HomeCase CommentaryA Textualist Critique of Bostock
The Roberts Court, November 30, 2018. Seated, from left to right: Justices Stephen G. Breyer and Clarence Thomas, Chief Justice John G. Roberts, Jr., and Justices Ruth Bader Ginsburg and Samuel A. Alito. Standing, from left to right: Justices Neil M. Gorsuch, Sonia Sotomayor, Elena Kagan, and Brett M. Kavanaugh. Photograph by Fred Schilling, Supreme Court Curator's Office.

A Textualist Critique of Bostock

The Supreme Court of the United States’s recent decision in Bostock v. Clayton Country, Georgia has already elicited a great deal of controversy and scholarly commentary. I typically refrain from commenting on U.S. decisions as I am not an expert on U.S. law. However, the decision in Bostock turned entirely on the application of the principles of statutory interpretation, which has long been an interest of mine.

What makes Bostock particularly noteworthy is that both the majority and dissent adopt a textualist approach to the issue of whether discrimination on the basis of sexual orientation comes within the ambit of Title VII’s prohibition on sex discrimination. This has caused some to criticize the textualist methodology on various grounds: as being unable to deliver the predictability it purports to offer, as a veil for results-oriented reasoning, or, worse still, as a betrayal of the conservative legal movement. Mark Mancini has dealt aptly with the latter criticism in a recent post for the Double Aspect blog arguing, correctly in my view, that the purpose of textualism is not to produce conservative outcomes in every instance, but to ensure that the venue for political change is the legislatures not the courts. With respect to the second criticism, there is every indication that Justice Gorsuch, the decision’s author, was motivated by an honest belief in the proper textual interpretation of the relevant legislative provision. The fact that his interpretation may appear counter-intuitive at first glance does not justify Justice Alito’s scathing critique that the decision is like a “pirate ship” in that it “sails under a textualist flag, but..actually represents…a theory of statutory interpretation…that courts should ‘update’ old statutes so that they better reflect the current values of society.” As I intend to argue, the majority decision is flawed, but not egregiously so. Its failure is not that it represents disingenuous judicial policy-making, but that is is not a proper application of the textualist method.

And this brings me to the first criticism, that Bostock demonstrates textualism’s inability to achieve certainty or predictability. This is the only legally serious criticism of the textualist method, but it is still misguided in my view. As I argued in my conversation with Justice Stratas at the Runnymede Society’s Law & Freedom Conference in 2019, the virtue of textualism is not that it removes uncertainty entirely, but that it substantially reduces it by narrowing the ambit of legally permissible interpretations – often, but not always, to one. As I intend to make clear below, Bostock illustrates this well. Both the majority and dissent (especially that of Justice Kavanaugh) agreed on the rules of the game and where the goalposts ought to be placed, and this enables us to assess their disagreement and evaluate their respective positions with a great deal of objectivity. And when we do so, it becomes clear that Gorsuch J.’s opinion is flawed both as a matter of law and logic.

As a matter of law – ordinary versus plain meaning

Bostock is wrong as a matter of law essentially for the reasons set out by Kavanaugh J. Both the majority and dissent agreed that the court was required to interpret the original meaning of the law (a principle that is noncontroversial when dealing with ordinary statutes, both in Canada and the U.S., but which becomes highly controversial in the constitutional realm). The problem with Gorsuch’s opinion is that he looks at the original plain or literal meaning of the law, not its original ordinary meaning. It is uncontroversial that courts ought typically to apply the latter and not the former. The ordinary meaning rule recognizes that phrases have meanings that are often different from their constituent words. The plain and ordinary meanings will often cohere, but where they do not, the ordinary meaning prevails. To use a Canadian example, a court examining whether a law passed by Parliament falls within its residual discretion does not separately ascertain whether the law furthers the “peace,” the “order” and the “good government” of Canada; it rather recognizes “peace, order and good government” to be a 19th century legal term that speaks to the broad power to make laws in the national interest.

To this end, the issue is not what the words “discriminate,” “because of” and “sex” meant when Title VII was enacted in 1964, but the meaning of the phrase “discriminate…because of sex.” There can be little doubt that when the law was enacted discrimination because of sex did not mean or include discrimination because of sexual orientation. This is not simply because the vast majority of the legislators at the time were staunchly homophobic – though of course they were – but because in common parlance the two terms had nothing to do with each other. And even today they remain separate concepts in common parlance. More importantly in my view, the terms have separate legal meanings. Since 1964, Congress (along with many states) has enacted various measures that have expressly distinguished between between sex discrimination and sexual orientation discrimination. Congress has, in other words, established two legal concepts with two distinct legal meanings. It is well established that terms in related statutes are to be interpreted together wherever possible as if they were one law, and similarly that courts are to presume Congress has not adopted superfluous language (see Scalia & Garner, Reading Law at 174-179 & 252-255). Treating sexual orientation as a necessary subset of sex would run afoul of these basic textualist principles.

To be sure, there is a good argument that transgender discrimination would in fact be captured by the law, but this may ultimately depend on how the term transgenderism is defined or understood. If gender is inextricably linked to sex, then discrimination against a transgender person would arguably be discrimination because of sex. If, on the other hand, gender is fluid and is independent of sex as some have argued, then discrimination because of gender would not be tantamount to discrimination because of sex. Intuitively, it does seem difficult to divorce sex from gender. To take an example that could have certainly arisen in 1964, if a woman is fired for coming to work wearing a suit and men’s dress shoes (instead of a dress and high heels) or for otherwise being insufficiently feminine, we would be inclined to say that this is discrimination because of sex since it was only because she was female that she was forced to dress and act in a feminine manner. However, the same obvious connection does not exist between sex and sexual orientation.

Mark Mancini argues that Gorsuch J.’s opinion is an example of what Professor Ruth Sullivan calls “dynamic” interpretation – that is, allowing new phenomena to come within the ambit of the original meaning. But dynamic interpretations are only to be applied where the text, context and purpose of the provision suggest they are appropriate. And even then, it is a spectrum – the level of dynamism being subject to these factors. Here, we actually have a relatively “static” legislative term in “sex”. The term “sex” is included in a list alongside other grounds including race, color, religion and national origin. Each of these terms had a particular meaning both in law and in common parlance in 1964. Admittedly, the legislative context does suggest some level of dynamism – there is no indication that sex is meant to be confined only to women, for example. And the absence of “solely” or “primarily” before the words “because of” suggests a more relaxed causation standard. But, conversely, there is no indication of an open-ended phrase that is intended to encompass new phenomena as they arise from time to time. More to the point, sexual orientation is not a phenomenon that has arisen since 1967.

Title VII could have been drafted much differently to prohibit discrimination based on “any immutable personal characteristic.” Had Congress enacted this law, then of course sexual orientation would now fall within its ambit. This phrase would mandate a dynamic interpretation and it would be of no consequence that Congress did not “intend” to prohibit sexual orientation. And this also underscores why the original ordinary meaning analysis is distinct from, and should not be conflated with, the “intent” of the legislators. If the law had been worded in this way, it would make no difference that the vast majority of the legislators in 1967 were homophobic; it would not even matter that they considered sexual orientation to be a lifestyle choice rather than an immutable personal characteristic. From the perspective of 2020, we would properly conclude that those individuals were simply incorrect as an empirical matter. And thus, even though the legislators would not have intended to protect gay people from discrimination, they in fact would have done so when they included such broad language, which at the time of enactment encompassed sexual orientation as a matter of fact.

Some scholars, such as Professor Ilya Somin, have analogized the dissent’s reasoning in Bostock to past laws banning interracial sex and marriage. The argument is that while these laws are obviously discriminatory, many argued at the time that those laws did not discriminate based on race, since blacks could still marry blacks and whites could still marry whites. With respect, this is a dubious analogy. Prohibiting interracial unions is primarily – and indeed solely – discrimination on the basis of race. There is no recognized phenomenon of “racial orientation” either in common parlance or in law. The idea of racial unions has always been inextricably tied up with race itself. Moreover, the logic of the seminal decision in Brown v. Board of Education of Topeka is that separating people based on race is prima facie discriminatory because the minority race – blacks – will invariably be treated unequally. Prohibiting interracial unions is therefore no different from school segregation in its discriminatory effects. By contrast, it has never been demonstrated that discriminating on the basis of sexual orientation invariably leads to the inequality of women (or men), and this is because sex and sexual orientation are two separate and distinct concepts. One could easily conceive of a society that has perfect sexual equality but that also discriminates on the basis of sexual orientation; but a society in which people of different racial backgrounds are forcibly separated could never be equal and would therefore always discriminate on the basis of race.

As a matter of logic – “but for” causation

Even if we accept Justice Gorsuch’s more literal approach, it is still wrong as a matter of logic as the conclusion does not follow from the premises (at least insofar as sexual orientation is concerned; as noted above, I grant there is a good argument that transgender discrimination would be captured by the law). The thrust of his argument is that the term “because of” creates a but for standard of causation. On this standard, sex need not be the only or even the primary motivating factor – it can account for as little as 1% of the discrimination. Following from this premise, he considers the example of two employees, one man and one woman, both of whom are attracted to men. Since only the man gets fired, this is discrimination “because of” sex.

This, however, is to beg the question. The analogy assumes what it sets out to prove – that sexual orientation discrimination necessitates sex discrimination. On a “but for” standard, an element need not be sufficient to cause the discrimination on its own, but it must be a necessary element. To determine if an element is necessary, we must isolate it to determine if the discrimination still occurs without it being present. To offer an unrelated but illustrative example, if a person gets into a serious car accident and is then taken to the hospital where he receives negligent medical treatment and suffers permanent brain damage, the court adjudicating the medical malpractice claim will assess whether the brain damage would have still occurred but for, meaning in the absence of, the medical negligence – whether, in other words, the medical negligence was a necessary element of the ultimate injury.

Coming back to the Title VII context, it is only if we change the variable of sex and all other elements remain the same, including sexual orientation, that we can logically conclude that discrimination was because of sex. However, the analogy offered by Gorsuch J. changes both the sex and the sexual orientation of the employee, from a gay man to a straight woman. Saying that Bill and Sue both like men misstates the issue as it treats “attraction to men” as the constant rather than “same sex attraction,” which is the basis for the discrimination. Thus the proper comparison is not between a gay man and a straight woman, but between a gay man and a lesbian. It is only if those two individuals are treated differently that we can fairly and logically conclude that the discrimination was because of sex.

For example, if a homophobic male employer refuses to hire gay men fearing that they will proposition him, this would be discrimination both on the basis of sexual orientation and sex. The prospective gay male employee would face discrimination in relation to his straight male colleagues and would also face discrimination in relation to his female colleagues, including lesbian colleagues. Sexual orientation discrimination, in other words, may very well involve sex discrimination in particular instances, but it does not follow that one necessarily occurs alongside the other.

Here, again, the racial analogy breaks down. If John, a black man, is fired for dating Jill, a white woman, he has been discriminated against solely “because of” his race. If John were white and all other elements about him remained the same, he would not have been fired. By contrast, in the Bill/Sue analogy, if we change a gay employee’s sex but not any other variable including sexual orientation, that person would still be fired.


On policy grounds, I am in complete agreement with the result reached in Bostock. A modern Western society should not still be debating whether it is permissible to discriminate against people simply because they are gay and it is truly a travesty that Title VII has not been amended to reflect changing values on this subject. The fault of this, however, lies with Congress and not the courts. It is not the judiciary’s role to “update” out of date statutes, and while I do not believe this was the intent of the majority opinion (or at least of its author, Justice Gorsuch), this is certainly its effect.

On the other hand, the decision in Bostock is a win for textualism and the judicial function more generally. Justice Scalia used to lament that the members of his court could not even agree on what it was they were supposed to be doing. Here, there is broad agreement between the majority and dissent. My sense is that while a majority of the justices of the U.S. Supreme Court may not yet be textualists, a plurality has emerged, and this will help ensure that disagreements going forward are rooted in textual and doctrinal analysis as opposed to far more sweeping normative disagreements.

Bostock is also an important reminder that a proper textualist approach should not be confused with a literal or plain meaning approach. The function of the court is not to determine what individual words mean, but what the law means. In undertaking the latter, courts will invariably become immersed in the former, but not to exclusion of a proper contextual approach that considers other statutory provisions, related terms in other statutes, and terms with defined legal meanings. Sexual orientation discrimination does not logically necessitate sex discrimination, but even if it did, the two terms are legally distinct and have been for years. Contrary to many social conservative laments, the problem with Bostock is not textualism; it’s bad textualism.