Article

Literalism in Statutory Interpretation

What Is It and What Is Wrong with It?

In two recent decisions—Bostock v. Clayton County and Niz-Chavez v. Garland—a majority of the Supreme Court claimed to apply a textualist approach to statutory interpretation, and a dissent charged the majority with applying “literalism” instead. But what is literalism and what, if anything, is wrong with it? This Essay borrows a few ideas from the philosophy of language to try to pin down a more precise sense in which the majority opinions in Bostock and Niz-Chavez were arguably literalistic. The opinions may have been literalistic in the sense that they failed to consider how context pragmatically enriched what the relevant statutes asserted by fixing the operative sense of a polysemous word. If that is right, then one problem with such a literalist approach is that it pushes controversial interpretive choices underground rather than giving a linguistic (or any other sort of) argument for those choices.

Introduction

Twice in roughly the last year, a majority of the Supreme Court has adopted a purportedly textualist interpretation of a statute and a dissent has disparaged that interpretation for being overly literalistic. Last term, in Bostock v. Clayton County,1 the Court held that Title VII of the Civil Rights Act of 1964 prohibits employers from discriminating based on an individual’s sexual orientation or gender identity. Justice Gorsuch, writing for the majority, reasoned that such discrimination was necessarily “discrimina[tion] … because of … sex” according to the ordinary meaning of that phrase in 1964.2 Justice Kavanaugh dissented, arguing that the majority’s conclusion rested not on Title VII’s ordinary meaning but on “literal meaning.”3

This term, in Niz-Chavez v. Garland,4 the Court held that the Illegal Immigration Reform and Immigrant Responsibility Act of 1996 (IIRIRA) requires notice of a removal hearing to be sent in a single document to trigger the statute’s “stop-time rule.” Justice Gorsuch, writing again for the majority, reasoned that the ordinary meaning of the phrase “a notice to appear” mandated this result.5 Justice Kavanaugh dissented, claiming once more that the majority erred by relying on “literal meaning.”6 This tactic of charging an opponent with being overly literalistic is neither new nor limited to disputes among the justices.7 Still, given the number of card-carrying textualists now on the Court, we are likely to see more and more intra-textualist disputes like these, where one side accuses the other of mistaking literal for ordinary meaning.

So, it is worth asking: What is literalism and what, if anything, is wrong with it? A simple answer is the following. Textualism is a theory of statutory interpretation that directs judges to enforce statutes in accord with their ordinary meaning—in accord with how a reasonable reader would understand their text in context. By contrast, literalism is not a theory at all, or at least not a theory that anyone claims to hold. It is a strawman of textualism, one that would have judges look only to statutes’ text regardless of their context. The problem with literalism is that it has little in common with how we normally understand texts: we normally understand texts in light of, not regardless of, their context. Literalism thus risks running afoul of both the intentions of the legislators who enacted a statute and the expectations of the people who are subject to it.

This simple answer seems correct as far as it goes, but it is too coarse-grained to fully capture the nature of the justices’ disputes in Bostock and Niz-Chavez. Justice Gorsuch claimed in those cases to be interpreting the relevant statutes in accord with how a reasonable reader would understand their text in context, and he clearly took at least some of the statutes’ context into account.8 Perhaps Justice Kavanaugh’s complaint was that Justice Gorsuch did not consider enough of the statutes’ context. But then, what context did he overlook and why would that context change how a reasonable reader would understand the statutes? In the rest of this Essay, I will borrow a few ideas from the philosophy of language to try to pin down a more precise sense in which the majority opinions in Bostock and Niz-Chavez were arguably literalistic.

I. Semantic and Asserted Contents

To begin, we can distinguish a sentence’s semantic content from what a speaker uses that sentence assert. Semantic content is a property of sentences considered in the abstract rather than on any occasion of their use. A sentence’s semantic content is a function solely of the conventional meanings of its words and the rules of syntax.9 It is roughly what you would get by looking up the definition of each word in a sentence in a dictionary and then combining those definitions according to the sentence’s syntactic structure. Take the following sentence:

(1) I have nothing to wear.

The semantic content of (1) is the conventional meanings of “I,” “have,” “nothing,” “to,” and “wear” combined according to the rules of syntax. It is simply that the speaker (whoever the speaker is) has nothing to wear (in some, undetermined sense of “wear”).

Asserted content, by contrast, is a property of particular uses, or utterances, of sentences. An utterance’s asserted content is the complete proposition that it asserts in the context in which it is uttered. Notice how the semantic content of (1) does not exhaust everything that a speaker who utters (1) normally asserts by it. In most contexts, the asserted content of an utterance of (1) is that the speaker has nothing suitable to wear for this occasion. If a friend invites me to his party, and I respond by saying “I have nothing to wear,” what I mean (and what my friend, no doubt, understands that I mean) is that I have nothing suitable to wear for his party. Of course, we can imagine other, unusual contexts in which the asserted content of an utterance of (1) is rather that the speaker has nothing at all to wear (e.g., where someone utters (1) upon exiting a locker-room shower and finding all of his clothes missing).

When an utterance’s context augments what a speaker asserts in this way, philosophers of language say that the utterance’s asserted content is pragmatically enriched.10 Much of what we assert in everyday speech is pragmatically enriched, as can be seen in the following examples (the pragmatic enrichment is bracketed):

(2) I haven’t had any breakfast [yet today].

(3) You’re not going to die [from this injury/illness].

(4) Sally and Karla went to Chicago [together].

(5) Aiden was late to work, and his boss fired him [because he was late].

As we saw with (1), utterances of (2)–(5) may assert different content in different contexts; although most utterances of (2)–(5) assert the bracketed content, not all do so.

Why do most utterances of (2)–(5) assert the bracketed content? We can distinguish two conceptions of asserted content that would answer this question differently. On a subjective, or “speaker meaning,” conception of asserted content, what a speaker intends to assert by uttering a sentence in some context is what the speaker asserts in that context.11 On this conception, an utterance’s asserted content is entirely a function of the speaker’s communicative intention. It follows that utterances of (2)–(5) typically assert the bracketed content because speakers who utter those sentences ordinarily intend to communicate that content. Since, however, this is an essay about textualism, I want to focus on a different, more objective conception of asserted content.

Let us stipulate that an utterance’s objective asserted content is the complete proposition that a reasonable hearer (or reader) would infer that a speaker intends to assert by an utterance in the context in which it is uttered. Objective asserted content is a function of both (a) the semantic content of the sentence uttered and (b) contextual beliefs that a reasonable hearer would have. Those contextual beliefs may include, for instance, beliefs about the social setting of the utterance, the speaker’s purpose in speaking, how words or phrases are commonly used, and conversational norms directing how people in the speaker’s position are expected to behave. It is these contextual beliefs that, on this view, explain why and when utterances of (2)–(5) assert the bracketed content.

A crucial lesson from our discussion so far is that, even when an utterance is non-figurative and non-ironic, the semantic content of the sentence uttered may underdetermine the utterance’s objective asserted content. When this happens, the semantic content of the sentence uttered still constrains what the utterance’s objective asserted content can be (there is no context in which I could use “I have nothing to wear” to objectively assert in standard English that the Tampa Bay Buccaneers won the Super Bowl). But the semantic content of the sentence uttered is consistent with the utterance’s objective asserted content being multiple propositions, and it is a reasonable hearer’s contextual beliefs that determine which of those propositions it is (if anything does—an utterance’s objective asserted content sometimes will be indeterminate).

II. Pragmatic Enrichment and Polysemy

Statutes, I will assume, are speech by a legislature that function roughly as everyday speech does. By this, I mean that the sentences in a statute have semantic content; the legislature’s utterance of those sentences (the statute itself) has asserted content; and those sentences’ semantic content may underdetermine the statute’s asserted content. I hasten to add, however, that whereas we can ordinarily assume that our interlocutors are fully cooperative, legislatures cannot make that assumption. A legislature’s audience may be motivated to deny or resist understanding what the legislature intended to communicate (to look for “loopholes”). As a result, legislatures tend to craft their utterances much more explicitly than we do, such that the semantic content of the sentences in statutes tends to more closely approximates statutes’ objective asserted content.12

This is not to say that pragmatic enrichment never occurs in legislative speech but only that it is less common and generally more subtle. One sort of pragmatic enrichment that we sometimes see in legislative speech results from a statute’s context fixing the operative sense of a polysemous word. A word is lexically ambiguous just in case it has two conventional meanings. For instance, the word “bank” is lexically ambiguous because it can refer to a type of financial institution or to the side of a river. Polysemy, by contrast, concerns subtle shades of meaning. A word is polysemous just in case it has multiple related senses of the same conventional meaning. “Bank” is again a good example: it can refer to a banking company or to the physical building in which such a company is housed (two senses of the same financial-institution meaning).

Take Smith v. United States,13 in which John Smith offered to trade a gun for cocaine, and the question for the Court was whether he was subject to a statutory sentencing enhancement for anyone who “uses a firearm” in a drug trafficking crime.14 The majority held that “uses a firearm” meant using a firearm in any way, such that Smith was subject to the enhancement.15 Justice Antonin Scalia argued in dissent that “uses a firearm” meant using a firearm for its usual purpose as a weapon, such that Smith was not subject to the enhancement.16 He gave a characteristically colorful example: “[If] someone asks, ‘Do you use a cane?,’ he is not inquiring whether you have your grandfather’s silver-handled walking stick on display in the hall; he wants to know whether you walk with a cane”—whether you use a cane for its usual purpose.17

The word “uses” is polysemous: it can refer broadly to employing an object for any purpose or narrowly to employing an object for its usual purpose.18 When a speaker utters a phrase of the form “uses an x” without qualifying how x is being used, a reasonable reader will often infer that the speaker intends the narrower sense. That is because, in most contexts, a reasonable reader would believe that a speaker who knew that x was being employed for an unusual purpose would have said so. Failing to say so would often violate a conversational norm by omitting information relevant to the conversation. Thus, when Congress uttered the phrase “uses a firearm” without qualifying how the firearm in question was to be used, a reasonable reader would perhaps have inferred that Congress meant using a firearm for its usual purpose as a weapon.

The key point for our purposes is that we cannot tell just by looking to the conventional meanings of the words in the relevant statute and the rules of syntax whether the broad or narrow sense of “uses” is operative. Context, however, may pragmatically enrich and thereby render more determinate the statute’s objective asserted content. Given that polysemy is a ubiquitous feature of language, we should not be surprised that a similar pattern emerges in many other cases:

In Nix v. Hedden,19 the issue was whether tomatoes were subject to a tax imposed on “[v]egetables” but not “fruits.”20 The Court noted that, “[b]otanically speaking,” tomatoes are fruits but, “in the common language of the people,” they are vegetables.21 The Court held that the statute used these words in their colloquial sense, such that tomatoes were vegetables and subject to the tax.22 The words “vegetables” and “fruits” were polysemous, but the statute’s context likely fixed their colloquial sense as operative.

In McBoyle v. United States,23 the question was whether someone who flew a stolen plane across state lines had violated a statute that prohibited transporting a stolen “motor vehicle” in interstate commerce.24 The Court observed that it is “etymologically” possible for the term “motor vehicle” to refer to any form of conveyance but, “in everyday speech,” it refers only to conveyances “moving on land.”25 The term “motor vehicle” is polysemous, but the statute’s context likely fixed the sense of motor vehicle that moves on land as operative.

In Taniguchi v. Kan Pacific Saipan,26 the issue was whether a prevailing party was entitled to recoup the costs of translating written documents for litigation under a statute directing the clerk of court to “tax as costs … compensation of interpreters.”27 The Court reasoned that, while “the word ‘interpreter’ can reasonably encompass” persons who translate written documents, there was no indication that Congress intended to “embrace the broadest possible meaning that the definition of the word can bear.”28 The word “interpreters” is polysemous, but the statute’s context likely fixed the sense of oral interpreters as operative.29

These are all at least arguable examples of a statute’s context pragmatically enriching what the statute objectively asserts by fixing which sense of a polysemous word is operative. In each case, contextual beliefs concerning how a word or phrase is commonly used—e.g., “interpreters” is commonly used to refer to oral interpreters—would perhaps lead a reasonable reader to infer that Congress intended to use the word or phrase in that sense.

III. Textualism and Literalism

Before turning to textualism and how it differs from literalism, let me first say a word about the statutory-interpretation debate. The debate, as I understand it, aims to give courts generally applicable instructions on how they should, morally and politically speaking, enforce statutes in “hard” cases—cases where what a statute communicates is either unclear or morally or politically problematic. The debate is thus not a descriptive debate about what statutes in fact communicate or about what the law in fact is. It is rather a normative debate about how courts generally should behave when what a statute communicates is unclear or otherwise insufficient to compel judicial agreement over how to resolve the case at hand.30 Textualism must therefore be a normative claim about how courts generally should enforce statutes to resolve hard cases.

Consider what some of textualism’s more prominent advocates have to say. Justice Scalia argues that courts should “look for a sort of ‘objectified’ intent—the intent that a reasonable person would gather from the text of the law, placed alongside the remainder of the corpus juris.”31 He and his Reading Law coauthor Bryan Garner state that textualism looks to “how a reasonable reader … would have understood the text at the time it was issued.”32 John Manning writes that “textualists want to know the way a reasonable user of language would understand a statutory phrase in the circumstances in which it is used.”33 All of these textualists stress the role of context: Justice Scalia and Garner state that “context is as important as sentence-level text,”34 and Manning notes that language is intelligible only in virtue of “conventions for understanding words in context.”35

Based on these remarks, I take textualism’s core claim to be that courts should strive to enforce statutes’ objective asserted content not only when such content is determinate but even when it is underdetermined. That is, even when what a statute communicates is unclear, courts should strive to enforce what a reasonable reader would be most likely to infer that the legislature intended to assert. Or put another way, textualism restricts the set of admissible arguments in hard cases. It tells courts to confine their reasoning to arguments about what a reasonable reader would be most likely to infer that the legislature intended to assert rather than arguments about what the legislature really intended or which interpretation best advances the statute’s purpose or maximizes social welfare.

Lastly, how does textualism differ from what textualists derisively call “literalism”?36 We can understand pure literalism as instructing courts to look to just the semantic content of the sentences in a statute, without asking how context pragmatically enriches what the statute uses those sentences to objectively assert. Of course, no one is or ever has been a pure literalist. Statutory interpretation cannot even get off the ground without considering some context: one must at least know that the sentences at issue were spoken by a legislature with the intention of making law (rather than written in a novel or graffitied on a highway overpass). Literalism, in practice, is always a matter of degree. Legal interpreters take a literalist approach insofar as they disregard how context pragmatically enriches a statute’s objective asserted content.

IV. Two Senses of “Discriminate”

With these ideas in mind, let us return to Bostock and Niz-Chavez. Title VII renders it unlawful for an employer to “discriminate against any individual … because of such individual[’s] . . . sex.”37 The majority in Bostock held that this provision prohibits employers from discriminating based on an individual’s sexual orientation or gender identity. Its reasoning proceeded in two steps. First, it identified the conventional meaning at the time of enactment of each word in the phrase “discriminate because of sex.”38 It found that “discriminate” meant treating an “individual worse than others who are similarly situated;” that “because of” referred to the legal standard of but-for causation; and that “sex” meant classification as male or female.39 Putting these pieces together yielded the following interpretation:

Interpretation 1. It is unlawful for an employer to treat an individual disadvantageously if the employer would not have done so but for the individual being male or female.40

Second, the majority applied this interpretation to cases where an employer fires an employee for being gay or transgender.41 An employer who fires a male employee because he is attracted to men, but who would not have fired a female employee who is attracted to men, takes an adverse action against the male employee that it would not have taken but for the employee being male. Likewise, an employer who fires a male employee who identifies as a woman, but who would not have fired a female employee who identifies as a woman, takes an adverse action against the male employee that it would not have taken but for the employee being male. In short, “When an employer fires an employee for being homosexual or transgender, it necessarily and intentionally discriminates against that individual in part because of sex.”42

This is a powerful argument: it gives a semantically permissible interpretation of Title VII that coheres with the Court’s sex-stereotyping precedents and directs what I and many others believe to be an equitable result in the instant case.43 But the question that interests us here is whether it is a textualist argument: Did the majority enforce what a reasonable reader would have been most likely to infer that Congress intended to assert by Title VII? According to Justice Kavanaugh, the answer is no. He stressed that, when it comes to phrases like “discriminate because of sex,” courts must look to “the phrase as a whole” because the phrase as a whole “may have a more precise or confined meaning than the literal meaning of the individual words.”44 This, he said, was something that the majority’s “literalist approach to interpreting phrases” failed to do.45

Understanding Justice Kavanaugh’s position is complicated by the fact that neither his dissent nor the other dissent in the case ever put forward its own interpretation of Title VII.46 Both dissents contented themselves with making the purely negative argument that, whatever Title VII means, it cannot mean what the majority took it to mean. I suspect that this was a strategic choice—one that the dissents made because there is no interpretation of Title VII’s objective asserted content that both supports their position and fits easily with the Court’s precedents interpreting Title VII. The result is that it is not obvious why Justice Kavanaugh believed that looking to the phrase “discriminate because of sex” as a whole should have made a difference to the majority’s interpretation. What follows is my own charitable reconstruction of his argument.

Recall that, in Smith, the word “uses” was polysemous, and the question was whether “uses a firearm” referred broadly to using a firearm in any way or narrowly to using a firearm for its usual purpose. Likewise, “discriminate” is polysemous: it can refer broadly to any adverse differential treatment (as the majority presupposed) or narrowly to adverse differential treatment that, either intentionally or unintentionally, disadvantages one social group relative to another.47 When a speaker utters a phrase of the form “discriminate against a because of a’s x” and x is an immutable trait, a reasonable hearer might infer that the narrow sense of “discriminate” is operative. That is, a reasonable hearer might infer that the speaker intends to assert something about adverse differential treatment of a that disadvantages the class of x to which a belongs relative to other classes of x.

Why might a reasonable reader make that inference with respect to Title VII? I can only gesture in the direction of an answer here. One reason might be contextual beliefs about how this kind of phrase is commonly used in everyday speech.48 Another reason might be contextual beliefs about events leading to up Title VII and what those events suggest that Congress’s purpose in enacting Title VII was. In this regard, Justice Alito’s dissent surveyed antidiscrimination laws predating Title VII and concluded that, by 1964, the concept of discrimination “because of” sex was understood to be “part of the campaign for equality that had been waged by women’s rights advocates for more than a century, and what it meant was equal treatment for men and women.”49 So perhaps a reasonable reader in 1964 would have inferred that Congress intended to assert the following by Title VII:

Interpretation 2. It is unlawful for an employer to take any adverse action against an individual that, either intentionally or unintentionally, disadvantages females relative to males or vice versa.

When an employer fires an employee for being gay or transgender, the employee’s sex is a but-for cause of the employer’s action; the employer discriminates in the broad sense (the majority was certainly right about that). But the employer does not, or at least does not obviously, discriminate in the narrow sense.50 Thus, if Interpretation 2 is the most plausible interpretation of Title VII’s objective asserted content, then textualism likely requires courts to dismiss claims like those brought in Bostock. Of course, whether Interpretation 2 is the most plausible interpretation of Title VII’s objective asserted content is debatable; this is a complex and fraught issue,51 which we cannot hope to resolve in the short span of this Essay.52

My point is not that this is a conclusive argument for what Title VII’s objective asserted content is (and certainly not that it is a conclusive argument for what the Court should, all things considered, have decided). My point is rather that the majority opinion gave hardly any argument for why Title VII’s context determines that the statute’s objective asserted content is Interpretation 1 and not Interpretation 2.53 From a textualist perspective, the majority opinion starts in the right place by looking to the conventional meanings of the words in Title VII. But then it essentially stops there, without considering the role of context in selecting among subtly different senses that those words can bear. This failure to address how context might have pragmatically enriched Title VII’s objective asserted content by fixing the operative sense of a polysemous word is the way in which the majority opinion is susceptible to the charge of literalism.

V. Two Senses of “Notice”

Niz-Chavez is a simpler case in that it concerns a statutory provision that has been subject to far less judicial scrutiny and commentary than the provision of Title VII at issue in Bostock, so I will have less to say about it. A noncitizen subject to removal from the United States may request a form of relief known as cancellation of removal.54 To be eligible for such relief, the noncitizen must have been continuously present in the country for at least ten years.55 The clock on those ten years starts when the noncitizen enters the country. The more difficult question is when it stops. Under IIRIRA’s stop-time rule, the clock stops when the noncitizen “is served a notice to appear” that states the time and date of a removal hearing and certain other information.56 The petitioner in Niz-Chavez received this information in two separate documents.57

The Court held that, for purposes of triggering the stop-time rule, the government must send a notice to appear in a single document. Justice Gorsuch’s majority opinion resolved the case based largely on the indefinite article “a.” He stated that “[t]o an ordinary reader … ‘a’ notice would seem to suggest … ‘a’ single document containing the required information, not a mishmash of pieces with some assembly required.”58 Justice Kavanaugh dissented, contending that “[a]s a matter of ordinary parlance … the word ‘a’ is not a one-size-fits-all word,” since the word is sometimes “used to modify a single thing that can be delivered in multiple installments.”59 “Context is critical to determine the proper meaning of ‘a’ in a particular phrase,” and in this case, “the better reading of the article ‘a’ is that it does not require delivery in only one installment.”60

I think that the majority and Kavanaugh were both confused about the word “a”: the sense of “a” is no different in either of their interpretations; in both interpretations, “a” operates as an indefinite article that denotes singularity. The word that actually makes a difference here is “notice.” Much as “bank” can refer to a banking company or to the physical building in which such a company is housed, “notice” can refer to a notification (an abstract thing) or to a document containing such a notification (a piece of paper, a concrete thing). Justice Kavanaugh gave some reason to think that IIRIRA’s context fixed the abstract sense of “notice” as operative by pointing to how similar terms are commonly used.61 We can sensibly speak of “a job application” being sent in multiple installments or “a contract” being formed through successive documents. There is no reason why a notice in the abstract sense could not be split among multiple documents.

Perhaps Justice Gorsuch’s majority opinion in Niz-Chavez is open to the charge of literalism for a similar reason as in Bostock, namely that it failed to explain why IIRIRA’s context fixed the concrete, as opposed to the abstract, sense of “notice” as operative. The opinion did place inordinate emphasis on the word “a.” Yet the opinion also considered some contextual beliefs that a reasonable reader would have. In particular, Justice Gorsuch noted that rules describing other case-initiating pleadings speak of “an indictment” or “a complaint,” and yet “[n]o one contends those documents may be shattered into bits.”62 Those rules at least arguably employ “indictment” or “complaint” in their concrete sense, and perhaps a reasonable reader would infer that IIRIRA employed “notice” in the same way. So while I agree that Justice Gorsuch put too much stock in the word “a,” his majority opinion in Niz-Chavez may not be quite as susceptible to the charge of literalism as in Bostock.

Conclusion

I have argued that textualism is the normative claim that judges should strive to enforce the most plausible interpretation of statutes’ objective asserted content—to enforce what a reasonable reader would be most likely to infer that the legislature intended to assert by a statute. Literalism, by contrast, would have judges look to just semantic content. While no judge is a pure literalist, judges exhibit literalistic tendencies insofar as they focus so much on the semantic content of statutes’ language that they fail to see how context pragmatically enriches what the legislature used that language to assert. This may occur where a statute contains a polysemous word, and a judge looks only to the conventional meaning of that word, ignoring how context might fix one of the word’s multiple related senses as operative. It seems that the majority opinions in Bostock, and to a lesser extent in Niz-Chavez, were literalistic in just this way.

Let me conclude by returning to one of my title questions: What is wrong with literalism? As suggested at the outset, literalism may direct outcomes that contravene both the intentions of the legislators who enacted a statute and the expectations of the people who are subject to it. More interestingly, however, our discussion above also shows how literalism can conceal controversial interpretive choices. Many words, like “discriminate” or “notice,” have subtly different senses of the same conventional meaning, and when a statute uses those words, some choice has to be made about which sense is operative.63 The problem with literalism is that it tends to push this sort of choice underground rather than raising it to the surface by giving a linguistic or policy-based argument for treating the statute as employing one sense over another. Such an approach may have the strategic advantage of making some cases seem easier than they are, but that is a temptation that textualists must avoid if their theory is to have any normative respectability.

a Ph.D. Candidate (Philosophy), Cornell University; J.D. 2014, University of Chicago; B.A. 2011, University of Saint Thomas. I am grateful to Danielle Limbaugh, Andrei Marmor, Anthony Sangiuliano, and Bianca Waked for their comments on earlier drafts of this Essay.

1. 140 S. Ct. 1731 (2020).

2. Id. at 1740–44.

3. Id. at 1824–26 (Kavanaugh, J., dissenting).

4. 141 S. Ct. 1474 (2021).

5. Id. at 1486.

6. Id. at 1491 (Kavanaugh, J., dissenting).

7. See, e.g., Campbell v. Allied Van Lines, Inc., 410 F.3d 618, 623 (9th Cir. 2005) (O’Scannlain, J., dissenting) (“[T]he majority adheres to a decontextualized literalism that even the staunchest defenders of textualism eschew.”).

8. See Bostock, 140 S. Ct. at 1750; Niz-Chavez, 141 S. Ct. at 1486.

9. I am adopting an austere view of semantic content that is not uncontroversial. Compare Kent Bach, Context ex Machina, in Semantics versus Pragmatics 15, 22–29 (Zoltán Gendler Szabó ed., 2005), with Jeffrey C. King & Jason Stanley, Semantics, Pragmatics, and the Role of Semantic Content, in Semantics versus Pragmatics, supra, at 111, 113. This controversy, however, is merely tangential to my argument here. As long as we bear in mind how I am defining “semantic content” and “asserted content,” it should make no difference.

10. See, e.g., Scott Soames, The Gap Between Meaning and Assertion: Why What We Literally Say Often Differs from What Our Words Literally Mean, in Philosophical Essays: Natural Language—What It Means and How We Use It 278, 282 (2008).

11. Cf. Paul Grice, Studies in the Way of Words ch. 5 (1991) (discussing speaker, or utterer’s, meaning).

12. See Andrei Marmor, The Language of Law 34 (2014) (“In ordinary conversations, pragmatic enrichment is the norm, not the exception; in statutory law, it is the exception.”).

13. 508 U.S. 223 (1993).

14. Id. at 227–28.

15. Id. at 225.

16. Id. at 242 (Scalia, J., dissenting).

17. Id.

18. Cf. Marmor, supra note 12, at 122–24 (discussing Smith and polysemy).

19. 149 U.S. 304 (1893).

20. Id. at 306.

21. Id. at 307.

22. Id.

23. 283 U.S. 25 (1931).

24. Id. at 26.

25. Id. at 26–27.

26. 566 U.S. 560 (2012).

27. Id. at 562–65.

28. Id. at 568–69.

29. The same pattern emerges in still other cases. See, e.g., Andrus v. Charlestone Stone Products Co., 436 U.S. 604, 606 (1978); Babbitt v. Sweet Home Chapter of Cmtys. for a Great Or., 515 U.S. 687, 690 (1995).

30. Cf. Richard H. Fallon Jr., The Statutory Interpretation Muddle, 114 Nw. Univ. L. Rev. 269, 278 (2019) (“[T]he deep dispute between textualists and purposivists involves questions of moral and political legitimacy.”); John F. Manning, What Divides Textualists from Purposivists?, 106 Colum. L. Rev. 70, 96 (2006) (“[T]he choice between [textualism and purposivism] must rest on political theory.”).

31. Antonin Scalia, Common-Law Courts in a Civil-Law System: The Role of United States Federal Courts in Interpreting the Constitution and Laws, in A Matter of Interpretation: Federal Courts and the Law 3, 17 (Amy Gutman ed., 1997).

32. Antonin Scalia & Bryan A. Garner, Reading Law: The Interpretation of Legal Texts 33 (2012).

33. Manning, supra note 30, at 81.

34. Scalia & Garner, supra note 32, at 323.

35. Manning, supra note 30, at 79.

36. See Scalia, supra note 31, at 24; Amy Coney Barrett, Assorted Canards of Contemporary Legal Analysis: Redux, 70 Case W. Res. L. Rev. 855, 856–57 (2020).

37. 42 U.S.C. § 2000e-2(a)(1).

38. Bostock v. Clayton Cnty., 140 S. Ct. 1731, 1739–40 (2020).

39. Id.. The majority adopted this sense of “sex” solely for the sake of argument. Throughout this Essay, I use “male” and “female” to refer to biological sex at birth. Cf. Foundational Concepts and Affirming Terminology Related to Sexual Orientation, Gender Identity, and Sex Development, Harv. Med. Sch., https://lgbt.hms.harvard.edu/terminology (May 1, 2020) [https://perma.cc/KSU2-4CG8].

40. Bostock, 140 S. Ct. at 1739–40.

41. Id. at 1742–43.

42. Id. at 1744.

43. See, e.g., Price Waterhouse v. Hopkins, 490 U.S. 228, 251 (1989) (plurality opinion).

44. Bostock, 140 S. Ct. at 1826 (Kavanaugh, J., dissenting). The examples that Justice Kavanaugh gave in this regard were inapt. He noted that “cold war” refers to conflict short of open warfare and that “washing machine” refers to a machine that washes clothes. Id. at 1826. But “cold war” and “washing machine” are open-compound words (each would likely have its own entry in a dictionary). By contrast, “discriminate because of sex” is a genuine phrase.

45. Id. at 1828.

46. Id.; id. at 1783 (Alito, J., dissenting).

47. What about “sex”? It seems that “sex” in 1964 was also polysemous: it could refer narrowly to classification as male or female or broadly to all that is culturally associated with being male or female. See id. at 1784–89 (Alito, J., dissenting) (collecting definitions of “sex”). Yet it is obvious that Title VII’s context fixed the narrow sense of “sex” as operative: a reasonable reader in 1964 would have known that Congress was nowhere near progressive enough to assert a prohibition against discrimination because of sex in the broad sense.

48. See id. at 1828–33 (Kavanaugh, J., dissenting).

49. Id. at 1769 (Alito, J., dissenting).

50. There is an argument that discrimination motivated by sexual orientation or gender identity disadvantages females relative to males by entrenching patriarchal gender roles. See William N. Eskridge, Jr., Title VII’s Statutory History and the Sex Discrimination Argument for LGBT Workplace Protections, 127 Yale L.J. 322, 380 (2017).

51. See generally Kevin Tobia & John Mikhail, Two Types of Empirical Textualism, 86 Brooklyn L. Rev. (forthcoming 2021); James Macleod, Finding Original Public Meaning, 56 Ga. L. Rev. (forthcoming 2021); William N. Eskridge Jr., Brian G. Slocum & Stefan Th. Gries, The Meaning of Sex: Dynamic Words, Novel Applications, and Original Public Meaning, 119 Mich. L. Rev. 1503 (2021).

52. Among other things, we would have to consider the “motivating factor” language added to Title VII in 1991. The majority chose not to focus on this language, see Bostock, 140 S. Ct. at 1739, but some have argued that it supports the majority’s conclusion, see, e.g., Eskridge, supra note 50, at 340–41.

53. The majority did briefly consider something like Interpretation 2 but rejected it because Title VII uses the word “individual.” See Bostock, 140 S. Ct. at 1740. The problem is that this just repeats the error of looking only to the conventional meanings of the words in Title VII without regard to the context in which Congress used those words. Title VII grants a cause of action to individuals and so, of course, speaks in terms of individuals, but that does not tell us whether the cause of action is based on discrimination against such individuals in the broad or narrow sense.

54. See 8 U.S.C. § 1229b(b)(1).

55. Id.

56. Id. at § 1229b(d)(1).

57. Niz-Chavez v. Garland, 141 S. Ct. 1474, 1479 (2021).

58. Id. at 1480.

59. Id. at 1491 (Kavanaugh, J., dissenting).

60. Id. at 1491–92.

61. Id. at 1492.

62. Id. at 1482 (majority opinion) (quoting Fed. R. Crim. P. 7; Fed. R. Civ. P. 3).

63. For this reason, I disagree with Tara Leigh Grove, who suggests that the Bostock majority’s approach—which she calls “formalistic textualism”—constrains judicial discretion. Tara Leigh Grove, Which Textualism?, 134 Harv. L. Rev. 265, 270 (2020). That approach does not constraint discretion; it only pushes discretion out of sight by assuming without argument that one sense of a polysemous word is operative rather than another.

The full text of this Article is available to download as a PDF.