I’m as skeptical of Justice Gorsuch’s New Textualism as they are.¹¹³   I don’t agree that the result in Bostock is “one big point for textualism,”¹¹⁴  even though I worked for that result.  It is more like the fact that a broken clock is absolutely accurate twice a day.

We play the cards we are dealt.  Professor Eskridge and I knew that Justice Gorsuch was our audience when we wrote our amicus brief, so we had to work within a New Textualist framework.  The formal sex discrimination argument does not rely on the more controversial argument based on antisubordination (which I do not believe Justice Gorsuch would have accepted).

As a general matter, New Textualism betrays its promise to constrain judges: it turns out that language is usually more manipulable when it is read out of context.¹¹⁵   Bostock was one of those rare cases where the text, standing alone, gave the interpreter no wiggle room.  Our aim was to show Justice Gorsuch that he had no wiggle room.  And he saw it.  And, to our enormous surprise, we won.

The antisubordination argument has two problems.  One is that it is not the law.

Consider the origins of the now-familiar rule that racial classifications are presumptively suspect and subject to strict scrutiny.  The most famous of the interracial marriage cases is Loving v. Virginia.¹¹⁶   There the Court declared that prohibitions of such marriages were “measures designed to maintain White Supremacy,”¹¹⁷  which “violates the central meaning of the Equal Protection Clause.”¹¹⁸   That language leads many to think that the interracial marriage question is properly resolved by antisubordination reasoning.

McLaughlin, not Loving, was however the groundbreaking case that laid to rest the argument that such statutes were valid because they applied equally to black and white alike.¹¹⁹   McLaughlin, not Loving, is the crucial precedent on which the sex discrimination argument relies.  McLaughlin did not make any claims about the legislature’s motivations in enacting the law or about the class that was harmed by the law.  It is the germinal decision that first laid down the now-familiar rule that racial classifications are automatically subject to strict scrutiny.¹²⁰   The sex discrimination argument for protecting gays from discrimination requires nothing more.

Whatever one thinks about the antisubordination/anticlassification debate concerning the interpretation of the Fourteenth Amendment, one ought to notice that the text of Title VII, like McLaughlin, takes an anticlassification approach.  The plain language of the text might be appropriately defeated (though textualists would probably disagree) if what is covered is no part of the mischief that the law aims to remedy, even more so if in some contexts racial classifications might themselves help to remedy that mischief.¹²¹   But classifications are what the statute focuses on.

The second problem with the antisubordination argument is that it is routinely invoked by defenders of antigay discrimination.  They don’t accept the sociological claim that the condemnation of homosexuality reinforces gender hierarchy.  There is no reason to doubt their sincerity.  Many people, including at least one Supreme Court Justice, honestly don’t see it.¹²²   And so the rejection of the sociological claim is offered as a defeater of the sex discrimination argument.¹²³   That is why, from the time I first made the argument, I carefully emphasized that it does not depend at all on the sociological claim, even though I believe that claim to be accurate.¹²⁴ 

CONCLUSION

This entire exchange is a bit bizarre.  Berman and Krishnamurthi are not textualists, and neither am I.  We are debating what follows from textualist premises that none of us really believe.  I claim that they have misread the text, thus yielding the bogus bogusness of Bostock.

As a bare textual matter, one discriminates “because of sex” when one fires men for specific conduct that would be tolerable if done by women.  The violation is not ameliorated if one also fires women for specific conduct that would be tolerable if done by men.  Parallel discriminations, bundled together by the linguistic happenstance of the term “homosexuality,” cannot defeat the sex discrimination claim when either of them alone would admittedly be sex discrimination.

But, as I said at the outset, they are both first-class scholars whose view must be taken seriously.  It is hard to confidently assert that the statute can’t reasonably be read their way when they in all sincerity think as they do.  Perhaps our disagreement simply displays the deepest flaw of the new textualism: that reading the words of a statute without regard to context yields deep indeterminacy, and so betrays the promise to constrain judicial discretion.

---

© 2022 Andrew Koppelman.  Individuals and nonprofit institutions may reproduce and distribute copies of this Essay in any format at or below cost, for educational purposes, so long as each copy identifies the author, provides a citation to the Notre Dame Law Review Reflection, and includes this provision in the copyright notice.

* John Paul Stevens Professor of Law and Professor (by courtesy) of Political Science, Department of Philosophy Affiliated Faculty, Northwestern University.  Thanks to Issa Kohler-Hausmann for comments on an earlier draft and to Josh Kleinfeld and Matt Spitzer for helpful conversations.  Special thanks to Mitchell Berman and Guha Krishnamurthi for generous comments, which rescued me from many errors.

¹ 539 U.S. 558 (2003).

² Id. at 562.  TEX. PENAL CODE ANN. § 21.06(a) (2003) provided that “[a] person commits an offense if he engages in deviate sexual intercourse with another individual of the same sex.”  The statute defined “[d]eviate sexual intercourse” to mean “(A) any contact between any part of the genitals of one person and the mouth or anus of another person; or (B) the penetration of the genitals or the anus of another person with an object.”  § 21.01(1), quoted in Lawrence, 539 U.S. at 563.

³ In fact, the defendants probably never had sex, and the police probably lied about what they saw.  See DALE CARPENTER, FLAGRANT CONDUCT: THE STORY OF LAWRENCE V. TEXAS 61–104 (2012).  I leave this complication aside and treat the case as the Supreme Court did, stipulating the alleged facts to be true.  Lawrence, 539 U.S. at 562–64.

⁴ Lawrence did in fact challenge the statute on this basis.  CARPENTER, supra note 3, at 156–57.  The claim was rejected by the lower court.  See Lawrence v. State, 41 S.W.3d 349, 357–59 (Tex. App. 2001); but see id. at 367–73 (Anderson, J., dissenting).  The Supreme Court litigators relegated it to a footnote.  CARPENTER, supra note 3, at 196.

⁵ Miss. Univ. for Women v. Hogan, 458 U.S. 718, 724 (1982) (quoting Kirchberg v. Feenstra, 450 U.S. 455, 461 (1981)); see also United States v. Virginia, 518 U.S. 515, 531 (1996) (quoting Hogan, 458 U.S. at 724).  In keeping with the Court’s practice, I here use “gender” and “sex” interchangeably.

⁶ Virginia, 518 U.S. at 533 (citing Hogan, 458 U.S. at 724).

⁷ See Andrew Koppelman, Essay, Bostock, LGBT Discrimination, and the Subtractive Moves, 105MINN. L. REV. HEADNOTES 1 (2020).

⁸ 140 S. Ct. 1731 (2020).

⁹ Mitchell N. Berman & Guha Krishnamurthi, Bostock Was Bogus: Textualism, Pluralism, and Title VII, 97 NOTRE DAME L. REV. 67 (2021).

¹⁰ Id. at 67–68.

¹¹ See id. at 112–13.

¹² Id. at 114.

¹³ ANDREW KOPPELMAN, THE GAY RIGHTS QUESTION IN CONTEMPORARY AMERICAN LAW 53–71 (2002); ANDREW KOPPELMAN, ANTIDISCRIMINATION LAW AND SOCIAL EQUALITY 146–76 (1996) [hereinafter KOPPELMAN, ANTIDISCRIMINATION LAW]; Andrew Koppelman, The Supreme Court Made the Right Call on Marriage Equality—But They Did It the Wrong Way, SALON (June 29, 2015, 3:15 PM), https://www.salon.com/2015/06/29/the_supreme_court_made_the_right_call_on_marriage_equality_—_but_they_did_it_the_wrong_way/ [https://perma.cc/6ADJ-Q8ZB]; Andrew Koppelman & Ilya Somin, Gender, the Gay Marriage Fight’s Missing Piece, USA TODAY (Apr. 20, 2015, 9:59 AM), https://www.usatoday.com/story/opinion/2015/04/19/supreme-court-same-sex-marriage-constitutionality-discrimination-column/70225124/ [https://perma.cc/2VW7-S6QQ]; Andrew Koppelman, Beyond Levels of Scrutiny: Windsor and “Bare Desire to Harm”, 64 CASE W. RSRV. L. REV. 1045, 1053–58 (2014); Andrew Koppelman, Response, Sexual Disorientation, 100 GEO. L.J. 1083 (2012); Andrew Koppelman, Discrimination Against Gays Is Sex Discrimination, in MARRIAGE AND SAME-SEX UNIONS: A DEBATE 209 (Lynn D. Wardle et al. eds., 2003); Andrew Koppelman, Reply to “The Constitutionality of Legal Preferences for Heterosexual Marriage”, in MARRIAGE AND SAME-SEX UNIONS: A DEBATE 227, 241, 242; Andrew Koppelman, Defending the Sex Discrimination Argument for Lesbian and Gay Rights: A Reply to Edward Stein, 49 UCLA L. REV. 519 (2001), reprinted in 1 DUKEMINIER AWARDS: BEST SEXUAL ORIENTATION & GENDER IDENTITY L. REV. 49 (2002) [hereinafter Koppelman, Defending the Sex Discrimination Argument]; Andrew Koppelman, The Miscegenation Analogy in Europe, or Lisa Grant Meets Adolf Hitler, in LEGAL RECOGNITION OF SAME-SEX PARTNERSHIPS: A STUDY OF NATIONAL, EUROPEAN AND INTERNATIONAL LAW 623 (Robert Wintemute & Mads Andenæs eds., 2001) [hereinafter Koppelman, The Miscegenation Analogy in Europe]; Andrew Koppelman, Three Arguments for Gay Rights, 95 MICH. L. REV. 1636 (1997) (reviewing ROBERT WINTEMUTE, SEXUAL ORIENTATION AND HUMAN RIGHTS (1995)); Andrew Koppelman, Why Discrimination Against Lesbians and Gay Men Is Sex Discrimination, 69 N.Y.U. L. REV. 197 (1994) [hereinafter Koppelman, Discrimination Against Lesbians and Gay Men]; Andrew Koppelman, Note, The Miscegenation Analogy: Sodomy Law as Sex Discrimination, 98 YALE L.J. 145 (1988) [hereinafter Koppelman, Sodomy Law as Sex Discrimination]; cf. Andrew Koppelman, The Miscegenation Precedents, in SAME-SEX MARRIAGE, PRO AND CON: A READER 333 (Andrew Sullivan ed., 2d ed. 2004); Andrew Koppelman, Same-Sex Marriage and Public Policy: The Miscegenation Precedents, 16 QUINNIPIAC L. REV. 105 (1996).

¹⁴ Brief of Amici Curiae William N. Eskridge Jr., Bruce A. Ackerman, Daniel A. Farber & Andrew Koppelman in Support of Respondents, Hollingsworth v. Perry, 570 U.S. 693 (2013) (No. 12-144); Brief Amicus Curiae of Legal Scholars Stephen Clark, Andrew Koppelman, Sanford Levinson, Irina Manta, Erin Sheley & Ilya Somin, Obergefell v. Hodges, 576 U.S. 644 (2015) (No. 14-556); Brief of William N. Eskridge Jr. & Andrew M. Koppelman as Amici Curiae in Support of Employees, Bostock v. Clayton Cnty., 140. S. Ct. 1731 (2020) (No. 17-1618).

¹⁵ Koppelman, supra note 7.

¹⁶ Andrew Koppelman, Bostock, LGBT Discrimination and the Subtractive Moves (Nw. P.L. Rsch. Paper No. 19-19, 2019), https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3424326.  At the time, I was memorializing points that wouldn’t fit in the amicus brief that Bill Eskridge and I had coauthored.

¹⁷ See Bostock v. Clayton Cnty., 140 S. Ct. 1731, 1773 (Alito, J., dissenting).

¹⁸ One error I describe that they do not make is the reliance on prototypical meaning, which they disavow.  Berman & Krishnamurthi, supra note 9, at 92–94.  They are mistaken, however, when they claim that other critics of the sex discrimination article do not commit this error.  See Koppelman, supra note 7, at 13–16 (collecting examples).

¹⁹ Mitchell N. Berman, Originalism Is Bunk, 84 N.Y.U. L. REV. 1 (2009).

²⁰ Guha Krishnamurthi & Peter Salib, Bostock and Conceptual Causation, YALE J. ON REGUL.: NOTICE & COMMENT (July 22, 2020), https://www.yalejreg.com/nc/bostock-and-conceptual-causation-by-guha-krishnamurthi-peter-salib/ [https://perma.cc/7GGE-5SM9].

²¹ Its argument is in deep tension with the new work of Berman and Krishnamurthi.  See infra notes 38, 72.

²² In fairness, even a legal theorist as distinguished as John Gardner fell into similar errors when assessing the sex discrimination argument.  Koppelman, The Miscegenation Analogy in Europe, supra note 13, at 629–31.

²³ Berman & Krishnamurthi, supra note 9, at 92.

²⁴ Koppelman, supra note 7, at 1.

²⁵ Id. at 8.  Thus, Justice Gorsuch writes in Bostock: “To ‘discriminate against’ a
person . . . would seem to mean treating that individual worse than others who are similarly situated.”  Bostock v. Clayton Cnty., 140 S. Ct. 1731, 1740 (2020) (quoting Burlington N. & Santa Fe Ry. Co. v. White, 548 U.S. 53, 59 (2006)).

²⁶ See William N. Eskridge, Jr., The New Textualism and Normative Canons, 113 COLUM. L. REV. 531, 532 (2013) (reviewing ANTONIN SCALIA & BRYAN A. GARNER, READING LAW: THE INTERPRETATION OF LEGAL TEXTS (2012)).  Berman and Krishnamurthi observe: “Statutory textualism, like standard versions of constitutional originalism, is a monistic thesis.  It’s a claim about the sole determinant of legal content, or the sole target of appropriate or legitimate judicial interpretation.”  Berman & Krishnamurthi, supra note 9, at 121.

²⁷ Oncale v. Sundowner Offshore Servs., Inc., 523 U.S. 75, 79 (1998).

²⁸ See infra Part V.

²⁹ 42 U.S.C. § 2000e-2(a) (2018).

³⁰ Id. § 2000e-2(m).

³¹ City of L.A. Dep’t of Water & Power v. Manhart, 435 U.S. 702, 711 (1978) (quoting Developments in the Law: Employment Discrimination and Title VII of the Civil Rights Act of 1964, 84 HARV. L. REV. 1109, 1170 (1971)).  Because the sex discrimination argument is about reasons for action rather than causes, the concept of “conceptual causation,” which Berman and Krishnamurthi take to be a (fallacious) part of some formulations of that argument and possibly even part of Justice Gorsuch’s reasoning, Berman & Krishnamurthi, supra note 9, at 88–89, 88 n.112, is irrelevant.  So is the problem of counterfactuals and possible worlds.  See id. at 111–16.

³² Bostock v. Clayton Cnty., 140 S. Ct. 1731, 1739 (2020) (quoting Univ. of Tex. Sw. Med. Ctr. v. Nassar, 570 U.S. 338, 350 (2013)).

³³ Id. (citation omitted) (quoting Nassar, 570 U.S. at 347) (citing Gross v. FBL Fin. Servs., Inc., 557 U.S. 167, 176 (2009)).

³⁴ Berman & Krishnamurthi, supra note 9, at 99.

³⁵ Berman and Krishnamurthi think that this argument does not protect bisexuals, because firing a bisexual is not “because of such individual’s sex.”  Id. at 108.  But if Dana, a bisexual cisman, is fired for dating men, while a woman would not be fired for dating men, then Dana is fired because of his sex.  The fact that Dana also dates women is neither an element of his claim nor a defense against it.  The terms “homosexual” and “bisexual” beget confusion in this area of the law.  An employer who fires a white employee for dating blacks is liable for discrimination.  It doesn’t matter whether the fired employee also sometimes dates whites.

³⁶ Bostock, 140 S. Ct. at 1737.

³⁷ Berman & Krishnamurthi, supra note 9, at 72.

³⁸ Id. at 82 (quoting Hively v. Ivy Tech Cmty. Coll. of Ind., 853 F.3d 339, 363 (7th Cir. 2017) (en banc) (Sykes, J., dissenting)).

³⁹ Koppelman, supra note 7, at 16–17 (footnotes omitted) (first quoting Zarda v. Altitude Express, Inc., 883 F.3d 100, 119 n.17 (2d Cir. 2018), aff’d sub nom. Bostock, 140 S. Ct. 1731; and then quoting Zarda, 883 F.3d at 147 (Lynch, J., dissenting)).  The possibility of the linguistic happenstance is considered in Krishnamurthi & Salib, supra note 20.  Their conclusion, that it would excuse invidious discrimination, is persuasive and is not answered in Berman and Krishnamurthi.

⁴⁰ This happenstance does all the work in this hypothetical:

Suppose the legislatures in North and South Textalia are both considering whether to ban employment discrimination and, if so, on what bases.  Each legislature is considering three options: (1) prohibiting discrimination “because of an individual’s sex or because of their sexual orientation”; (2) prohibiting discrimination “because of an individual’s sex”; or (3) enacting no prohibition.  After much debate, North Textalia enacts option (1) and South Textalia enacts option (2).  On Justice Gorsuch’s textualist analysis, the law is actually the same in both jurisdictions because discrimination “because of an individual’s sexual orientation” is discrimination “because of an individual’s sex” as a matter of legal meaning.  This is so even though the fact that the legislators vigorously debated the choice between (1) and (2) might seem near-conclusive evidence that they accorded the phrases different meanings.

Berman and Krishnamurthi, supra note 9, at 107.  Test this “surprise” against a statute prohibiting discrimination “because of an individual’s race or because of their participation in an interracial marriage.”  Or, for that matter, a statute that clearly specifies any result that was not anticipated by the legislature enacting the statute.

⁴¹ On this function of antidiscrimination law, see KOPPELMAN, ANTIDISCRIMINATION LAW, supra note 13, at 1.

⁴² Perhaps he believes, as the 1956 Southern Manifesto claimed, that in the world before desegregation, there was “friendship and understanding” between the races.  See 102 CONG. REC. 4460 (1956).

⁴³ Elspeth Reeve, Yep, ‘Uppity’ Is Racist, ATLANTIC (Nov. 22, 2011), https://www.theatlantic.com/politics/archive/2011/11/yep-uppity-racist/335160/ [https: //perma.cc/3W2D-HL24]; John Ridley, How Bad is ‘Uppity’?, NPR (Sept. 16, 2008, 5:00 PM), https://www.npr.org/sections/visibleman/2008/09/how_bad_is_uppity.html [https: //perma.cc/X3EV-VCDW].

⁴⁴ Price Waterhouse v. Hopkins, 490 U.S. 228, 234 (1989) (plurality opinion).

⁴⁵ Id. at 250.

⁴⁶ Phillips v. Martin Marietta Corp., 400 U.S. 542, 544 (1971) (per curiam).

⁴⁷ Berman & Krishnamurthi, supra note 9, at 105 (footnote omitted).

⁴⁸ This is the parallel-discriminations move that I consider in Part III.

⁴⁹ Berman & Krishnamurthi, supra note 9, at 72.

⁵⁰ Id. at 82.

⁵¹ Berman and Krishnamurthi note that in one survey, sixty percent of respondents said that firing somebody because she is pregnant is not firing her “because of her sex.”  Id. at 105 n.189 (citing Kevin Tobia & John Mikhail, Two Types of Empirical Textualism, 86 BROOK. L. REV. 461, 480 (2021)).

⁵² Phillips v. Martin Marietta Corp., 411 F.2d 1, 4 (5th Cir. 1969), rev’d, 400 U.S. 542 (1971).

⁵³ Bostock v. Clayton Cnty., 140 S. Ct. 1731, 1773 (Alito, J., dissenting).

⁵⁴ See CONSTANTIN STANISLAVSKI, AN ACTOR PREPARES 12–17 (1936).

⁵⁵ See Koppelman, supra note 7, at 28–30.

⁵⁶ Berman and Krishnamurthi observe that textualists are still trying to figure out how to specify the role of context in interpretation.  Berman and Krishnamurthi, supra note 9, at 84 n.94.  Since textualism’s defining commitment is to decontextualization, it is unsurprising that this is a persistent headache.

⁵⁷ Bostock, 140 S. Ct. at 1769 (Alito, J., dissenting).

⁵⁸ Brief of William N. Eskridge Jr. & Andrew M. Koppelman as Amici Curiae in Support of Employees at 17, Bostock v. Clayton Cnty., 140 S. Ct. 1731 (2020) (No. 17-1618).

⁵⁹ I am of course paraphrasing Saint Augustine’s description of his own youthful
resistance to religious salvation, asking God: “Give me chastity and continence, but not yet.”  SAINT AUGUSTINE, CONFESSIONS 169 (R.S. Pine-Coffin trans., Dorset Press 1961) (c. 400 A.D.).

⁶⁰ Berman & Krishnamurthi, supra note 9, at 94.

⁶¹ Id. at 105.

⁶² See Koppelman, supra note 7, at 16–17.

⁶³ Pace v. Alabama, 106 U.S. 583, 583 (1883).

⁶⁴ Id.

⁶⁵ Id. at 585.  This reasoning probably reflects the original intent of the Framers of the Fourteenth Amendment, who offered similar arguments.  Just as proponents of the Federal Equal Rights Amendment denied opponents’ allegations that sex equality would require legal recognition of gay marriage, see Note, The Legality of Homosexual Marriage, 82 YALE L.J. 573, 583–88 (1973), proponents of the Fourteenth Amendment denied opponents’ allegations that racial equality would require legal recognition of interracial marriage, see Alfred Avins, Anti-Miscegenation Laws and the Fourteenth Amendment: The Original Intent, 52 VA. L. REV. 1224, 1255 (1966).  This interpretation of equality predates the Fourteenth Amendment, appearing in President Andrew Johnson’s defense of his (later overridden) veto of the Civil Rights Act of 1866.  See CONG. GLOBE, 39th Cong., 1st Sess. 1858 (1866) (“I do not say this bill repeals State laws on the subject of marriage between the two races, for as the whites are forbidden to intermarry with the blacks, the blacks can only make such contracts as the whites themselves are allowed to make, and therefore cannot, under this bill, enter into the marriage contract with the whites.”).  The lower court explained that the law aimed to prevent “the amalgamation of the two races, producing a mongrel population and a degraded civilization, the prevention of which is dictated by a sound public policy affecting the highest interests of society and government.”  Pace v. State, 69 Ala. 231, 232 (1881), aff’d 106 U.S. 583 (1883).  This was enough to persuade Justice Harlan, who famously dissented in Plessy but who joined the majority in Pace.  See Plessy v. Ferguson, 163 U.S. 537, 552 (1896) (Harlan J., dissenting).

⁶⁶ Jones v. Commonwealth, 80 Va. 538, 542 (1885).  See generally Koppelman, Sodomy Law as Sex Discrimination, supra note 13, at 149–51.

⁶⁷ MO. REV. STAT. § 566.090.1(3) (1986).

⁶⁸ State v. Walsh, 713 S.W.2d 508, 509 (Mo. 1986) (en banc) (quoting an unpublished trial court opinion).

⁶⁹ Id. at 510.  This argument was ubiquitous in the caselaw.  Koppelman, Discrimination Against Lesbians and Gay Men, supra note 13, at 209 n.40.

⁷⁰ Bostock v. Clayton Cnty., 140 S. Ct. 1731, 1758 (2020) (Alito, J., dissenting).

⁷¹ Shorthand Operators, EMORY UNIV., http://www.mathcs.emory.edu/~cheung/Courses/170/Syllabus/04/shorthand.html [perma.cc/UKV6-GWZ9].

⁷² The possibility of such shorthand operators, and their discriminatory character, is acknowledged in Krishnamurthi & Salib, supra note 20.

⁷³ Berman & Krishnamurthi, supra note 9, at 94.

⁷⁴ Strauder v. West Virginia, 100 U.S. 303, 307 (1880).

⁷⁵ On the other hand, like the statute invalidated in Strauder, the law in Pace was “practically a brand upon [African Americans], affixed by the law, an assertion of their inferiority, and a stimulant to that race prejudice which is an impediment to securing to individuals of the race that equal justice which the law aims to secure to all others.”  Id. at 308.

⁷⁶ 379 U.S. 184 (1964).

⁷⁷ Id. at 188.

⁷⁸ Id. at 192.

⁷⁹ Id. at 192–93.

⁸⁰ Id. at 198 (Stewart, J., concurring).  Similarly, in the states that specifically prohibited homosexual sex before the Supreme Court invalidated those laws on privacy grounds, the defendant’s own sex evidently was one of the essential elements of the crime that the prosecution must prove.  Koppelman, Defending the Sex Discrimination Argument, supra note 13, at 523; id. at n.20 (citing statutes).

⁸¹ I relied heavily on McLaughlin when I explained the fallacy of this particular subtractive move.  Koppelman, supra note 7, at 18–19.  Professors Berman and Krishnamurthi respond, in conversation, that McLaughlin is not relevant because it is interpreting a different text, and Bostock is a textualist decision.  But the McLaughlin Court understood the Equal Protection Clause to prohibit the state from classifying by race, thereby making it in this context the functional equivalent of Title VII.

⁸² Berman & Krishnamurthi, supra note 9, at 90.

⁸³ Id.

⁸⁴ Id.

⁸⁵ Price Waterhouse v. Hopkins, 490 U.S. 228, 242 (1989) (plurality opinion).

⁸⁶ Koppelman, supra note 7, at 19–20 (quoting Bostock v. Clayton Cnty, 140 S. Ct. 1731, 1748 (2020)).

⁸⁷ 42 U.S.C. § 2000e-2(m) (2018).

⁸⁸ Berman & Krishnamurthi, supra note 9 at 102 (footnote omitted).  This argument is not in my catalogue of subtractive moves, but it was offered by some of the lower-court judges who rejected the sex discrimination argument.  See, e.g., Hively v. Ivy Tech Cmty. Coll. of Ind., 853 F.3d 339, 366 (7th Cir. 2017) (en banc) (Sykes, J., dissenting).

⁸⁹ Berman & Krishnamurthi, supra note 9, at 106.

⁹⁰ Id. at 113.

⁹¹ See supra note 70 and accompanying text.

⁹² Nor would the claim be blocked if the employer encoded the discriminatory standard in the job application, so that the applicant had to disclose whether he had the discriminated-against characteristic, thus triggering automatic rejection.  See Berman & Krishnamurthi, supra note 9, at 115 n.218.  This is just another shorthand operator, another way to delegate the decision to discriminate.  Berman and Krishnamurthi write that “it is quite easy to know whether somebody is gay without knowing their sex.”  Id. at 94.  It is also easy to know that someone is in an interracial marriage without knowing their race.  But it is not possible to apply this category without knowing that.  Similarly with homosexuality, or persons who do not comport themselves in a manner consistent with the traditional understanding of their gender.  See Koppelman, supra note 7, at 17–20.

⁹³ Global-Tech Appliances, Inc. v. SEB S.A., 563 U.S. 754, 769 (2011).

⁹⁴ Id. at 769 n.9 (quoting United States v. Schnabel, 939 F.2d 197, 203 (4th Cir. 1991)).

⁹⁵ Krishnamurthi & Salib, supra note 20.

⁹⁶ It is one of the principal defects of textualism that it can lead to this result.  “It invites perverse readings of statutes that defeat the purposes for which they were enacted.”  Koppelman, supra note 7, at 25 (citing King v. Burwell, 135 S. Ct. 2480 (2015) (Scalia, J., dissenting)).

⁹⁷ Berman & Krishnamurthi, supra note 9, at 89.

⁹⁸ Id.

⁹⁹ Id. at 90.

¹⁰⁰ Koppelman, supra note 7, at 25.

¹⁰¹ Plessy v. Ferguson, 163 U.S. 537, 550 (1896).  If the encoding of the shorthand operator in this formulation is less effective than that of “homosexual,” it is only because of the linguistic happenstance that this encoding is not compacted into a single word.

¹⁰² Id.

¹⁰³ The example is adapted from Berman & Krishnamurthi, supra note 9, at 102–03.

¹⁰⁴ Id. at 113.

¹⁰⁵ Id. at 115.

¹⁰⁶ Id. at 106 (emphasis omitted).

¹⁰⁷ Id. at 122.

¹⁰⁸ Id. at 124.  This is the basis for my claim, at the beginning of this paper, that by Berman and Krishnamurthi’s logic, Pace v. Alabama is correct: because the statute applied equally to both races, it was not racially discriminatory.  See supra notes 64–65 and accompanying text.

¹⁰⁹ Berman & Krishnamurthi, supra note 9, at 124.

¹¹⁰ Id.

¹¹¹ Id. at 124–25.

¹¹² See supra notes 13-14 and accompanying text.

¹¹³ I say so repeatedly in my defense of Bostock, and in Andrew Koppelman, Passive Aggressive: Scalia and Garner on Interpretation, BOUNDARY, Summer 2014, at 227 (reviewing ANTONIN SCALIA & BRYAN A. GARNER, READING LAW: THE INTERPRETATION OF LEGAL TEXTS (2012)).  The coauthor of my amicus brief, William Eskridge, is one of the most prominent critics of the New Textualism.  See, e.g., William N. Eskridge, Jr., The New Textualism and Normative Canons, 113 COLUM. L. REV. 531 (2013) (reviewing ANTONIN SCALIA & BRYAN A. GARNER, READING LAW: THE INTERPRETATION OF LEGAL TEXTS (2012)).

¹¹⁴ Berman and Krishnamurthi, supra note 9, at 69.

¹¹⁵ This is shown in considerable detail in VICTORIA NOURSE, MISREADING LAW, MISREADING DEMOCRACY (2016).

¹¹⁶ 388 U.S. 1 (1967).

¹¹⁷ Id. at 11.

¹¹⁸ Id. at 12.

¹¹⁹ McLaughlin v. Florida, 379 U.S. 184, 198 (1964).  The significance of McLaughlin is argued in Ariela R. Dubler, From McLaughlin v. Florida to Lawrence v. Texas: Sexual Freedom and the Road to Marriage, 106 COLUM. L. REV. 1165 (2006).

¹²⁰ On the crucial place of McLaughlin in the evolution of modern strict scrutiny, see RICHARD H. FALLON JR., THE NATURE OF CONSTITUTIONAL RIGHTS: THE INVENTION AND LOGIC OF STRICT JUDICIAL SCRUTINY 22–26 (2019).

¹²¹ See, e.g., United Steelworkers v. Weber, 443 U.S. 193 (1979).

¹²² Andrew Koppelman, Marriage Equality and the Sex Discrimination Argument, BALKINIZATION, (Oct. 3, 2020, 9:30 AM), https://balkin.blogspot.com/202010/marriage-equality-and-sex.html [https://perma.cc/KYW3-D9B6].

¹²³ See, e.g., David Orgon Coolidge, Playing the Loving Card: Same-Sex Marriage and the Politics of Analogy, 12 BYU J. PUB. L. 201 (1998); Richard F. Duncan, From Loving to Romer: Homosexual Marriage and Moral Discernment, 12 BYU J. PUB. L. 239 (1998); Jay Alan Sekulow & John Tuskey, Sex and Sodomy and Apples and Oranges—Does the Constitution Require States to Grant a Right to Do the Impossible?, 12 BYU J. PUB. L. 309 (1998); Lynn D. Wardle, A Critical Analysis of Constitutional Claims for Same-Sex Marriage, 1996 BYU L. REV. 1 (1996); Craig M. Bradley, The Right Not to Endorse Gay Rights: A Reply to Sunstein, 70 IND. L.J. 29 (1994).

¹²⁴ Koppelman, Discrimination Against Lesbians and Gay Men, supra note 13, at 220.  Nan D. Hunter, responding to this problem, observes: “[H]owever favored by progressive scholars, anti-subordination theory is not the law.  The anti-subordination language of Loving was dicta; the reliance on color-blindness and formal neutrality in constitutional jurisprudence has increased, not decreased, since that decision.  If courts were to assert the inadequacy of anti-subordination reasoning as a doctrinal bar to sex discrimination claims in gay marriage cases, that rationale would reek of intellectual dishonesty.”  Nan D. Hunter, The Sex Discrimination Argument in Gay Rights Cases, 9 J.L. & POL’Y 397, 411 (2001) (footnotes omitted).