What is it that a judge interprets in a statutory interpretation case? This Article shows that the answer to this question is surprisingly complex. First, the text that a judge interprets is not simply given. Rather, judges must select texts to interpret. Second, the background against which a judge views that text is also not given. Rather, judges situate the texts they interpret within unique, case-specific contexts that they construct from a diverse and unpredictable variety of factors. Selecting and situating: these form the infrastructure of interpretation. Each requires judges to exercise creativity and choice. But opinions tend to present each instead as predictable and inevitable: not a creative and agonistic process, but a basis for assertions of determinate meaning. This Article illuminates how contestation and indeterminacy permeate legal interpretation even as judicial opinions seek to fix and finalize meaning.

How do judges explain why they select the texts they do? How do they justify situating those texts in some factors but not others? How do they substantiate the way they characterize the factors they choose? Asking how opinions address their selecting and situating choices reveals how very unevenly they fulfill their basic obligation of giving reasons for their conclusions. Recognizing selection and situation opens up other lines of analytic and normative inquiry as well. For instance, it facilitates evaluations and comparisons that do not depend on judges’ expressed preferences or commitments. It also provides a robust way to analyze judges’ statutory interpretations with respect to the normative questions that interest a particular commentator, instead of the normative values prepackaged by prominent theories like textualism and purposivism.

Indeed, my approach highlights the limitations of those theories. Purposivism and textualism do not recognize that judges select text to interpret and drastically oversimplify how judges situate that text, leaving judges with little guidance about the very choices on which interpretation is based. This failure may not be too surprising: these theories prescribe what interpreters ought to do, rather than explain what they, in fact, do. In contrast, my contribution helps us understand the practices through which legal actors justify interpretations, claim legitimacy, and set the terms of valid legal argument.


I.  Reframing Interpretation

Interpretation requires an object: a text, an act, a concept, a something to be interpreted. An interpreter must pick out that object. How is that picking out accomplished? Studies of legal interpretation rarely ask. Commentators argue about interpretive approaches, but tend to treat interpretation’s object as given. It is not. Legal interpreters select the objects they interpret. They are not entirely free in making such selections, of course. Like all communicators, they work within the constraints of their genre; and like all legal actors, they are subject to society’s understanding of what constitutes legitimate action. But they are not entirely bound, either. As communicators, they can creatively deploy and combine a variety of rhetorical moves. And as legal actors, they help shape the very parameters of legitimation to which they are subject.

An object of interpretation, moreover, needs a background to make its contours visible. Where does this background come from? Commentators agree that context matters for interpreting legal language, but rarely ask how that context comes to be. They sometimes talk as though interpreters were limited to a small set of preexisting contexts, each tied to some set of objects. Far from it. To situate the thing they interpret against a background that gives it shape and meaning, legal interpreters cobble together diverse factors with unpredictable contours. They attribute to these factors particular attributes, and claim—explicitly or not—that each is relevant to the interpretive project at hand. In this work also, they are both creative and constrained: as bound and as free as they are in the work of selecting the object of interpretation in the first place.

In what follows, I elaborate on these claims. Focusing on statutory interpretation opinions written by Supreme Court justices,1 I show how judicial opinions select text to interpret and how they situate that text within contexts they create. In the genre of the judicial opinion, these creative moves are often presented as inevitable or obvious: expressions of indisputable fact rather than the claims and arguments that they really are. I suggest instead that these two conceptual moments—selecting and situating—are the constitutive forces of interpretation. The assertions and conclusions that opinions call interpretation are merely their precipitates.

I take judicial opinions as ethnographic objects: artifacts that both reflect and affect cultural values and norms. In other words, I analyze not what judges say about themselves, but how judicial opinions create their effects. To select what to interpret, opinions pick and choose among statutory phrases and concepts raised in litigation. Their selected texts open up some interpretive arguments while closing off others. This process, moreover, happens independently of a justice’s preference for a particular theory of interpretation; selecting is, in a sense, conceptually prior.2

To situate the texts they have selected, opinions draw on a wide range of sources, both legal and nonlegal, to create contexts unique to each case. Legal sources turn out to be sufficiently diverse to support divergent interpretations.3 And, despite the implications of doctrine, nonlegal sources turn out to be no less decisive.4 Moreover, when an opinion situates a text, it makes claims about relevance and reality. It posits that the factors it uses are the most useful to interpretation, and that its characterization of those factors is accurate. These claims are central to an opinion’s reason-giving role, yet the classic opinions I discuss here vary dramatically in how they support their claims about both relevance and reality.5

The prevailing theories of statutory interpretation, textualism and purposivism, provide little analytic purchase on these fundamental interpretive practices.6 Neither recognizes that judges select texts to interpret. While they acknowledge that judges situate text in context, each specifies prefabricated contexts for judges to use—an offer the opinions examined here consistently turn down. These theories prescribe how judges ought to interpret, rather than analyzing how they actually do so.7 The problem with textualism and purposivism, I suggest, is not just that most judges adopt a little of both. It is that these theories ignore the infrastructure of interpretation. This makes it especially unfortunate that so much scholarship focuses on these theories rather than on how opinions structure and support their interpretations.

In contrast, recognizing selecting and situating as key conceptual moments in statutory interpretation helps explain how opinions stake their claims. It also provides a firm, flexible basis for a range of normative evaluations not hemmed in by the predetermined stances of interpretive theories. Rather than worrying about whether an opinion conforms to a predetermined set of prescriptive tenets, commentators can use my approach to evaluate the thoroughness of an opinion’s reason-giving, the propriety of its relative valuation of legal and nonlegal sources, the desirability of its image of the law, and much more. Directly recognizing selecting and situating allows for normative evaluation of—and normative debate about—the underlying values that adjudication serves.8

Finally, drawing on scholarship in linguistics, anthropology, and political theory, my approach illuminates how judicial opinions exert effects on the world. Selecting constitutes a particular text as the focus of interpretation; situating constitutes particular factors as relevant to that interpretation. Presenting these things as predetermined, judicial opinions routinely give the impression that their conclusions are inevitable or determinate, as though opinions could escape from the ongoing semiosis of the common-law system and of democratic practice itself. This false sense of finality suggests that legal language forms a discourse apart: a separate sphere that is somehow spared the uncertainties and instabilities that characterize other forms of communication.

Selecting and situating help us see the speech act at the heart of legal interpretation.9 Just as arguing over meaning expresses “the impulse to keep the contest going,” attempts to fix meaning express the countervailing “impulse[ ] of political life . . . to be finally freed of the burdens of contest.”10 My analysis brings both impulses into view at the same time. And it suggests that it is precisely through the indeterminacy of meaning that people claim, contest, and set the conditions for legal legitimacy.

  • 1. I choose this focus for several reasons. Supreme Court opinions set not just the doctrine but the tone for the judiciary, displaying modes of selecting and situating that can be taken up by other judicial interpreters. Their frequent dissents and concurrences help me highlight these practices’ creative nature. And the familiarity of many of the cases I discuss allows me to consider why selecting and situating have not previously been recognized as building blocks of interpretation. Of course, my focus is necessarily partial. The Supreme Court is a unique institution; lower courts may display their own selecting and situating patterns. Statutes are just one thing that legal actors interpret; selecting and situating may work differently with other legal objects. And courts are certainly not the only interpreters of statutes; administrative agencies, among others, play a central interpretive role in the modern state. See generally Jerry L. Mashaw, Norms, Practices, and the Paradox of Deference: A Preliminary Inquiry into Agency Statutory Interpretation, 57 Admin L Rev 501 (2005); Peter L. Strauss, When the Judge Is Not the Primary Official with Responsibility to Read: Agency Interpretation and the Problem of Legislative History, 66 Chi Kent L Rev 321 (1990). Still, I hope that my discussion here contributes both to the literature on statutory interpretation and to an evolving understanding of legal interpretation more generally.
  • 2. This is why selecting different focal texts is one way that opinions that seem to follow the same interpretive approach can arrive at different interpretive conclusions. See Part II.A.
  • 3. See Part III.A.
  • 4. See Part III.B.
  • 5. See Part III.C.
  • 6. From the vast literature elucidating and debating these theories, I draw particularly on, for example, Richard H. Fallon Jr, The Meaning of Legal “Meaning” and Its Implications for Theories of Legal Interpretation, 82 U Chi L Rev 1235 (2015); Robert A. Katzmann, Judging Statutes (Oxford 2014); Richard H. Fallon Jr, Three Symmetries between Textualist and Purposivist Theories of Statutory Interpretation—and the Irreduc­ible Roles of Values and Judgment within Both, 99 Cornell L Rev 685 (2014); John F. Manning, What Divides Textualists from Purposivists?, 106 Colum L Rev 70 (2006); Lawrence M. Solan, The New Textualists’ New Text, 38 Loyola LA L Rev 2027 (2005); Antonin Scalia, A Matter of Interpretation: Federal Courts and the Law (Princeton 1997) (Amy Gutmann, ed); William N. Eskridge Jr, Dynamic Statutory Interpretation (Harvard 1994); and T. Alexander Aleinikoff, Updating Statutory Interpretation, 87 Mich L Rev 20 (1988).
  • 7. See Part IV.A. See also William N. Eskridge Jr and Philip P. Frickey, Statutory Interpretation as Practical Reasoning, 42 Stan L Rev 321, 322 (1990) (“[E]ach theory rests upon and subserves important values that should be considered when interpreting statutes.”).
  • 8. See Part IV.B. Professors Daryl J. Levinson and Mark Kelman have each shown how expanding or contracting a view along one axis of interpretation, such as time, can change a legal outcome in ways doctrine does not recognize. See Daryl J. Levinson, Framing Transactions in Constitutional Law, 111 Yale L J 1311, 1316–18, 1326–32 (2002); Mark Kelman, Interpretive Construction in the Substantive Criminal Law, 33 Stan L Rev 591, 600–16 (1981). Like Levinson and Kelman, I hope to “expos[e] the substantive issues and judgments buried beneath empty and anodyne doctrinal [and theoretical] rhetoric.” Levinson, 111 Yale L J at 1314 n 2 (cited in note 8). And I, too, show how legal interpretation depends on an interpreter’s frame. The frames I present, however, do not come preconstructed: there is no one axis—time, group, and so on—along which a frame becomes larger or smaller. Rather, I highlight the essential bricolage of interpretation.
  • 9. See Part IV.C. Utterances that constitute the conditions they refer to are known as creative utterances. Michael Silverstein, Shifters, Linguistic Categories, and Cultural Description, in Keith H. Basso and Henry A. Selby, eds, Meaning in Anthropology 11, 33–34 (New Mexico 1976) (distinguishing between utterances in which an “aspect of the speech situation [is] presupposed by the sign token,” such that one cannot understand a word without some shared knowledge about its situation of use, and a creative usage, which “make[s] explicit and overt the parameters of structure of the ongoing events” and brings some aspect “into sharp cognitive relief”). The most widely known kind of creative utterance is the performative or speech act. See J.L. Austin, How to Do Things with Words 4–7 (Harvard 1962); John R. Searle, Speech Acts: An Essay in the Philosophy of Language 16–19 (Cambridge 1974).
  • 10. Bonnie Honig, Political Theory and the Displacement of Politics 2, 14 (Cornell 1993).

II.  Selecting: Choosing a Focal Text to Interpret

Judicial opinions often state that interpretation starts with the statutory text.11 But how do judges know what that text is? That is, how do they determine the proper object for their interpretation? For the most part, legal analysis does not ask. It treats this process as unproblematic: a predictable precipitate of the disagreement that gave rise to the lawsuit in the first place.12 The subject of a case supposedly makes it obvious what it is judges should interpret. In any event, it is litigants, not judges, who put statutory terms at issue. Commentary and doctrine tend to assume that judges take the case as they find it, interpreting the terms they are presented with.

This easy assumption turns out to be wrong. As this Part shows, considerable judicial labor can go into determining the object of statutory interpretation. Sometimes acknowledged, sometimes obscured, this labor involves choices about what text to interpret. At a minimum, judges choose whether to focus on statutory text at all; which swatch of text to focus on; which part of a swatch to emphasize; and how to parse a part. Text selection specifies the focal text whose meaning is to be determined. It is thus the condition of possibility for interpretation.

Text selection also has profound implications for the ensuing interpretative process. Legal texts come heavily populated with meanings: they are always intertextually interacting with discursive trajectories that imbue them with implications and place them within worlds of reference.13 Because of this intertextual interplay, selecting a focal text opens up some arguments and closes off others; makes some conclusions obvious and others absurd; and helps determine what might be relevant to interpretation. Selecting the focal text is the axis on which interpretation turns.

A.    How Opinions Select Focal Texts

The way judges select their focal texts usually goes unacknowledged. Yet it is sometimes in plain view and even forms an explicit subject of debate.

1.   Selecting statutory phrases.

In the statutory interpretation classic Babbitt v Sweet Home Chapter of Communities for a Great Oregon,14 for instance, the majority interpreted the word “harm,” while the dissent instead focused on the word “take.15 The Endangered Species Act of 197316 made it “unlawful for any person” to “take any” endangered wildlife.17 The Act defined “take” as “to harass, harm, pursue, hunt, shoot, wound, kill, trap, capture, or collect, or to attempt to engage in any such conduct.”18 Fish and Wildlife Service regulations implementing the statute interpreted the term harm in the statute’s definition of take as “an act which actually kills or injures wildlife,” including through “significant habitat modification or degradation.”19 Commercial users of forests challenged this regulation, arguing that the prohibition should be “limited . . . to direct applications of force against protected species” with intent to injure them, and should not extend to habitat modifications that affected species indirectly.20

The majority opinion, written by Justice John Paul Stevens for six justices, evaluated this argument by asking whether the statutory term harm must be limited to direct force. The major­ity concluded that it did not have to be. The opinion noted that dictionary definitions do not limit harm to direct force,21 and that the statute’s “broad purpose” and Supreme Court precedent suggested that direct force was not required for the prohibition to kick in.22 Further, imputing an intent requirement to harm would make it overlap with other terms in the definition, rendering it superfluous.23

In contrast, Justice Antonin Scalia’s dissenting opinion, for three justices, identified the text at issue not as harm but as take: “The term ‘harm’ in [the statutory definition] has no legal force of its own . . . , for the only operative term in the statute is to ‘take.’”24 For instance, “[a]n indictment or civil complaint that charged the defendant with ‘harming’ an animal protected under the Act would be dismissed as defective,” because it is taking that is prohibited.25 The dissent took take as the thing to interpret.

Harm versus take: What difference does it make? In Sweet Home, the difference between reversing and affirming. The two sides’ divergent focal texts justified divergent interpretations. For the dissent, take came prepopulated with centuries of meaning. A “term [ ] as old as the law itself,” take was “a term of art deeply embedded in the statutory and common law concerning wildlife.”26 This history imbued it with an enduring definition: “when applied to wild animals, [to take] means to reduce those animals, by killing or capturing, to human control.”27 Scalia recognized that the statute provided its own definition of the term: “If ‘take’ were not elsewhere defined in the Act, none could dispute what it means.”28 But take’s historical pedigree took definitional primacy. On the dissent’s approach, the Court’s job was to interpret the common-law term take as it appeared in the statute. Being defined in the statute, in other words, could not remove its common-law signification. Harm, on this reading, could not enlarge or alter the meaning of take; rather, harm itself had to be read in light of the common-law meaning of take.29

Focusing on harm, the majority was not so constrained. It drew, for instance, on a 1982 statutory amendment authorizing the agency to issue a permit for an otherwise-prohibited taking “if such taking is incidental to, and not the purpose of, . . . an otherwise lawful activity.”30 To the majority, this authorization indicated that, for the purposes of the Endangered Species Act, taking need not always be purposeful: there would be no need to specially permit “incidental” taking if it were not prohibited to begin with. Thus, “Congress understood [the original provision] to prohibit indirect as well as deliberate takings.”31 For the majority, the statutory definition, with its inclusion of harm, modified the common-law implications of take. Even if it had no operative “legal force of its own,”32 harm was an operative definition: understanding what take meant in the statute required interpreting harm. So harm formed the proper focus for interpretation, and its range of meanings informed the meaning of take, which the majority concluded spanned “indirect as well as deliberate takings.”33 For the dissent, in contrast, there was no such thing as “indirect[ ]” takings.34 The common law lacked a concept of incidental taking; including harm in the statutory definition could not change that.

Note how selecting the focal text in Sweet Home largely determined the path that interpretation would take. Take’s history as a common-law term of art emphasized the willful, individualized aspects of human relations to wildlife based on the prototypes of hunting and trapping. Harm’s diffuse range of meanings better supported an interpretation independent of human will or intent. After all, we often speak of someone being harmed accidentally, or doing harm without meaning to—indeed, the concept of negligence is based on this possibility. Each opinion considered the relation between the statute’s prohibition on taking and its definition of that term. But the way each opinion figured that relationship—with harm taking precedence in one and take taking precedence in the other—predicted the interpretation it would reach.

I do not mean to suggest that these connections are necessary or inevitable. A judge selecting take as the focal text could surely still reason her way to an interpretation that did not require intentional injury. Nonetheless, particular terms exist within webs of historical, social, and legal relationships. These intertextual connections can link terms to particular spheres of meaning and orientations to the world.35 Such connections do not necessarily define a term conclusively, but they can make some ways of characterizing it easier, more natural, than others. Through such intertextual resonance, a text can seem to suggest a path for its own interpretation. This is why selecting the focal text of interpretation is so important: because legal texts never exist in isolation but sit within larger spheres of meaning, understanding, and normative commitment, each option opens up certain arguments while making others more difficult.

Judicial opinions rarely acknowledge that they select focal text, nor do they usually discuss the reasoning behind their selection. It is likely that, at some level, underlying worldviews and normative commitments guide text selection. Note, though, that neither selecting focal text nor exploring the commitments that motivate that selection is comfortably captured by the two most prominent approaches used to discuss statutory interpretation: textualism and purposivism.36 A textualist deciding Sweet Home might reasonably focus on take, the operative legal term that the statutory definition defines. But she might also focus on harm, the operative definition that Congress enacted in the statute. A purposivist might focus on harm as the term that reveals what Congress wanted the taking restriction to accomplish; but she might also focus on take as expressing Congress’s overarching goals for the restriction. Each choice could be compatible with each theory. Neither theory’s tenets motivate, or explain, the selection through which each side justifies its interpretive conclusions.

2.   Selecting parts of statutory phrases.

Selecting entirely different sections of the statute—such as take as opposed to harm—is not the only way judges choose different focal texts. Opinions may emphasize different parts of the same phrase in ways that make certain interpretations more or less natural. King v Burwell,37 for instance, asked whether the Affordable Care Act38 made federal tax credits available only to those who purchased health insurance on a state-run market, or also to those who used a federally run market.39 The statute required states to organize such markets, or “Exchange[s],” but also provided that the federal government would organize markets for states that did not.40 The statute provided for tax credits for eligible individuals who purchased health insurance on “an Exchange established by the State under section 1311 of the Patient Protection and Affordable Care Act [codified at 42 USC § 18031].”41 Both the majority and dissenting opinions at the Supreme Court took this statutory phrase as their object of interpretation. But the dissent focused on the phrase established by the State, while the majority focused on Exchange . . . under section 1311.

Written for a six-justice majority, Chief Justice John Roberts’s opinion presented the statute as dealing with a single type of thing—“an Exchange”—rather than with a federal Exchange and a state Exchange. “[T]he Act requires the creation of an ‘Exchange’ in each State” and “provides that the Federal Government will establish the Exchange if the State does not.”42 Thus, the statute established one kind of marketplace—an Exchange—that could be run by different bodies. “This case is about whether the Act’s interlocking reforms apply equally in each State no matter who establishes the State’s Exchange.”43 Whether run by the state or the federal government, an insurance marketplace is “the Exchange” in a given state: “the State’s Exchange.”44

The background understanding that the statute contemplated only one kind of marketplace—an Exchange—allowed the majority opinion to weave statutory provisions together to conclude that a federally run Exchange was “an Exchange established by the State under section 1311.”45 Section 1311 required each state to form an Exchange.46 If a state failed to do so, however, § 1321(c)(1) of the Act (codified at 42 USC § 18041(c)(1)) required the federal agency to “establish and operate such Exchange within the State.’’47 The use of the term such Exchange, the majority explained, indicated that the federally operated Exchange was the same thing as a state-operated Exchange. It was an Exchange established under § 1311, by way of § 1321’s requirement that the federal government establish such Exchange under certain circumstances.

On this reading, the notion of Exchange was a single statutory creation. Established by the State was not a stand-alone phrase but one modified by the words that followed it: under section 1311. The phrase under section 1311, in turn, could be given meaning only by looking to § 1311 itself. That section created the concept of an Exchange and provided one way of establishing one; that way of establishing an Exchange was supplemented by § 1321, to which § 1311 was therefore tied.48 The majority, in other words, focused on those portions of the statutory phrase most closely tied to the statute: concepts and processes the statute itself created. Treating an Exchange as one type of thing that came in different flavors—all connected through statutory provisions—made it easier to argue that all Exchanges were subject to the same strictures, irrespective of which institution ran them.

Scalia’s dissent, for three justices, selected a different part of the same phrase: established by the State. It mocked as “absurd” the majority’s conclusion that “when the [statute] says ‘Exchange established by the State’ it means ‘Exchange established by the State or the Federal Government.’”49 The important thing, on this approach, was neither the nature of an Exchange nor the function of § 1311. It was the meaning of established by the State. And that, the dissent explained, was quite obvious: “Words no longer have meaning if an Exchange that is not established by a State is ‘established by the State.’”50 Focusing on established by the State made it easier to argue that different kinds of objects, perhaps subject to different strictures, were at issue.

Moreover, because established by the State is a phrase that could be found in everyday speech, it was easier for the dissent to imply that the answer to the interpretive question was obvious, not requiring a complex analysis of how parts of the statute worked together.51 In other words, the availability of a simple and straightforward reading of established by the State allowed the dissent to suggest that the entire statutory phrase should be simple and straightforward. In contrast, the majority’s focus on terms whose meanings depended on the statute itself suggested that something other than everyday usage might be at play.

The King dissent thus made a textualist argument for the way it interpreted established by the State. But was there a textualist argument for selecting that part of the phrase in the first place? In Sweet Home, a textualist could reasonably select not the definitional term harm but the verb it defined, take. Similarly, in King, a textualist might reasonably select not the modifying phrase established by the State but the noun that it modified, Exchange.52 Textualist tenets do not preclude selecting established by the State; but they do not necessitate it, either. Purposivism might have more predictive power here: a background understanding of the statute’s overarching goals and the way the tax credit provision fit into them could help motivate a focus on Exchange and section 1311. But the majority’s selection is also compatible with textualism: it takes into account an interlocking set of statutory provisions, makes sense of language that treats all Exchanges as the same sort of thing, and coheres this phrase with enacted statutory text specifying how the system should function.

Textualism and purposivism, then, may describe the attitudes or self-reported commitments of the justices who wrote or voted for the opinions. But they are at best blunt instruments for analyzing how and why the opinions select their focal texts. Worse yet, the theories’ analytic weaknesses may bolster the ability of each opinion to frame its selection as obvious and indisputable, rather than explaining it through underlying beliefs or commitments that others could evaluate.

3.   Parsing selected phrases.

In Sweet Home, the two opinions selected separate focal texts: harm in the definitions section, and take in the statutory provision. In King, the two opinions selected different portions of the same phrase: Exchange . . . under section 1311 as opposed to established by the State.53 But judges can also select the same words and still arrive at different interpretations by parsing their selected text in different ways. For instance, in Smith v United States,54 a statute imposed a mandatory minimum prison sentence on a person who “use[s] [ ] a firearm during and in relation to a drug trafficking crime.”55 The defendant, John Smith, had offered to give someone a gun in exchange for two ounces of cocaine.56 Did that constitute “use of a firearm”?

The majority opinion, written by Justice Sandra Day O’Connor for six justices, asked whether what Smith did with his firearm constituted a use of it. O’Connor reasoned that to use something means simply to “employ” it for some purpose.57 On this reading, one can just as easily “use” a gun to trade for drugs as “use” money to do so. “Surely petitioner’s treatment of his [firearm] can be described as ‘use’ within the everyday meaning of that term.”58 Focusing on the word use, the majority thus parsed the statutory phrase to separate out the verb from its object. The verb use, O’Connor implied, meant the same thing wherever it appeared. The fact that it appeared here in connection with a firearm did not alter or delimit its meaning.

This majority opinion took what I call an additive approach to the statutory phrase. An additive approach treats each word as retaining its full range of meaning irrespective of the other words it co-occurs with. It interprets a phrase by adding together the independent meanings of its constituent words, as O’Connor did here, putting together the full range of meanings for use with that for firearm.

Writing for three justices, Scalia dissented. He did not claim that the majority’s interpretation of use was wrong. Rather, he argued that the majority was interpreting the wrong term. The term at issue was not use, as applied to anything from money to firearms to tobacco. Instead, the proper object of interpretation was “the phrase ‘use[ ] a firearm.’”59 While use was an “elastic” word with a wide “range” of meanings,60 use a firearm was considerably narrower: “to speak of ‘using a firearm’ is to speak of using it for its distinctive purpose, i.e., as a weapon.”61 On this approach, use did not have the same meaning wherever it appeared. The things it appeared with could indicate its proper range of meaning. Although one can use a cane as a decoration, Scalia analogized, “[w]hen 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.”62

In contrast to the majority’s additive approach, the dissent here took what I call a consolidating one. It treated the words as an integral whole—a set phrase or an idiom—and not a collection of words to be put together.63 On this view, a phrase can mean something less, more, or other than the sum of its parts. Approaching the statutory text in a consolidating way, as an idiom, enabled the dissent to argue that use a firearm should be interpreted to mean what that idiom normally means: to actualize some dangers or characteristics particular to a firearm, such as shooting or threatening to shoot someone. The majority’s choice of an additive approach, in turn, enabled its conclusion that any kind of use at all would satisfy the standard. By choosing to see the same statutory text as two different types of expression, the opinions arrived, quite naturally, at opposing conclusions.64

Note that both opinions in Smith fit comfortably within the tenets of textualism, each sticking closely to the focal text and situating it in examples drawn from ordinary speech and dictionary definitions.65 Yet sharing fundamental interpretive commitments did not prevent them from reaching opposite conclusions. From the vantage point of textualism as well as purposivism, the divergence of opinions in Smith is quite mysterious. That is because neither theory recognizes the crucial fact that allows for Smith’s opposing opinions: judges select focal texts to interpret. Still less can these theories illuminate the patterns that structure selection or the creativity that suffuses it.

4.   Selecting concepts raised by the litigation.

The examples so far show how opinions can engage in the process of selecting—and thereby foreshadow different interpretive conclusions—by choosing different focal texts; emphasizing different portions of a focal text; and parsing focal text differently. Other times, judges can instead select an issue, concept, or subject matter that lies outside the statutory text. Food and Drug Administration v Brown & Williamson Tobacco Corp,66 which asked whether the Food and Drug Administration (FDA) had the authority to regulate tobacco products,67 exemplifies this approach.

Under the 1938 Federal Food, Drug, and Cosmetic Act68 (FDCA), the FDA regulated “articles . . . intended to affect the structure or any function of the body.”69 Thus, the agency had jurisdiction over articles that both affected the structure and function of the body and were marketed with the intent to do so.70 From its inception, the FDA had declined to regulate tobacco, at least in part because of its inability to show both of these factors.71 But in 1996, the agency asserted such jurisdiction, citing evidence confirming both that “cigarettes . . . are ‘drug delivery devices’” for nicotine, and that their effects on the structure and functions of the body were “‘intended’ under the FDCA because they ‘are so widely known and foreseeable that [they] may be deemed to have been intended by the manufacturers.’”72 Tobacco manufacturers sued.

Writing for a five-justice majority, O’Connor’s opinion first discussed the “essential purpose” of “the FDCA as a whole,” which it described as “ensur[ing] that any product regulated by the FDA is ‘safe’ and ‘effective’ for its intended use.”73 This overarching purpose “generally requires the FDA to prevent the marketing of any . . . device where the ‘potential for inflicting death or physical injury is not offset by the possibility of therapeutic benefit.’”74 The FDA, however, had determined that “tobacco products [we]re unsafe.”75 Thus, the majority concluded, the FDA could exercise jurisdiction over tobacco products in a way that comported with the statute’s essential purpose only by banning them entirely.76 This it could not do, however, because several other statutes had made it clear that Congress did not wish to prohibit tobacco products entirely.77 It would be incongruous, the majority decided, for Congress both to regulate tobacco directly and to give the FDA jurisdiction over it.78

What is the focal text that the Brown & Williamson majority interpreted? It is a bit difficult to say. The bulk of the opinion—roughly eighteen pages—discussed statutes enacted “since 1965,” that is, decades after the FDCA.79 The opinion also spent roughly eight pages discussing the course of conduct that other provisions of the FDCA would impose on the FDA, but these provisions become relevant only in light of the congressional intent imputed from subsequently enacted statutes.80 To the extent that the FDCA’s text played a role in the majority’s interpretation, the majority did not focus on the provision of the FDCA that determined the agency’s scope of authority: “articles . . . intended to affect the structure or any function of the body.” In fact, the majority specifically declined to interpret that jurisdiction-granting provision with respect to tobacco.81 Instead, the opinion viewed other provisions in light of later statutes to derive the FDCA’s “essential purpose.”82

The import of subsequently enacted statutes for understanding the FDCA’s provisions lies not in some particular statutory text, but in the concept of tobacco. The question the majority set out to answer, in other words, was not what a particular statutory phrase or provision meant, but whether Congress intended for the FDA to regulate tobacco. Selecting this focus allowed the majority to resist interpreting the provision that defined the FDA’s jurisdiction and focus instead on later-enacted statutes. The Brown & Williamson majority’s approach eschews a focus on statutory text in favor of an extrastatutory one: a focal object or concept suggested by the subject of the litigation, in light of which the statute’s general implications can be interpreted.

The dissent disputed the inferences that the majority drew from both the FDCA and subsequent laws.83 But it first asserted that looking to subsequent laws was incorrect: the “language and purpose” of the statute were “sufficient to establish that the FDA ha[d] authority to regulate tobacco.”84 Justice Stephen Breyer first noted that the majority “nowhere denie[d]” that “tobacco products . . . fall within the scope of th[e] statutory definition” of the FDA’s jurisdiction; that is, that they were intended to affect the structure or function of the body.85 The dissent’s interpretation selected that definition as its focal text. On this approach, the lawsuit’s central question was not whether Congress intended for the FDA to regulate tobacco individually, but whether tobacco was an “article . . . intended to affect the structure or any function of the body.”

Selecting the statutory text allowed the dissent to posit that the FDCA gave the agency jurisdiction not over individual objects, but over a category of objects. Whether Congress specifically intended for the agency to regulate tobacco was, on this approach, immaterial. The jurisdictional provision authorized the agency to regulate objects insofar as they possessed certain characteristics.86 This suggested a dynamic category: objects could move in or out of jurisdictional scope depending on how manufacturers treated them and what scientists knew about them. Selecting the jurisdictional provision made specific congressional intent about any given object irrelevant.

Note again the limits of textualism and purposivism in analyzing these opinions. Textualism would argue strongly for focusing on the statute’s jurisdictional grant and stopping there. Allowing the FDA jurisdiction over tobacco may have seemingly absurd results, such as forcing the FDA to ban tobacco products, but textualists have a high tolerance for absurd results.87 In any event, those results are absurd only in light of statutes enacted decades later, and later statutes do not reveal what the provision meant at the time it was enacted. Textualist tenets thus do favor selecting a particular text—the one chosen by Breyer in dissent. Purposivist tenets would argue for looking to the goals of the enacting Congress by reading the FDCA’s legislative history, considering the fate of earlier drug regulation, and relating the jurisdictional definition to other provisions. Purposivist te­nets thus also argue for a particular selection—and again it is the one selected by Breyer’s dissent. The majority instead looked to later-enacted statutes’ implications for the FDA’s ability to ban tobacco, a selection that neither textualist nor purposivist tenets support. The utility of these theories for explaining or predicting the selections in Brown & Williamson is, as with the other cases I have discussed, quite limited.

This Section shows that selecting focal text is a primary way opinions stake their claims, frame their arguments, and make their conclusions seem inevitable. My discussion also indicates that purposivism and, especially, textualism are of limited benefit in predicting, assessing, or even recognizing selection practices. Purposivism might tell judges to look to the legislature’s goals, and textualism might tell them to enforce the textual deals the legislature struck. These exhortations make sense only with reference to some focal text: What was the goal of the provision? What deal did this phrase represent? Yet, as this Section has shown, a commitment to an interpretive theory is at best a weak predictor of how a judge will select that focal text. The decisive moment of text selection thus goes unacknowledged.

Seeing selection clearly, in contrast, helps statutory interpretation analysis catch up to litigation practices: litigants’ briefs often suggest particular text selections to judges.88 Judges may be receptive to some such suggestions out of a commitment to a particular interpretive theory. But as this Section shows, these theories often have at best an oblique relation to text selection practices. Recognizing this can push us to consider what other underlying values and views—about the relations among laws, government institutions, or policies—might motivate text selection.89

B.    Selecting Constitutes a Case

As my examples demonstrate, judges have a range of options when selecting the text “at issue” in a case. Should they select a concept or object raised by the litigation (tobacco products in Brown & Williamson’s majority)? Or select statutory text (FDA jurisdiction over products “intended to affect the human body” in Brown & Williamson’s dissent)? Should they select a swatch that resonates with legal concepts outside the statute (“take” as a common-law term in the Sweet Home dissent)? One that resonates with everyday speech (“established by the State” for the King dissent)? Or one grounded in the statute itself (“take” as “harm” per the statutory definition for the Sweet Home majority; “an Exchange . . . under section 1311” for the King majority)? Should they treat a phrase as a consolidated set piece (Smith’s “use-a-firearm” dissent)? Or take each word as a separate unit of meaning and add their meanings together (Smith’s “use + firearm” majority)? Choices about selecting focal text abound.

Judicial opinions show a range of attitudes toward their text selection choices. Some directly address the issue. The Sweet Home dissent excoriated the majority for focusing on harm when the “operative” word was take,90 while the majority dismissed the criticism by noting that harm was how “Congress explicitly defined” that “operative term.”91 The Brown & Williamson majority explicitly declined to address the operative statutory text, while the dissent argued that that text sufficed to decide the issue. Other opinions acknowledge text selection indirectly. The Smith dissent characterized the majority’s focus on use in isolation from the rest of the idiom as a misguided interpretation, rather than a parsing choice.92 Yet other opinions simply do not acknowledge how disagreements about meaning are structured around underlying divergences in the text to which meaning is ascribed. The two sides in King, for instance, debate the meaning of the focal text without noting that their differing focal text selections underlie the reasoning of their conclusions. But whether or not they discuss it, authors of opinions must select text to interpret.

The fact that text selection is often tacit, in other words, should not suggest that it is not at the same time active. One can think of text selection as a technique of interpretation along the lines of the sociologist Professor Marcel Mauss’s classic Techniques of the Body.93 Mauss notes that people perform seemingly intuitive bodily movements differently in different cultures. The way one walks, runs, throws—for “every attitude of the body[,] [e]ach society has its own special habits.”94 These “techniques” are the “effective and traditional” ways of doing things among a particular group of people.95 Such techniques of the body, although learned, can come to feel so intuitive that they seem natural. Such naturalization can help individuals fit comfortably into the societies that practice these techniques: they come to naturally walk the way they have been taught to.96 So, similarly, judges educated in the tradition of American opinion writing can select texts intuitively, without meaning to or realizing it. Yet, like walkers, they still participate in this effective and traditional technique of interpretation.

Moreover, judges do not select from a predetermined set of choices but display creativity in selecting focal texts. Here, I have outlined a few recurring types of choices that judges make when writing opinions. These moves do not form a strict taxonomy but a dynamic range; there are surely examples of other choices, as well as examples that blur the boundaries between these options in creative ways. The point is that writers of opinions select what text to interpret whether or not they acknowledge, or even realize, that they are doing so.97

I am sure readers have themselves noticed some of what I point out in this Part; anyone who teaches these cases is familiar with their contours. But I believe I am the first to recognize the selection of focal text as a pervasive factor in statutory interpretation opinions. My contention finds support in anthropological and linguistic studies of communication, which show that being the focus of discussion is not a natural or inherent feature of an event, object, or text. Rather, the process of communication itself leads things to occupy that focal role.98 This means that being focal is always a contingent and potentially temporary state. Something that emerges as a focal point in one communicative situation does not necessarily do so in another. What participants treat as focal depends on a range of factors: the particularities of the communicative interaction; its larger social setting; and all the constraints of communicative genre, social structure, and power relations that go along with them.99

Moreover, the emergence of a focal object has a dynamic relationship to the background within which it is understood. Text and context create one another in the communicative encounter. A “focal event” emerges through communication as “the phenomenon being contextualized.”100 Just as the background interaction allows a particular object to emerge as a joint focus of attention, it is the emergence of an object as a focus of attention that allows everything else to be characterized as background. What participants “treat as ‘focal’ and what as ‘background’” is not predictable;101 each constructs the other in the process of communication.

To give a feel for how this co-constitutive relation works, imagine someone driving down a street. She may be thinking about a range of things—a lecture she just attended, what to have for dinner, how brightly the sun is shining. At the corner stands a stop sign. She feels the little jolt the car gives when it stops and continues across the street. Is the stop sign a focal text? Perhaps not. Today, it may just slip into other background considerations: lecture, dinner, sun. But if a police officer pulls the driver over for failing to come to a complete stop, the stop sign might click into focus. She might insist that she did, too, come to a complete stop. Suddenly, that little jolt she felt becomes meaningful, pointing in a new way to the focal stop sign. She might wonder if she was distracted by her thoughts about the lecture, or blinded by the setting sun, so that she did not fully notice the stop sign. A range of factors comes together, configured in a particular constellation around the focal text, as the interaction progresses. Or, again, maybe there is no police officer and the stop sign never makes the leap from background to focus. Perhaps instead her thoughts coalesce around dinner, and hunger takes center stage.

This way of moving something from background to object of focus is called entextualization.102 The entextualizing “process [ ] render[s] discourse extractable” and “mak[es] a stretch of linguistic production into a unit—a text—that can be lifted out of its [ ] setting.”103 Neither stop sign nor dinner is inherently focal. But they may become entextualized through a particular communicative experience (with another or with oneself). The concept of entextualization highlights the dynamic, interactive, and creative nature of meaning production. A text becomes focal through the very act of people focusing on it: “[a] text . . . is discourse rendered decontextualizable.”104

Judicial opinions present a particular form of communication. Their specific conventions differentiate them from other communicative genres, as does the fact that they arise from, and resolve, disputes.105 Yet, for all their particularities, judicial opinions also present a form of communication. One can therefore expect that at least some patterns structuring most forms of communication will also appear in judicial opinions. Insights from studies of language and communication cannot predict the analysis of legal texts, but they can certainly guide it.

This Part shows that the active selection of focal text is indeed a central—though underappreciated—aspect of how judicial opinions rationalize their communications and exert their effects on the world. Negotiations over focus and background in ordinary communication happen within accepted rhetorical genres and cultural understandings. Similarly, judicial opinions’ selecting of focal text is constrained by conventions of legal argument and influenced by how litigants and lower courts frame a case. Nevertheless, just as with ordinary communication, selecting focal text is a necessary, and sometimes contested, practice.

Showing that judges must select the text they interpret illuminates how opinions help create the objects of their own analysis. When an opinion asserts, “This case is about . . . ,” it does not merely present a neutral description. Rather, it constitutes the case as being about a particular question; or at least, it attempts to do so. This phrase is thus properly understood not as a description, but as an argument: an argument phrased as a description. It is a kind of “hidden polemic,” which utters its own “word with a sideward glance at someone else’s hostile word,”106 in this case, at other things that someone else could say the case was about.

The assertion that a case is “about” something is thus a creative utterance, a speech act: an efficacious communication that does not merely describe, but helps create, the world it refers to.107 But, being phrased as merely a passive description, it is a speech act that disguises its own efficacious quality. This veneer of nonefficacy allows judges’ choices about what they will interpret to appear as passive descriptions of, or references to, objective facts that are not in their control. Perhaps it is this seeming objectivity that has obscured the central importance of selecting to statutory interpretation.

III.  Situating: Creating a Context for Focal Text

I began the preceding Part with a question: How do judges know what text to interpret? The answer, I showed, was that judges actively select text to interpret. But to interpret a text, one needs to situate it in a broader context: “The assignment of a local meaning to [a] given utterance depends on more global contextual relations which are selectively invoked and activated by discourse participants.”108 How, then, do judges situate a focal text within an encompassing context through which it gains meaning? Commentators often act as though texts naturally carried their own contexts. Attention tends to focus on how a judge interprets the implications of such a context, rather than on where that context comes from. A closer look, however, reveals that situating focal text is far from a natural, obvious, or presumable process. Like selecting text to begin with, situating it requires creative judicial labor.

To situate a text, judges must find some aspects of the world relevant to interpreting it. They are of course constrained by the conventions of legal reasoning. But even within those constraints, the factors available to situate a focal text far exceed what any opinion could actually use. So judges must choose factors within which to situate a text. Is a given legal text best understood when situated against the background of related statutory provisions, or facts about the physical world the statute addresses? Should one understand it by reference to principles that pervade the law or patterns that characterize ordinary conversation? Opinions tend to make the answers to such questions seem more obvious than they are.109

Because a range of options is always available, opinions situating focal text stake a claim that their preferred factors are the ones relevant to interpretation. Relevance, in other words, is made, not born.110 Making something relevant to interpretation is a “dynamic and contingent” social process in which “conflict and misunderstanding” play a “regular role,” putting various interpretive possibilities “‘at risk’ at any given moment in the unfolding discursive activity.”111 Relatedly, when invoking a situational factor, an opinion must attribute some characteristics to it: a related statutory provision says this; a physical fact is that way; a general legal principle holds something; ordinary conversation works a certain way. When creating contexts within which to situate texts, in other words, opinions stake claims about both relevance and reality.

The following Sections show how judges situate focal text within a range of sources. While I divide my discussion into the overarching categories of legal and nonlegal sources, this Part illustrates that each category contains a diversity of possibilities for situating text.112 Judges creating contexts may consciously consider why particular factors seem more relevant than others, and they may actively learn about the characteristics of the factors they invoke. But they also may not.113 As with selecting text, judges must make choices about situating text whether they realize it or not. And as this Part shows, opinion-writing conventions impose few constraints on judges to explain why the factors they invoke are the most pertinent or to substantiate their characterizations of those factors—few constraints, that is, on judicial claims about either relevance or reality. This leaves judges with a great deal of freedom when choosing how to situate focal text.

A.    How Opinions Situate in Legal Sources

Not surprisingly, judicial opinions often turn to legal strictures to help interpret statutory terms. But the law is a multifarious thing, and opinions travel widely in it. They look within the statute they interpret and outside it; they turn to legal understandings that precede the statute, postdate it, and coincide with its enactment; and they draw on overarching, abstract principles. Law’s internal diversity gives writers a range of legal factors so broad that opinions can use legal factors to support opposing interpretations. Situating statutory text against the background of “the corpus juris114 thus does not necessarily constrain interpretation or make it predictable. Because opinions do not often directly challenge one another’s claims about the relevance of a legal factor, however, each can present its chosen sources as the exclusive keys to interpretation. Picking and choosing within the vast legal realm, opinions present their claims about relevance as facts, rather than as arguments.

1.   Intrastatutory and extrastatutory legal sources.

Think back to Sweet Home, in which the majority and the dissent selected different words in the statute as the focal text for interpretation.115 What factors did each opinion consider relevant to interpreting its text? The majority looked first within the statute itself. Recall that the Act made it “unlawful for any person . . . to . . . take” endangered animals.116 The statute defined “take” as “harass, harm, pursue, hunt, shoot, wound, kill, trap, capture, or collect.”117 The agency said that “harm” in this definition included incidental habitat modification, and challengers objected. The majority opinion focused on the word “harm.” Responding to the dissent’s claim that “take” should be the focus of interpretation because it was the legally “operative” word,118 the majority directed attention back to the statute: “Congress explicitly defined the operative term ‘take’ in the [statute], . . . thereby obviating the need for us to probe its meaning.”119

The majority took what I call an intrastatutory approach. It situated its focal text within the context of the statute. An opinion that uses intrastatutory factors to situate a focal text might invoke fairly explicit statutory statements, such as definitions or commands that specifically address the focal text. It might also invoke fairly implicit statutory indications by drawing inferences from other provisions. The Sweet Home majority did both. It drew on the explicit statutory definition of “take,” which included “harm.” It also drew on an implication of the 1982 amendment allowing permits for incidental takings, from which the majority inferred that incidental takings were otherwise prohibited by the statute. Explicitness and implicitness, of course, are not absolute or binary; they are merely useful notions for exploring the range of intrastatutory factors that opinions draw on.

The Sweet Home dissent, in contrast, took what I call an extrastatutory approach. Although it focused on the statutory term “take,” the dissent situated that word in earlier law. The legal term “take,” the dissent insisted, was “as old as the law itself.”120 Its relevant context was, accordingly, also old. The dissent therefore looked to an older, historically situated extrastatutory understanding: that of the common law, a historical period that preceded the statute’s enactment.

Returning to Sweet Home highlights how the selection of focal text sets an opinion’s interpretive orientation. Focusing on take makes it easy to invoke common-law understandings, because the word has such a long pedigree. Focusing on harm makes it easy to invoke more modern and less intentional understandings, because the word lacks those common-law connections and has a broad range of meaning. Thus, each focal text facilitates the making-relevant of a different range of factors.

Crucially, the factors that both opinions draw on are legal: both make relevant a context of legal pronouncement. But their shared invocation of legal understandings does not lead them to shared conclusions. That is because the legal factors they draw on come from different conceptual and historical locations: one intrastatutory, the other extrastatutory in a period that precedes the statute. That difference allows the opinions to invoke factors that are equally legal in character to bolster completely different interpretations.

2.   Sources preceding or contemporaneous with the statute.

It may seem that the Sweet Home opinions were able to support contrasting interpretations because one looked inside the statute while the other looked outside it. But extrastatutory legal sources alone can also bolster opposing conclusions. In Great-West Life & Annuity Insurance Co v Knudson,121 for instance, the majority situated the text in an extrastatutory context that preceded the statute’s enactment, while the dissent looked to legal meanings that were contemporaneous with the statute.

In Knudson, an injured person received money through a settlement; her insurer sought to recoup the funds to offset money it had paid out for her.122 The insurer claimed that the Employee Retirement Income Security Act of 1974123 (ERISA) allowed this recovery when it authorized a party “to obtain [ ] appropriate equitable relief . . . to enforce any provisions of the [insurance] plan.”124 Justice Scalia, writing for five justices, concluded that the “equitable” relief permitted by ERISA included only “those categories of relief that were typically available in equity” as it was “[i]n the days of the divided bench.”125 Although the majority opinion did not specify a precise time period from which the authoritative meaning of “equitable” should be derived, this approach clearly limited the relevant context to an era that preceded the 1938 unification of law and equity in the Federal Rules of Civil Procedure.126

Four dissenting justices found it “fanciful” to claim that the relevant time period for understanding ERISA’s use of “equitable” preceded the statute’s enactment by decades.127 There was little reason “to assume that in 1974 Congress intended to revive the obsolete distinctions between law and equity” as they existed before 1938.128 On the contrary, Justice Ruth Bader Ginsburg wrote in dissent, legal understandings of equity continued to develop after the courts of equity were abolished. “[E]quity [ ] was and should remain an evolving and dynamic jurisprudence.”129 The dissent interpreted equitable with reference to the meaning that legal professionals would have ascribed to it at the time the statute was written.130 Thus, majority and dissent agreed that the proper context for interpretation lay in extrastatutory legal sources: pronouncements outside ERISA itself. But rather than invoking the relevance of developments that preceded the statute, the dissent invoked understandings contemporaneous with the statute’s passage. Each opinion thus claimed the relevance of a different time period’s legal understandings.

Observers can disagree over which interpretation in Knudson is more correct. But it is difficult to take a stand on the matter without also having an opinion about the relative relevance of the historical eras each opinion invokes to support its interpretation. It is worth asking, then, whether the dominant theories of statutory interpretation provide a basis on which to form such an opinion.

Textualism urges judges to interpret a statutory term by following “the intent that a reasonable person would gather from the text . . . , placed alongside the remainder of the corpus juris.”131 To do that, a judge in Knudson might place the term alongside the corpus juris preceding the statute and the Federal Rules of Civil Procedure, as the majority did. But she might just as well place it alongside the corpus juris of the time when the statute was passed, as did the dissent. Each choice facilitates a different outcome; yet textualist tenets could support either.132 The corpus juris of textualist exhortation is a more static and unified object than the world of legal pronouncement that actual opinions consider. More tellingly, textualism does not even recognize that this act of situating requires a decision.

Purposivism can explain the choices in Knudson somewhat better, because the range of sources it validates is not as static or confined. It can potentially incorporate this choice into its general goal of finding how the legislature understood a statute. Purposivism could thus motivate invoking the contemporaneous law of equity by noting that the bulk of the enacting legislature’s legal experience postdated the unification of the divided bench. Still, purposivism, like textualism, does not explicitly recognize the choices judges must make when invoking extrastatutory factors.

The debate in Knudson was thus not between using the corpus juris and ignoring it. It was between making one part of the corpus juris relevant, as opposed to another. Looking at the particular factors that the Knudson opinions invoke, as I have here, reveals how judges claim relevance for distinct factors within what are sometimes described as internally undifferentiated fields.133 It also helps assess how well a prescriptive theory of statutory interpretation predicts or explains a given opinion’s choices—that is, how useful a given theory is in a particular case.

3.   Sources postdating the statute.

I have discussed how opinions can draw on sources that precede the statute at issue, like the Sweet Home dissent and the Knudson majority; and on sources that coincide with it, like the Knudson dissent. Opinions can also look later, at events that postdate the statute’s enactment. Recall how the Brown & Williamson majority determined that, irrespective of whether tobacco products were “articles . . . intended to affect . . . the body,” Congress did not intend for the FDA to regulate them.134 To reach that conclusion, the majority determined that laws passed from 1965 onward had implications for the FDCA, which was enacted in 1938. The majority thus claimed primary relevance for legal assertions that postdated the statute’s passage.

The dissent, in contrast, drew on contemporaneous expressions of congressional attitude from the FDCA’s legislative history to argue that the statute conferred broad and dynamic jurisdiction on the FDA. This fairly standard use of legislative history to deduce congressional understanding has been criticized for imputing a unitary mind to a multimember body—treating Congress as an “it” rather than a “they.”135 The Brown & Williamson majority did not turn to legislative history, but it did attribute unity to a multimember body—and not just one multimember body but many different Congresses over time.136 The opinion did not explain how subsequent laws revealed the meaning the FDCA had at the time it was enacted; nor did the majority claim that the preferences of later Congresses took precedence over those of the one that enacted the FDCA. Rather, it treated Congress like the ship of Theseus: maintaining a unitary character despite continuous change to its component parts.137

Would textualism have predicted this recourse to subsequent statutes? Probably not. Then again, neither would purposivism. As Professor T. Alexander Aleinikoff has written, both purposivism and textualism are usually deployed as “archeological” approaches that aim to uncover what a statute meant at the time it was enacted, not “nautical” approaches that see statutes developing as the legal landscape around them shifts.138 Breyer’s dissent, in contrast, started from the jurisdiction-granting text of the statute, in accordance with both textualist and purposivist principles, and proceeded to the legislative and statutory history, in accordance with purposivist tenets. The point, of course, is not to argue about whether a particular opinion is in fact textualist or purposivist. On the contrary, this discussion suggests that evaluating these opinions in the terms of textualism and purposivism provides quite limited analytic purchase. In contrast, determining which legal factors they claim as relevant helps show how the opinions build their interpretive argument. It also, I argue, provides more substance to a normative evaluation of their conclusions than would a debate about the extent to which either adheres to a prescriptive theory of interpretation.139

4.   General legal principles.

Aside from particular historical periods, opinions also sometimes situate focal texts within general legal principles that they present as timeless or fundamental. The opinions in Green v Bock Laundry Machine Co140 exemplify this recourse to general legal values.141 They also show that judges do not situate texts in such general principles only when more specific indications are unavailable. Rather, judges can choose to situate focal text within overarching, extrastatutory principles even when intrastatutory sources are available.

A prisoner on work release at a car wash sued the manufacturer of a machine that tore off his arm. The manufacturer successfully introduced evidence of the prisoner’s convictions. Federal Rule of Evidence 609(a) mandated that a judge admit evidence of a witness’s prior conviction “only if” its probative value outweighed “its prejudicial effect to the defendant.”142

This provision made sense in the criminal context, in which a jury’s inappropriate prejudice against a witness with a criminal record poses more of a threat to defendants than to prosecutors. But the defendant in Green was a corporation accused of causing an injury. The Rule thus created a strange asymmetry in civil litigation: the court had to admit any prior conviction evi­dence prejudicial to the plaintiff, but could admit prior conviction evidence prejudicial to the defendant only if its probative value outweighed that prejudicial effect.

Each of the Court’s three opinions frankly admitted that it departed from the text to avoid this anomaly.143 But each situated the Rule in different legal factors to bolster that departure. The majority, written by Justice Stevens for five justices, invoked a general legal principle: procedural equality between parties to a civil suit. “No matter how plain the text of the Rule may be, we cannot accept an interpretation that would deny a civil plaintiff the same right to impeach an adversary’s testimony that it grants to a civil defendant.”144 The Constitution grants special protections to criminal defendants,145 the opinion noted, but not to “civil litigants,” who “share equally the protections of the Fifth Amendment’s Due Process Clause,” who “almost always must testify,” and who often take the role of plaintiff or defen­dant more or less by “happenstance[,] based on which party filed first or on the nature of the suit.”146 Given this, “[i]t is unfathomable why a civil plaintiff—but not a civil defendant—should be subjected to th[e] risk” of mandatory admission of damaging evidence.147 With the principle of civil party equality in the background, the majority provided a thorough survey of communications surrounding and preceding the Rule’s passage, concluding that Congress intended that the Rule protect only criminal defendants.148

Concurring, Scalia agreed with the majority’s interpretation.149 He also situated the Rule within a general background legal principle; but he chose a different one. Instead of a proscription on arbitrary distinctions between civil litigants, the concurrence invoked “the policy of the law in general and the Rules of Evidence in particular of providing special protection to defendants in criminal cases.”150 Against this background, it would make sense for the Rule to give criminal defendants protections that it denied others.151

Neither opinion cited any sources for the general legal principles it invoked. The majority’s fleeting references showed that the Constitution created no distinctions between civil parties, not that it prohibited such distinctions.152 The concurrence, likewise, called the Rule’s literal meaning “absurd, and perhaps unconstitutional,”153 but did not specify which constitutional provision it might violate. It asserted that reading the Rule to single out criminal defendants accorded with “the policy of the law in general and the Rules . . . in particular,”154 but identified no text articulating such a policy, either in the Rules or elsewhere.155 Each opinion thus situated the Rule within a general legal principle, but provided no textual basis for its principle. Neither explained why its chosen principle, and not another, should determine the interpretation.

Justice Harry Blackmun’s dissent, for three justices, opted instead for intrastatutory factors. Rule 102 “specif[ied] that [the Rules] ‘shall be construed to secure fairness in administration . . . to the end that the truth may be ascertained and proceedings justly determined’ in all cases.”156 Moreover, because “the Rules . . . by their terms govern both civil and criminal proceedings,”157 the dissent “d[id] not approach the Rules . . . with the presumption that their general provisions should be read to ‘provid[e] special protection to defendants in criminal cases’” or otherwise distinguish between criminal and civil litigants.158 For the dissent, the literal reading of the Rule was unacceptable not because it arbitrarily distinguished between civil parties or because it failed to treat criminal defendants better than others, but because it risked undermining the fairness, truth, and justice that the Rules themselves identified as their principles. Based on these intrastatutory indications, the dissent read the Rule to require judges to determine whether the probative value of prior conviction evidence outweighed its prejudicial effect on a party, rather than only a criminal defendant.159 Intrastatutory factors, then, may be available even when a majority of justices choose not to draw on them to situate the text.

A close look at Green reveals again the limited analytic benefit of textualism and purposivism. The majority’s use of legislative history is guided by adherence to a background principle that it grounds in neither text nor indications of purpose. The concurrence rejects the majority’s use of legislative history on textualist grounds, but itself relies on an assertion about law in general that is not rooted in any textual source.160 Neither mentions any intrastatutory indications of meaning. The dissent, in turn, grounds its analysis in the text of the Rules themselves—and specifically in a part of the text that expresses the Rules’ purpose. It is thus both the most textualist and the most purposive of the opinions.

Is the opinions’ failure to comport to prescriptive theories a problem? If so, I would suggest it is a problem for the theories. While they may provide a way for opinion writers to declare their allegiances, the terms of textualism and purposivism do little to help explain how these opinions create contexts for the text they interpret. And as Green indicates, the theories do little to guide these choices. My analysis, in contrast, illuminates the way judges situate text within a range of contextual factors; how they claim those factors’ relevance for interpretation; and the extent to which they substantiate their depictions of those factors.

5.   Multiple sources.

A final example demonstrates how opinions can creatively combine various legal sources when situating a focal text. In Bob Jones University v United States,161 a statute exempted nonprofit “[c]orporations . . . organized and operated exclusively for religious, charitable . . . or educational purposes” from tax liability.162 For much of the twentieth century, the Internal Revenue Service (IRS) had applied this exemption to all schools, but in 1970 it announced that schools that discriminated on the basis of race were no longer eligible for tax exemption.163 Two schools sued, claiming they were entitled to tax exemption irrespective of their racial policies.164

The majority opinion, written by Chief Justice Warren Burger for seven justices,165 situated the provision within common-law understandings, asserting that the educational tax exemption “was intended to express the basic common law concept of ‘charity,’”166 which required that an institution must “serve a public purpose and not be contrary to established public policy.”167 Like the Sweet Home dissent, the Bob Jones majority thus saw common-law understandings as constraining the meanings of words in the statute.

To figure out what “established public policy” comprised, however, the opinion looked neither to common-law notions nor to the time at which the statute was enacted, but to its own contemporary setting—the time at which the IRS had instituted its policy. Canvassing recent judicial decisions, congressional legislation, and executive action regarding racial discrimination, the majority found that, by the 1970s, “there [could] no longer be any doubt that racial discrimination in education violate[d] deeply and widely accepted views of elementary justice.”168 This approach viewed the common-law notion of public policy as itself historically dynamic. The Bob Jones majority did not explain or even discuss the situating choices that supported its interpretation; rather, it treated them as statements of the obvious.

Alone in dissent, Justice William Rehnquist noted that the statute mentioned neither race nor public policy, and that it listed educational institutions as a separate category from charitable ones.169 The dissent thus rejected the relevance of common-law terminology along with the contemporaneous policy analysis it led to. Instead, Rehnquist would have drawn the bounds of relevance more narrowly, at intrastatutory relations among the statute’s words. This approach fit textualist prescriptions to focus on the words of the statute. It also resembled the approach of the Sweet Home majority, which saw take as defined not by the common law but by the statutory definition. Rehnquist was not in the majority in Sweet Home, though. He joined the dissent, agreeing that take should be defined by its common-law heritage. That dissent, recall, was written by Scalia, the Court’s foremost textualist.

So it appears that textualism countenances both, on the one hand, excluding common-law influences by limiting the bounds of relevance to the statute; and, on the other hand, expanding the bounds of relevance by letting the common law determine the statute’s valence. The differences between the opposing opinions in Bob Jones and Sweet Home hinge on the choice between these options. But textualism’s prescriptions allow a judge to go either way, even while textualists claim that the theory cabins judicial discretion and leads judges to predictable interpretive outcomes.170

Purposivism encourages interpreters to seek signs of congressional goals in a variety of sources and tolerates a more pragmatic and variable interpretive process than textualism. Unlike textualists, purposivists do not generally make strong claims for the predictability and constraints of their approach. The eclectic relevance-making practices of the opinions discussed here, and the fact that a judge could follow a theory’s prescriptive tenets to opposite interpretive conclusions, thus pose less of a challenge to purposivist claims than to textualist ones. Nonetheless, purposivism, like textualism, does not acknowledge many of the factors judges draw on to situate texts. The wide range and internal differentiation of those factors is part of what I illuminate here. While purposivism’s greater prescriptive flexibility may make it more realistic than textualism, this lack of recognition leaves it analytically limited.

Asserting or assuming the relevance of various time periods or principles, the opinions in this Section created contexts that led quite naturally to their particular interpretations. When two opinions reached different conclusions, moreover, it was not because one situated its text in legal factors while the other looked elsewhere. It was because legal factors could support opposing conclusions. This suggests that those who hope to constrain judges by insisting that they interpret statutory text in light of “the corpus juris171 may ascribe more determinacy to that corpus than actual opinions find there.

Opinions, like commentators, rarely remark on the wide range of legal factors they draw on to create contexts for interpretation. They tend to phrase disagreements in terms of interpretive correctness. As this Section’s examples indicate, they rarely challenge one another’s claims of relevance directly. This lack of direct confrontation—or even recognition—may help interpretive claims pose as statements of fact, rather than as the arguments they really are. The next Section shows, moreover, that nonlegal sources play an equally important role in situating focal text.

B.    How Opinions Situate in Nonlegal Sources

Statutes usually apply to circumstances outside legal discourse itself. Accordingly, even judicial opinions that interpret law often situate it within the wider circumstances to which that law applies. Though no discussion could be exhaustive, this Section aims to highlight the broad range of nonlegal sources that opinions use to create contexts for focal texts. Because of the breadth of this range, opinions must always make choices about what nonlegal factors to draw on. An opinion must thus stake a claim—implicit or explicit—that the factors it chooses are the ones needed to interpret the statute. And it must argue—again, implicitly or explicitly—that it describes those factors accurately. My discussion emphasizes the way that opinions drawing on nonlegal factors make claims about both relevance and reality.

1.   Relative relevance.

When the opinions in Chapman v United States172 considered what constituted “a ‘mixture or substance containing a detectable amount of lysergic acid diethylamide (LSD),’”173 the majority situated the text in physical facts while the dissent invoked sociological ones.174 Pure LSD weighs so little that, to be sold, it is “dissolved in a solvent such as alcohol, and [ ] the solution” is used to impregnate a heavier and more easily divisible substance like paper, gelatin, sugar cubes—or something else.175 This “carrier” is then divided and sold. Petitioners argued that the medium on which LSD was sold should be excluded when calculating the drug’s weight for the purposes of a statute that mandated a five-year minimum sentence for distributing more than one gram of a “mixture or substance” containing LSD.176

The majority opinion, written by Rehnquist for seven justices, rejected the petitioners’ argument.177 The opinion related two dictionary definitions of mixture to physical facts about LSD. Webster’s Third New International Dictionary described a mixture as “two or more components . . . that however thoroughly commingled are regarded as retaining a separate existence,”178 while the Oxford English Dictionary (OED) said it comprised “two substances blended together so that the particles of one are diffused among the particles of another.”179 The majority explained that once a carrier is impregnated with the LSD-solvent solution, the solvent evaporates, leaving “[t]he LSD crystals [ ] inside of the” carrier, “commingled with it,” even though “the LSD does not chemically combine with” it.180 Conforming to Webster’s Third, the drug thus “retains a separate existence”: it “can be released by dropping the [carrier] into a liquid or by swallowing” it.181 At the same time, per the OED, “[t]he LSD is diffused among the fibers of the paper”: “[l]ike cutting agents used with other drugs” such as heroin and cocaine, the carrier “cannot be distinguished” from LSD “nor easily separated from it.”182

The majority opinion thus asserted, in the same paragraph, both that LSD could “be released [from its carrier] by dropping the [carrier] into a liquid,” and that LSD “cannot be distinguished . . . nor easily separated from” its carrier.183 The dissent did not dwell on the puzzle of how both of these physical facts could be true. Instead, it situated the statute in another arena of nonlegal factors: sociological facts about “how LSD is sold.”184 The dissent did not have to look far for that information, because the majority itself stated it: unlike most drugs, “LSD is sold by dose, rather than by weight.”185 Indeed, the weight of the drug itself is negligible: “an average dose” of pure LSD “weighs 0.05 milligrams.”186 How much it weighs when it is sold depends primarily on the weight of the carrier, not the pure LSD, and has no effect on the price. “[W]hether one dose of LSD is added to a glass of orange juice or to a pitcher of orange juice, it is still only one dose that has been added.”187 Dose and price, moreover, are both set by the amount of pure LSD, not the amount of orange juice. Thus, while people buying or selling a drug like heroin understood the dosage to include any cutting or diluting agent, people buying or selling LSD understood the dosage to be independent of the carrier.

Drawing on this sociological fact, the dissent argued that the majority’s interpretation would result in arbitrary differences in sentencing among LSD dealers. “A person who sells five doses of LSD on sugar cubes is not [ ] worse [ ] than a manufacturer of LSD . . . caught with 19,999 doses in pure form, but [the sugar-cube seller] is subject to a ten-year mandatory minimum [ ] sentence,” while the manufacturer “is not even subject to the five-year minimum” because pure LSD is so light.188 It would also lead to dramatically longer sentences relative to other drugs: one would have to sell “between one and two million doses” of heroin to receive the same sentence as someone selling just twelve thousand doses of LSD on blotter paper.189 Yet no one had presented a reason to punish LSD dealers more severely than dealers of other drugs.

While both Chapman opinions situated the text in nonlegal factors, then, one invoked physical facts while the other rested on sociological ones. Note that neither claimed that the other’s facts were not true or insufficiently documented. The dissent did not even mention that the majority’s claims that LSD both could, and could not, be easily be separated from its carrier were mutually contradictory. Rather, each opinion presented its own implicit claims about what factors were relevant to interpreting the statute. The majority readily admitted that LSD was sold differently than other drugs and that counting the carrier in the total weight would produce disparate sentences for different carriers. But it did not agree that these facts held the key to interpretation. The relevant factor was instead the physical characteristics of LSD, which led it to become a “mixture” with its carrier. For the dissenters, physical facts about how LSD interacted with its carrier did not resolve the ambiguity of the term “mixture.” Instead, the dissent viewed sociological facts about how dealers and buyers treated the drug as the relevant factors within which to situate the text. Although each opinion drew on a different sphere of nonlegal facts, in other words, facts were not the contested terrain. Relevance was.

2.   Contested realities.

Opinions can also reach opposing conclusions when drawing on similar kinds of nonlegal sources. For instance, the plurality and dissent in Rapanos v United States190 each invoked physical facts about how materials affect water to interpret a Clean Water Act191 provision.192 The Act required landowners to acquire permits before “discharg[ing] [ ] dredged or fill material into [ ] navigable waters.”193 The Army Corps interpreted the statute to require permits before discharging such materials into wetlands adjacent to waters that were themselves navigable, or their tributaries.194 Landowners claimed that this interpretation stretched the Corps’s jurisdiction too far.195

The plurality opinion, written by Scalia for four justices, situated the text against the background fact that “‘dredged or fill material’ . . . does not normally wash downstream.”196 In this empirical context, a landowner discharging such material into a wetland would not normally be considered to release it into “navigable waters.”197 The dissent, written by Stevens for another four justices, also drew on a fact. But it was the opposite of the fact that the majority invoked: “fill material,” the dissent explained, “has the potential to move downstream and degrade the quality of [ ] navigable waters.”198 Situated in this different empirical context, the Corps’s interpretation seemed more modest. A long line of case law had held the notion of “navigable waters” to encompass bodies of water that affected the quality of waters that were in fact navigable.199 If fill material in wetlands moved and entered navigable waters downstream, it would be reasonable to conclude that the Act gave the Corps jurisdiction over those wetlands. Plurality and dissent thus each asserted its own, opposing, truths about the physical world.200

How did the Rapanos opinions substantiate their incompatible claims about physical facts? The plurality, as Professor Todd Aagaard has written, based its factual assertion on “statements made in three amicus briefs,” each of which “relied on [the same] package of training materials” posted “on a private corporation’s web site.”201 These materials, like the plurality opinion itself, did not cite any scientific studies on the topic; nor did either give any other indication that a scientific community accepted its assertion.202 For its claim about the mobility of fill material, the dissent relied on an amicus brief and on a lower court opinion (which itself did not provide a basis for its assertion).203 Neither opinion, in short, “could cite a reliable basis for the [position] it expressed as to th[e] facts” about fill material’s mobility.204

In addition to its factual assertion, the dissent also challenged the plurality’s implicit claim that the key to interpreting the statute lay in determining whether fill material moved downstream. Instead, the dissent proposed a wider sphere of possibly relevant factors. Because the statute aimed to maintain the viability of navigable waters, the really relevant issue was not whether fill material moved, but whether it harmed those waters. And it could do so even without moving downstream: “fill can harm the biological integrity of downstream waters even if it largely stays put upstream” by, for instance, creating “excessive sediment” that interferes with fish reproduction, as shown in a published scientific paper about fish reproduction.205

The plurality, however, responded neither to the dissent’s contrary factual assertion nor to its alternative relevance claim. In this, Rapanos demonstrates the freedom that opinion writers often enjoy when situating legal texts. Even opinions that assert opposing facts about the same thing do not necessarily challenge one another to substantiate their factual statements. And opinions disagreeing on the proper sphere of relevance do not necessarily push one another to justify their respective purviews. From Rapanos, it seems that judges have wide latitude both in deciding what to find relevant and in how to characterize it. Moreover, neither textualism nor purposivism has much to say about how judges should, or do, use nonlegal sources. Yet the way that opinions situated statutory text in nonlegal context determined the outcomes of the opinions in Chapman and Rapanos.

3.   The relevance and reality of language use.

Opinions situating focal text also often draw on assertions about language. Language, of course, is the opinion’s medium of communication. But it can also be a kind of context: opinions often use assertions about language use, history, and perception to situate focal texts. Perhaps most famously, opinions may draw on dictionary definitions.206 Some applaud the use of dictionaries as an objective guide to meaning, but the Chapman majority’s use of somewhat conflicting definitions of “mixture” indicates the range of options writers have among entries, publishers, and time periods.207 In accordance with their selected texts, for instance, the Sweet Home majority drew on a dictionary roughly contemporaneous with the statute’s enactment,208 while the dissent turned to older ones.209

Opinions also draw on other language facts to situate focal text. The Rapanos dissent, for instance, noted that “the very existence of words like ‘alluvium’ and ‘silt’ in our language . . . suggest[ed] that at least some fill [material] ma[de] its way downstream.”210 Scalia’s plurality opinion derided this reasoning as “philological,”211 but such inferences from language history are common. For instance, in Whitfield v United States,212 the Court considered whether a fleeing bank robber who “guided” a resident from one room in a house to another had “force[d] [a] person to accompany him,” triggering a mandatory minimum sentence.213 The defendant argued that the enhanced sentence for forced accompaniment applied only to “substantial” movement, not to the few feet he had walked with the resident.214 Scalia’s unanimous opinion rejected this argument, holding that “accompany” simply meant to “go with” someone, irrespective of distance.215

Whitfield situated its focal text within a number of sources, including passages from Jane Austen’s Pride and Prejudice and Charles Dickens’s David Copperfield.216 Why would novels published by speakers of British English in 1813 and 1850, respectively, be authoritative guides to ordinary American English usage in 1934, when the statute was enacted? Perhaps Scalia in Whitfield took the “philological” approach he had rejected in Rapanos,217 deriving the meanings of words in one place and time from their appearance in another. The opinion itself does not say. Similarly, the opinion quotes a 1995 Supreme Court opinion’s use of “accompany,” but does not explain how that could indicate what the provision meant to a reader when the statute was passed in 1934.218 Perhaps the point was that a bank robber could not reasonably claim that he was not on notice of the punishment because the meaning of “accompany” had become ambiguous through historical change.219 But the opinion again does not say. Although it is widely accepted that language use differs over time and speech community,220 Whitfield did not advert to that possibility.

Whitfield’s silence about how its sources relate to its focal text is unexceptional. Opinions routinely invoke works from eras and areas other than those in which the statute was written to support claims about ordinary language use.221 But they rarely discuss how, and whether, the works they quote support their inferences about the text they interpret. Instead, they situate focal text in contexts whose bearing on interpretation they never clarify.

4.   Weighting legal and nonlegal sources.

Assertions about physical, sociological, and linguistic facts often appear alongside legal factors when opinions situate focal texts. In Brown & Williamson, for instance, the majority situated the statutory command within legal sources postdating the statute’s enactment; the dissent brought in realities outside the law. Brown & Williamson, recall, asked whether the FDA had authority to regulate tobacco products under the statutory grant of jurisdiction over “articles . . . intended to affect the . . . body.” Neither the majority nor the dissent questioned the FDA’s evidence that tobacco products were intended to affect the body.222 But the majority did not find this factor relevant to the question it posed itself, which was whether Congresses over the twentieth century had meant for tobacco to be regulated.

The dissent, in contrast, contended that the law itself made such nonlegal factors decisive. The statute gave the agency jurisdiction over “articles . . . intended to affect the . . . body.”223 Whether a particular Congress meant for some particular thing to be regulated was less relevant than whether that thing belonged to the category that the enacting Congress had laid out in the statute. And that was largely a nonlegal question: particular objects could move into or out of the category as the circumstances of their production and sale changed. For the dissent, the central question was whether something was an article intended to affect the body; answering it required evaluating empirical facts, not congressional pronouncements that postdated the statute.

From Brown & Williamson, one might be tempted to think that legal factors generally win out over nonlegal ones. Perhaps legal practitioners naturally treat legal sources as more important, and perhaps there are good doctrinal reasons for doing so. Yet nonlegal sources are not just common but often decisive. “Supreme Court . . . opinions are chock-full of . . . general statements of fact about the world”224 that not only are “used rhetorically . . . [to] make[ ] [a] position more persuasive,” but also “go to the practical consequences of [ ] decisions” and form “a critical part of the doctrinal inquiry.”225 Given that judges are generally experts in law rather than in any given factual situation to which law applies,226 one might expect that the nonlegal factors within which they situate their interpretations will involve findings in cases below or administrative proceedings. One might also expect that the practice of invoking empirical realities would be constrained by doctrine. Neither is the case.227 Judicial opinions’ factual assertions are not recognized by doctrine, and they draw on a wide range of sources, including administrative records, litigant briefs, amicus submissions, and independent research by justices and clerks.228

Many, moreover, are not grounded in any clear source but are best described as expressions of judges’ beliefs. Opinions sometimes present, as though they were uncontestable facts, nonlegal factors that are culturally, historically, and even individually contingent. For instance, an opinion might assert that “[w]hether to have an abortion requires a difficult and painful moral decision.”229 In fact, “in many countries, abortion is an accepted part of family planning” that is not treated as a difficult moral issue.230 The kind of decision abortion requires is thus an empirical question, not a general fact. Similarly, opinions sometimes repeat contested views about social practice as though they were facts, such as that “[r]eligious organizations exist to foster the interests of persons subscribing to the same religious faith.”231 The purposes that religious organizations serve is a complicated question that scholars have debated for centuries, and “[f]ew who study religion would agree with [Ginsburg’s] statement.”232 Opinions sometimes also base conclusions on logical inferences from premises that are neither acknowledged nor substantiated, such as the claim that market competition among privately run prisons “mean[s] . . . that a firm whose guards are too aggressive will face damages that raise costs, thereby threatening its replacement” by a different firm.233 This inference assumes that governments outsource efficiently to the safest and most effective private prison firms and that prisoners are able to obtain damage awards on a scale that would endanger the economic viability of a private prison corporation. But the opinion provides no reason to believe that either of these assumptions is correct. On the contrary, the Department of Justice under President Barack Obama determined that private prisons provide neither a safe nor an efficient method for incarceration.234

Finally, sometimes it is the very decision whether to treat something as a legal or nonlegal factor that is crucial.235 Words often lead double lives in the law and outside it. A person described as having “actual malice” in an ordinary conversation might be presumed to actively harbor ill intent toward another. But a publisher being sued for libel can be said to have acted with actual malice for much less: merely having a “reckless disregard of whether [a statement] was false” suffices.236 The term means something different in everyday speech than in libel law.237 And a judge in a libel suit who thought that a publisher would be liable only if it actively harbored ill intent toward the plaintiff would easily come to different conclusions than one who thought that reckless disregard of possible falsity sufficed. Whether to situate text in legal terms of art can thus be decisive.

Yet it can seem strangely awkward for practitioners to address this choice straightforwardly. Take, for instance, the recent oral argument in United States v Texas.238 Several states challenged the Department of Homeland Security (DHS) policy Deferred Action for Parents of Americans (DAPA),239 which allowed some people not legally authorized to be in the United States to request that DHS refrain from removing them and grant them work authorization and other benefits.240 Those receiving such benefits are said to have “lawful presence” within the meaning of the benefits-authorizing statutory provisions, but the benefits do not confer a right to remain in the country. “Lawful presence” thus resembles “actual malice”: in its particular area of law, each means something different than it does in everyday speech.

Nonetheless, at oral argument some justices purported not to recognize the concept of legal terms of art.241 The government’s brief described DAPA beneficiaries as both “lawfully present” and “present in violation of the law.”242 “[T]hat must have been a hard sentence to write,” said Chief Justice Roberts.243 “I mean they’re—they’re lawfully present, and yet, they’re present in violation of the law.”244 While the solicitor general replied, Roberts repeated with incredulity that jumps off the written page: “Lawfully present does not mean you’re legally present in the United States. . . . Lawfully present does not mean you’re legally present.”245 Justice Samuel Alito was even clearer on his lack of clarity: “I’m just talking about the English language. I just don’t understand it.”246 “How can [it] be . . . lawful to work here but not lawful to be here?”247 The solicitor general, in turn, seemed to find it oddly difficult to remind the justices that expressions can have technical legal meanings: “It’s—let me just go through the reality here, and—and I’ll give you some sense of just how disruptive a ruling would be to accept [the plaintiffs’] theory on—on who can lawfully work in the United States.”248 Perhaps part of the difficulty lies in the lack of articulation about interpretive practices that characterizes so many of the cases discussed here. Opinions are rarely explicit about choosing to situate texts in legal terms of art, and that very reluctance to topicalize the choice may make it difficult to directly confront it.

As this Part shows, judicial opinions situate focal texts within a wide range of sources. This is not to say, of course, that judges can situate texts any old way they want. In the lead-up to Bob Jones, for instance, the IRS had maintained that it lacked authority to revoke the tax exemption of racially discriminatory schools; it changed course when a court seemed poised to enjoin it to do so.249 Rehnquist’s dissent tells this story disparagingly.250 The majority does not mention it at all. Neither considers the impetus of litigation for agency policy as legitimately relevant to a court’s statutory interpretation.251 Such reluctance indicates the limits that the genre places on relevance. But as this Part shows, those limits are broad and internally diverse to an extent generally not recognized in discussions of statutory interpretation.

In this Section alone, opinions have situated statutory text within the physical world, physiology, psychology, social and economic practices, and linguistic patterns. The preceding Section, moreover, showed opinions situating texts in legal sources that included the statute at issue; historical times before, during, and after it; and general principles. This list, though long, does not exhaust the range of factors within which judges can situate legal texts. Quite likely, no list could. Law is internally diverse and touches on all aspects of human life. And judges writing opinions can pick situating sources anew in each case, claiming relevance for new sources or combining sources in new ways. The possibilities for situating focal text are potentially endless.

Strikingly, however, many practices explored in this Part remain unnoted by both doctrine and commentary. Opinion authors situate texts within factors that seem relevant and valid to them. But they often do not face a call to explain why those, and not other, factors are relevant, or to substantiate how they characterize them. This allows opinion writers to present their situating choices as natural and factual—not as the argumentative claims that they really are.

C.    Situating Makes Claims about Relevance and Reality

The opinions discussed in this Part show how contexts are not given in the world but created by legal interpreters. Judges draw on a wide and differentiated range of factors to creatively produce relevant contexts within which to interpret their chosen texts, and to creatively undermine others’ interpretations. This process, moreover, is studded with silent claims.

First, opinions make claims about relevance. By invoking some factors and not others, they formulate arguments—often implicit ones—about what and who counts. Which speakers, writers, or communities should they look to for clues about how language works? Which facts about the physical, social, or psychological world should they lean on to make their decisions? Which legal eras, concepts, and texts should determine their views? In creating context, opinions take the position that the factors they present are the ones most relevant to interpretation, and they reject the relevance of other factors.

Second, opinions make claims about reality. As they assemble the factors they use into a context, opinions also express positions—often implicit ones—about those factors’ characteristics. How do people, in fact, use a phrase? How does the physical, social, or psychological world actually work? What position does a legal era, concept, or text take? To make them speak to the questions at issue, opinions must characterize the factors they draw on.

Situating text, then, involves characterizing background factors and claiming that they are relevant to interpretation. How, though, do opinions convince readers to accept their claims? How, for instance, does an opinion justify its contention that some factors, but not others, are the most relevant to interpretation? The opinions discussed here demonstrate a startlingly wide range of approaches. Some lay out the basis on which they claim relevance, as when the Green dissent explains that it takes the Federal Rules of Evidence’s own stated principles of purpose and applicability as guidance for interpreting an individual rule. Others simply present evidence without specifying how it sheds light on interpretation, as when the unanimous Whitfield opinion quotes Austen to show what a word meant in 1930s America. Relevance claims, then, run the gamut from explicit and explained to unstated and assumed.

We can ask the same question about the way opinions characterize the factors they draw on. How does an opinion substantiate its depiction of the factors it assembles? How does it convince readers that its view of the world is accurate? Here we see an equally wide range of approaches. Some opinions ground their assertions about their chosen factors in voluminous evidentiary records, as when the Brown & Williamson dissent draws on the FDA’s scientific findings about tobacco. Others opt for thinner evidence, as when the Rapanos plurality substantiates its claim about the physical nature of fill material with an unsupported assertion on a corporate website. Still others provide no evidence at all, as when the Smith dissent lays down its rule for interpreting idioms.

Within this wide range, the cases sampled here—admittedly a small set—suggest a pattern worth exploring. These cases tend to explain the relevance of the factors they use and substantiate their characterization of them most carefully when they stick closest to the statute at issue. Intrastatutory, contemporaneous, and poststatutory factors tend to be carefully justified and substantiated.252 As these opinions move further out from the statutory text, though, they often get less specific.253 When it comes to substantiating their assertions about nonlegal factors, moreover, the opinions surveyed here differ vastly, using everything from peer-reviewed findings to personal beliefs. Strikingly, an opinion’s ability to command a majority of votes, or to avoid challenge on a particular point, seems not to depend much on how it substantiates its claims about the nonlegal factors it invokes or how it justifies their relevance to the interpretation.

If the opinions surveyed here are any indication, it may be that the genre of judicial opinion has evolved to require the most from judges when they discuss the things that they are best versed in—the statute and its surrounding laws—and less when they invoke things they are less expert in, like empirical facts. That is, the opinions I discuss here suggest that American legal culture might credit judicial beliefs about the nonlegal world more than judicial statements about the law, even though judges will generally be more expert in the law than they are in the nonlegal world. To what extent this pattern holds for opinions beyond those discussed here deserves further research. I hesitate to make a broad claim about opinions in general based on my small set of examples; but it is worth noting that most of the cases I discuss are classics of the genre, used in teaching materials and discussed at length. If not strictly representative, then, they are at least important examples.

A rationalization of this phenomenon might posit that opinions will properly be more careful in discussing factors closer to the statutory text because those factors are most authoritative. Because legal factors carry greater weight in judicial interpretation, on this reasoning, judges are right to be more thorough in substantiating their factual assertions and justifying their relevance claims. The opinions here, however, contradict this hypothesis: they do not consistently give greater weight to legal, or even intrastatutory, factors. The Sweet Home dissent does not retreat in the face of the majority’s intrastatutory factors; it insists on the primacy of extrastatutory traditions. The Green majority and concurrence find extrastatutory principles not tied to any source decisive, even while the dissent quotes intrastatutory indications of applicability. The Rapanos dissent substantiates its factual assertion about fill material’s mobility by quoting a lower court finding of fact; the plurality ignores this legally validated factual assertion in favor of a different assertion by a source that is neither legal nor scientifically validated. Legal factors may be entitled to greater weight in principle, but the cases discussed here indicate that they do not systematically receive greater weight in practice.

The previous two Sections outlined the kinds of factors that judicial opinions tend to draw on when situating focal text. I divided these factors into two broad categories, legal and nonlegal, but noted that the borders between them are not fixed or ontological. Rather, opinions sometimes have leeway in characterizing something as belonging to one or the other category. More­over, each category is internally differentiated—so highly differentiated, in fact, that we can probably never list all of its possible parts. This internal diversity emerges from the diver­sity of law and the things it addresses. But it emerges just as much from the fact that it is opinion writers themselves who identify and delimit the factors they draw on, in the act of drawing on them.

This Section has suggested that, as we analyze how judicial opinions situate focal text in legal and nonlegal factors, we can productively evaluate them through two related inquiries. We can ask how an opinion justifies its claim that certain factors are relevant to interpretation, and how it substantiates its depiction of those factors’ characteristics. One way such inquiry can be productive, I have suggested, is in bringing to light patterns that reveal underlying values structuring the genre of the judicial opinion. I have pointed out that the opinions I discuss here demonstrate a rather counterintuitive pattern: requiring the least justification and substantiation in judges’ areas of least expertise.

This suggestion is, of course, impressionistic. But I hope it indicates the kind of inquiry that my approach can open up. By looking in detail at the kinds of factors opinions invoke to situate text, and asking how they justify and substantiate their recourse to those factors, one stands to learn new things about the patterns of judicial practice. This in turn allows scholars to look further, probing both the conduct of judges and the legal culture within which they operate. On the one hand, one can ask what values and presuppositions may tilt opinion writers toward these practices. On the other, one can ask what these practices reveal about the legal culture’s conditions for legitimate legal assertion and argumentation.

The next Part expands on these benefits. My approach illuminates constitutive elements in judicial opinions that are analytically prior to the commitments identified by prescriptive theories of interpretation. Grounded in widely shared commitments and focused on practices rather than ideological self-reports, it provides a firm basis for the normative evaluation of judicial opinions. And it highlights the creative, interactive, and iterative nature of judicial meaning-making.

IV.  The Infrastructure of Interpretation: A Constitutive Analysis

The preceding Parts illuminate the infrastructure of judicial interpretation. Opinions select texts to interpret by parsing statutory phrases, emphasizing some statutory texts over others, or looking outside the statute altogether. These are ways of framing the selected text as focal, relegating everything else to background. To situate the text they select, opinions weave together factors delineated anew in every case. To do that, opinions depict the factors they choose as having certain characteristics, and treat those factors as the ones most relevant to interpretation. Situating brings together some few backgrounded items and puts them in a particular light to serve as the immediate context within which to view the focal text. Selecting and situating: this is how judicial opinions constitute what the law is.

This way of describing things departs from the common terminology provided by prescriptive theories of statutory interpretation, especially textualism and purposivism. This Part explains the benefits of that departure. My approach provides better tools for analysis because it illuminates the steps judges actually take to make their arguments, rather than focusing on the normative commitments judges proclaim. I argue that the prescriptive theories are statements of commitment rather than methods of analysis. They make appropriate objects of scholarly analysis, but should not set its terms.254

Adopting my more realistic analytic approach, however, does not mean abandoning the normative evaluation that often characterizes legal scholarship.255 On the contrary, my approach provides a more solid grounding for normative inquiry than the prescriptive theories can offer. It allows evaluation to directly address core values in American democracy while dynamically incorporating different ideas as to what those core values are.256 Finally, my approach illuminates some of the key things that make judicial opinions both efficacious in, and representative of, the legal system. It shows the deeply creative, or speech-act, nature of judicial opinions, which help construct the world they inhabit. It highlights their attempts to stabilize and finalize that world by creating a seemingly closed discourse of semiotic final­ity. And it elucidates how this stability is always illusory, its closed discourse permeable, its finalized meanings provisional.257

A.    The Limitations of Prescriptive Theories of Interpretation

Parts II and III periodically considered how the predominant theories of statutory interpretation, textualism and purposivism, would analyze my selecting and situating examples. Textualism, recall, instructs the interpreter to draw on the focal text’s “semantic context,” that is, “evidence about the way a reasonable person conversant with relevant social and linguistic practices would have used the words,”258 often drawn from dictionaries and judges’ intuitions.259 These rules and norms of communication, moreover, are to be viewed against the backdrop of the rest of the statute and of other legal strictures and traditions—the “corpus juris.”260 The idea is that because only words actually enacted into law give a legal command, interpretation should stay within the bounds of the language the statute uses. One common justification is that this approach best enforces the “deals” or compromises that legislators inevitably make when passing a law.261

Purposivists look to the statute’s language and legal surround as well,262 but also look for clues about how the legislature that enacted the statute expected it to work, and what goals it meant the statute to achieve.263 Legislative history such as “committee reports, conference committee reports, and the joint statements of conferees who drafted the final bill” can provide those clues.264 The idea is that Congress passes a law for a reason, and judges should heed the ways Congress “makes its purposes known.”265 On this view, affording Congress lawmaking primacy means taking Congress’s cue and “using the interpretive materials the legislative branch thinks important to understanding its work,” including legislative history.266

Both of these approaches ask judges to be “faithful agents” of the Congress that passed a statute.267 But they disagree about what constitutes a valid indication of the meaning of a statutory term: Is it a statute’s words in their linguistic and legal context, or is it that plus the legislative situation in which the statute was enacted? That disagreement, in fact, may be the linchpin distinguishing the two theories, which otherwise have so much in common.268 In other words, it is the theories’ differing attitudes about how judges should situate focal text that differentiates them from one another.269 One might expect, then, that each theory would have an elaborated approach to situating.

In fact, these theories treat the linchpin notion of context rather simplistically. They acknowledge—sometimes grudgingly—that contextualization plays a role in statutory interpretation.270 But they conceive of context in large, internally undifferentiated chunks: statutory text, legal background, linguistic usage, legislative history.271 As my examples demonstrate, actually existing opinions do not confine themselves to these few, discrete, predetermined realms. First, opinions situating text in these contexts tend to parse the world more finely, carving out areas within each category to juxtapose: a preceding legal background versus a contemporaneous one, my linguistic usage against yours. Second, opinions routinely ignore the boundaries of these categories, invoking with abandon things like empirical facts and developments that postdate the statute—kinds of context that neither theory acknowledges. The prominent theories’ descriptions of how judges situate text are thus both underspecified and underinclusive.

Additionally, these theories purport to predetermine the allowable contexts within which opinions may situate statutory terms. Indeed, their main dispute is about which of their context categories judges ought to invoke. But, as I showed in Part III, actual opinions do not use prefab contexts. They situate text in particular contexts that they create for particular cases by making claims about particular factors. The factors they invoke, moreover, are not predictable: judges can carve out aspects of the law and the world in new and different ways, or put together preexisting factors to create a new context. When it comes to situating text, these theories offer narrow, blunt, and static options, while judges choose broad, nuanced, and dynamic ones.

When it comes to selecting text, the theories offer even less. Neither recognizes that judges select the text they interpret. Further, as Part II showed, these theories often do not give judges a reason to select one text over another. Both a purposivist and a textualist could side with Justice Scalia in Sweet Home, selecting take because it was the operative legal term and thus expressed congressional purpose. Both a purposivist and a textualist could also join Justice Stevens to select harm, because it was the operative qualifier and expressed what Congress wanted take to mean. The same goes for King: established by the State could be taken as the operative qualifier that expressed what Congress wanted Exchange to mean; or Exchange . . . under section 1311 could be construed as the operative legal term that expressed congressional purpose.272 Often, the theories do not motivate selecting one text over another.273

These theories and the commitments they express thus enter the picture only after the crucial decision has already been made. Say one thinks of textualism as a method for enforcing the legislative deals expressed in statutory language.274 A judge striving to achieve that goal in Sweet Home still must decide what text in the statute expresses the relevant deal. Is it take or harm? How would she know where the relevant deal lies? The interpretation of the statute hinges on that choice, but textualism gives a judge no principled way to make it. Perhaps a textualist judge selects the text intuitively, out of a general feeling or understanding about what the statute means and what legislative deals it embodies. But if so, she acts to fulfill textualist commitments by violating textualist tenets, because textualism does not countenance interpretation based in general feelings or understandings rather than in text. In other words, textualist commitments might rationalize how a judge approaches the text she has selected. But they do not rationalize selecting one text as opposed to another.

Purposivism may be more able to recognize and motivate text selection. A purposivist in Sweet Home would look for the purpose of statutory text—but which text should it be? Selecting take might suggest a different purpose than selecting harm would. A judge might, in keeping with purposivist tenets, consider general statutory purposes and other clues to help decide which text to select. Purposivism’s greater flexibility and tolerance for overall assessment would allow it to more easily incorporate the fact that judges select text, and it can also help rationalize particular selections. At the same time, purposivism, like textualism, does not account for the fact that opinions select text, and therefore provides no way to analyze selection.

From a certain perspective, this inability to analyze the infrastructure of interpretation may not be that much of a failure after all. It makes sense once we consider the function these theories serve. Textualism and purposivism declare what the underlying goals of interpretation should be and prescribe how people should go about fulfilling them. But, as life experience suggests, telling people what they ought to do is not the same as explaining what people, in fact, do. As prescriptive theories, textualism and purposivism present philosophical commitments, not analytical methods. Their stances on how people ought to behave are grounded in overarching convictions about what constitutes political legitimacy and how political processes should work. In this sense, these theories are classically ideological products.

By ideology, I mean rationalizations or explanations through which people cohere their worldviews, guide their practices, and align themselves with others: those “more or less coherent patterns of meaning which are felt to be so commonsensical that they are no longer questioned, thus feeding into taken-for-granted interpretations of activities and events.”275 These systematized expressions of worldviews relate social practice with social structure in ways that explicate, but also systematically obscure, important aspects of the world to adherents.276 Such explanations are an integral part of social life: this is how people make sense of the world and their place in it.

At the same time, ideology works primarily to cohere, justify, and judge, not to analyze or explain. So one should not be too surprised when textualism and purposivism fail to acknowledge or explicate crucial aspects of statutory interpretation. These theories express self-understandings through which members of the American legal community describe, distinguish, and validate themselves. As ideological statements, these theories should be objects of scholarly analysis. But this does not mean that they should provide the categories of scholarly analysis. On the contrary, it suggests the opposite: “It is a social science commonplace that the ways the natives talk about behavior must be recognized as different from the analysts’ accounts of how and why they act the way they do.”277 And indeed, they turn out to be a poor source for categories of analysis, insofar as they do not explain—or in some cases even recognize—actual interpretive practices.

Despite their explicitly prescriptive, ideological character, these theories of statutory interpretation are sometimes treated as though they were tools for analysis, rather than for normative evaluation. As I have shown, this mistake allows the infrastructure of interpretation to go unrecognized. It also allows prescription to sneak into the limelight of scholarly inquiry. Much statutory interpretation scholarship, for instance, focuses on these prescriptive theories. Scholars debate which prescriptive approach is better; gloss their intricacies and implications; critique them; or offer alternative, better, prescriptions. But all of this tells us more about prescriptive theories than about judicial practices.

Allowing the analysis of interpretation to be hemmed in by the terms of prescriptive theories, in other words, may make scholars less attentive to what opinion writers do. And what opinion writers do turns out to be much more complex and creative than the prescriptive theories allow for. The analytic purchase that selecting and situating give for explaining the argumentative structure of opinions thus warrants departing from the rubric offered by textualism and purposivism.

B.    Selecting and Situating as a Basis for Normative Evaluation

Focusing on how judges select and situate text, moreover, does not mean abandoning normative judgment. On the contrary, taking selecting and situating into account could provide firm and flexible bases for the normative evaluations of how opinions should fulfill their reason-giving obligations; what role different factors ought to play in interpretation; and what overarching visions of the law opinions ought to further. It can also help scholars evaluate the strength and consistency of the prescriptive theories themselves.

This may seem surprising, because my approach does not involve any normative strictures. Rather than starting from premises about what judges ought to do, I have started from what judges actually do to build up to their interpretive conclusions.278 Unlike the prescriptive theories, my aim has not been to evaluate which focal text is the “correct” one or which contextualizing factor is “really” relevant.279 As Parts II and III showed, opinions produce texts and contexts as social effects. Determining the correct text or real relevance that precedes an opinion’s treatment is therefore often not possible. There is no objectively correct option that exists outside those creative practices.

I have also eschewed asking whether opinion writers honestly represent their inner convictions or reveal their true motivations—for instance, whether their interpretations are simply masks for their policy preferences. Selecting and situating structure how inner states such as convictions and motivations are publicly expressed and legitimated to their audiences. In this sense, I have taken a social, rather than individual, view of opinions, taking them as artifacts that reveal something about their society as much as about individual authors. The conventions and the creativity of selecting and situating illuminate what counts as valid legal argument in American legal culture, and show how legal actors claim legitimacy while both harnessing and modifying its parameters.

Having no normative views built into the analytical structure, however, does not preclude normative evaluations. On the contrary, it provides a flexible basis for nuanced and grounded normative judgment. Perhaps more importantly, it opens the door for discussions and debates about a wide range of normative considerations—not just those that the prescriptive theories of statutory interpretation choose to foreground.

First, my approach allows us to evaluate how, and how well, opinions justify their interpretive conclusions. In the American tradition, reason-giving is one of the judiciary’s fundamental obligations. A legitimate opinion does not merely hand down a fiat; it should reach its conclusions through rational considerations that it articulates so that its audiences can utilize, evaluate, and contest them.280 Because selecting and situating focal text are the building blocks of statutory interpretation, rationalizing those choices is crucial to giving reasons for an opinion’s interpretive conclusions.

I have suggested that we can productively evaluate this aspect of reason-giving by asking two related questions. One, how does an opinion justify its claim that the factors it invokes are the most relevant to interpretation? Two, how does it substantiate its characterization of those factors? Justifying claims of relevance and substantiating claims about reality are central to an opinion’s rational articulation of its interpretation. Yet doctrine and theory impose no constraints on these practices. And my examples show that opinions vary wildly on this score. If anything, I suggest, the opinions surveyed here display a rather counterintuitive tendency to justify and substantiate the most in areas in which judges have the most expertise, such as statutes, and the least in areas in which judges are least expert, such as empirical facts about the nonlegal world.

The uneven, and possibly perverse, way that opinions support claims about relevance and reality calls out for normative evaluation. Is it rational, or desirable, to have judges create contexts using empirical claims for which they provide no basis? Or to treat as authoritative assertions about language use whose relation to the speakers they purport to describe is never explained? In a judicial system oriented around reason-giving, the answers seem self-evidently negative. Yet reading the cases discussed here, it is hard to avoid the conclusion that American legal culture at times tolerates, and may even encourage, such unreasonable reasoning.

One can evaluate text selection in a similar way. Opinions, as I have shown, have considerable latitude in choosing what they will interpret, and the prescriptive theories have little to say about it. Judges can thus easily choose not to explain why they select the text they do, leaving one central step of interpretation out of their reason-giving altogether.

Evaluating how opinions justify and substantiate their selecting and situating choices thus provides a solid foundation for normative judgments about how well opinions fulfill their reason-giving obligations. Some may claim that judges select and situate on the basis of their adherence to one or another prescriptive theory of interpretation. Yet, as my discussions here have shown, the theories often offer little guidance about selecting and situating. That leaves judges on their own in the leap from normative orientation to efficacious interpretation. It seems only reasonable to ask how these—or other—normative commitments lead judges to select and situate text as they do. It seems even more reasonable to ask judges to explain that process themselves. Such evaluation is all the more important, moreover, because judges will inevitably lack expertise in many of the sources on which they draw for interpretation. Independently substantiating every characterization of every factor opinions draw on could well bring adjudication to a standstill.281 To allow the public to evaluate the strength of their reasoning, then, judges should be maximally clear about the reasons behind their selecting and situating choices. My approach allows readers to evaluate how well judges fulfill this reason-giving obligation.

Second, by sketching a provisional map of the ways opinions select and situate text, my approach allows commentators to evaluate the relative weight opinions give to different kinds of authority. Does an opinion stick close to the statute or range further afield? Does it hinge on legal sources or the nonlegal world? Some commentators may subscribe to a hierarchy of sources. For instance, they may believe that considerations closest to the statute should carry the greatest weight and that legal considerations should have greater authority than nonlegal ones.282 By focusing attention on the range of factors opinions employ, my approach provides a way to assess the extent to which opinions adhere to any such hierarchy of sources in statutory interpretation.

Third, while I do not attempt to probe judges’ inner states, my analysis facilitates evaluating the attitudes that their opinions suggest or comport with. For instance, the majority opinion in Green invoked equality between civil litigants as an underlying principle guiding the law. The concurring opinion invoked the special treatment of criminal defendants as such an underlying principle. Should we accept either proposition? Which one? And why? These opinions invoke these principles as presumable background considerations. But the very fact that one opinion invokes one principle while the other invokes another suggests that the legal system may benefit from a clearer elucidation of the role these commitments should play.

Similarly, the majority and dissent’s different selections in Sweet Home comport with different understandings of the role of regulatory statutes. If one sees environmental regulation as a break with traditional property law, it makes sense to treat a statute as providing its own definition for a legal term: harm. If, on the other hand, one sees regulatory rights and obligations as constrained by older, private law legal norms, it makes sense to treat the statute as incorporating, not replacing, common-law property concepts: take. I do not claim that this is what Stevens or Scalia thought when writing their opinions; I do not have access to their inner states. But something more important than inner state is at play here. These different text selections present readers with competing conceptions of the proper role of different kinds of law. Recognizing this allows others to identify, and weigh in on, an important normative debate—one that centers not on whether take or harm is the proper object of interpretation, but on the proper relationship between administrative and common law.283 Bringing such values to the surface, my approach encourages scholars and judges to engage in debate about the underlying values that these opinions express or further.

Moreover, the range of positions we can inquire into in this way is flexible. It is circumscribed by the concerns not of textualism or purposivism, but of those doing the evaluating.284 I don’t stake a normative position, in other words; but I provide tools for nuanced and dynamic normative assessments.

Finally, my approach can provide a basis on which to evaluate prescriptive theories themselves. Each theory rests on an image of legislation and a normative position about how judges ought to relate to it. We can debate each theory’s external claims, such as whether its image of legislation is accurate.285 But we can also ask about its internal coherence: How well do each theory’s aspirations fit its prescriptions? Textualists, I have suggested, can enforce only the legislative deals they find in text they have selected. But textualism gives little guidance as to how to select the text that best expresses the legislative deal, or to determine which text expresses the most important legislative deal. While it gives some instructions for situating text, it drastically underspecifies the variety of legal sources judges must choose from, gives judges no way to evaluate their own assertions about language, and does not recognize the importance, and availability, of nonlegal sources. Purposivism may bring its aspirations closer to its prescriptions, allowing judges to use the statute and legislative history as a guide to selecting the text that fits best with the overall purposes they find there and to situating it in the factors closest to those goals. Yet purposivism also falls short of recognizing these practices or instructing judges to justify and substantiate claims about relevance and reality.

My point is not to give a full-fledged evaluation of these theories. Rather, I hope to show that whatever their interpretive commitments, they will be carried out through selecting and situating text. So assessing how well a theory’s commitments comport with the guidance it gives on selecting and situating text should provide a useful way of evaluating the theory. Similarly, relating what interpreters do to what theories tell them to do gives us more indication of the theory’s power than does asking whether a judge claims adherence to it.

C.    Judicial Opinions as Efficacious and Unpredictable Utterances

I have argued that the concepts I sketch out here—selecting and situating, justifying and substantiating—give us tools for analyzing judicial opinions that should at least supplement, and possibly replace, those provided by the prescriptive theories of interpretation. I have also suggested that my approach can provide a more grounded, nuanced, and flexible basis for the normative evaluation of opinions. In this Section, I argue that my approach also sheds light on another critical feature of judicial opinions: their effects in the world, and the way those are tied up with how they present their reasoning.

Judicial opinions—it hardly needs saying—are an important locus of legal action in American society. Most obviously, holdings directly affect the world around them: they determine legal rights and obligations and then mobilize state power to effectuate them. In this sense, judicial holdings are classic performative utterances, or speech acts. Performatives are statements that create the state of affairs they describe.286 When I promise to buy milk, I do not just describe a situation. I create one. I create my obligation to buy milk, or, perhaps more accurately, I create a new world, a world in which I have obligated myself to buying milk. Similarly, when a judge pronounces a couple married, she creates a world in which the couple is, in fact, married.

As Professor Marianne Constable has recently written, speech acts permeate the law.287 Oaths, contracts, objections—these are all efficacious ways of uttering.288 A court holding that trading a gun for drugs is using a firearm in a drug trafficking crime makes it so. An opinion asserting that a case is “about” something constitutes that thing as the case’s subject. Opinions exert this constitutive power even when drawing on facts about the nonlegal world. First, they create relations of relevance: a Brown & Williamson opinion invoking tobacco’s effect on the body establishes tobacco’s effect on the body as significant. Second, they constitute nonlegal factors as facts for the purposes of the opinion—even if not for the purposes of real life. When a Rapanos opinion asserts that fill material does not wash downstream, it creates a world in which that is so for legal purposes, irrespective of how such material behaves in the physical world.289 Or, at least, it attempts to do so. If Justice Anthony Kennedy had joined either side in Rapanos, that side would have successfully created its world as a matter of legal, if not of physical, fact.290

That is the power of the performative: even when it fails as accurate representation, it may create its own truth for the purposes of its social situation. My promise to buy milk may fail as a description of truth. I may forget to buy milk. I may secretly have had no intention of buying milk. But by promising, I created a somewhat separate, no less powerful, truth: the truth of having obligated myself. Similarly, an opinion may fail to represent the physical truth about fill material and yet, at the same time, create a truth for its particular, legal, situation.

But it can do so only if others accept its act as legitimate. Speech acts have their effects only when uttered under appropriate “felicity conditions.”291 A judge who pronounced passing strangers married would not have a world-changing effect: not being uttered under felicitous conditions, those words would lack the speech act’s efficacious power. The felicity conditions under which we consider an utterance efficacious can also reveal shared values and understandings, that is, cultural convictions.292 Knowing the felicity conditions for a marriage pronouncement, for instance, can deepen our understanding of how a culture constructs the ability to enter into associations and the authority to change others’ legal status.

Speech acts, moreover, are just one particularly visible instantiation of the creative function of language use, in contrast to its presupposing function.293 The creative function helps affect, create, or show in a new light the state of affairs in which the utterance appears. “We,” for instance, is often the sort of term through which “speakers in speaking create a social group around them, including some members of [society] and excluding others.”294 That is, while the term “we” refers to a group, saying it can also help constitute a dispersed bunch of people as a group. Like a speech act but more subtly, that kind of creative utterance can have an effect in the world by helping to create the situation it refers to.295 As this discussion suggests, creative utterances also depend on felicity conditions of sorts. Saying “we” would not constitute any old bunch of people as a group; they must be somehow amenable to being groupified. While these sorts of conditions may be difficult to define abstractly, they are no less revealing of underlying cultural convictions and conventions.296

Judicial opinions are replete with creative expression, often masked as presupposing. Selecting text and situating it appear to rely on existing, independently verifiable, obviously relevant factors, but in fact are creatures and claims of opinion writers. Through this creativity, opinions can spread ideas about law, society, and justice that go beyond the outcome of any given case. This is one reason I focus as much on dissents as on majority opinions here. Both instantiate accepted practices of legal argumentation, and both can nudge those practices in new directions.

The creative power of legal utterances can thus both consolidate and change relations of authority and conditions of legitimacy—sometimes at the same time. For a famous example, think of the words from the Declaration of Independence, “We hold.” Who is that “we,” and how can it commit itself to holding? That statement, Professor Hannah Arendt theorized, “constitutes a free coming together” of a community that it also “empowers.”297 Moreover, in the same moment that it conjures the community, it issues “a promise and a declaration” of “a shared agreement [for] . . . the community’s subsequent being together,” that is, a shared agreement about the ongoing parameters of legitimated political action in that community.298 This beginning is Arendt’s exemplary instantiation of “a singularly human capacity: that of world-building,” whose “source of power . . . is the speech act itself.”299

For Arendt, this kind of creativity—grounded in the creative community rather than in religious or traditional authority, and expressed through original language rather than through routine or force—is the very definition of the political action appropriate to a democracy.300 And while Arendt discusses a particularly famous phrase, the efficacious quality of language she discusses—its world-building power—characterizes creative utterances more generally. We see this power with selecting and situating, utterances that do not call attention to their world-building capacity but nonetheless help constitute the world they refer to.

This constituting power of action through language, how­ever, is also inherently unpredictable. Utterances live on as others take them up, argue with them, redeploy them, or reinterpret them over time.301 We see this quite overtly in judicial opinions, for which the accrual of commentary and reason-giving provides a rich resource. Dictum can make its way into doctrine, making law out of offhand remarks.302 Concurrences can come to overshadow their majorities.303 Dissents can foreshadow—or provide—arguments that appear in later majorities.304 Supreme Court opinions affect the reasoning of lower courts, but lower courts also take them up in their own ways. And judicial statements can have broader cultural influence, presenting or cementing ways of seeing the world beyond legal doctrine. The opinions I discuss here demonstrate this unpredictability in their contestation about the proper focus and sources for legal action.305 This lack of closure—the dynamic way that participants in a polity continually wrangle their way toward a never-completed “shared agreement [for] . . . the community’s [ ] being together”306—is, for Arendt and others, the essence of democracy.307

Yet, even as judicial opinions engage in their creative, agonistic, and unpredictable work of meaning-making, they present this action as something else: presupposing, uncontroversial, obvious. They try to fix meanings even as they unsettle them. This process, I suggest, is one way that “the mundane details of the legal process . . . are arranged to produce the effect that the law exists as a formal framework, superimposed above social practice,” and obscure the fact that legal processes like interpretation are themselves “social practices.”308 The way that opinions hide their creativity bolsters this illusion that law is ontologically distinct from the society it governs. And it allows opinions to talk as though their legal interpretations were not subject to ongoing reinterpretation and contestation—as though they could be final or determinate. The genre of judicial opinion, then, seems to push participants to ignore the inherent instability and contestability of legal meaning.

Some have suggested that when courts deny the availability of alternative meanings, claiming that “a statute is plain and therefore needs no interpretation,” they do so through subterfuge or delusion: “They have already interpreted, and they then declare that so interpreted the statute needs no further interpretation.”309 From a psychological perspective, judges may indeed find some texts sufficiently clear not to require interpretation. Professor Richard Fallon, for instance, has posited that judges are pushed to interpret by an “interpretive dissonance” in which they sense a disjuncture between the “first-blush meaning” of statutory words and what one would expect from a “reasonable legislature” writing “well-written legislation.”310 On this view, that creative, agonistic, and unpredictable meaning-making takes place only at the margins. And indeed, ordinary interactions suggest that communication does not always require creativity and agonism. Sometimes a stop sign is just a stop sign.

Yet, at the same time, intuitions—about first-blush meanings, about well-written legislation, and about stop signs—are structured by cultural characteristics and individual experiences.311 Whether a term is obvious or ambiguous, then, depends to some extent on the questions people pose about it. Agonism, creativity, and unpredictability do not have to accompany every assertion of meaning; but they are always potentially available. This is especially so given the premises of the adversarial litigation system, which allows parties to subject statutory meanings to interpretive arguments. Ambiguity, in other words, is not something that inheres in a term or an utterance so much as something that is created through social processes. One key place in which legal ambiguity is created is in the lawsuit, in which each adversary presses its own interpretation. One can say that, in an adversarial system, it is precisely arguing about things that makes them arguable.

The cases I have discussed here show hints of this. The meaning of the Affordable Care Act’s tax credit provision was widely seen as obvious—until it became an issue of heated debate in King. What “accompany” means might seem obvious once it is defined by a unanimous Supreme Court, but the fact that certiorari was granted on the question in Whitfield undermines that easy dismissal. To say that an utterance is obvious, intuitive, understood, or not in need of interpretation, in other words, involves an assertion about social realities that are subject to change. Whether we feel dissonance may depend on our communicative intuitions, but those intuitions are in turn influenced by the arguments that others make about the text.

Perhaps one reason not to acknowledge this is the worry that admitting that some interpretive question has no single, definitive answer would delegitimate judicial power. In American legal theory, after all, the claim that meaning is indeterminate has been seen to “undermine[ ] the legitimacy of [any given] legal interpretation.”312 This poses a deep problem for a discipline that has long—or at least periodically—recognized that “[a] statute . . . is essentially ambiguous”313 because “words which are applicable to things . . . always have many meanings.”314

In contrast, I have suggested here that it is through this very indeterminacy that people create—that is, claim, contest, and set conditions for—the legitimacy of legal texts.315 Legitimation is always a socially constituted process. In law, it is carried out through linguistic expression and interpretation. A constitutive analysis of legal interpretation reveals how people deal with tensions and uncertainties in their cultural norms and structures of authority—whether that means changing, illuminating, or burying them out of sight. How opinions communicate these choices is not a separate issue in this process; it is a central participant.316 As Constable writes, “Language is not like a window through which we look to an outside of which we are not part. Our speech is akin rather to the paths we walk as we make our way through the wider world.”317 Rather than undermining a judicial opinion’s legitimacy, the indeterminacy of meaning is the condition for its legitimation.

Recognizing the creativity through which judicial opinions interpret the law, moreover, helps us understand the standards for legitimate action that the legal community imposes. We see through legal discourse’s implicit claims to closure to the dynamic production of an interpretation structured by sometimes surprising social norms. The approach I present here, in other words, can teach us not just about judicial interpretation but also about ourselves.


I have argued that the text a judicial opinion interprets is not given but selected, and that the context within which the opinion situates this text is not predetermined but created by the opinion itself. As with so much communicative labor, these practices also reflect back on their practitioners. As judges select and situate, they also claim legitimacy for their approaches and set the parameters for legitimate legal utterances over time. This recognition provides new tools with which to analyze how opinions achieve effects and attract adherents. It provides a firm basis for a flexible range of normative evaluations. And it shows how opinion-writing practices themselves participate, in sometimes hidden ways, in the arguments about what matters and who counts that characterize democratic politics.

These insights are especially helpful given the grip that prescriptive theories of statutory interpretation hold on scholarly analysis. Those theories fail to recognize text selection, and mistakenly treat context as a guided choice within a limited and static buffet of predetermined frames. In contrast, I show that selecting the object of interpretation is a crucial constituent in legal interpretation, and demonstrate that judges create contexts by combining diverse and unpredictable factors in unique ways. Just as importantly, I explain why these theories make their mistakes: their prescriptive orientation is ill-suited to analysis. Such normative tenets should not be mistaken for either representations or explanations of how judges actually act through their writing.

Recognizing the constitutive nature of selecting and situating thus opens up a range of inquiries about legal interpretation and the theories that purport to guide it. It allows one to evaluate how opinions comport to a range of normative ideals and commitments.318 It allows one to survey how different opinions build their arguments without depending on practitioners’ self-reports about interpretive commitments, and without being boxed in by the narrow range of normative concerns identified by prescriptive theories. In the same way, my approach provides a basis on which to compare sets of opinions. Say, for instance, one wants to know whether interpretive approaches have changed over time or differ among courts. Instead of trying to shoehorn opinions into the broad, predetermined commitments of the prescriptive theories, my approach gives a commentator tools with which to track subtle differences and similarities across opinions by looking for patterns in the way they select and situate text. It thus provides a nuanced, detailed way to pursue a comparative inquiry free of normative preconceptions. Similarly, if one wants to assess whether judges or courts have been consistent in their statutory interpretation, one can track their selecting and situating practices, rather than their claims of adherence to a particular approach.

As an example of the kind of patterning one might unearth, I noted above that the opinions discussed here tend to provide the least rationalization when they roam furthest from the statutory text.319 Whether this tendency in fact characterizes opinions in general, or opinions at any particular time, deserves further study. If the trend holds, it may indicate one way that statutory interpretation has not quite caught up to the administrative state. While legislation now encompasses a vast world of complex empirical facts studied by a host of expert disciplines, statutory interpretation may still treat the facts that law addresses as relatively simple, available in an unmediated way to “judicial experience and common sense.”320 In other words, assessing patterns and contrasts in selecting and situating can help us understand how statutory interpretation has developed and how it relates to its contemporary legal surround. It may also spur further inquiry into judicial competence, to evaluate whether judges would benefit from training or outside assistance when dealing with particular substantive areas.321 One could also consider whether doctrine ought to recognize the need to rationalize selecting and situating choices, perhaps even prescribing such explanations.

This Article has focused on Supreme Court statutory interpretation opinions, but the approach here suggests many other places to look and questions to ask. One can look elsewhere in the judicial system. Do lower courts select and situate in the same way, or do they display different patterns and practices? One can hypothesize that, having a different place in the judicial hierarchy, lower courts might make their choices differently as well. Do specific areas of statutory interpretation—tax law, say, or employment—tend toward particular patterns of selection and situation? It is possible that opinions in a particular area of law serve not only as legal precedents but also as models of selection and situation for later opinions.

Using my approach, one can also look beyond the judiciary. The other primary interpreters of statutes are administrative agencies.322 How do agency administrators select and situate text? Because of their particular institutional characteristics, they may well select and situate differently from judges.323 One can also look to the producers of statutes. How do those who draft legislation understand the selection and situation choices of those who interpret it, and how do they incorporate those choices into their drafting?324

Furthermore, one can extend my approach beyond the statutory realm. Do these moves also capture the key moments of constitutional adjudication? Because they are closely connected to interpretive practices in other areas of communication, I suspect that selecting and situating will be key aspects of any legal interpretation. But I also suspect that the way they manifest may differ over subject areas, institutions, and types of law. All these suspicions, of course, are empirical questions that call for further investigation of legal texts.

Questions remain about my approach as well. I have described selecting and situating as co-constitutive conceptual moments: each depends on the other, and there is no hard line separating the two. It is not entirely clear, though, whether one is more important to interpretation than the other. From one perspective, once a judge has selected a text, she has effectively chosen the context within which to situate it. From another, it is the choice of context that leads a judge to a particular text. Is one move primary, or are they equal? Are they sequential, or simultaneous? The precise parameters and relationships at play remain to be explored.

At a more abstract level, this Article also implies a position on analytic methods. In asking readers to focus on the micropractices of interpretation, I have asked them to reproduce those practices here. That is, I have asked readers to select particular judicial practices and interpretive moments as objects of analysis, and to situate them within the opinions that use them and the theories that address them. Selecting and situating, I have suggested, largely determine how a judicial interpretation will go. In a similar way, choosing a method may largely determine how an analysis will go. Opinions present judicial interpretation as a closed discourse of obvious choices. In a similar way, I present my approach as a better way to explain that discourse: both a way to show that it is neither closed nor obvious, and a way to understand how opinions present it as being so. In the end, I hope that this Article has fulfilled its own burden of giving reasons to treat selection and situation as central to statutory interpretation.

  • 11. See, for example, King v Burwell, 135 S Ct 2480, 2489 (2015) (“We begin with the text of [the statutory provision].”); Milner v Department of the Navy, 562 US 562, 569 (2011) (“Our consideration of [the statutory provision’s] scope starts with its text.”); Tapia v United States, 564 US 319, 326 (2011) (“Our consideration . . . starts with the text of [the statute].”); Muscarello v United States, 524 US 125, 127 (1998) (“We begin with the statute’s language.”); King v St. Vincent’s Hospital, 502 US 215, 218 (1991) (“We start with the text of [the statute].”).
  • 12. This lack of attention to the process by which a focal statutory text is selected contrasts with a more developed discussion around the related issue of the “level of generality” at which a legal principle or individual right should be stated. See Herbert Wechsler, Toward Neutral Principles of Constitutional Law, 73 Harv L Rev 1, 15–19 (1959) (discussing levels of generality in constitutional law); Anthony O’Rourke, Substantive Due Process for Noncitizens: Lessons from Obergefell, 114 Mich L Rev First Impressions 9, 15–16 (2015) (discussing level-of-generality issues in substantive due process). The level of generality question asks: “[A]t what level of generality should the Court describe the right previously protected and the right currently claimed? The more abstractly one states the already-protected right, the more likely it becomes that the claimed right will fall within its protection.” Laurence H. Tribe and Michael C. Dorf, Levels of Generality in the Definition of Rights, 57 U Chi L Rev 1057, 1058 (1990). Analyzing how opinions select focal text involves something related, but not identical. The level of generality question assumes a concentric regime in which every successive level of abstraction encompasses its less general articulation of rights. In selecting text, however, judges do not need to move along a single scale of abstraction but can find their focal text in a variety of different places. Determining the appropriate level of generality is like determining how wide to open the aperture; selecting text is more like deciding where to point the camera.
  • 13. See M.M. Bakhtin, The Dialogic Imagination: Four Essays 278 (Texas 1981) (Michael Holquist, ed) (Caryl Emerson and Michael Holquist, trans). Mikhail M. Bakhtin wrote that, although literary theory preferred to treat written texts in isolation, “[f]or the writer,” any “object reveals first of all precisely the socially heteroglot multiplicity of its names, definitions and value judgments,” a “multitude of routes, roads and paths that have been laid down in the object by social consciousness” and the “unfolding of social heteroglossia surrounding the object,” that is, the texts that have gone before and that coexist with the writer. Id. This “dialogic orientation,” moreover, “is a phenomenon that is, of course, a property of any discourse.” Id at 279.
  • 14. 515 US 687 (1995). Justice John Paul Stevens wrote the majority opinion, joined by Justices Sandra Day O’Connor, Anthony Kennedy, David Souter, Ruth Bader Ginsburg, and Stephen Breyer. O’Connor also wrote a concurring opinion. Justice Antonin Scalia dissented, joined by Chief Justice William Rehnquist and Justice Clarence Thomas. Id at 688.
  • 15. Id at 694–708 (Stevens); id at 717–18 (Scalia dissenting).
  • 16. Pub L No 93-205, 87 Stat 884, codified as amended at 16 USC § 1531 et seq.
  • 17. Sweet Home, 515 US at 690–91, quoting 16 USC § 1538(a)(1).
  • 18. Sweet Home, 515 US at 691, quoting 16 USC § 1532(19) (emphasis added).
  • 19. Sweet Home, 515 US at 691, quoting 50 CFR § 17.3.
  • 20. Sweet Home, 515 US at 692–93, 696–97.
  • 21. Id at 697–98.
  • 22. Id at 698–700.
  • 23. Id at 697–98.
  • 24. Sweet Home, 515 US at 717 (Scalia dissenting).
  • 25. Id (Scalia dissenting).
  • 26. Id at 717–18 (Scalia dissenting).
  • 27. Id at 717 (Scalia dissenting).
  • 28. Sweet Home, 515 US at 717 (Scalia dissenting).
  • 29. Professor Karin P. Sheldon points out that the dissent, while conceding that “take” was “elsewhere defined in the Act,” id (Scalia dissenting), effectively ignored this fact. Karin P. Sheldon, “It’s Not My Job to Care”: Understanding Justice Scalia’s Method of Statutory Interpretation through Sweet Home and Chevron, 24 BC Envir Aff L Rev 487, 530 (1997) (“‘Harm’ is subordinate to and limited by the historic and common meaning of ‘take,’ Justice Scalia concluded, because it has no legal force of its own.”). See also Sweet Home, 515 US at 697 n 10 (Stevens) (“Congress explicitly defined the operative term ‘take’ in the ESA, . . . thereby obviating the need for us to probe its meaning.”). On this criticism, rather than taking Congress’s word, the dissent held Congress to a meaning “as old as the law itself.” Sweet Home, 515 US at 717 (Scalia dissenting). See also Robin Kundis Craig, The Stevens/Scalia Principle and Why It Matters: Statutory Conversations and a Cultural Critical Critique of the Strict Plain Meaning Approach, 79 Tulane L Rev 955, 1021 (2005) (“[T]he import of Justice Scalia’s argument is that Congress cannot rid itself of the meaning and limitations of a well-established term of art, no matter how Congress defines the term in the statute, because no definition can escape the historical baggage that accompanies the operative term.”).
  • 30. Sweet Home, 515 US at 691, quoting 16 USC § 1539(a)(1)(B).
  • 31. Sweet Home, 515 US at 700.
  • 32. Id at 717 (Scalia dissenting).
  • 33. Id at 700.
  • 34. Id at 718 (Scalia dissenting).
  • 35. “[I]ntertextual relationships between a particular text and prior discourse (real or imagined) play a crucial role in shaping form, function, discourse structure, and meaning; . . . and in building competing perspectives on what is taking place.” Charles L. Briggs and Richard Bauman, Genre, Intertextuality, and Social Power, 2 J Linguistic Anthropology 131, 147 (1992). Intertextual relations set up audience expectations: “[a]s soon as we hear a generic framing device, such as ‘once upon a time,’ we unleash a set of expectations regarding narrative form and content.” Id. But they also “pertain[ ] crucially to negotiations of identity and power—by invoking a particular genre, producers of discourse assert (tacitly or explicitly) that they possess the authority needed to decontextualize discourse that bears these historical and social connections and to recontextualize it in the current discursive setting.” Id at 148. When old texts carry great weight, for instance, “creating links with traditional genres is often the most powerful strategy for creating textual authority,” id, as the Sweet Home dissent does in its invocation of the common law. “We can say, thus, that [ ] intertextuality affords great power for naturalizing both texts and the cultural reality that they represent.” Id.
  • 36. For examples of articles that discuss textualism and purposivism but ignore the issue of selecting text, see Aleinikoff, 87 Mich L Rev at 21–22 (cited in note 6) (noting that “[t]raditional debates about statutory interpretation have usually been intramural disputes” between the theories of “textualism (or plain meaning) and intentionalism (or purpose analysis)”); Fallon, 82 U Chi L Rev at 1237 (cited in note 6) (stating that “[t]he law reviews and judicial opinions both teem with debates about theories of legal interpretation,” including textualism and purposivism); Jonathan R. Siegel, Textualism and Contextualism in Administrative Law, 78 BU L Rev 1023, 1024 (1998) (“The current debate about statutory interpretation is often characterized as a battle between textualists and intentionalists.”).
  • 37. 135 S Ct 2480 (2015). Chief Justice John Roberts wrote the majority opinion, joined by Kennedy, Ginsburg, Breyer, Justice Sonia Sotomayor, and Justice Elena Kagan. Scalia dissented, joined by Thomas and Justice Samuel Alito. Id at 2484.
  • 38. Patient Protection and Affordable Care Act, Pub L No 111-148, 124 Stat 119 (2010).
  • 39. King, 135 S Ct at 2485.
  • 40. Id.
  • 41. Id at 2487 (emphasis omitted).
  • 42. Id at 2485 (emphasis added).
  • 43. King, 135 S Ct at 2485 (emphasis added).
  • 44. Id.
  • 45. Id at 2487 (emphasis omitted).
  • 46. Id at 2489.
  • 47. King, 135 S Ct at 2489.
  • 48. See id at 2491.
  • 49. Id at 2496 (Scalia dissenting). Another way of putting it is that the dissent rejected the notion—suggested by the majority—that Exchange is a legal term of art whose definition derives exclusively from the statute. See Part III.B.
  • 50. King, 135 S Ct at 2497 (Scalia dissenting).
  • 51. See, for example, id at 2496 (Scalia dissenting) (calling the majority’s interpretation “quite absurd”); id at 2496–97 (Scalia dissenting) (“You would think the answer would be obvious—so obvious there would hardly be a need for the Supreme Court to hear a case about it. . . . [A]n Exchange established by the Secretary is not an Exchange established by the State.”); id at 2497 (Scalia dissenting) (“It is hard to come up with a clearer way to limit tax credits to state Exchanges than to use the words ‘established by the State.’”). See also Part III.B (discussing invocations of ordinary language usage).
  • 52. Focusing on Exchange would make an ordinary language analysis more elusive, because Exchange in the sense of health insurance marketplace is not a term that existed in ordinary language before the Affordable Care Act created it. But that would not be a reason to reject focusing on it under the principles of textualism.
  • 53. These are, of course, not absolute distinctions: one could characterize the King opinions as focusing on different texts as well. The strategies I discuss here are often fuzzy-bordered and overlapping. My point is not to create a strict taxonomy but to point out a range of moves that opinions make and to show that they are systemically related.
  • 54. 508 US 223 (1993). O’Connor wrote the majority opinion, joined by Rehnquist, Justice Byron White, Justice Harry Blackmun, Kennedy, and Thomas. Blackmun also wrote a concurring opinion. Scalia dissented, joined by Stevens and Souter. Id at 224.
  • 55. Id at 225, quoting 18 USC § 924(c)(1) (quotation marks, ellipsis, and brackets omitted). See also Solan, 38 Loyola LA L Rev at 2033–36 (cited in note 6) (analyzing the debate in Smith).
  • 56. Smith, 508 US at 225–26.
  • 57. Id at 228–29, quoting Webster’s New International Dictionary 2806 (Merriam 2d ed 1950).
  • 58. Smith, 508 US at 228.
  • 59. Id at 243 (Scalia dissenting).
  • 60. Id at 241–42 (Scalia dissenting), quoting Webster’s New International Dictionary at 2806 (cited in note 57) (arguing that the term “use” is “elastic” insofar as its “meanings range all the way from ‘to partake of’ (as in ‘he uses tobacco’) to ‘to be wont or accustomed’ (as in ‘he used to smoke tobacco’)”).
  • 61. Smith, 508 US at 242 (Scalia dissenting).
  • 62. Id (Scalia dissenting).
  • 63. See Lawrence B. Solum, Originalism and the Natural Born Citizen Clause, 107 Mich L Rev First Impressions 22, 24 (2008) (“The notion that phrases acquire meanings that are not reducible to the meanings of the constituent words is familiar to any competent speaker of a natural language. . . . We sometimes call such phrases ‘idioms.’”).
  • 64. Professor Samuel L. Bray has explored a related phenomenon: hendiadys, a rhetorical trope “in which two terms separated by a conjunction work together as a single complex expression.” Samuel L. Bray, “Necessary and Proper” and “Cruel and Unusual”: Hendiadys in the Constitution, 102 Va L Rev 687, 688 (2016). In the terms I introduce here, hendiadys is one version of a consolidating reading, as opposed to an additive one that is “read like a telegram—a word said, then ‘Stop,’ then another word, then another ‘Stop.’” Id at 763. Bray argued for approaching his two focal phrases in a consolidating, or hendiadic, way, as a matter of original understanding. Id at 694. Bray may be right about the original understanding of these two phrases. But I believe that both hendiadic figures and other phrases susceptible to idiomatic reading are often characterized by a fundamental indeterminacy: often there will be no clearly correct, or incorrect, option. Rather, judges exercise judgment—often unacknowledged—to make such decisions.
  • 65. On textualism, see Scalia, A Matter of Interpretation at 17 (cited in note 6) (arguing that a judicial interpreter of an unclear statutory term should discern “the intent that a reasonable person would gather from the text of the law, placed alongside the remainder of the corpus juris”); Solan, 38 Loyola LA L Rev at 2029–30 (cited in note 6) (explaining that, for evidence about the meanings of unclear terms, textualists eschew pronouncements by those who wrote and voted on the terms and consider instead what those terms mean to idiomatic speakers of American English); Manning, 106 Colum L Rev at 91 (cited in note 6) (noting that “[t]extualists give primacy to the semantic context” of statutory terms) (emphasis omitted). Neither opinion in Smith took a purposivist stance: neither based its conclusions on claims about what the legislature that enacted the statute was trying to achieve, nor did either draw on sources surrounding the statute’s enactment. On seeking statutory purpose and using legislative history, see Katzmann, Judging Statutes at 31 (cited in note 6) (arguing that judges’ “fundamental task . . . is to determine what Congress was trying to do,” which requires considering “the statute’s purpose(s) as enacted by legislators,” which can be gleaned in part from “legislative materials that reliably contribute to understanding the statute’s meaning”); Abner J. Mikva, A Reply to Judge Starr’s Observations, 1987 Duke L J 380, 386 (“If judges are to make congressional primacy meaningful, they cannot afford to ignore those obvious tools which members of Congress use to explain what they are doing and to describe the meaning of the words used in the statute.”); Victoria F. Nourse, A Decision Theory of Statutory Interpretation: Legislative History by the Rules, 122 Yale L J 70, 91, 94–95 (2012) (explaining what legislative history is and how it is produced); Victoria F. Nourse, Elementary Statutory Interpretation: Rethinking Legislative Intent and History, 55 BC L Rev 1613, 1615 (2014) (arguing that “Congress has the functional equivalent of intent”).
  • 66. 529 US 120 (2000). O’Connor wrote the majority opinion, joined by Rehnquist, Scalia, Kennedy, and Thomas. Breyer dissented, joined by Stevens, Souter, and Ginsburg. Id at 123.
  • 67. Id at 125–26.
  • 68. 52 Stat 1040 (1938), codified as amended at 21 USC § 301 et seq.
  • 69. FDCA § 201(g)(3), 52 Stat at 1041, codified as amended at 21 USC § 321(g)(1)(C).
  • 70. Brown & Williamson, 529 US at 170–72 (Breyer dissenting) (discussing congressional hearings at which tobacco manufacturers denied that nicotine is addictive).
  • 71. See id at 125; Theodore W. Ruger, The Story of FDA v. Brown & Williamson: The Norm of Agency Continuity, in William N. Eskridge Jr, Philip P. Frickey, and Elizabeth Garrett, eds, Statutory Interpretation Stories 335, 346 (Foundation Press 2011) (“The newfound evidence of [tobacco companies’] nicotine manipulation [to increase cigarettes’ nicotine content], and therefore ‘intent,’ was the trigger to bring nicotine within the FDCA’s ambit. . . . Knowledge of that factual predicate did not exist until after 1992.”).
  • 72. Brown & Williamson, 529 US at 127 (brackets in original). See also Food and Drug Administration, Regulations Restricting the Sale and Distribution of Cigarettes and Smokeless Tobacco to Protect Children and Adolescents, 61 Fed Reg 44396, 44418 (1996), amending various sections of CFR Title 21. Additionally, the FDA had amassed considerable evidence that tobacco manufacturers sold nicotine products with the intent to affect human bodies, despite their public claims to the contrary. See Brown & Williamson, 529 US at 172–73 (Breyer dissenting). See also generally Ruger, The Story of FDA v. Brown & Williamson (cited in note 71).
  • 73. Brown & Williamson, 529 US at 133, citing 21 USC § 393(b)(2) (Supp III 1994). The current version of this provision reads: “The Administration shall . . . protect the public health by ensuring that . . . there is reasonable assurance of the safety and effectiveness of devices intended for human use.” 21 USC § 393(b) (paragraph breaks omitted).
  • 74. Brown & Williamson, 529 US at 134, quoting United States v Rutherford, 442 US 544, 556 (1979).
  • 75. Brown & Williamson, 529 US at 134, quoting 61 Fed Reg at 44412 (cited in note 72).
  • 76. Brown & Williamson, 529 US at 143 (“[I]f tobacco products were within the FDA’s jurisdiction, the Act would require the FDA to remove them from the market entirely.”).
  • 77. Id at 137 (“Congress, however, has foreclosed the removal of tobacco products from the market.”).
  • 78. Id at 126 (concluding that the FDA’s assertion of “authority is inconsistent with the intent that Congress has expressed in the FDCA’s overall regulatory scheme and in the tobacco-specific legislation that it has enacted subsequent to the FDCA”).
  • 79. See id at 137–39, 143–59.
  • 80. See Brown & Williamson, 529 US at 133–37, 139–43.
  • 81. See id at 131–32:

    We need not resolve this question [of whether the FDA may impute an intent to affect bodily structure or function], however, because assuming, arguendo, that a product can be “intended to affect the structure or any function of the body” absent claims of therapeutic or medical benefit, the FDA’s claim to jurisdiction contravenes the clear intent of Congress.

  • 82. Id at 133.
  • 83. Id at 163 (Breyer dissenting) (“The inferences that the majority draws from later legislative history are not persuasive, since . . . one can just as easily infer from the later laws that Congress did not intend to affect the FDA’s tobacco-related authority at all.”).
  • 84. Brown & Williamson, 529 US at 163 (Breyer dissenting).
  • 85. Id at 162 (Breyer dissenting).
  • 86. See id at 165 (Breyer dissenting).
  • 87. See John F. Manning, The Absurdity Doctrine, 116 Harv L Rev 2387, 2391 (2003) (“If one accepts the textualist critique of strong intentionalism, it is difficult to sustain the absurdity doctrine.”).
  • 88. For instance, the respondents’ brief in Sweet Home argued, “While much attention has focused on the meaning of ‘harm,’ the operative statutory term is ‘take.’ It is the ‘take’ of listed wildlife, as defined in the ESA, that § 9 forbids.” Brief for Respondents, Babbitt v Sweet Home Chapter of Communities for a Great Oregon, Docket No 94-859, *18 (US filed Mar 24, 1995) (available on Westlaw at 1995 WL 130541).
  • 89. I elaborate on this suggestion in Part IV.B.
  • 90. Sweet Home, 515 US at 717 (Scalia dissenting) (emphasis omitted).
  • 91. Id at 697 n 10 (Stevens).
  • 92. Smith, 508 US at 242 (Scalia dissenting).
  • 93. See generally Marcel Mauss, Techniques of the Body, 2 Economy & Society 70 (1973).
  • 94. Id at 71–72.
  • 95. Id at 75.
  • 96. Mauss quotes an ethnography describing how Maori mothers “drilled their daughters” in walking with the “gait . . . admired” in their society. Id at 74. Although this was how women walked in that society, “[t]his was an acquired, not a natural way of walking.” Id. In fact, Mauss continues, “there is perhaps no ‘natural way’ for the adult” to walk—all walks are learned in particular social settings. Id.
  • 97. The extent to which text selection choices are conscious is a question that lies beyond the scope of this Article, and would often not be discernible from the text of the opinions analyzed here.
  • 98. See Charles Goodwin and Alessandro Duranti, Rethinking Context: An Introduction, in Alessandro Duranti and Charles Goodwin, eds, Rethinking Context: Language as an Interactive Phenomenon 1, 3 (Cambridge 1992).
  • 99. Id at 6–9.
  • 100. Id at 3.
  • 101. Id.
  • 102. Richard Bauman and Charles L. Briggs, Poetics and Performance as Critical Perspectives on Language and Social Life, 19 Ann Rev Anthropology 59, 73 (1990).
  • 103. Id.
  • 104. Id.
  • 105. For my purposes here, two characteristics stand out as particularly important. First, judicial opinions are performative, or creative, not only in the everyday senses of those words, but in the technical linguistic sense that they help constitute the very things they refer to in the world. Part IV elaborates on this assertion. Second, judicial opinions in our system of adjudication ground their legitimacy in reason-giving. Mathilde Cohen, The Rule of Law as the Rule of Reasons, 96 Archiv für Rechts- und Sozialphilosophie [Archives Phil L & Soc Phil] 1, 2 (2010) (“[L]egal reason giving is one of the essential properties of the concept of the rule of law, if not the essential one.”). Part IV explains how my approach helps evaluate this function.
  • 106. Mikhail Bakhtin, The Heteroglot Novel, in Pam Morris, ed, The Bakhtin Reader: Selected Writings of Bakhtin, Medvedev and Voloshinov 88, 107–08 (Edward Arnold 1994) (contrasting “[o]vert polemic . . . directed at another’s discourse, which it refutes,” with “hidden polemic,” in which “discourse is directed toward an ordinary referential object, . . . only indirectly striking a blow at the other’s discourse, clashing with it, as it were, within the [referential] object itself”).
  • 107. For a fuller discussion of speech acts and creative utterances, see Part IV.C.
  • 108. Paul J. Thibault, Communicating and Interpreting Relevance through Discourse Negotiation: An Alternative to Relevance Theory–a Reply to Franken, 31 J Pragmatics 557, 560 (1999). See also, for example, Scalia, A Matter of Interpretation at 37 (cited in note 6) (“In textual interpretation, context is everything.”); Fallon, 99 Cornell L Rev at 687 (cited in note 6), citing Manning, 106 Colum L Rev at 73, 79–80 (cited in note 6) (stating that textualists and purposivists “concur . . . [that] the meaning of the words of a statute . . . depends on context”); Siegel, 78 BU L Rev at 1027–28 (cited in note 36) (noting that textualists agree with intentionalists that “text is not self-interpreting and that text takes its meaning from and must be understood in context”); Antonin Scalia and Bryan A. Garner, Reading Law: The Interpretation of Legal Texts 16, 32–33 (Thomson/West 2012).
  • 109. I suggest why the genre is oriented this way in Part IV. The extent to which individual judges strive to make their situating choices seem inevitable, and whether they do so consciously or simply as part of a technique of interpretation, are interesting questions that lie beyond the scope of this Article. Here, I address not the individual psychology of judges but the efficacy and structure of judicial opinions.
  • 110. See Fallon, 99 Cornell L Rev at 695 (cited in note 6) (“There . . . could be[ ] no purely non-normative criterion precisely marking what should be deemed relevant or irrelevant.”). This is as true for textual interpretation as it is for constructing group identities, which ask people to interpret themselves and their place in the world around them with reference to factors that are highly influenced by local cultural norms, understandings of selfhood, and desiderata of legitimacy. See Anya Bernstein, Parameters of Legitimation and the Environmental Future of a Taipei Neighborhood, in Sylvia Hood Washington, Paul C. Rosier, and Heather Goodall, eds, Echoes from the Poisoned Well: Global Memories of Environmental Injustice 311, 315–17 (Lexington 2006).
  • 111. Thibault, 31 J Pragmatics at 561 (cited in note 108). Although I talk a lot about relevance here, I do not subscribe to what is known as “relevance theory,” which places ultimate meanings in the speaker’s thoughts rather than in the discursive elaboration of utterances. Id (“Relevance theory [ ] assumes that . . . the speaker’s thought[ ] is privileged in the interpretation of utterances.”). In other words, I see communication and meaning-making as a fundamentally social, interactive practice.
  • 112. In a classic article, Professors William Eskridge and Philip Frickey argued that the Supreme Court “considers a broad range of textual, historical, and evolutive evidence when it interprets statutes.” Eskridge and Frickey, 42 Stan L Rev at 322 (cited in note 7). I add to their analysis by both broadening and specifying that range, showing both how opinions often go further afield and how they cut up the categories Eskridge and Frickey identify in varied ways.
  • 113. The psychological processes of situating a text—whether it is conscious or purposeful, what judges feel about it—reveal little about whether these practices occur and how they structure judicial opinions. Evidence for such psychological processes, more­over, resides in the invisible inner states of individual judges. I eschew discussion about how judges think or what they feel. Rather, I base my analysis on the written products that judges have made available. These may not reveal judges’ inner states, but they do reveal something more important: the practices and structures that shape these efficacious legal utterances.
  • 114. See Scalia, A Matter of Interpretation at 17 (cited in note 6). See also Siegel, 78 BU L Rev at 1025 (cited in note 36) (discussing this passage).
  • 115. See text accompanying notes 21–34.
  • 116. Sweet Home, 515 US at 690–91, quoting 16 USC § 1538(a)(1).
  • 117. Sweet Home, 515 US at 691, quoting 16 USC § 1532(19).
  • 118. Sweet Home, 515 US at 717 (Scalia dissenting) (emphasis omitted).
  • 119. Id at 697 n 10.
  • 120. Id at 717 (Scalia dissenting).
  • 121. 534 US 204 (2002). Justice Scalia wrote the majority opinion, joined by Chief Justice Rehnquist and Justices O’Connor, Kennedy, and Thomas. Justice Stevens dissented. Justice Ginsburg also filed a separate dissenting opinion, joined by Stevens and Justices Souter and Breyer. Id at 206.
  • 122. Id at 207–09.
  • 123. Pub L No 93-406, 88 Stat 829.
  • 124. Knudson, 534 US at 209 (ellipsis omitted), quoting ERISA § 502(a)(3), 88 Stat at 891, codified as amended at 29 USC § 1132(a)(3).
  • 125. Knudson, 534 US at 210–12, quoting Mertens v Hewitt Associates, 508 US 248, 256 (1993).
  • 126. See FRCP 2 (“There is one form of action—the civil action.”); Stephen N. Subrin, How Equity Conquered Common Law: The Federal Rules of Civil Procedure in Historical Perspective, 135 U Pa L Rev 909, 912 (1987) (arguing that equity principles endure in, and perhaps dominate, the Federal Rules of Civil Procedure).
  • 127. Knudson, 534 US at 221–22 (Stevens dissenting) (characterizing Ginsburg’s dissenting opinion).
  • 128. Id (Stevens dissenting).
  • 129. Id at 233 (Ginsburg dissenting).
  • 130. The Ginsburg dissent (and scholars) also challenged the majority’s understanding of what the term would have meant in the days of the divided bench. See id at 229–30 (Ginsburg dissenting) (“The[ ] cases establish what the Court . . . cannot dispute: Restitution was within the recognized power . . . of a court of equity.”) (quotation marks omitted); Samuel L. Bray, The Supreme Court and the New Equity, 68 Vand L Rev 997, 1000 (2015) (“Despite all [its] appeals to history and tradition [in Knudson and other equity-related cases], what the Court is offering is not something that most historians would recognize, for it does not reflect the complexity and contingency of equity’s past.”); Tracy A. Thomas, Justice Scalia Reinvents Restitution, 36 Loyola LA L Rev 1063, 1064 (2003) (“[T]he Supreme Court distorted history and equity to reach its result [in Knudson].”). Although it is not central to my point here, it is certainly worth noting that invoking the relevance of a given historical era does not settle the interpretive question. An opinion must still make contestable assertions about what that historical era implied for a given legal text.
  • 131. Scalia, A Matter of Interpretation at 17 (cited in note 6).
  • 132. To the extent that textualism gives judges reason to prefer one option, it would seem to favor looking to the law contemporaneous to the statute’s enactment, insofar as readers at that time would be likely to read the statute’s words as having the meanings most familiar to them. Scalia and Garner, Reading Law at 33 (cited in note 108). But this was the option rejected in Knudson by the avowedly textualist Scalia. Knudson, 534 US at 212–13.
  • 133. See Adrian Vermeule, The Judiciary Is a They, Not an It: Interpretive Theory and the Fallacy of Division, 14 J Contemp Legal Issues 549, 567–69 (2005).
  • 134. See text accompanying notes 68–78.
  • 135. See generally Kenneth A. Shepsle, Congress Is a “They,” Not an “It”: Legislative Intent as Oxymoron, 12 Intl Rev L & Econ 239 (1992). This criticism has itself been criticized as overly individualistic and insufficiently cognizant of the dynamics of group action. See Nourse, 55 BC L Rev at 1615 (cited in note 65) (“Congress has the functional equivalent of intent.”).
  • 136. See Brown & Williamson, 529 US at 155 (“Taken together, these actions by Congress over the past 35 years preclude an interpretation of the FDCA that grants the FDA jurisdiction to regulate tobacco products.”).
  • 137. See Andre Gallois, Identity over Time (Stanford Encyclopedia of Philosophy, Oct 6, 2016), online at http://plato.stanford.edu/archives/win2016/entries/identity-time (visited Mar 26, 2017) (Perma archive unavailable). For a more detailed introduction to the philosophical puzzle of the ship of Theseus, see generally Graeme S. Cumming and John Collier, Change and Identity in Complex Systems, 10 Ecol and Society 29 (June 2005).
  • 138. Aleinikoff, 87 Mich L Rev at 21 (cited in note 6).
  • 139. See Part IV.B.
  • 140. 490 US 504 (1989). Stevens wrote the majority opinion, joined by Rehnquist, Justice Byron White, O’Connor, and Kennedy. Scalia wrote an opinion concurring in the judgment. Justice Harry Blackmun dissented, joined by Justices William Brennan and Thurgood Marshall. Id at 505.
  • 141. Green concerned a provision in the Federal Rules of Evidence, which fall squarely within the realm of legal pronouncement, though they are not statutes. See id. For a discussion of the way that doctrine and scholarship generally do not distinguish Federal Rules from statutes, but why they ought to, see Lumen N. Mulligan and Glen Staszewski, Civil Rules Interpretive Theory, 101 Minn L Rev *7–8 (forthcoming 2017), archived at http://perma.cc/39BV-ZGKP.
  • 142. Green, 490 US at 509 (emphasis added), quoting FRE 609(a) (1975). The Rule read, in relevant part:

    For the purpose of attacking the credibility of a witness, evidence that the witness has been convicted of a crime shall be admitted . . . only if the crime [ ] was punishable by death or imprisonment in excess of one year . . . , and the court determines that the probative value of admitting this evidence outweighs its prejudicial effect to the defendant.

    Green, 490 US at 509, quoting FRE 609(a) (1975).

  • 143. See Green, 490 US at 509 (Stevens) (“[The] literal reading would compel an odd result in a case like this.”); id at 527 (Scalia concurring in the judgment) (“[I]f interpreted literally, [the provision] produces an absurd, and perhaps unconstitutional, result.”); id at 530 (Blackmun dissenting) (“The majority concludes that Rule 609(a)(1) cannot mean what it says on its face. . . . I fully agree.”).
  • 144. Id at 510.
  • 145. These include the Fifth Amendment’s option “not to testify at trial” and the Sixth Amendment’s grant of “certain fair trial rights not enjoyed by the prosecution.” Id at 510–11.
  • 146. Id at 510.
  • 147. Green, 490 US at 510–11.
  • 148. Id at 522–24.
  • 149. See id at 527–30 (Scalia concurring in the judgment).
  • 150. Id at 529 (Scalia concurring in the judgment). The concurrence thus found it “[q]uite obvious[ ]” that the majority’s reading of the Rule to give special protection only to criminal defendants “does the least violence to the text.” Id.
  • 151. Green, 490 US at 529 (Scalia concurring in the judgment).
  • 152. The majority did not, for instance, consider whether some procedural inequalities between civil parties might be allowed, and under what circumstances. How, for instance, would the majority approach a supplemental jurisdiction provision that allows defendants, but not plaintiffs, in a civil suit to join parties who would destroy diversity of citizenship? See 28 USC § 1367(b). Such inequality may be justified with reference to other values, but the majority opinion did not consider what kind of justification might be required. Similarly, one could imagine a legislative policy of discouraging unnecessary lawsuits by providing extra protections for civil defendants. The majority opinion neither considered such a possibility nor evaluated the constitutional issues it might raise.
  • 153. Green, 490 US at 527 (Scalia concurring in the judgment).
  • 154. Id at 529 (Scalia concurring in the judgment). In keeping with the general nature of this allusion to legal principles, the concurrence cited no legal stricture or Rule provision expressing special solicitude for criminal defendants.
  • 155. Similarly, the concurrence stated that its goal was to find the reading that was the most “compatible with the surrounding body of law” without referring to any particular text or aspect of that body of law. Id at 528 (Scalia concurring in the judgment).
  • 156. Id at 533 (Blackmun dissenting) (emphasis omitted and ellipsis in original), quoting FRE 102 (1975).
  • 157. Green, 490 US at 533 (Blackmun dissenting).
  • 158. Id (Blackmun dissenting), quoting id at 529 (Scalia concurring in the judgment).
  • 159. See id at 533 (Blackmun dissenting). The dissent cited support in the conference committee’s report, which found the possibility that prior conviction evidence would damage the integrity of the lawsuit intolerable, but declined to ask judges to consider the “danger of prejudice to a . . . [nondefendant] witness’ reputation.” Id at 531–32 (Blackmun dissenting), quoting Federal Rules of Evidence Conference Report, HR Conf Rep No 93-1597, 93d Cong, 2d Sess 9–10 (1974), reprinted in 1974 USCCAN 7098, 7103. The report thus focused not on maintaining equality between parties or protecting criminal defendants, but on excluding evidence that “present[ed] a danger of improperly influencing the outcome of the trial.” Green, 490 US at 532 (Blackmun dissenting), quoting HR Conf Rep No 93-1597 at 9 (cited in note 159). Green, in which evidence of a conviction not relevant to the liability claim stood to damage the integrity of the trial, presented precisely this kind of danger.
  • 160. Scalia wrote separately to object to the majority’s detailed legislative and prelegislative history, which he found irrelevant to determining the meaning of the text:

    I think it entirely appropriate to consult all public materials . . . to verify that what seems to us an unthinkable disposition (civil defendants but not civil plaintiffs receive the benefit of weighing prejudice) was indeed unthought of, and thus to justify a departure from the ordinary meaning of the word “defen­dant” in the Rule. For that purpose, however, it would suffice to observe that counsel have not provided, nor have we discovered, a shred of evidence that anyone has ever proposed or assumed such a bizarre disposition.

    Green, 490 US at 527 (Scalia concurring in the judgment).

  • 161. 461 US 574 (1983). Chief Justice Warren Burger wrote the majority opinion, joined by Brennan, White, Marshall, Blackmun, Stevens, and O’Connor. Justice Lewis F. Powell Jr joined in Part III of the majority opinion and also wrote an opinion concurring in part and concurring in the judgment. Rehnquist dissented. Id at 576.
  • 162. Id at 585 (ellipses and brackets in original), quoting 26 USC § 501(c)(3).
  • 163. Bob Jones, 461 US at 577–78.
  • 164. The case involved two religious educational institutions. Bob Jones University began accepting African Americans married “within their race” in 1971. Id at 580. After a Supreme Court decision prohibited racial exclusion from private schools in 1975, Bob Jones began admitting unmarried African Americans, but its student code of conduct prohibited engaging in or advocating interracial relationships. Id at 580–81. Goldsboro Christian Schools was a K–12 private school that had a policy of accepting only whites or, sometimes, students with one white parent. Id at 583.
  • 165. Powell added an eighth vote to Part III of the majority opinion, which rejected the schools’ argument that the denial of a tax exemption constituted religious regulation in violation of the First Amendment. See id at 606 (Powell concurring in part and concurring in the judgment); id at 602–04 (Burger) (discussing appellants’ religious freedom arguments). The concurrence seemed to accept the majority’s invocation of common-law charity and its interpretation of public policy in this case, but opined that the majority went too far in its description of the IRS’s power to determine what kinds of activities served or violated public policy. Id at 606, 611 (Powell concurring in part and concurring in the judgment).
  • 166. Id at 579 (brackets omitted), quoting Rev Rul 71-447, 1971-2 Cum Bull 230.
  • 167. Bob Jones, 461 US at 586. This interpretation, proposed by the appeals court below, had been formalized by the IRS in Revenue Ruling 71-447. See id at 578–79; Rev Rul 71-447, 1971-2 Cum Bull 230.
  • 168. Bob Jones, 461 US at 592.
  • 169. Id at 613–14 (Rehnquist dissenting).
  • 170. See Scalia, A Matter of Interpretation at 17–18 (cited in note 6).
  • 171. Id at 17.
  • 172. 500 US 453 (1991). Rehnquist wrote the majority opinion, joined by White, Blackmun, O’Connor, Scalia, Kennedy, and Souter. Stevens dissented, joined by Marshall. Id at 455.
  • 173. Id, quoting 21 USC § 841(b)(1)(B)(v).
  • 174. Chapman affirmed the decision in United States v Marshall, 908 F2d 1312 (7th Cir 1990) (en banc); both the majority and dissent in Chapman drew heavily on the opinions in Marshall.
  • 175. Chapman, 500 US at 457.
  • 176. Id at 457–58.
  • 177. Id at 455 (“We hold that it is the weight of the blotter paper containing LSD, and not the weight of the pure LSD, which determines eligibility for the minimum sentence.”).
  • 178. Id at 462, quoting Webster’s Third New International Dictionary 1449 (Merriam 1986).
  • 179. Chapman, 500 US at 462, citing 9 Oxford English Dictionary 921 (Clarendon 2d ed 1989).
  • 180. Chapman, 500 US at 462.
  • 181. Id.
  • 182. Id.
  • 183. Id.
  • 184. Chapman, 500 US at 475 (Stevens dissenting), quoting Marshall, 908 F2d at 1333 (Posner dissenting).
  • 185. Chapman, 500 US at 458 (Rehnquist).
  • 186. Id at 457.
  • 187. Id at 475 (Stevens dissenting).
  • 188. Id at 474 (Stevens dissenting), quoting Marshall, 908 F2d at 1333 (Posner dissenting).
  • 189. Chapman, 500 US at 475 n 12 (Stevens dissenting), quoting Marshall, 908 F2d at 1334 (Posner dissenting).
  • 190. 547 US 715 (2006). Scalia wrote the plurality opinion, joined by Chief Justice Roberts, Justice Thomas, and Justice Samuel Alito. Roberts also wrote a concurring opinion, and Kennedy wrote an opinion concurring in the judgment. Stevens dissented, joined by Souter, Ginsburg, and Breyer. Breyer also filed a separate dissenting opinion. Id at 718.
  • 191. Federal Water Pollution Act Amendments of 1972, Pub L No 92-500, 86 Stat 816, codified as amended at 33 USC § 1251 et seq.
  • 192. See generally Rapanos, 547 US 715. For a fuller discussion of the role of facts in this case, on which my discussion of Rapanos draws, see Todd S. Aagaard, Factual Premises of Statutory Interpretation in Agency Review Cases, 77 Geo Wash L Rev 366, 368–69, 372–73 (2009).
  • 193. Rapanos, 547 US at 723 (Scalia) (plurality), quoting 33 USC § 1344(a), (d).
  • 194. Rapanos, 547 US at 724 (Scalia) (plurality), citing 33 CFR § 328.3(a)(7). The Act itself defined “navigable waters” as “waters of the United States.” Rapanos, 547 US at 723 (Scalia) (plurality), quoting 33 USC § 1362(7).
  • 195. Rapanos, 547 US at 730–31 (Scalia) (plurality).
  • 196. Id at 744 (Scalia) (plurality). As Professor Todd Aagaard notes, “The plurality distinguished dredged or fill material from ‘traditional water pollutants’ inasmuch as dredged or fill material ‘are solids that do not readily wash downstream.’” Aagaard, 77 Geo Wash L Rev at 369 n 15 (cited in note 192), quoting Rapanos, 547 US at 723 (Scalia) (plurality); Rapanos, 547 US at 744 (Scalia) (plurality) (“‘[D]redged or fill material[ ]’ [ ] is typically deposited for the sole purpose of staying put.”).
  • 197. Rapanos, 547 US at 723 (Scalia) (plurality), quoting 33 USC § 1344(a), (d).
  • 198. Rapanos, 547 US at 807 (Stevens dissenting), quoting United States v Deaton, 332 F3d 698, 707 (4th Cir 2003).
  • 199. See Deaton, 332 F3d at 707 (collecting cases).
  • 200. Rapanos yielded three other opinions. Kennedy concurred with the judgment of the plurality, but was not convinced by its factual assertion. To him, it “seem[ed] plausible” that fill material “could travel downstream.” Rapanos, 547 US at 774 (Kennedy concurring in the judgment). Breyer joined the dissent and also wrote separately to emphasize the propriety of the Corps’s assertion of jurisdiction based on Congress’s delegation of authority, and to encourage the agency to pass a regulation interpreting the term. Id at 811–12 (Breyer dissenting). Roberts joined the plurality and also wrote separately to castigate the Corps for failing to “develop[ ] some notion of an outer bound to the reach of [its] authority” through rulemaking, and for instead “adher[ing] to its essentially boundless view of the scope of its power.” Id at 758 (Roberts concurring).
  • 201. Aagaard, 77 Geo Wash L Rev at 369–70 (cited in note 192).
  • 202. The plurality itself did not cite any empirical analyses. The plurality stated: “[A]mici cite multiple empirical analyses . . . including one which concludes that ‘the idea that the discharge of dredged or fill material into isolated waters, ephemeral drains or non-tidal ditches will pollute navigable waters located any appreciable distance from them lacks credibility.’” Rapanos, 547 US at 744 n 11 (Scalia) (plurality) (brackets omitted), quoting Robert J. Pierce, Technical Principles Related to Establishing the Limits of Jurisdiction for Section 404 of the Clean Water Act *34–40 (Wetland Science Applications Inc, Apr 2003), archived at http://perma.cc/NX5U-3AP9. For the amicus briefs that cited this source, see Brief of the International Council of Shopping Centers, National Multi Housing Council, National Association of Industrial and Office Properties, Real Estate Roundtable, Associated General Contractors of America, American Resort Development Association, and National Association of Real Estate Investment Trusts as Amici Curiae in Support of Petitioners, Rapanos v United States, Docket Nos 04-1034 and 04-1384, *26–27 (US filed Dec 2, 2005) (available on Westlaw at 2005 WL 3294931); Brief of Pulte Homes, Inc., Centex Homes, Hovnanian Enterprises, Inc., KB Home, Lennar Corporation and M.D.C. Holdings, Inc. as Amici Curiae in Support of Petitioners, Rapanos v United States, Docket Nos 04-1034 and 04-1384, *20–21 (US filed Dec 2, 2005) (available on Westlaw at 2005 WL 3308790); Brief of Amici Curiae Foundation for Environmental and Economic Progress, National Association of Realtors®, Utility Water Act Group, and Chamber of Commerce of the United States of America in Support of Petitioners, Rapanos v United States, Docket Nos 04-1034 and 04-1384, *29 & n 53 (US filed Dec 2, 2005) (available on Westlaw at 2005 WL 3322931) (“Fill material does not migrate.”). The plurality cited this Wetland Science Applications, Inc, document as available on the website of the Wetland Training Institute. Rapanos, 547 US at 744 n 11 (Scalia) (plurality) (citing the piece as “available at http://www.wetlandtraining.com/tpreljscwa.pdf”). The Wetland Training Institute, Inc, is a for-profit entity that describes itself as “an association of more than twenty individuals from industry, academia, and government” that “was formed . . . to meet the educational and technical resource needs of professionals involved in all aspects of water resource conservation, management, and regulation.” About Wetland Training Institute (Wetland Training Institute, Inc), archived at http://perma.cc/9ALL-K7U9. Its website does not indicate any vetting or review process for its material or factual assertions.

    The quoted Wetland Science Applications, Inc, document, similarly, did not cite a scientific source for its contention about fill material. Nor does this sole-authored document indicate that it received any kind of scientific vetting or peer review. The section cited by the plurality, moreover, contains primarily legal analysis. For instance, the sentence quoted by the plurality is followed by this explanation:

    The discharge of fill material as has been classically regulated by the COE, results in the “discharge” of a “pollutant” that is far different from any other regulated by the CWA. Congress recognized the difference in the nature of discharges of dredged and fill material when it specifically carved-out Section 404 and named the Secretary of the Army to administer it.

    Pierce, Technical Principles at *34 (cited in note 202). See also Aagaard, 77 Geo Wash L Rev at 369–70 (cited in note 192).

  • 203. See Rapanos, 547 US at 807 (Stevens dissenting).
  • 204. Aagaard, 77 Geo Wash L Rev at 370 (cited in note 192).
  • 205. Rapanos, 547 US at 807 (Stevens dissenting), citing generally Don C. Erman and Vernon M. Hawthorne, The Quantitative Importance of an Intermittent Stream in the Spawning of Rainbow Trout, 105 Transactions Am Fisheries Society 675 (1976).
  • 206. See Samuel A. Thumma and Jeffrey L. Kirchmeier, The Lexicon Has Become a Fortress: The United States Supreme Court’s Use of Dictionaries, 47 Buff L Rev 227, 248–60 (1999) (discussing the prevalence of dictionary usage in Supreme Court opinions); Stephen C. Mouritsen, Note, The Dictionary Is Not a Fortress: Definitional Fallacies and a Corpus-Based Approach to Plain Meaning, 2010 BYU L Rev 1915, 1921–41 (discussing a number of incorrect assumptions about dictionaries that pervade Supreme Court opinions).
  • 207. See Chapman, 500 US at 462, quoting Webster’s Third at 1449 (cited in note 178) and 9 OED at 921 (cited in note 179).
  • 208. Sweet Home, 515 US at 697:

    The dictionary definition of the verb form of “harm” is “to cause hurt or damage to: injure.” Webster’s Third New International Dictionary 1034 (1966). . . .

    . . . [T]he dictionary definition does not include the word “directly” or suggest in any way that only direct or willful action that leads to injury constitutes “harm.”

  • 209. Id at 717 (Scalia dissenting):

     To “take,” when applied to wild animals, means to reduce those animals, by killing or capturing, to human control. See, e.g., 11 Oxford English Dictionary (1933) (“Take . . . To catch, capture (a wild beast, bird, fish, etc.)”); Webster’s New International Dictionary of the English Language (2d ed. 1949) (take defined as “to catch or capture by trapping, snaring, etc., or as prey”).

    (ellipsis in original).

  • 210. Rapanos, 547 US at 807 (Stevens dissenting).
  • 211. Id at 744 n 11 (Scalia) (plurality).
  • 212. 135 S Ct 785 (2015). Scalia wrote the opinion for a unanimous Court. Id at 787.
  • 213. Id, quoting 18 USC § 2113(e).
  • 214. Whitfield, 135 S Ct at 788, quoting Brief of Appellant, United States v Whitfield, Appeal No 10-5217, *50–52 (4th Cir filed July 27, 2011) (available on Westlaw at 2011 WL 3159889).
  • 215. Whitfield, 135 S Ct at 788, citing Oxford English Dictionary 60 (Clarendon 1st ed 1933).
  • 216. Whitfield, 135 S Ct at 788.
  • 217. Rapanos, 547 US at 744 n 11 (Scalia) (plurality).
  • 218. Whitfield, 135 S Ct at 788 n 4, quoting Vernonia School District 47J v Acton, 515 US 646, 650 (1995). See also Whitfield, 135 S Ct at 788 (“Congress enacted the forced-accompaniment provision in 1934 after an outbreak of bank robberies committed by John Dillinger and others.”) (quotation marks omitted).
  • 219. See Zachary Price, The Rule of Lenity as a Rule of Structure, 72 Fordham L Rev 885, 885 (2004) (“[T]he rule of lenity . . . directs courts to construe statutory ambiguities in favor of criminal defendants.”).
  • 220. See, for example, J.M. Balkin and Sanford Levinson, Constitutional Grammar, 72 Tex L Rev 1771, 1773 (1994) (“At some point, misuse becomes common use, and common use becomes accepted use.”); William Labov, Principles of Linguistic Change: Cognitive and Cultural Factors 2–5 (Wiley-Blackwell 2010). For an indication that some words used by Austen have acquired new meanings over time, see Jane Austen, The Annotated Emma 228 n 33 (Anchor 2012) (David M. Shapard, ed) (explaining that the phrase “making violent love to her” meant “professing or demonstrating passionate love for her” and stating that “‘making love’ had no further meaning then”) (emphasis omitted).
  • 221. A good example is Muscarello v United States, 524 US 125 (1998), in which the majority cited, among other sources, The King James Bible (translated in Britain between 1604 and 1611), Robinson Crusoe (originally published in Britain in 1719), and Moby-Dick (originally published in America in 1851), to interpret an American statute enacted in 1968. Muscarello, 524 US at 129.
  • 222. Brown & Williamson, 529 US at 162 (Breyer dissenting) (“[T]he majority nowhere denies [that] . . . tobacco products . . . fall within the scope of th[e] statutory definition, read literally.”).
  • 223. 21 USC § 321(g)(1)(C).
  • 224. Allison Orr Larsen, Factual Precedents, 162 U Pa L Rev 59, 61 (2013). See also Stuart Minor Benjamin, Stepping into the Same River Twice: Rapidly Changing Facts and the Appellate Process, 78 Tex L Rev 269, 273 (1999) (“Judicial opinions are filled with assertions about the state of the world.”).
  • 225. Allison Orr Larsen, Confronting Supreme Court Fact Finding, 98 Va L Rev 1255, 1263 (2012). Scholars like Aagaard, Professor Allison Orr Larsen, and Professor Stuart Minor Benjamin have viewed such judicial fact invocation skeptically, arguing for procedures to make the use of research and fact more transparent, predictable, and reliable, and to limit the power of judicially asserted facts. See generally Aagaard, 77 Geo Wash L Rev 366 (cited in note 192); Larsen, 98 Va L Rev 1255 (cited in note 225); Benjamin, 78 Tex L Rev 269 (cited in note 224).
  • 226. Aagaard finds that “courts often interpret statutes based on factual premises that are outside of the judges’ expertise and experience.” Aagaard, 77 Geo Wash L Rev at 371 (cited in note 192).
  • 227. See, for example, id at 372–73 (arguing that, because “courts . . . make a mess of the factual premises of their statutory interpretation decisions,” agencies should have the authority to “revisit factual premises underlying judicial decisions interpreting statutes administered by agencies”); Larsen, 98 Va L Rev at 1262 (cited in note 225) (“[T]his new approach to fact finding is troubling and requires us to rethink our entire process for judicial evaluation of legislative facts.”).
  • 228. Larsen, 98 Va L Rev at 1263 (cited in note 225) (showing that “virtually all of the Justices” engage in independent fact-finding, creating empirical contexts for their interpretive claims irrespective of “whether they are traditionally labeled liberal or conservative, and [that] they cite authorities on a wide range of subject matters (from biology to history to golf)”).
  • 229. Gonzales v Carhart, 550 US 124, 159 (2007).
  • 230. Wen-Shing Tseng and Jon Streltzer, Cultural Competence in Health Care: A Guide for Professionals 119 (Springer 2008).
  • 231. Burwell v Hobby Lobby Stores, Inc, 134 S Ct 2751, 2795 (2014) (Ginsburg dissenting).
  • 232. Winnifred Fallers Sullivan, The Impossibility of Religious Freedom (The Immanent Frame, July 8, 2014), archived at http://perma.cc/5AV7-TXPD.
  • 233. Richardson v McKnight, 521 US 399, 409 (1997) (“Competitive pressures . . . also [mean] that a firm whose guards are too timid will face threats of replacement by other firms with records that demonstrate their ability to do both a safer and a more effective job.”).
  • 234. See generally Sally Q. Yates, Memorandum for the Acting Director, Federal Bureau of Prisons: Reducing Our Use of Private Prisons (DOJ, Aug 18, 2016), archived at http://perma.cc/QDY9-9TZX (laying out a plan to eliminate the Department of Justice’s use of privately run prisons on the ground that they are neither cost-effective nor safe). Notably, the Department of Justice memo did not suggest that prisoner lawsuits could bring private prisons into line; instead, it outlined an administrative plan to cease using private prisons. See generally id. As attorney general under President Donald Trump, Jefferson B. Sessions recently rescinded this memorandum. See generally Jefferson B. Sessions III, Memorandum for the Acting Director, Federal Bureau of Prisons: Rescission of Memorandum on Use of Private Prisons (DOJ, Feb 21, 2017), archived at http://perma.cc/5SAK-94J6. This notice of rescission, however, did not contradict the conclusions or analysis of the Yates memorandum; nor did it contain any analysis to suggest that private prisons were indeed safe or efficient. It merely stated that the Yates memorandum “changed long-standing policy and practice, and impaired the Bureau’s ability to meet the future needs of the federal correctional system.” Id at *1. See also Jon D. Michaels, Privatization’s Pretensions, 77 U Chi L Rev 717, 725–47 (2010) (arguing that the privatization of government functions allows executives to aggrandize powers that are otherwise outside their legal prerogative).
  • 235. The idea that some terms may be within the specialized meaning-giving domain of some expert community is widely accepted in discussions of statutory interpretation. See, for example, Ian Haney López, White by Law: The Legal Construction of Race 1–7 (New York 2006) (discussing the contestation between everyday and purportedly scientific notions of race in Supreme Court immigration cases). See also Anya Bernstein, Differentiating Deference, 33 Yale J Reg 1, 22–23 (2016) (discussing the relation between terminology based in legal expertise and that based in other kinds of expertise). Favorite cases like Nix v Hedden, 149 US 304 (1893)—in which the Supreme Court had to determine whether a tomato was a fruit or a vegetable for purposes of a customs statute—illustrate how judges must sometimes choose between everyday and technical meanings. See id at 306–07. Nix recognized that tomatoes were botanically classified as fruits, but held that they constituted vegetables for the purposes of a tariff statute. The opinion invoked American cultural habits, such as broadly accepted times and manners for eating, to support its conclusion:

    Botanically speaking, tomatoes are the fruit of a vine, just as are cucumbers, squashes, beans and peas. But in the common language of the people, whether sellers or consumers of provisions, all these are vegetables, which are grown in kitchen gardens, and which, whether eaten cooked or raw, are . . . usually served at dinner . . . and not, like fruits generally, as dessert.

    Id at 307. In Nix, the Court considered technical usage in the field of botany. What I discuss here differs slightly: the question is whether a term should be considered as a technical usage in the field of law.

  • 236. New York Times Co v Sullivan, 376 US 254, 279–80 (1964).
  • 237. The disagreement in the Sweet Home opinions can be described along these lines: the dissent treated take as a legal term of art that came to the statute predefined, while the majority viewed it as a more ordinary sort of word qualified by other ordinary words in the statute. Sweet Home, 515 US at 718 (Scalia dissenting) (“[T]ake . . . [is] a term of art . . . in the statutory and common law concerning wildlife.”) (quotation marks omitted); id at 697 n 10 (Stevens) (“Congress explicitly defined the operative term ‘take’ in the ESA . . . thereby obviating the need for us to probe its meaning.”).
  • 238. 136 S Ct 2271 (2016) (per curiam).
  • 239. See Texas v United States, 787 F3d 733, 743–44 (5th Cir 2015), citing Jeh Charles Johnson, Exercising Prosecutorial Discretion with Respect to Individuals Who Came to the United States as Children and with Respect to Certain Individuals Who Are the Parents of U.S. Citizens or Permanent Residents *3–4 (Department of Homeland Security, Nov 20, 2014), archived at http://perma.cc/2A25-778J.
  • 240. See 2014 Executive Actions on Immigration (Department of Homeland Security, US Citizenship and Immigration Services, Apr 15, 2015), archived at http://perma.cc/PSX4-54K4; 8 USC § 1611(b)(2) (granting a range of benefits to otherwise-nonqualified aliens insofar as they are “lawfully present in the United States as determined by the Attorney General”).
  • 241. Others have noted this troubling exchange. Linda Greenhouse, When Smart Supreme Court Justices Play Dumb (NY Times, Apr 28, 2016), online at http://www.nytimes.com/2016/04/28/opinion/when-smart-supreme-court-justices-play-dumb
    .html (visited Mar 27, 2017) (Perma archive unavailable):

    Any lawyer knows that a word that means one thing in ordinary conversation can be deployed as a term of art and assume a separate meaning in the context of legal analysis. Yet this garden-variety insight seemed to elude Chief Justice John G. Roberts Jr. and Justice Samuel A. Alito Jr. during last week’s argument. . . . It turns out that the phrase “lawful presence” . . . doesn’t have the obvious meaning it would have in everyday speech. . . . Is that really so hard for two of the top lawyers in the United States to understand?

  • 242. Transcript of Oral Argument, United States v Texas, Docket No 15-674, *26–27 (US Apr 18, 2016), archived at http://perma.cc/VE2J-3ESA (“Texas Transcript”).
  • 243. Id at *27.
  • 244. Id.
  • 245. Id at *27–28 (paragraph breaks omitted).
  • 246. Texas Transcript at *28 (cited in note 242).
  • 247. Id.
  • 248. Id at *29.
  • 249. Establishing racially discriminatory private schools was a widespread response to the judicial invalidation of segregation in public schools. Olatunde Johnson, The Story of Bob Jones University v. United States: Race, Religion, and Congress’ Extraordinary Acquiescence, in Eskridge, Frickey, and Garrett, eds, Statutory Interpretation Stories 127, 131 (cited in note 71).
  • 250. Bob Jones, 461 US at 619–20 (Rehnquist dissenting):

    Then in the midst of this litigation, and in the face of a preliminary injunction, the IRS changed its position and adopted the view of the plaintiffs.

    . . . [In a subsequent ruling, t]he IRS then concluded that a school that promotes racial discrimination violates public policy and therefore cannot qualify as a common-law charity. The circumstances under which this change in interpretation was made suggest that it is entitled to very little deference.

    (citations, brackets, and quotation marks omitted).

  • 251. Courts, of course, themselves interpret statutes only under that impetus. While judicial interpretations are one-time events motivated by lawsuits, agency interpretations can be motivated by a wide range of impetuses at different temporal spans, from lawsuits to ongoing projects with long-term development. It seems plausible that these differences in interpretation’s temporality and motivation might lead to different modalities of interpretation as well. See Jerry L. Mashaw, Between Facts and Norms: Agency Statutory Interpretation as an Autonomous Enterprise, 55 U Toronto L J 497, 519 (2005) (proposing a number of factors that may lead administrative agencies to employ approaches to statutory interpretation that differ from those used or recognized by courts). I pursue this possibility in related research on agency statutory interpretation.
  • 252. For instance, the King majority was at pains to explain the way that the provision it interpreted interacted with other provisions. King, 135 S Ct at 2489–90. The Green majority traced the development of FRE 609 in great detail, while the dissent opted for an equally detailed view focused on one particular contemporaneous document. See notes 140–60 and accompanying text. The majority and dissent in Brown & Williamson and the Bob Jones majority discussed in detail a range of legal enactments and policies that postdated the statutes at issue and explained how they supported particular interpretations of the statutory language. See text accompanying notes 71–87 and 161–69.
  • 253. For instance, the Knudson majority invoked principles of equity without specifying which period of equity it looked to. See text accompanying notes 121–26. And the Green majority and concurrence drew on legal principles they described as general without connecting them to any cited source. See notes 140–55 and accompanying text.
  • 254. See Part IV.A.
  • 255. See Annelise Riles, Representing In-Between: Law, Anthropology, and the Rhetoric of Interdisciplinarity, 1994 U Ill L Rev 597, 601 (noting that the “movement between normative and reflexive genres . . . is a highly salient aspect of contemporary legal thought”).
  • 256. See Part IV.B.
  • 257. See Part IV.C.
  • 258. Manning, 106 Colum L Rev at 91 (cited in note 6). Professor John Manning goes on to describe purposivists as “giv[ing] precedence to policy context—evidence that goes to the way a reasonable person conversant with the circumstances underlying enactment would suppress the mischief and advance the remedy.” Id. The suggestion that purposivists strive to discern how a generic reasonable person familiar with the legislative process would suppress the mischief is not accurate. Purposivists ask judges to look for clues about the way that the particular legislators enacting the statute expected, or wanted, it to work. Manning’s phrasing draws a parallel between textualism and purposivism at one of the few points at which they diverge: in textualism’s preference for general contextualizations versus purposivism’s preference for specific contextualizations. See Solan, 38 Loyola LA L Rev at 2028 (cited in note 6) (“Gone largely unnoticed in the battles between [the textualist and purposivist] camps . . . is the fact that both sides . . . agree upon almost everything when it comes to statutory interpretation.”). Because Manning is a prominent proponent of textualism, I take his presentation of textualism as fairly authoritative, even though I question the accuracy of his accompanying characterization of purposivism.
  • 259. Solan, 38 Loyola LA L Rev at 2053 (cited in note 6).
  • 260. Scalia, A Matter of Interpretation at 17 (cited in note 6). On textualism, see Solan, 38 Loyola LA L Rev at 2029–30 (cited in note 6) (explaining that, for evidence about the meanings of unclear terms, textualists eschew pronouncements by those who wrote and voted on the terms and consider instead what those terms mean to idiomatic speakers of American English); Manning, 106 Colum L Rev at 91 (cited in note 6) (arguing that textualists should look to a term’s “semantic context”) (emphasis omitted).
  • 261. For this view, see, for example, Manning, 116 Harv L Rev at 2390 (cited in note 87):

    [T]he precise lines drawn by any statute may reflect unrecorded compromises among interest groups, unknowable strategic behavior, or even an implicit legislative decision to forgo costly bargaining over greater textual precision. . . . Textualists [ ] believe that the only safe course for a faithful agent is to enforce the clear terms of the statutes that have emerged from that process.

  • 262. I use the term “purposivism” to encompass both purposivism and intentionalism. Both are generally opposed to textualism insofar as they attempt to discern the will of the enacting legislature, and they are often used together or interchangeably. See Aleinikoff, 87 Mich L Rev at 22 (cited in note 6) (describing intentionalism as “purpose analysis”).
  • 263. Katzmann, Judging Statutes at 31–35 (cited in note 6). See also generally Heydon’s Case, 76 Eng Rep 637 (Ex 1584). Given its emphasis, it may make sense to describe purposivism as a search for legislators’ understanding of statutory meaning as much as, or even more than, their purposes in enacting the statute. A group of legislators may, for instance, have diverse reasons for voting for a statute and desires about what it should accomplish, yet share a roughly similar understanding of what the words mean and how the statute will work.
  • 264. Katzmann, Judging Statutes at 19 (cited in note 6). For an explanation of what legislative history is and how it is produced, see Nourse, 122 Yale L J at 92–97 (cited in note 65).
  • 265. Katzmann, Judging Statutes at 4 (cited in note 6). See also Mikva, 1987 Duke L J at 386 (cited in note 65) (“If judges are to make congressional primacy meaningful, they cannot afford to ignore those obvious tools which members of Congress use to explain what they are doing and to describe the meaning of the words used in the statute.”).
  • 266. Katzmann, Judging Statutes at 29 (cited in note 6).
  • 267. Paul W. Kahn and Kiel Brennan-Marquez, Statutes and Democratic Self-Authorship, 56 Wm & Mary L Rev 115, 118–20 (2014) (explaining the faithful agent model and arguing that it is wrong).
  • 268. See Fallon, 99 Cornell L Rev at 687 (cited in note 6) (noting that textualism and purposivism both acknowledge the relevance of context to statutory interpretation); Solan, 38 Loyola LA L Rev at 2028 (cited in note 6).
  • 269. Some legal theorists have taken a different tack, arguing that statutory interpretation should take a more present-oriented approach. See, for example, Aleinikoff, 87 Mich L Rev at 49–61 (cited in note 6); Eskridge, Dynamic Statutory Interpretation at 107–11 (cited in note 6); Richard A. Posner, Pragmatic Adjudication, 18 Cardozo L Rev 1, 4–5 (1996). These approaches tend to have a broader and more realistic view of how judges use context to interpret statutes.
  • 270. See Fallon, 99 Cornell L Rev at 687 (cited in note 6).
  • 271. Legislative history presents a partial exception to the theories’ oversimplification of context. There has been some discussion—leading to some consensus—about what kinds of legislative history exist and which are more reliable indicators of congressional purposes. See generally, for example, Nourse, 122 Yale L J 70 (cited in note 65).
  • 272. See Abbe R. Gluck, Symposium: The Grant in King–Obamacare Subsidies as Textualism’s Big Test (SCOTUSblog, Nov 7, 2014), archived at http://perma.cc/44Y7-FKR9.
  • 273. The failure to recognize selecting and the oversimplification of situating may be one reason that others have found that neither of these theories “can systematically produce determinate results in the ‘hard cases.’” Eskridge and Frickey, 42 Stan L Rev at 322 (cited in note 7).
  • 274. See Manning, 116 Harv L Rev at 2390 (cited in note 87).
  • 275. Jef Verschueren, Notes on the Role of Metapragmatic Awareness in Language Use, 10 Pragmatics 439, 450 (2000) (describing what “are usually called ideologies”).
  • 276. See Carol J. Greenhouse, Courting Difference: Issues of Interpretation and Comparison in the Study of Legal Ideologies, 22 L & Society Rev 687, 706 (1988) (arguing that ideologies “reveal the terms in which a society organizes its own contests over the universe of its imagined alternatives”). This use should not be confused with the narrower, and more negative, use of “ideology” in much legal scholarship, in which it tends to mean something more akin to political party membership or position on specific political issues.
  • 277. Michelle Z. Rosaldo, The Things We Do with Words: Ilongot Speech Acts and Speech Act Theory in Philosophy, 11 Language Society 203, 227 (1982).
  • 278. Of course, much more goes into interpretation than judicial opinions. Parties’ briefs, behind the scenes exchanges among judges, and the research, writing, and conversations of judicial clerks are primary sites for selecting and situating text. Judges’ educational, political, and personal backgrounds and beliefs play a vital role as well. This Article looks at just the public product of this complex process—the opinion itself.
  • 279. See Riles, 1994 U Ill L Rev at 601 (cited in note 255) (lauding an anthropological perspective on law, which situates legal reasoning “as one particular cultural reality” to make “apparent and explicit” a “movement between normative and reflexive genres” that characterizes “contemporary legal thought”).
  • 280. See Cohen, 96 Archiv für Rechts- und Sozialphilosophie at 2 (cited in note 105).
  • 281. This is, perhaps, why procedures for arriving at adjudicative facts are so much more elaborated than those for arriving at legislative facts.
  • 282. Professors William Eskridge and Philip Frickey’s well-known “funnel of abstraction” suggests this hierarchy, though it does not address nonlegal sources. Eskridge and Frickey, 42 Stan L Rev at 353–54 (cited in note 7) (arguing that sources closer to the statutory text tend to carry more interpretive weight than sources further from it: an “interpreter will value more highly a good argument based on the statutory text than a conflicting and equally strong argument based upon the statutory purpose” because the purpose is higher up on the scale of abstraction, that is, more removed from the most concrete level, statutory text).
  • 283. In this sense, my position here echoes Professor Levinson’s argument that constitutional adjudication suffers from too much emphasis on finding common-law analogues to public law situations and too little discussion of its underlying substantive values. Levinson, 111 Yale L J at 1375–76 (cited in note 8).
  • 284. See Eskridge and Frickey, 42 Stan L Rev at 322 (cited in note 7) (“[A]lthough each [prescriptive] theory rests upon and subserves important values that should be considered when interpreting statutes, no theory persuades us that its cluster of underlying values is so important as to exclude all others.”). The normative grounding I provide here allows scholars to construct—and make a case for—alternative clusters of underlying values.
  • 285. See generally, for example, Abbe R. Gluck, What 30 Years of Chevron Teach Us about the Rest of Statutory Interpretation, 83 Fordham L Rev 607 (2014) (explaining that neither textualism nor purposivism presents a realistic understanding of how federal legislation is produced).
  • 286. Michael Silverstein, Cultural Prerequisites to Grammatical Analysis, in Muriel Saville-Troike, ed, Linguistics and Anthropology 139, 144 (Georgetown 1977) (describing speech acts as utterances that “create those ‘facts’” necessary to understanding the utterances “in the very context in which they are uttered, by convention”). The term “performative utterance,” drawn from Professor J.L. Austin’s classic 1955 lectures, is often used interchangeably with “speech act,” drawn from Professor John Searle’s Speech Acts. See Austin, How to Do Things with Words at 4–7 (cited in note 9); Searle, Speech Acts at 16–19 (cited in note 9).
  • 287. Marianne Constable, Our Word Is Our Bond: How Legal Speech Acts 21 (Stanford Law 2014) (arguing that legal speech pervasively “depend[s] for [its] success as law not only on the meaning of words but also on the circumstances in which they are” conveyed).
  • 288. Id at 1–16.
  • 289. See Rapanos, 547 US at 744 (Scalia) (plurality).
  • 290. See Constable, Our Word at 23 (cited in note 287).
  • 291. See Austin, How to Do Things with Words at 14 (cited in note 9) (explaining that, for a speech act to be successful, “[t]here must exist an accepted conventional procedure having a certain conventional effect”).
  • 292. See Rosaldo, 11 Language Society at 228–29 (cited in note 277):

    Searle uses English performative verbs as guides to something like a universal law. I think his efforts might better be understood as an ethnography—however partial—of [ ] views of human personhood and action as these are linked to culturally particular modes of speaking. . . .

    . . . [U]nderstanding of linguistic action always, and necessarily, demands much more than an account of what it is that individuals intend to say: . . . the “force” of acts of speech depends on things participants expect; and . . . such expectations are themselves the products of particular forms of sociocultural being.

  • 293. See, for example, Silverstein, Shifters, Linguistic Categories, and Cultural Description at 33–34 (cited in note 9) (distinguishing between utterances in which an “aspect of the speech situation [is] presupposed by the sign token,” such that one cannot understand a word without some shared knowledge about its situation of use, and a creative usage, which “make[s] explicit and overt the parameters of structure of the ongoing [context]” and brings some aspect “into sharp cognitive relief”); Silverstein, Cultural Prerequisites at 142 (cited in note 286) (noting that the “entities” referred to by personal pronouns such as “I,” which designate “speech-event roles, are not ‘out there’ in any sense; they are created by speech itself”).
  • 294. Bernard Weissbourd and Elizabeth Mertz, Rule-Centrism versus Legal Creativity: The Skewing of Legal Ideology through Language, 19 L & Society Rev 623, 626 (1985).
  • 295. In contrast, saying “that chair is green” does not help constitute the chair. As a presupposing utterance, it builds on things that are already recognized: “[T]he chair exists apart from the utterance.” Id. See also Silverstein, Cultural Prerequisites at 142 (cited in note 286) (“Presupposition is [ ] a relationship [ ] between signal token and context: some proposition must be establishable as true on grounds independent of the signal token itself in order for the signal token to have some specifiable functional effect.”). Presupposition and creativity are best understood not as distinct types of utterances but as functions that utterances can serve in communication. See Silverstein, Cultural Prerequisites at 142 (cited in note 286) (noting that these functions run along a “scale” from more to less). Legal writing often “fail[s] to do justice to the creativity of speech, concentrating instead on presupposed . . . elements of the speech situation.” Weissbourd and Mertz, 19 L & Society Rev at 627 (cited in note 294). So do many everyday understandings of how language works. Research indicates that speakers, in general, are more sensitive to and articulate about the presupposing aspects of language use (the referents that exist independently of the communication) than its creative ones (the things that communication creates or highlights in a new way). See Michael Silverstein, The Limits of Awareness, in Alessandro Duranti, ed, Linguistic Anthropology: A Reader 382, 382 (Blackwell 2001). The creative power of speech often falls outside “[t]he [l]imits of [a]wareness.” Id.
  • 296. For example, linguists and anthropologists have shown how choices between different forms of address—for example, the French tu and vous—can both signal and affect social relations in subtle ways. See, for example, Roger Brown and Albert Gilman, The Pronouns of Power and Solidarity, in Thomas A. Sebeok, ed, Style in Language 253, 257–59 (MIT 1960) (showing that across European languages, tu forms tend to be used for subordinates and among equals who are in “solidarity,” while vous forms tend to be used for the more powerful and the less solidary); Paul Friedrich, Social Context and Semantic Feature: The Russian Pronominal Usage, in John J. Gumperz and Dell Hymes, eds, Directions in Sociolinguistics: The Ethnography of Communication 270, 287–95 (Holt, Rinehart and Winston 1972) (discussing how choices in pronouns can contribute to “dynamic relations” among speakers, in part by “express[ing] idiosyncratic impulses or the peculiarities of a situation”).
  • 297. B. Honig, Declarations of Independence: Arendt and Derrida on the Problem of Founding a Republic, 85 Am Polit Sci Rev 97, 101 (1991), citing Hannah Arendt, On Revolution 175 (Viking 1963).
  • 298. Honig, 85 Am Polit Sci Rev at 101 (cited in note 297).
  • 299. Id.
  • 300. Hannah Arendt, The Human Condition 198 (Chicago 1958). See also id at 25 (naming this kind of efficacious yet contingent action and speech as the two categories of human activity “out of which rises the realm of human affairs”).
  • 301. Id at 237. Arendt sees truly political action as both irreversible and unstable: action has real effects whose meanings and implications exceed our control. Our only “redemption” from these terrifying characteristics are, on the one hand, the capacity of “promising,” and, on the other, the “faculty of forgiving.” Id.
  • 302. For instance, the notion that courts should consider “special factors” in Bivens actions, which are now central to the constitutional tort inquiry, began with offhand dictum. Anya Bernstein, Congressional Will and the Role of the Executive in Bivens Actions: What Is Special about Special Factors, 45 Ind L Rev 719, 731 (2012).
  • 303. For instance, “Justice Jackson’s concurrence” in Youngstown Sheet & Tube Co v Sawyer, 343 US 579 (1952), “has emerged as the Court’s controlling paradigm for confronting [ ] separation of powers issues.” Laura A. Cisneros, Youngstown Sheet to Boumediene: A Story of Judicial Ethos and the (Un)Fastidious Use of Language, 115 W Va L Rev 577, 578 (2012).
  • 304. For example, Scalia’s dissent in Lawrence v Texas, 539 US 558 (2003), argued that the majority’s invalidation of statutes prohibiting same-sex sexual activity would logically imply invalidating statutes prohibiting same-sex marriage; the Court went on to do just that. Compare id at 604–05 (Scalia dissenting) (arguing that the Lawrence majority “opinion dismantles the structure of constitutional law that has permitted a distinction to be made between heterosexual and homosexual unions, insofar as . . . marriage is concerned”), with Obergefell v Hodges, 135 S Ct 2584, 2597 (2015) (holding that constitutionally guaranteed “liberties extend to . . . intimate choices” such as marriage).
  • 305. The American pragmatists similarly thought that meaning-production had a dynamic quality of ongoing reinterpretation and contestation. Charles Sanders Peirce famously described communication as involving a sign “which stands to somebody for something in some respect or capacity.” Charles Sanders Peirce, Logic as Semiotic: The Theory of Signs, in Justus Buchler, ed, Philosophical Writings of Peirce 98, 99 (Dover 1955). The sign stands for its object. Its interpretant represents the relationship between sign and object. But the interpretant is itself a sign, a “translation, explanation, or conceptualization of the sign/object relation in a subsequent sign representing the same object.” Richard J. Parmentier, Peirce Divested for Non-intimates, 7 Recherches Sémiotiques [Semiotic Inquiry] 19, 20 (1987). The inherently social cycle of semiosis thus continues indefinitely, each “‘interpretant becoming in turn a sign, and so on ad infinitum’ . . . so that semiosis can only be ‘interrupted’ but never ended.” Barton Beebe, The Semiotic Analysis of Trademark Law, 51 UCLA L Rev 621, 638 (2004).
  • 306. Honig, 85 Am Polit Sci Rev at 101 (cited in note 297).
  • 307. Arendt, The Human Condition at 144 (cited in note 300) (“Action, though it may have a definite beginning, never . . . has a predictable end.”).
  • 308. Timothy Mitchell, Society, Economy, and the State Effect, in George Steinmetz, ed, State/Culture: State-Formation after the Cultural Turn 76, 90 (Cornell 1999). Professor Timothy Mitchell calls on scholars to illuminate how distinctions between seemingly ontological categories like “state” and “society” are produced. Id at 76–77. Disciplinary and discursive practices, such as regimes of time, space, and bodily control—and, I would add, genre, framing, and argumentation—are among “those modern techniques that make the state appear to be a separate entity that somehow stands outside society.” Id at 85.
  • 309. Max Radin, Statutory Interpretation, 43 Harv L Rev 863, 869 (1930).
  • 310. Fallon, 99 Cornell L Rev at 688–89 (cited in note 6).
  • 311. As Professor Thomas S. Kuhn famously suggested in the realm of science, there is no truly pretheoretical understanding: “[P]aradigms give form to the scientific life.” Thomas S. Kuhn, The Structure of Scientific Revolutions 109 (Chicago 4th ed 2012).
  • 312. Guyora Binder and Robert Weisberg, Literary Criticisms of Law 113 (Princeton 2000) (noting that this claim is frequently made by “legal hermeneutic scholars”).
  • 313. Radin, 43 Harv L Rev at 868 (cited in note 309).
  • 314. Charles P. Curtis, A Better Theory of Legal Interpretation, 3 Vand L Rev 407, 419 (1950) (arguing that interpreters are never limited to any “one meaning, . . . neither the meaning which [the] author intended, nor the meaning which a reasonable person would ascribe to [the words], nor the meaning which either party expected of the other”).
  • 315. In this I echo Charles P. Curtis’s valorization of multivalence: “[W]ords which are applicable to things . . . always have many meanings. Their author, I say, expects them to, wants them to, indeed uses, and even comes to admire, this above all their other virtues.” Id.
  • 316. Professors Guyora Binder and Robert Weisberg, for instance, have expressed impatience with commentary that “ma[kes] ‘language’ and ‘meaning’ into philosophical pseudo-problems,” allowing scholars to avoid confronting the “specific . . . sources of incoherence in [ ] American” law, such as the distance between America’s “history of racial hierarchy and exclusion” and its “textual promise of racial equality.” Binder and Weisberg, Literary Criticisms of Law at 22 (cited in note 312). My analysis suggests that just how judicial opinions constitute the relation between social structure and “textual promise” remains relevant to the analysis of both.
  • 317. Constable, Our Word at 15 (cited in note 287).
  • 318. See Part IV.B.
  • 319. See Part III.C.
  • 320. Ashcroft v Iqbal, 556 US 662, 679 (2009).
  • 321. See Bernstein, 33 Yale J Reg at 34–35 (cited in note 235).
  • 322. See Strauss, 66 Chi Kent L Rev at 321 (cited in note 1); Mashaw, 55 U Toronto L J at 502–03 (cited in note 251).
  • 323. See, for example, Mashaw, 55 U Toronto L J at 519 (cited in note 251) (explaining how agencies differ from courts in ways that are likely to affect their interpretive practices). Professor Christopher J. Walker has done a crucial initial study on this question. See generally Christopher J. Walker, Inside Agency Statutory Interpretation, 67 Stan L Rev 999 (2015). I also pursue this question in separate qualitative research.
  • 324. Professors Abbe R. Gluck and Lisa Schultz Bressman’s groundbreaking study of legislation drafters demonstrates the importance of their views of legal interpretation. See generally Abbe R. Gluck and Lisa Schultz Bressman, Statutory Interpretation from the Inside—an Empirical Study of Congressional Drafting, Delegation, and the Canons: Part I, 65 Stan L Rev 901 (2013); Lisa Schultz Bressman and Abbe R. Gluck, Statutory Interpretation from the Inside—an Empirical Study of Congressional Drafting, Delegation, and the Canons: Part II, 66 Stan L Rev 725 (2014).