The Canon Wars

Book Review - Volume 97 - Issue 1

Canons are taking their turn down the academic runway in ways that no one would have foretold just a decade ago. Affection for canons of construction has taken center stage in recent Supreme Court cases[3] and in constitutional theory. Harvard Dean John Manning and originalists Will Baude and Stephen Sachs have all suggested that principles of “ordinary interpretation”[4]including canons[5]should inform constitutional interpretation. Given this newfound enthusiasm for canons, and their convergence in both constitutional and statutory law, it is not surprising that we now have two competing book-length treatments of the canons—one by Justice Scalia and Bryan Garner, Reading Law, and the other by Yale Law Professor William N. Eskridge, Interpreting Law. Both volumes purport to provide ways to use canons to read statutes and the Constitution. In this Review of Interpreting Law, we argue that this contemporary convergence on canons raises some significant interpretive questions about judicial power and the very idea of a canon.


What counts as a “canon” of statutory interpretation? Is any interpretive principle articulated by the U.S. Supreme Court a canon? Or does canonical status require something more in the way of historical pedigree, longevity, regularity of use, or some other measure? Can and should all interpretive tools, including legislative history, statutory precedents, and administrative agency interpretations, be considered part of the canon—or are these separate resources standing apart from, and perhaps in hierarchical tension with, the canons of construction per se?

These are just some of the fascinating questions raised by Bill Eskridge’s new book, Interpreting Law: A Primer on How to Read Statutes and the Constitution.[6] Styled as a treatise, the book is, like all Eskridgean work, a delight to read. Eskridge puts what he calls “the gentle reader” through her interpretive paces, asking her to apply the hypothetical “No Vehicles in the Park” statute to a variety of vehicular items—from an 1897 Léon Bollée Voiturette to a Bock Otto SuperFour Motorized Wheelchair.[7] The reader-friendly veneer is fitting, as the book is labeled a “primer” for students, scholars, lawyers, and judges. Eskridge’s chief goal is to offer a robust and important alternative to the late Justice Antonin Scalia and Professor Bryan A. Garner’s treatise-like tome on statutory interpretation, Reading Law: The Interpretation of Legal Texts.[8] Even the names of the two books are—intentionally, we think—similar.

I. Theories of the Canons: Static v. Dynamic

If the names of the books are similar, the implied—and sometimes express—theories of canons reflected therein are distinctly different. Scalia and Garner’s Reading Law offers fifty-seven valid canons that it urges “provide greater certainty in the law, and hence greater predictability and greater respect for the rule of law”[9] by focusing on the statute’s text.[10] By contrast, Eskridge’s Interpreting Law eschews the idea that a handful of canons applied mechanically can provide a complete answer to the interpretive questions at issue in a case.[11] Eskridge makes a significant bow to text,[12] insisting that the foundation of interpretation includes ordinary meaning, but the implicit theory of canons the book produces is overtly dynamic—as one might expect from the author of Dynamic Statutory Interpretation.[13] To Eskridge, canons are a regime, a regime that is inherently normative and nonmechanical and likely to change over time.

Interpreting Law is an incredibly rich and illuminating contribution to the fields of statutory and constitutional interpretation.[14] Eskridge’s encyclopedic knowledge of the fields is revealed in his treatment of numerous interpretive canons and his nuanced, honest discussions about the advantages, disadvantages, and intricacies confronted in applying each of the canons he presents. He gives texture to a vast array of canons that textualists hold dear, emphasizing the importance of ordinary meaning among many other textual canons. Unlike the more dogmatic Justice Scalia volume, however, Eskridge gives us both sides of the debate. For example, consider the question of dictionary usage. Eskridge is admirably restrained. He knows that despite judicial enthusiasm for the practice, linguists find dictionary usage entirely unhelpful because the meanings in dictionaries are acontextual.[15] So, he proposes a helpful amendment to allow for use of dictionaries to limit, rather than determine, meaning and goes on to ask various questions about the canon’s scope.[16] Such a style is dramatically different from Scalia and Garner’s more rigid approach and appears throughout Eskridge’s volume, as he thoughtfully considers and interrogates a broad range of textual and substantive canons in both statutory and constitutional interpretation.

The bottom line is this: even if you are interested in only the conventional textual canons emphasized by Justice Scalia, you should read Eskridge’s treatise. Like treatises generally should do, Eskridge’s treatise gives you a balanced view of the limits and advantages of the canons, along with reams of citations. Although a pragmatist himself, Eskridge comprehensively explains and treats with respect the ordinary meaning canon, the rule of the last antecedent, the canon of negative implication, and the whole act canon, among other textualist favorites.[17] What is more, in addition to the conventional textual canons discussed by Justice Scalia, Eskridge’s volume also addresses substantive issues about which he and Justice Scalia disagree—such as Justice Scalia’s dynamic anticanon against legislative history[18] and his rejection of the ancient Blackstonian rule against considering absurd consequences in statutory interpretation.[19] At each turn, Interpreting Law expands our knowledge about canons, evaluating the tradeoffs in terms of stability, predictability, and democratic accountability.[20] Many devotees of the Scalia–Garner treatise will find (no doubt to their surprise) that Eskridge agrees with Scalia and Garner on such things as the fact that some canons can have a stabilizing effect.

However balanced, readers familiar with the canon literature will note some Eskridgean canons that will strike them as new or strange. (To be fair, this is also true of some of the canons and anticanons in Justice Scalia’s book.) This is, in part, because Eskridge’s account of canons in Interpreting Law repurposes several traditional statutory interpretation tools as canons—including legislative history,[21] statutory precedents,[22] and standards governing deference to administrative interpretations.[23] This raises important methodological concerns for us. As we argue in the next Part, it creates a substantive ordering problem because, in our view, precedent and legislative history should take precedence over rules like noscitur a sociis. For this Part, however, the point is that Eskridge’s implied theory of canons is highly dynamic—it is not limited to the canons we know existed at the Founding; it is not limited to the traditional textual canons; rather, it includes ideas found in modern cases, ideas embraced by the law professoriate, and in some cases, ideas that are simply aspirational.[24]

In many ways, this more dynamic view reflects reality: the Supreme Court has in the past twenty-five years embraced a variety of new substantive canons. The Scalia–Garner treatise is dynamic as well; it may focus on text, but it includes interpretive rules that few would describe as canons as opposed to judicial doctrines, like rules about preemption and implied rights of action. Moreover, the Scalia–Garner treatise goes out of its way to thumb its nose at canons that one might think well-established: rejecting the notion that remedial statutes are to be interpreted liberally and the idea that the “purpose of interpretation is to discover intent.”[25] On these and other matters, Professor Eskridge’s treatise is in some respects more traditional than Justice Scalia’s. In fact, one way to look at Eskridge’s treatise is as a dynamic response to Justice Scalia’s own dynamism. In our view, Eskridge was right when he criticized Scalia and Garner’s Reading Law as picking and choosing the authors’ favorite canons.[26] In some cases, Eskridge has replied with his own countercanons.[27] The difference is that Eskridge openly admits that he is not describing a static world.

All of this raises the central question of this Review: What counts as a canon of construction? This is a notoriously unresolved question in statutory and constitutional theory, one that Eskridge himself highlighted in a review he wrote of Reading Law shortly after it was published.[28] Interpreting Law never directly tackles this question, but its own capacious list of canons contains an implicit answer. That answer appears to be that a canon is any judicial principle or method of reasoning that the Supreme Court can use, should use, or has used (even once) in construing a statute.

This implicit definition makes sense given Eskridge’s project of recording in one place all of the interpretive rules the Court has used—particularly in a universe in which no established consensus exists regarding the criteria for achieving canon status.[29] We nevertheless resist a super-dynamic-canon theory on the ground that “any interpretive rule or method of reasoning” is simply too capacious a definition. We are worried that the “dynamic canon” theory elides important distinctions and fails to answer basic questions. First, the definition ignores important differences between language and substantive canons, which we believe should be treated separately rather than lumped together under a universal umbrella. Second, the dynamic definition papers over crucial questions about the role of canons in statutory interpretation more generally, such as: Have canons now subsumed all interpretation or, as we believe, do they play a subsidiary role? How do we know a real canon when we see one? And so on.

After exploring such questions in some detail, we propose our own preliminary answers in the Parts that follow.

II. An Ordering Problem

Eskridge begins the volume with a healthy moderation in the great canon debates. He rejects the realist but cynical Llewellyn view that the canons are mere window-dressing as well as the formalist view that canons are rule-like:

A thesis of this volume is that the canons are neither mechanical rules that by their own force assure the rule of law or democratic accountability or good governance, nor cynical instruments for result-oriented judges to decorate judicial opinions like ornaments on a Christmas tree. Rather, the canons constitute an interpretive regime, namely, a set of conventional considerations relevant to statutory interpretation that ought to be laid out systematically in one volume available to students, attorneys, judges, agencies, and legislative drafting offices.[30]

We admire this moderation but worry about the idea of an “interpretive regime.” The term “regime” suggests a consistency, force, and primacy that the canons do not exhibit.[31] In our view, the canons do not constitute an interpretive regime but, rather, are one type of interpretive resource—a set of judicial assumptions and presumptions—that can guide statutory interpretation when other, more authoritative, interpretive resources fail to fill a gap or render clarity illusory. Canons generally should be a last resort, not a first one. Why? Because, as a general rule, canons are judicial assumptions about meaning—default rules. Default rules are second-best guesses or policies that apply when all first-best evidence fails.

In other words, the ordering problem is this: By treating legislative history, precedent, administrative practice, and other interpretive resources as canons, Interpreting Law inadvertently (we think) places them on the same level in the hierarchy of interpretive tools as judicially created maxims, such as the presumption that tax statutes should be narrowly construed or that all statutes should be construed to avoid redundancy. But that is not how judges should—or in practice actually do—treat precedent, administrative practice, and legislative history when construing statutes. At various points, Eskridge appears to recognize that statutory interpretation is rife with ordering problems.[32] He argues that some materials take precedence over others.[33] Precedents should control against a new plain meaning analysis.[34] And judges should “consider relevant legislative history even if the judge believes there is or might be an ordinary or plain meaning.”[35] At the same time, however, he treats stare decisis, legislative history, and administrative constructions in rhetorical ways—i.e., calling them canons—that suggest that these materials are on a par with other more well-known canons.

A. The Canonization of Precedent, Legislative Evidence, and Administrative Interpretations

First, consider Interpreting Law’s treatment of precedent. Eskridge is a huge fan of precedential forces in statutory interpretation, explaining that “adherence to stare decisis ‘marks an essential difference between statutory interpretation on the one hand and [common] law and constitutional interpretation on the other.’”[36] In the next sentence, however, he equates this principle with traditional canons of construction: “Like the ordinary meaning rule and the whole act rule, this super-strong presumption of correctness for statutory precedents purports to be a foundation for the application of federal statutes . . . .”[37] Eskridge goes on to build up a hefty subset of stare decisis canons: a canon of relaxed stare decisis for common law statutes; an “acquiescence” canon for legislative acquiescence to a judicial determination; and a “reenactment” canon for judicial interpretations Congress reenacts.[38] Other “precedent-based canons”[39] include a canon for judicial interpretation of common law “terms of art,” a canon for statutes that borrow from other acts (what Eskridge charmingly dubs stare de statute, following Frank Horack),[40] and the “shadow precedents canon” (which we discuss in more detail below).[41]

Now consider the book’s treatment of legislative history. Compared to stare decisis, Eskridge is not as enthusiastic about legislative history (or what one of us calls “legislative evidence”),[42] but he seems very enthusiastic about canonizing the approaches he recommends (including the recommendations of one of the authors of this Review). For example, he offers up the “committee report canon,” which notes that House and Senate committee reports and explanations of the conferees are considered the most reliable and authoritative form of legislative history; the “sponsor’s statement canon,” which explains that the Supreme Court routinely considers statements made by the sponsor of the statute when relevant to the statutory question at issue; and the “subsequent legislative history canon,” which recognizes that the “current federal judicial understanding is that subsequent legislative history is generally not a reliable source for statutory meaning”—although Eskridge argues that courts should consider this form of legislative history.[43] Eskridge also recommends as canons interpretive rules advocated by one of us in a law review article, including a “backwards induction canon,” which reads legislative evidence from the last point of legislative decision rather than requiring a full-scale, front-forward history; the “rules of proceedings canon,” which holds that in cases of doubt Congress should resolve the doubt as would the rules of the House and the Senate; and the “sore losers canon,” which would bar the judiciary from citing legislative materials created by those who lost the vote, except in limited circumstances.[44]

Canonization continues in the area of deference to administrative interpretations of statutes. As Eskridge rightfully acknowledges, “the overwhelming weight of official statutory interpretation is by administrators and agencies, not by judges.”[45] This is an exceedingly important point and one that Justice Gorsuch’s recent confirmation hearings put into the spotlight given his concerns about Chevron deference.[46] Here, what most conventionally know as precedents are dubbed canons. There is the “Skidmore Canon” on the interpretive value of regulatory history and the “Chevron Rule” on judicial deference to agency lawmaking.[47] Under Chevron, Interpreting Law offers up the “Major Questions Canon,” the “Plain Meaning Rule (Chevron Step One),” the “Brand X Canon (Chevron Step Two),” and other deference canons, including the “Seminole Rock/‌Auer Canon” and the “Curtiss-Wright Super-Deference Canon.”[48] This may be a rhetorical tic, or a strategy to rebut Reading Law’s own lengthy list of fifty-seven canons (with its own imaginative use of the canon label), but, in the end, Interpreting Law leaves the impression of canons, canons everywhere.

B. Interpretive Priority

Now that we have seen the tendency of the book to canonize a wide variety of precedents, practices, and materials, it is possible to consider in greater detail what we mean by the ordering problem. We begin with a hypothetical we hope will illustrate the problem. Eskridge is known for his amusing and illuminating hypotheticals, so we respond in kind. The basic idea of the “garbled order” was first deployed by Judge Posner.[49]

Imagine soldier Bill searching for the meaning of a garbled order from General Dick.[50] Soldier Bill cannot hear the precise order but knows that he must decide whether to attack or retreat. Imagine that there is a norm among soldiers that one is a coward if one does not attack, known informally as the “Tally Ho” canon. As soldier Bill is pondering the garbled command, General Dick’s aide Colonel Victoria, who drafted the original order, appears at soldier Bill’s side telling him that General Dick ordered retreat. Would a rational soldier attack? Would he tell Colonel Victoria to bug off, because the Tally Ho canon controls? Now assume the General’s aide is nowhere to be found, but soldier Bill remembers a similar battle days earlier when General Dick’s order was clear: attack only when fired upon. It seems highly doubtful that soldier Bill would ignore that precedent and proceed willy-nilly forward, raising the Tally Ho flag and attacking without a shot fired.

If these intuitions are sound, then we can begin to see why over-canonization raises interpretive problems. By suggesting that legislative history and precedent amount to canons, Interpreting Law essentially assigns them the same authoritative value as more conventionally understood canons—like noscitur a sociis or the presumption against extraterritorial application. As a result, we may have no way of distinguishing actual evidence of meaning from informed estimates or sheer guesses about meaning. Taken to its extreme, the privileging of canons over actual evidence of context can yield irrational outcomes—e.g., the soldier attacking rather than retreating—because it privileges hypothetical over actual evidence of meaning.[51] Recall our story: in the first case, the soldier has evidence of the actual speaker’s meaning. Under Eskridge’s interpretive regime approach, however, if the actual evidence amounts to a canon (legislative evidence), it appears to have no priority in authoritativeness, and the soldier may attack even when the general wants him to retreat. Similarly, in the second case, the soldier has no actual evidence of the speaker’s meaning but has an actual experience to which he can analogize. If precedent amounts to a canon, the soldier is left weighing one canon telling him to attack against another telling him to wait—leaving our poor soldier-interpreter in no man’s land.

We doubt that Eskridge himself would accept at least some of the implications from our hypothetical. First, he has been most insistent about the value of precedent in deciding statutory cases.[52] He is at pains in Interpreting Law to argue that an authoritative precedent “is more immediately important than the ordinary meaning that today’s judge might have otherwise found.”[53] Second, although Justice Scalia has famously rejected legislative evidence, dubbing it an anticanon, Eskridge rejects that position.[54] He recognizes that “[f]or more than a century, federal judges have been willing to consider legislative history . . . .”[55] The questions we pose here are not about particular interpretive resources in isolation, however. They are about the value of interpretive resources relative to each other. We worry that dubbing a wide variety of materials as canons, or as a consistent “interpretive regime,” can yield serious problems regarding the relative authority of different interpretive resources. Rather than making something sacred, this labeling practice may reduce the level of importance of central legal principles.

To be fair, Professor Eskridge is doing nothing different than Justice Scalia and Bryan Garner did in Reading Law, which took general practices (e.g., preemption and implied rights of action) and turned them into canons and anticanons. Moreover, since Eskridge’s book is styled as a “primer,” the author may have felt it more important to leave major ordering problems to other work. On the other hand, this ordering problem has important consequences for interpreters, as the garbled order hypothetical demonstrates. One of the great principles of the Scalia–Garner treatise—the book to which Eskridge is responding—is its insistence on valid canons and invalid ones, the most important “invalid” one being the use of legislative history.[56] One of us has argued, at length, that Justice Scalia’s antipathy toward legislative evidence is antidemocratic and ignorant.[57] Here, our point is different: treating legislative history as a canon makes it appear as if it is on par with ejusdem generis, or the whole act rule. If that is what Interpreting Law recommends, Eskridge may well encourage that which he would otherwise reject: blindness to actual evidence of Congress’s meaning, administrative precedents about meaning, or even a controlling judicial precedent on meaning, because some other canon prevails in a feverish canon war.

III. Overcanonization

One of Interpreting Law’s most useful features is also one of its most nettlesome. That is, the book is a cornucopia of interpretive tools, rules, and maxims. It aims for thoroughness, and it admirably achieves its goal. But the book’s very thoroughness is also somewhat problematic, in our view. In reading through Interpreting Law’s exhaustive list of canons, we often found ourselves having the same reaction that Justice Scalia once described having to many of the canons in Karl Llewellyn’s famous list of “Thrusts” and “Parries”—i.e., “Never heard of it.”[58]

This is because, in addition to listing numerous canons that are well-established and accepted by all—e.g., the avoidance canon, ejusdem generis, the whole act rule—Interpreting Law labels as canons (i) methods or patterns of judicial reasoning that scholars have identified and even criticized but that none would call a canon;[59] (ii) statements made by the Supreme Court in a few cases that are neither well-known nor well-established;[60] and (iii) aspirational rules of interpretation that have been advocated by scholars.[61] Moreover, it inadvertently projects a false equivalency between the former and the latter by neglecting in some cases to acknowledge when a particular canon is one that has been only infrequently invoked, lacks consensus, represents a method of reasoning rather than an interpretive rule, or is merely aspirational.[62] This Part highlights a few such overcanonizations and concludes that they are problematic because labeling everything the Supreme Court says a canon emboldens judges to make things up as they go along—particularly if these canons are then allowed to trump legislative history and other interpretive resources.

A. Canons v. Patterns of Judicial Reasoning

Part of what makes Interpreting Law so rich is that it engages deeply with the literature in the statutory interpretation field. Eskridge weaves academic commentary throughout his discussion of the canons, and this contributes significantly to his presentation of the advantages and disadvantages of particular canons. But we fear that he may go a little too far in this weaving. Indeed, some of the canons he lists merely restate patterns or practices in judicial reasoning that scholars have identified in a handful of cases rather than rules or legal principles that courts have regularly announced.

Consider, for example, what Eskridge dubs the “Shadow Precedents Canon.”[63] The canon derives from a series of law review articles authored by Deborah Widiss that expose a surprising Supreme Court practice in employment discrimination cases: the Court sometimes continues to reason from its own past precedents even after Congress has enacted legislation overriding those precedents.[64] Widiss criticizes this practice—urging the Court to give full effect to Congress’s overrides and to cease reliance on superseded precedent cases—and suggests default interpretive rules that would help ameliorate this problem when it results from confusion about the scope of a congressional override.[65] Interpreting Law turns these instances of judicial misinterpretation into a canon of statutory construction: “Where Congress has only overridden the narrow result of a precedent, but not its underlying doctrinal structure, that decision is still citable as a shadow precedent.”[66] This, in our view, is not and should not be a canon for a number of reasons.

First, the “shadow precedents” precept is not a rule that the Court itself has announced, although that deficiency perhaps could be overcome if it were the only one.[67] Second, and more importantly, the practice of continuing to rely on shadow precedents is not a well-accepted or established interpretive rule. Indeed, it is not really an interpretive “rule” at all but, rather, a pattern of judicial behavior observed in a handful of cases in one area of the law. Widiss herself does not describe reliance on shadow precedents as an “interpretive rule”; instead, she characterizes it as a “mistake”[68] or as the result of understandable confusion about how to interpret an override statute.[69] Transforming this practice into a canon lends it an unwarranted sense of legitimacy and makes it seem like a more far-reaching practice—one that extends across statutory subject areas—than we necessarily know it to be.

In a similar vein is something Eskridge calls the “Administrability Canon.”[70] This canon draws in part on a law review article written by one of us[71] and dictates that “an interpretation that has been shown over a period of time to have been easy to administer will be preferred to one that is less time-tested and harder to administer.”[72] While we certainly agree that administrability is an important factor that the Supreme Court regularly takes into account when construing statutes, we disagree with Eskridge’s effort to canonize it. As one of us has argued elsewhere, administrability concerns are a subset of practical, consequences-based reasoning, not a canon or interpretive rule.[73] That is, when courts, including the U.S. Supreme Court, take administrability into account in interpreting a statute, they tend to discuss things like the practical difficulty of implementing a particular interpretation, the likely effect the interpretation will have on judicial or other public resources, or the clarity or predictability of the legal rule established by the interpretation.[74] These are all practical consequences that follow from an interpretation, and they demonstrate that judges are remarkably pragmatic in interpreting statutes rather than driven by mechanical rules or canons. Trying to turn judges’ administrability-based pragmatic reasoning into a rule of construction thus strikes us as a bit upside down.

Moreover, it is potentially dangerous. When we label an interpretive principle a canon, it inevitably becomes imbued with an aura of legitimacy. Canons are considered neutral legal principles, handed down over time, that transcend ideology and judicial policy preferences and that constrain judges. Despite numerous efforts to shatter this mythical vision,[75] it persists in at least some form and is the basis for Scalia and Garner’s effort to provide a list of valid canons for courts to employ. In fact, we suspect it is the reason Eskridge himself has sought to label so many different interpretive tools as canons. But in so doing, Interpreting Law runs the risk of creating a new problem—i.e., encouraging judicial power grabs. If anything the Supreme Court says in the course of explaining its reasoning in a case can be called a canon, then the Court may freely make things up as it goes along—inventing new canons, announcing caveats to existing ones, and even perhaps denouncing existing canons as it sees fit.

B. Occasional Canons

Still other canons listed in Interpreting Law appear to us to be aspirational, in that they set forth interpretive rules that Eskridge thinks are good rules but that have not necessarily been embraced by the Court. Oftentimes, we agree that these rules are either linguistically erudite or justified by attention to legislative drafting practices. Nevertheless, we worry that Interpreting Law has a very capacious standard for a canon—and, again, that this is dangerous because it encourages judges to simply make up canons that they think reflect good rules of thumb.

One such example is the rule that “[f]ailure of U.S. Attorneys to initiate criminal prosecutions in the past is evidence that the Attorney General’s current reading of the statute is too broad.”[76] We have not heard of such a canon. It appears to derive from one case, Lopez v. Gonzales,[77] in which the Court observed that the “failure of even a single eager Assistant United States Attorney to act on the Government’s interpretation of [the statute]” is telling evidence that “belies the Government’s claim that its interpretation is the more natural one.”[78] This may well be a logical inference that the Court in future cases wishes to invoke (we express no views on the merits of the rule), but as used in Lopez, it seems to us more of a case-specific comment than a rule designed to govern the construction of all criminal statutes.

Moreover, there are some logical and linguistic leaps between the Court’s inference that the government’s interpretation is not the “more natural one”[79] and Interpreting Law’s articulation that failure to initiate prosecutions is proof that the government’s reading is “too broad.”[80] Indeed, this example highlights two problems with turning stray Supreme Court comments into canons: (1) the approach leaves substantial room for idiosyncratic characterization of the interpretive rule and (2) it enables judges (or other would-be canonizers) to mistake what they think are good rules for what the Court has actually said or done in past cases.

One might, at this point, legitimately ask: What is wrong with including rules a scholar thinks ought to be canons in a comprehensive list of canons? Treatise writers can and do push the law in new directions. Indeed, the Scalia–Garner volume makes similar moves—urging, for example, that seeking justice in an individual case is an anticanon.[81] But that is precisely the point. Blurring the lines between established, universally accepted canons such as noscitur a sociis, expressio unius, the rule against superfluity, and the rule of lenity, on the one hand, and rules scholars think ought to be canons, on the other, can result in naked power grabs. Blurriness, combined with a lack of clear guidelines regarding what it takes for an interpretive principle to be considered a canon, opens the door for anyone—and judges in particular—to make up canons anytime they choose. Don’t like a particular canon? Make up an exception or limitation. Think a particular norm would help justify a favored interpretation of a statute? Make up a canon embodying that norm.

We do not mean to suggest that Interpreting Law’s approach to canons is going to usher in an entirely new form of judicial overreaching. Indeed, as Eskridge and Frickey have highlighted, the Rehnquist Court in the 1990s invented several new federalism clear statement rules in just this fashion.[82] But those clear statement rules were based on constitutional norms, not plucked out of thin air. And, importantly, they garnered significant criticism when first announced.[83] If statutory interpretation theory moves to a regime in which we begin to call everything the Court says in the course of interpreting a statute a canon, then we run the risk of opening the floodgates and emboldening judges to invent many more such canons in the future. What in the 1990s struck many as a problematic judicial power grab might in the 2020s and beyond become common judicial practice, without any constraining guidelines or criteria. (The problem is exacerbated, moreover, if and when such judicially invented canons are treated as equivalent in authority to legislative history or past precedents, or even used to trump such interpretive resources, as we noted in Part I.)

This, of course, brings us to a series of crucial—yet unresolved—questions in interpretation theory: What counts as a canon? From where do canons derive their legitimacy or authority? Does any comment made by the Supreme Court in a statutory interpretation case qualify? If not, how many times does the Court have to invoke an interpretive principle in order for it to become a canon? Or, on the other hand, is Supreme Court invocation the wrong test for what constitutes a canon? Must an interpretive principle be grounded in the common law, or date back to Blackstone or some other historical source, in order to be considered a canon? Must it accurately reflect how language works or how legislators draft statutes? We turn to these questions in the next Part.

IV. What Counts as a Canon?

In the previous two Parts, we have sketched out some interpretive tools that, in our view, clearly should not be considered canons of statutory construction—e.g., other interpretive resources including legislative history, precedents, administrative interpretations, patterns or practices of judicial reasoning, and aspirational principles that scholars think should be a canon or are cited rarely. In this Part, we turn to the more difficult question of what should count as a canon and consider several positive criteria that might be used to evaluate potential canons.

We begin with a threshold point. Scholars tend to think of the canons of statutory construction as falling into two distinct categories: language canons and substantive canons,[84] also sometimes referred to as descriptive and normative canons.[85] Language canons, as their name suggests, focus on the text of the statute and encompass rules of syntax and grammar, “whole act” rules about how different provisions of the same statute should be read in connection with each other (e.g., to minimize internal inconsistency, to avoid superfluity), and Latin maxims such as expressio unius est exclusio alterius and noscitur a sociis.[86] Substantive canons, by contrast, are policy-based principles and presumptions that derive from the Constitution, common law practices, or normative concerns related to particular subject areas.[87]

In its effort to characterize the canons as a coherent interpretive regime, Interpreting Law ignores important differences between these two categories of canons—and, indeed, compounds the problem by expanding the universe of canons to include other interpretive tools such as legislative history, precedents, and administrative interpretations. Any honest, useful attempt at defining what it takes for an interpretive rule to count as a canon must, in our view, acknowledge three salient differences between language and substantive canons. First, language canons are typically considered neutral or objective,[88] whereas substantive canons are viewed as policy-based and are thought to add a thumb on the scales[89] in favor of a particular outcome. Second, and related to the first, the two sets of canons derive their authority and legitimacy from different sources: language canons are thought to reflect rules of grammar, logic, sentence organization, or even congressional drafting; substantive canons are thought to reflect background norms established in the Constitution, the common law, or some other element of the legal system.[90] Third, at the U.S. Supreme Court level at least, language canons are used far more frequently than substantive canons,[91] although substantive canons tend to bear the brunt of scholarly criticism.

It strikes us that to be considered a canon, an interpretive rule must be well-established in the legal community. That is, judges and lawyers must be familiar with it. Moreover, a canon must derive from an authoritative source; it cannot simply be a rule that a party suggests or makes up in its brief. Beginning from that premise, we consider four potential tests or measures that might be used to determine whether a particular interpretive rule counts as a canon: (1) the frequency with which the rule has been invoked by the U.S. Supreme Court; (2) the rule’s longevity (i.e., when it was announced or how long it has been in place, and whether it has been adopted across different iterations or generations of the Supreme Court); (3) the justification for the rule; and (4) whether the Court definitively declared or announced the rule as one of general applicability (as opposed to treating it as a case-specific interpretive argument).

1. Frequency of Invocation.—One sign that an interpretive rule is well-established and even ingrained in the legal community is that it is invoked frequently, or at least regularly and consistently, over time in legal discourse. The number or frequency of citations to a rule or other legal source often is used as a proxy for its status in the legal community. In previous work on canonical dissents, for example, one of us has used the number of favorable citations to a dissenting opinion made by subsequent Supreme Court majority opinions, combined with the overruling of the original decision, as the measure for whether the dissenting opinion has been canonized.[92] Aaron-Andrew Bruhl similarly has compared the rate at which lower federal courts invoke particular linguistic canons and legislative history with the rate at which the Supreme Court invokes those same linguistic canons and legislative history; he has suggested that the comparison be used as a proxy for how closely lower federal courts follow the Supreme Court’s lead in embracing or rejecting particular interpretive tools.[93] And in a recent study that traces the evolution of several prominent substantive canons, Judge Amy Coney Barrett has argued that canons do not become “deeply entrenched” the moment they are announced by the Supreme Court but only after subsequent cases begin to invoke them regularly.[94]

Frequency of citation is an admittedly imperfect measure of canonical status. It is not necessarily the case, for example, that the most frequently invoked interpretive rule is also the most universally accepted. Nevertheless, frequency of judicial invocation does capture an important aspect of what it means to be well-established and entrenched in the legal community. That said, a couple of questions remain. First, which judicial forum should be used to measure frequency of use—the U.S. Supreme Court, all federal courts, federal courts of appeals, or state courts? Most commentators have taken for granted that citation by the U.S. Supreme Court should be the yardstick by which frequency of use of a particular canon is measured.[95] We agree for several reasons. First, there is some evidence that lower courts tend to follow the U.S. Supreme Court’s lead with respect to whether a particular canon is in or out of favor.[96] Second, the Supreme Court is the highest court in the land, so it is difficult to envision it following the lead of lower courts or state courts if such courts were to articulate a new canon or begin to employ an old one frequently. Moreover, the U.S. Supreme Court has the greatest visibility of any court in the country. Lawyers and judges in Oregon or Chicago may pay close attention to what the Oregon Supreme Court or the Seventh Circuit decide, but those living and practicing in other states may not. Last, given recent evidence that several state courts have fashioned their own unique interpretive regimes—which sometimes deviate significantly from the approach taken by the U.S. Supreme Court—it seems prudent to avoid state courts’ potentially idiosyncratic pronouncements as the benchmark for determining which interpretive rules have become entrenched in the broader legal community.[97]

The next question then becomes: What suffices to constitute frequent use by the U.S. Supreme Court? Any numerical threshold will be inherently arbitrary. If we were to set a floor somewhere in the ballpark of seven to ten citations in U.S. Supreme Court opinions since 1790, when the first iteration of the Court began deciding cases,[98] this would capture most canons that scholars and judges are familiar with and then some. It would eliminate some of the canons listed in Interpreting Law’s appendix—but that is appropriate because some of the listed rules are not, in our view, canons but rather other interpretive resources, and others have been cited only sparingly.[99] In any event, we do not mean for this number to be set in stone but merely suggest it as a starting point for discussion. Perhaps the ideal number should be lower for canons that are subject-matter specific—e.g., those calling for liberal construction of the Freedom of Information Act[100] or interpreting the Sherman Act to benefit consumers[101]—but should be at the higher end of the spectrum for generally applicable rules such as the maxim that a “precisely drawn, detailed statute preempts or governs a more general statute.”[102] In the end, the exact number of minimum citations is not what matters most; the minimum threshold should merely be a vehicle for ensuring that loose judicial commentary is not labeled a canon on the basis of one or two (or even three) stray utterances by the U.S. Supreme Court.

2. Longevity or Historical Pedigree.—How long an interpretive rule has been in effect may be one of the most important factors in determining whether it qualifies as a canon of statutory construction. When an interpretive rule has been around for a while, it is likely to be familiar to members of the legal community. It also is more likely to be cited or quoted in cases and to be listed in treatises. Latin maxims such as expressio unius and ejusdem generis, for example, seem to derive much of their authority from the mere fact that they have been on the books for a long time. And Justice Scalia famously once commented that the rule of lenity “is validated by sheer antiquity.”[103]

Perhaps just as importantly, the law is inherently backward-looking and preoccupied with continuity, consistency, and predictability. That is why courts look to the common law to fill in gaps left in statutes[104] and why Blackstone’s Commentaries on the Laws of England[105] are so widely cited by American courts even in the modern era.[106] In short, the longer a rule has been on the books, the more comfortable we are with it and the more we tend to trust that it must be a good rule—otherwise, how would it have endured? One of us has elsewhere called this an assumption of “soundness” and has noted its connection to the Burkean philosophical preference for tradition and longstanding understandings that pervades much of the American political and legal system.[107]

But Burkeanism and rule of law preferences for consistency and predictability aside, we are hesitant to treat longevity as a requirement—rather than merely an indicator—that an interpretive rule should be considered a canon. It is one thing to recognize longevity and consistency as signs that a rule is well-established; it is quite another to insist on longevity, perhaps in the form of a minimum number of years on the books, before a rule may be considered a canon. Indeed, such an approach would bar the recognition of “new” canons, such as the federalism clear statement rules that cropped up largely out of the blue and in quick succession during the 1980s–1990s.[108] Moreover, it would entrench old rules that the Court no longer uses simply because they were adopted in the Blackstone era or were uttered once by the Supreme Court and then forgotten. Such concerns lead us to the conclusion that while longevity may act as an important “plus” factor in helping to determine whether a particular interpretive rule qualifies as a canon, it should not be used as a dispositive measure that all canons must meet.

In other words, we believe that while longevity can lend weight to an interpretive rule’s claim to canonical status, it is not sufficient by itself to justify such status. In order for an interpretive rule to qualify as a canon, regular Supreme Court use also seems necessary. If the rule was invoked fleetingly—once or a few times during the nation’s early years—but never used regularly, then it should not be considered a canon. By contrast, if such use was frequent but then fell off over time, the canon should continue to be considered a canon, barring express later rejection by the Court. One corrective for the “fleeting” canon problem would be to consider not only longevity in the abstract but consistent usage across different courts. If an interpretive rule was employed by both the Peckham and Warren Courts, for example, it should have greater warrant to be dubbed a canon than a rule whose use is limited to a single Supreme Court generation. Just as we view certain cases[109] as canonical because they have survived over long periods and have been cited in diverse situations by diverse judges, canon status for interpretive rules should require similar indicia.

3. Justification.—A third factor that might be used to measure whether an interpretive rule should be considered a canon is the basis or justification for the rule. There are, unfortunately, a variety of theories and justifications of canons. Some canons, for example, have been justified on the theory that they accurately reflect how Congress drafts statutes or how ordinary people use language,[110] or that they promote coherence throughout the U.S. Code.[111] Others have been said to derive their authority from the Constitution or established background norms pervading our legal system.[112] Further, some scholars, including Eskridge, have argued that the canons reflect principles basic to all communication.[113]

It is worth asking, then, whether grounding in one of these justifications is necessary for an interpretive rule to be considered a canon. That is, must a language canon do one of the following in order to qualify: reflect legislative drafting practices, reflect rules of grammar and logic that ordinary people use, or promote coherence across the U.S. Code? Must a substantive canon promote constitutional values or at least be grounded in some fundamental tenet of the American legal system?

In theory, one would want something more than mere age and usage to solidify a canon’s legitimacy. But even the most plausible theories of canons have not fared well under scrutiny. Consider the theory that canons reflect how Congress drafts or how reasonable people use language. Recent empirical studies have shown that, in some cases, the canons actually contradict congressional staffers’ descriptions of common legislative drafting practices. For example, the well-established rule against superfluity dictates that statutes should be construed to avoid redundancy, so that when there are two overlapping terms, each should be construed to have an independent meaning.[114] Interviews with congressional staffers, however, reveal that they sometimes deliberately err on the side of redundancy in order to “capture the universe,” ensure coverage of key items, or satisfy particular legislators, constituents, or lobbyists who “want[] to see that word” included.[115] Similarly, the expressio unius canon, which instructs that the inclusion of one statutory term implies the intentional exclusion of another,[116] has many logical imperfections—most notably, that the legislator simply may not have contemplated the particular application at issue.[117] Yet there can be little doubt by any measure—frequency of use, longevity, or historical pedigree—that it is a canon.[118] And it is a canon even if there is no consensus among linguists that it is necessary to, or an accurate reflection of, everyday communication.[119]

Or consider the theory that canons should be grounded in a constitutional principle (which in theory would eliminate the language canons). On the one hand, the most commonly invoked substantive canons—e.g., avoidance, the rule of lenity, federalism clear statement rules, sovereign immunity waivers, preemption, and the presumption of nonretroactivity—are all connected to or based upon constitutional principles.[120] At the same time, however, the Constitution has nothing to say on many subjects that statutes regulate—including, for example, antitrust rules, relations with Indian tribes, and veterans’ benefits—yet we have numerous longstanding and frequently invoked canons about how to read antitrust, Indian-tribal, and veterans’ benefits statutes, among others.[121]

Thus, what counts as a canon must be about more—or perhaps, less—than accuracy regarding how words are used or a connection to a constitutional provision. In our view, the basic thread connecting the canons is (or should be) established convention. Longevity or historical pedigree, and perhaps a connection to the Constitution, can help demonstrate established convention, but for the reasons we have outlined above, the real, indispensable measure for such convention must be regular Supreme Court use across ideological divides. Usage is important because canons claim their status as authoritative not simply based on age but because they represent how a “language community” understands and uses terms.[122] It is not that these are rules every citizen must or does speak, but that they are terms lawyers learn to speak. As the canons’ most sophisticated supporters suggest, they are the lingua franca of the law.[123]

Basic communication, however, requires agreement to cooperate, as philosophers of language know quite well. Paul Grice famously wrote that a “Cooperative Principle” governs communication.[124] Canons that are ideologically divisive are not canons; they are not established as rules accepted by the linguistic community. They do not follow the most basic notion of cooperation. It is well to remember that canons hail to ancient Roman practice (it is called “canon law” after all). Emperor Theodosius was famous for uniting warring religious sects. The “compromise” was one of the most important in all Christendom: the Nicene Creed is perhaps the most famous example of religious canons as well as one of enduring compromise.[125] This example simply reflects what positive political theorists tell us—that the idea of “compromise” is essential to any regime’s stability.[126] Thus, in our view, the ultimate test of a canon is one that reflects the stability of compromise, by which we mean that the canon reflects the agreement of Supreme Court Justices appointed by different parties and across ideological divides.

4. Can Canons Be Undeclared?—Last, it is worth considering whether an interpretive rule can qualify as a canon only if the Supreme Court itself has consciously declared or understood itself to be adopting a canon (or generally applicable legal rule) as opposed to merely making a comment about its reasoning in the particular case in front of it. A “conscious declaration” requirement has the advantage of ruling out off-handed comments made by the Court that are case-specific—such as the so-called canon that U.S. Attorneys’ failure to advance a particular reading of a criminal statute is strong evidence that the reading is incorrect.[127] It also would make it much harder for anyone other than the members of the Court to identify or designate particular interpretive rules as canons. Depending on one’s perspective, this could be either a positive or a negative feature.

On the one hand, if regular Supreme Court use is the most sensible measure of a rule’s status as a canon, then it may make a lot of sense to limit the ability to “declare” a canon to the Court. On the other hand, doing so risks missing at least some of the Court’s gradual, perhaps understated, articulations of new canons. Not all canons are intended as such the moment they are uttered. Even if they are, the Court might not clearly declare them as canons, focusing instead on the application of the canon to the interpretive issue before it in the particular case. Indeed, the Court sometimes announces a rule or principle in one case with little thought about whether it will make sense in other cases, only to later confront a similar case in which the rule is useful and to invoke it in the second case as well. If the Court proceeds in this manner for several more cases, it becomes hard to argue that the rule is not a canon, even if the Court itself has never openly declared it one.[128] Accordingly, we would recommend that the nature and frequency of the Supreme Court’s reliance on an interpretive rule—rather than the act of openly declaring the rule to be a broadly applicable one—should drive the analysis of whether the rule rises to the level of an established canon of statutory construction.

In short, we reject the notion that any and all norms can count as a canon. We also reject the notion that only textually authoritative canons should count. The former is too broad and the latter too narrow. Instead, canon status should be awarded only to interpretive rules that are well-established in the legal community and that reflect what we call “the stability of compromise.”

* * *

We end this Review where we began, by comparing Interpreting Law to Justice Scalia and Bryan Garner’s Reading Law on the question, “What counts as a canon?” Under the test we have recommended in this Partthe nature, frequency, and stability of the Supreme Court’s use of an interpretive ruleInterpreting Law’s list of canons is a bit too long, and Justice Scalia and Bryan Garner’s list is quite a bit too short. Where Eskridge includes some interpretive rules that are not widely enough established to merit canonical status, Scalia and Garner exclude from the metacanon numerous interpretive rules that are widely established. In our view, Eskridge’s overinclusion is less troublesome because he is open about where the rules he lists derive from, citing cases to support each canon he identifies and citing the scholarly work from which he draws recommended canons. By contrast, Justice Scalia and Bryan Garner offer little justification, beyond the authors’ opinions, for expelling numerous established interpretive tools from the list of canons (or turning them into anticanons). Sins of omission may be more troublesome than overinclusion as readers unaware of the larger universe may reach inappropriate conclusions.


There is every reason to believe that the “canon wars” are likely to continue in the Supreme Court and in the legal literature. Theories of interpretation have been relentlessly moving toward textualism, and, with that move, interpreters have attempted to fill gaps with standard canonical practice, claiming that our “law of interpretation,” including our constitutional law, depends upon canons.[129] Such claims, however, leave unresolved the central question raised by both Eskridge’s Interpreting Law and Scalia and Garner’s Reading Law—What counts as a canon of construction? While we cannot hope to have answered that question definitively in this short Review, we have aspired to at least outline and evaluate some potential measures of canonical status. Others, we hope, will continue that conversation.

  1. * Mary C. Daly Professor of Law, St. John’s University School of Law. J.D., Yale Law School, 1999; A.B., with distinction, Phi Beta Kappa, Stanford University, 1996. We owe deep thanks to Aaron-Andrew Bruhl, William N. Eskridge Jr., and Kate Shaw for valuable comments on earlier versions of this Review. Special thanks also to Dean Michael A. Simons and St. John’s Law School for generous research support. All errors are our own.
  2. ** Jack N. Pritzker Distinguished Visiting Professor of Law, Northwestern University (2017–2018); Professor, Georgetown University Law Center. J.D., University of California, Berkeley School of Law, Order of the Coif, 1984; B.A., Stanford University, Phi Beta Kappa, 1980.
  3. .See, e.g., District of Columbia v. Heller, 554 U.S. 570, 576 (2008) (employing the ordinary meaning canon—the maxim that the Constitution’s words ought to be given their normal and ordinary meaning).
  4. .Jack Goldsmith & John F. Manning, The Protean Take Care Clause, 164 U. Pa. L. Rev. 1835, 1859–60 (2016) (using interpretive canons such as “expressio unius” to interpret the Constitution); John F. Manning, Separation of Powers as Ordinary Interpretation, 124 Harv. L. Rev. 1939, 2015, 2025–38 (2011) (applying various “rules of construction” to interpret the Constitution’s vesting clauses).
  5. .See generally William Baude & Stephen E. Sachs, The Law of Interpretation, 130 Harv. L. Rev. 1079 (2017) (analyzing at length the canons as part of the “law of interpretation”).
  6. .William N. Eskridge Jr., Interpreting Law: A Primer on How to Read Statutes and the Constitution (2016) [hereinafter Eskridge, Interpreting Law].
  7. .Id. at 33–34, 133–34.
  8. .Antonin Scalia & Bryan A. Garner, Reading Law: The Interpretation of Legal Texts (2012).
  9. .Id. at xxix.
  10. .See id. at 343–46, 369–90 (discussing the false notions that “the spirit of a statute should prevail over its letter” and that “committee reports and floor speeches are worthwhile aids in statutory construction”).
  11. .See Eskridge, Interpreting Law, supra note 4, at 9 (“Text and purpose are like the two blades of a scissors; neither does the job without the operation of the other. More important, statutory text and legislative purpose do not exhaust the context that is relevant for the proper application of statutes.”); id. (arguing, as an example, that “the rule of law requires the judge to consider practice and precedent before she confidently declares statutory meaning”); id. at 23 (“Any accurate description of statutory practice ought to include ordinary meaning, the whole statute and related statutes, statutory precedents, legislative history, administrative constructions, and constitutional and other background norms as relevant factors to consider.”).
  12. .Id. at 40–41 (asserting that the ordinary meaning of a statute should be “the anchor for statutory interpretation by judges”).
  13. .William N. Eskridge Jr., Dynamic Statutory Interpretation (1994).
  14. .Cf. Eskridge, Interpreting Law, supra note 4, at 10–11 (commenting that Justice Scalia’s judicial opinions “devoted many more pages of analysis to precedent and practice than to plain meaning”).
  15. . See id. at 57–58 (noting that linguists are critical of judicial decisions that quote a dictionary definition and describing the importance of context).
  16. .See id. at 59 (“[D]ictionaries are often useful in clarifying or narrowing our understanding of statutory terms.”).
  17. .See id. at 33–55, 67–68, 78–81, 102–17 (providing thorough explanations of the various canons).
  18. .As explained infra pp. 119–21, we call this anticanon “dynamic” because it reflects the way Justice Scalia believes legislative history should be treated, not the way the vast majority of courts currently treat it.
  19. .See Scalia & Garner, supra note 6, at 23438 (2012) (proposing that limiting conditions must be used when applying the absurdity doctrine so as to limit judicial revision of the text).
  20. .See Eskridge, Interpreting Law, supra note 4, at 160–62 (observing how precedent and the order of litigation impact predictability and democratic accountability).
  21. .See id. at 240–45 (describing, for example, a “Committee Report Canon”).
  22. .See id. at 174–76 (describing, for example, a “Canon of Relaxed Stare Decisis for Common Law Statutes”).
  23. .See id. at 299–301 (describing, for example, a “Curtiss-Wright Super-Deference Canon”).
  24. .See discussion infra subparts II(A)–(C).
  25. .Scalia & Garner, supra note 6, at 364–66 (arguing against the liberal construction of remedial statutes because of the difficulty in determining what constitutes a remedial statute and what constitutes liberal construction); id. at 391–96 (arguing that “further uses of intent in questions of legal interpretation [should] be abandoned”).
  26. .See William N. Eskridge Jr., The New Textualism and Normative Canons, 113 Colum. L. Rev. 531, 535–36 (2013) (reviewing Scalia & Garner, supra note 6) (arguing that Scalia and Garner’s attempt to set forth a collection of “valid canons” of statutory construction was unsuccessful).
  27. .See id. at 541–42 (discussing certain canons not analyzed by Scalia and Garner, such as the stare decisis canon).
  28. .See id. at 541 (focusing specifically on the question of what rules are “canonical”).
  29. .Indeed, one of us has followed a similar, only slightly less capacious, definition in previous work. See Anita S. Krishnakumar, Statutory Interpretation in the Roberts Court’s First Era: An Empirical and Doctrinal Analysis, 62 Hastings L.J. 221, 240 (2010) [hereinafter Krishnakumar, Roberts Court First Era] (defining substantive canons as “interpretive presumptions and rules based on background legal norms, policies, and conventions”); see also Anita S. Krishnakumar, Reconsidering Substantive Canons, 84 U. Chi. L. Rev. 825, 855–57, 857 n.148 (2017) [hereinafter Krishnakumar, Reconsidering] (comparing Krishnakumar’s versus Eskridge and Frickey’s definitions of substantive canons in empirical work measuring the Court’s reliance on such canons).
  30. .Eskridge, Interpreting Law, supra note 4, at 20.
  31. .See Nina A. Mendelson, Change, Creation, and Unpredictability in Statutory Interpretation: Interpretive Canon Use in the Roberts Court’s First Decade, 117 Mich. L. Rev. 71 (2018) (citing data from the Roberts Court’s first ten terms showing that judicial canon use is unpredictable and ever-changing).
  32. .See, e.g., Eskridge, Interpreting Law, supra note 4, at 141 (exemplifying the ordering problem through interpretation of the Park Safety Act).
  33. .Id.
  34. .Id.
  35. .Id. at 202.
  36. .Id. at 163 (quoting Edward H. Levi, An Introduction to Legal Reasoning, 15 U. Chi. L. Rev. 501, 540 (1948)).
  37. .Id.
  38. .See id. at 174–76 (discussing the canon of relaxed stare decisis for common law statutes); id. at 176–77 (discussing the acquiescence canon for cases in which Congress does not act); id. at 177–79 (discussing the reenactment canon for cases in which Congress reenacts a statute).
  39. .Id. at 179 (discussing other precedent-based canons beyond strict stare decisis).
  40. .Id. at 182 (citing Frank E. Horack Jr., The Common Law of Legislation, 23 Iowa L. Rev. 41 (1937)).
  41. .Id. at 180–82 (discussing a common law canon for terms of art); id. at 182–85 (discussing a borrowed act canon); id. at 185–86 (discussing a shadow precedents canon).
  42. .See Victoria Nourse, Misreading Law, Misreading Democracy 153 (2016) (referring to legislative history as “legislative evidence”).
  43. .Eskridge, Interpreting Law, supra note 4, at 240–48, 251–54.
  44. .Id. at 224–37. Given that one of us is the author of these putative canons, see Victoria F. Nourse, A Decision Theory of Statutory Interpretation: Legislative History by the Rules, 122 Yale L.J. 70 (2012), it seems distinctly ungracious to decline canonization, if by canonization one means something sacred. Our point is that canonization has the potential to reduce the importance of a principle that should take priority.
  45. .Eskridge, Interpreting Law, supra note 4, at 259.
  46. .See Ilya Somin, Gorsuch Is Right About Chevron Deference, Wash. Post: Volokh Conspiracy (Mar. 25, 2017), [] (highlighting public interest in Justice Gorsuch’s opinions regarding Chevron deference); see also Gutierrez-Brizuela v. Lynch, 834 F.3d 1142, 1149 (10th Cir. 2016) (Gorsuch, J., concurring) (articulating Justice Gorsuch’s opposition to Chevron deference).
  47. .Eskridge, Interpreting Law, supra note 4, at 269, 278.
  48. .Id. at 287–301.
  49. .See Richard A. Posner, Legal Formalism, Legal Realism, and the Interpretation of Statutes and the Constitution, 37 Case W. Res. L. Rev. 179, 189 (1986).
  50. 48. See id. (“The commander replies, ‘Go—’; but the rest of the message is garbled.”).
  51. .Irrationality is a strong statement, but one accepted by at least some political scientists. See McNollgast, Legislative Intent: The Use of Positive Political Theory in Statutory Interpretation, Law. & Contemp. Probs., Winter 1994, at 3, 23–25 (applying statistical decision theory to demonstrate that ignoring the legislative history increases the probability of judicial error).
  52. .See William N. Eskridge Jr., Overruling Statutory Precedents, 76 Geo. L.J. 1361, 1362–64 (1988) (arguing for an “evolutive” approach when analyzing statutory precedents).
  53. .Eskridge, Interpreting Law, supra note 4, at 141; see also id. at 163 (discussing the super-strong presumption of correctness for statutory precedents).
  54. .See id. at 191–204, 240–45 (recognizing that no source “has generated greater debate” than legislative history but, nevertheless, applying legislative evidence based on a conference committee report).
  55. .Id. at 198.
  56. .See Scalia & Garner, supra note 6, at 9, 69, 82, 341, 37778 (listing fifty-seven valid canons, including the ordinary-meaning and fixed-meaning canons, and distinguishing them from the various pitfalls of statutory interpretation, including reliance on legislative history).
  57. .See Nourse, supra note 40, at 161–81 (arguing that “constitutional skepticism about legislative history is unwarranted”). Justice Scalia famously called legislative history “garbage” but could not tell the difference between a committee report and a conference report, one of which happens at the beginning of the legislative process and the other at the end. See id. at 69–95 (discussing various Supreme Court “basic” mistakes in Congress 101).
  58. .Antonin Scalia, Common-Law Courts in a Civil-Law System: The Role of United States Federal Courts in Interpreting the Constitution and Laws, in Antonin Scalia, A Matter of Interpretation: Federal Courts and the Law 26–27 (1997).
  59. .See discussion infra subpart III(A) (describing the “Shadow Precedents Canon” and the “Administrability Canon”).
  60. .See Eskridge, Interpreting Law, supra note 4, at 118–21, app. at 410, 421 (coining a noscitur a legibus sociis canon, a drafting manuals canon, and a lower court consensus canon); see also infra subpart III(B) (highlighting the “problems with turning stray Supreme Court comments into canons”).
  61. .See Eskridge, Interpreting Law, supra note 4, app. at 440 (inventing U.S. Attorneys’ canon); id. app. at 409 (articulating a dictionary-rule caveat to the effect that “[b]y revealing variety in word use, dictionaries can suggest ambiguity”).
  62. .See supra notes 57–59.
  63. .Eskridge, Interpreting Law, supra note 4, at 185.
  64. .Deborah A. Widiss, Shadow Precedents and the Separation of Powers: Statutory Interpretation of Congressional Overrides, 84 Notre Dame L. Rev. 511, 513–18 (2009) [hereinafter Widiss, Shadow Precedents]; Deborah A. Widiss, Undermining Congressional Overrides: The Hydra Problem in Statutory Interpretation, 90 Texas L. Rev. 859, 860–66 (2012) [hereinafter Widiss, Overrides].
  65. .See Widiss, Shadow Precedents, supra note 62, at 560–74 (discussing problems that arise from reliance on shadow precedents and proposing interpretive reforms).
  66. .Eskridge, Interpreting Law, supra note 4, app. at 421. In the body of the book, the canon is formulated somewhat differently: “Depending on how broadly it is drafted, a statutory override of the precedent’s result may not negate the force of the precedent’s reasoning.” Id. at 185 (citing Widiss, Shadow Precedents, supra note 62).
  67. .See discussion infra Part IV (noting Eskridge’s failure to differentiate between “language canons” and “substantive canons”).
  68. .See Deborah A. Widiss, Still Kickin’ After All These Years: Sutton and Toyota as Shadow Precedents, 63 Drake L. Rev. 919 passim (2015) (identifying mistakes made by courts that relied on shadow precedents).
  69. .See, e.g., Widiss, Shadow Precedents, supra note 62, at 537–38 (discussing “doctrinal confusion” and “lack of analytic clarity” regarding the interpretation of overrides); id. at 551 (describing how narrow override language can lead to understandable confusion).
  70. .Eskridge, Interpreting Law, supra note 4, at 114–17.
  71. .See Anita S. Krishnakumar, The Anti-Messiness Principle in Statutory Interpretation, 87 Notre Dame L. Rev. 1465, 1469 (2012) (“[T]he anti-messiness principle reflects a judicial preference for simple, easy-to-administer interpretations.”).
  72. .Eskridge, Interpreting Law, supra note 4, at 114–15.
  73. .Krishnakumar, Roberts Court First Era, supra note 27, at 244–46.
  74. .Id. at 244–45.
  75. .The most famous is Karl Llewellyn’s list of twenty-eight pairs of canons and countercanons. Karl N. Llewellyn, Remarks on the Theory of Appellate Decision and the Rules or Canons About How Statutes Are to Be Construed, 3 Vand. L. Rev. 395, 401–06 (1950); see also Jonathan R. Macey & Geoffrey P. Miller, The Canons of Statutory Construction and Judicial Preferences, 45 Vand. L. Rev. 647, 647–48 (1992) (arguing that even if canons of construction were outcome determinative in every case, judges could choose to ignore them and invoke a different source of authority); Richard A. Posner, Statutory Interpretation—In the Classroom and in the Courtroom, 50 U. Chi. L. Rev. 800, 805–17 (1983) (agreeing with Llewellyn that every canon has an equal and opposite canon and arguing further that most canons are “just plain wrong”).
  76. .Eskridge, Interpreting Law, supra note 4, app. at 440.
  77. .549 U.S. 47 (2006).
  78. .Id. at 57–58.
  79. .Id. at 58.
  80. .Eskridge, Interpreting Law, supra note 4, app. at 440.
  81. .Scalia & Garner, supra note 6, at 347–48.
  82. .See, e.g., William N. Eskridge Jr. & Philip P. Frickey, Quasi-Constitutional Law: Clear Statement Rules as Constitutional Lawmaking, 45 Vand. L. Rev. 593, 597 (1992) (stating that the “most striking innovation” of the Rehnquist Court had been its creation of “a series of new ‘super-strong clear statement rules’ protecting constitutional structures, especially structures associated with federalism”).
  83. .See id. at 598 (comparing the new super-strong clear statement rules to a “‘backdoor’ version of the constitutional activism that most Justices on the current Court have publicly denounced”).
  84. .For detailed explanations of this dichotomy, see William N. Eskridge Jr. et al., Cases and Materials on Legislation and Regulation: Statutes and the Creation of Public Policy 643–761 (5th ed. 2014) (introducing language canons and substantive canons of construction); James J. Brudney & Corey Ditslear, Canons of Construction and the Elusive Quest for Neutral Reasoning, 58 Vand. L. Rev. 1, 12–14 (2005) (describing the differences between language and substantive canons); David L. Shapiro, Continuity and Change in Statutory Interpretation, 67 N.Y.U. L. Rev. 921, 927–41 (1992) (providing an overview of the range of interpretative cannons from the “purely linguistic” to the “substantive”).
  85. .See Stephen F. Ross, Where Have You Gone, Karl Llewellyn? Should Congress Turn Its Lonely Eyes to You?, 45 Vand. L. Rev. 561, 563 (1992) (summarizing the distinction between descriptive canons, which are based on particular uses of language, grammar, or syntax, and normative canons, which dictate that ambiguous text be construed in favor of certain judicially crafted policy objectives).
  86. .Eskridge Jr. et al., supra note 82, at 644–47, 658, 668, 674–79. The maxim expressio unius est exclusio alterius means the “expression . . . of one thing indicates exclusion of the other.” Id. at 668. The rule rests on a logical assumption of negative implication; if the legislature specifically enumerates certain items in a statute, this is taken to imply a deliberate exclusion of all other items. Id. For further explanation of the expressio unius est exclusio alterius canon, see 2A Norman J. Singer & Shambie Singer, Sutherland Statutes and Statutory Construction § 47:23 (7th ed.) [hereinafter Sutherland]. Noscitur a sociis translates to “it is known from its associates.” Id. § 47:16. The canon dictates that when a statute contains a list of two or more words, courts are to give each word in the list a meaning that is consistent with the meaning of other words in the list. Id.; see also Eskridge Jr. et al., supra note 82, at 658 (describing noscitur a sociis).
  87. .See, e.g., Eskridge Jr. et al., supra note 82, at 643 (asserting that substantive canons usually “derive from policy positions articulated by courts”).
  88. .That these are in fact “neutral” is not necessarily true. For example, one of us has argued that each of the Latin canons can be paired with another Latin canon to come out with different results. See William N. Eskridge Jr., Abbe R. Gluck & Victoria F. Nourse, Statutes, Regulation, and Interpretation: Legislation and Administration in the Republic of Statutes 91–93 (Supp. 2017) (noting that Latin canons have faced criticism for leading to counter canons). Similarly, linguists are not necessarily so sanguine about their neutrality. See Lawrence M. Solan, The Language of Judges 29 (1993) (noting that conflicting principles of interpretation have created a body of “mutually inconsistent legal rules,” enabling lawyers and judges to use them to support almost any position they choose).
  89. .See Scalia, supra note 56, at 29 (“Some of the rules, perhaps, can be considered merely an exaggerated statement of what normal, no-thumb-on-the-scales interpretation would produce anyway.” (emphasis added)).
  90. .See sources cited supra notes 82–83 (discussing generally various language and substantive canons of construction).
  91. .See, e.g., Krishnakumar, Reconsidering, supra note 27, at 849–50 (reporting, in Tables 1 and 2, relative rates of the Roberts Court’s references to language versus substantive canons); Mendelson, supra note 29, app. at 101 (tracking rates of engagement for all canons—both textual and substantive).
  92. .See Anita S. Krishnakumar, On the Evolution of the Canonical Dissent, 52 Rutgers L. Rev. 781, 784 n.11 (2000) (describing her methodology).
  93. .Aaron-Andrew P. Bruhl, Communicating the Canons: How Lower Courts React when the Supreme Court Changes the Rules of Statutory Interpretation, 100 Minn. L. Rev. 481, 496–502 (2015).
  94. .See Amy Coney Barrett, Substantive Canons and Faithful Agency, 90 B.U. L. Rev. 109, 138, 140–43, 151–52 (2010) (concluding that the “Charming Betsy” canon, articulated in 1804, was absent from the case law for the next century and did not become entrenched until the 1950s; noting that the avoidance canon was first clearly articulated by the U.S. Supreme Court in the 1830s but took several decades to become “a fixture in both case law and commentary”; observing that the “Indian canon” was articulated in an 1832 case but “lay dormant” for thirty-four years and that, given the “paucity of nineteenth century cases applying the canon,” it could not be called a “well-settled law” until much later). In a different but related context, Richard Posner has used the number of citations to a scholarly work as “an index to the influence, and less confidently to the quality, of the paper.” Richard A. Posner, The Learned Hand Biography and the Question of Judicial Greatness, 104 Yale L.J. 511 app. at 534 (1994) (book review).
  95. .See, e.g., Frank B. Cross, The Theory and Practice of Statutory Interpretation 99–101 (2009) (examining the Supreme Court’s use of canons); Eskridge, Interpreting Law, supra note 4, passim (referring to the Supreme Court’s use of various canons); Eskridge Jr. et al., supra note 82, app. B (listing canons invoked by the Rehnquist and Roberts Courts); Barrett, supra note 92, at 128–54 (tracing the history of substantive canons); Brudney & Ditslear, supra note 82, at 29–33 (reporting the Court’s reliance on different interpretive resources); Krishnakumar, Reconsidering, supra note 27, at 847–50 (reporting the frequency with which Justices on the Roberts Court referenced canons); Krishnakumar, Roberts Court First Era, supra note 27, at 221–24 (examining the Roberts Court’s reliance on canons in statutory cases); Nicholas S. Zeppos, The Use of Authority in Statutory Interpretation: An Empirical Analysis, 70 Texas L. Rev. 1073, 1075–76 (1992) (analyzing the Supreme Court’s use of authority in statutory cases). But see Cross, supra, at 180–200 (examining use of canons and other interpretive resources by federal courts of appeals).
  96. .Bruhl, supra note 91, at 496–503.
  97. .See Abbe R. Gluck, The States as Laboratories of Statutory Interpretation: Methodological Consensus and the New Modified Textualism, 119 Yale L.J. 1750, 1778, 1805–06 (2010) (describing approaches adopted by several state supreme courts).
  98. .The U.S. Supreme Court began its first sittings on February 2, 1790. Bernard Schwartz, A History of the Supreme Court 18 (1993).
  99. .Notably, several of the canons listed individually in Interpreting Law are narrower applications or formulations of a broader canon that is well-established (and that has been cited at least seven times by the Supreme Court). In such cases, even if the narrow application is not referenced the minimum number of times required to count as a canon on its own, it should be considered a subset of the broader canon. For example, the appendix lists a rule stating that “[a]mbiguities or uncertainties in criminal laws referenced in immigration statutes should be resolved in favor of noncitizens.” See Eskridge, Interpreting Law, supra note 4, app. at 442. This strikes us as a narrow application, or subset, of both the rule of lenity and the immigration-law-specific rule that ambiguities should be resolved in favor of the alien.
  100. .Id. app. at 441.
  101. .Id. app. at 436.
  102. .Id. app. at 435.
  103. .Scalia, supra note 56, at 29; see also Barrett, supra note 92, at 129–34 (calling lenity “an entrenched part of the English approach to statutory interpretation” and tracing its early adoption by American courts).
  104. .See, e.g., William N. Eskridge Jr., Public Values in Statutory Interpretation, 137 U. Pa. L. Rev. 1007, 1051 (1989) (explaining that courts use common law rules to fill in gaps in statutes because the common law offers a “readily accessible body of rules” that private parties already are familiar with and are accustomed to following).
  105. .William Blackstone, Commentaries on the Laws of England (1st ed. 1765).
  106. .A quick word search in Westlaw (Blackstone! /s comment!), for example, found 433 U.S. Supreme Court references to the Commentaries—356 of them made in 1902 or later.
  107. .See Anita S. Krishnakumar, Longstanding Agency Interpretations, 83 Fordham L. Rev. 1823, 1849–50 (2015) (citing Edmund Burke, Reflections on the Revolution in France and the Rights of Man (Dolphin ed. 1961) (1790)) (maintaining that an interpretation’s survival for a long period of time is evidence that it is sound, which is consistent with the Burkean preference for longstanding understandings).
  108. .See, e.g., Eskridge & Frickey, supra note 80, at 597 (calling the Court’s creation of these clear statement canons a “most striking innovation”).
  109. .E.g., Miranda v. Arizona, 384 U.S. 436 (1966); Gideon v. Wainwright, 372 U.S. 335 (1963); Brown v. Bd. of Educ., 347 U.S. 483 (1954); Marbury v. Madison, 5 U.S. (1 Cranch) 137 (1803).
  110. .See Green v. Bock Laundry Mach. Co., 490 U.S. 504, 528 (1989) (Scalia, J., concurring) (arguing that the meaning of a statute’s terms ought to be based on what is “most in accord with context and ordinary usage, and thus most likely to have been understood by the whole Congress which voted on the words of the statute”); James J. Brudney, Canon Shortfalls and the Virtues of Political Branch Interpretive Assets, 98 Calif. L. Rev. 1199, 1203 (2010) [hereinafter Brudney, Canon Shortfalls] (describing the theory held by some that judicial reliance on conventional usage when construing a statute’s terms “promote[s] greater predictability in statutory interpretation”); Brudney & Ditslear, supra note 82, at 12 (explaining that language canons aim to give effect to the plain meaning of the legislature’s language, “which in turn is understood to promote the actual or constructive intent of the legislature that enacted such language”); William W. Buzbee, The One-Congress Fiction in Statutory Interpretation, 149 U. Pa. L. Rev. 171, 226 (2000) (discussing Justice Scalia’s views about the judicial obligation to impose coherence on the U.S. Code); Elizabeth Garrett, Attention to Context in Statutory Interpretation: Applying the Lessons of Dynamic Statutory Interpretation to Omnibus Legislation, Issues Legal Scholarship, Nov. 2002, art. 1, at 7 (describing how Congress drafts statutes knowing that they will be interpreted according to certain norms and default rules); Scalia, supra note 56, at 16 (discussing tension between applying the plain meaning of a statute and attempting to give effect to the legislature’s intent).
  111. .See, e.g., Bock Laundry, 490 U.S. at 528 (Scalia, J., concurring) (“The meaning of terms on the statute books ought to be determined . . . on the basis of which meaning is . . . most compatible with the surrounding body of law into which the provision must be integrated . . . .”); Scalia & Garner, supra note 6, at 252–53 (describing how words or phrases in a statute should be construed not to clash with other provisions of that statute).
  112. .See, e.g., Brudney, Canon Shortfalls, supra note 108, at 1205 (explaining Frickey’s view that legal interpretation does not rely on conventional usage or ordinary meaning and emphasizing Frickey’s references to “evolving circumstances and extrinsic public-law values”); John F. Manning, Textualism and the Equity of the Statute, 101 Colum. L. Rev. 1, 125 (2001) (explaining the textualists’ practice of “reading statutes in light of established background conventions”); Scalia, supra note 56, at 29 (defending the rule of lenity and rules requiring a clear statement to eliminate state sovereignty or to waive the federal government’s sovereign immunity).
  113. .See, e.g., Eskridge, Interpreting Law, supra note 4, at 52–53 (explaining that conversations and statutory interpretation both operate under a cooperative principle); Geoffrey P. Miller, Pragmatics and the Maxims of Interpretation, 1990 Wis. L. Rev. 1179, 1220 (describing how certain canons are recognizable in everyday conversational settings).
  114. .See, e.g., Bailey v. United States, 516 U.S. 137, 146 (1995) (“We assume that Congress used two terms because it intended each term to have a particular, nonsuperfluous meaning.”).
  115. .Abbe R. Gluck & 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, 934 (2013).
  116. .See, e.g., Chevron U.S.A. Inc. v. Echazabal, 536 U.S. 73, 80–81 (2002) (declining to apply the expressio unius canon where other reasons existed for exclusion of a statutory term).
  117. .See Nat’l Petroleum Refiners Ass’n v. FTC, 482 F.2d 672, 676 (D.C. Cir. 1973) (noting that the canon “stands on the faulty premise that all possible alternative or supplemental provisions were necessarily considered and rejected by the [legislature]”); Sutherland, supra note 84, § 47:25 (discussing the limitations of the canon).
  118. .The canon appears to have first been referenced by the U.S. Supreme Court in 1806. See United States v. Grundy & Thornburgh, 7 U.S. (3 Cranch) 337, 353, 356d (1806) (referencing a lower court opinion employing the maxim “[e]xpressio unius est exclusio alterius”). Since then, it has been cited in at least another 130 cases. (A Westlaw search for “expressio unius” turned up 131 cases total.)
  119. .Geoffrey Miller wrote a fascinating article suggesting that the canons could be justified under Paul Grice’s theory of conversational cooperation. Miller, supra note 111, at 1191–92. Grice’s theory of cooperation, however, is quite controversial. Moreover, there is a significant question whether it applies in environments—within Congress or between Congress and courts—in which speakers have incentives not to cooperate. Similarly, David Shapiro has written a deservedly famous defense of the canons as favoring continuity as opposed to change. Shapiro, supra note 82. Eskridge cites Shapiro favorably, but one must wonder how “new” or “unconventional” canons preserve continuity.
  120. .See Krishnakumar, Reconsidering, supra note 27, at 856, 901–08 (reporting results of an empirical study finding that the vast majority of substantive canons invoked by the Roberts Court fell into one of these six categories).
  121. .These include, but are not limited to, canons instructing that the Sherman Act should be construed in light of its overall purpose of benefitting consumers; that Indian tribes cannot be sued without explicit congressional authorization; that veterans’ benefits statutes must be construed liberally for their beneficiaries; and that a presumption against the national “diminishment” of Indian lands exists. See Michigan v. Bay Mills Indian Cmty., 134 S. Ct. 2024, 2030 (2014) (invoking the principle that tribal sovereignty and immunity from suit can only be abridged through Congressional action); Henderson ex rel. Henderson v. Shinseki, 562 U.S. 428, 440–41 (2011) (holding that the Veterans’ Judicial Review Act should be “construed in the beneficiaries’ favor,” absent “clear indication” otherwise); Weyerhaeuser Co. v. Ross-Simmons Hardwood Lumber Co., 549 U.S. 312, 317, 319 (2007) (explaining that the Court has not permitted “recovery for above-cost price cutting” under the Sherman Act because it could chill “legitimate price cutting,” which benefits consumers); Hagen v. Utah, 510 U.S. 399, 411 (1994) (acknowledging that in diminishment cases, statutory ambiguities are resolved “in favor of the Indians”).
  122. .See John F. Manning, Continuity and the Legislative Design, 79 Notre Dame L. Rev. 1863, 1863 (2004) (“[P]roponents now emphasize that much like any other interpretive practice, a canon’s utility will depend on the interpreter’s capacity, at times, to identify how members of a linguistic community would ordinarily use that canon in context.”).
  123. .Id. at 1863, 1869–70.
  124. .For a detailed explanation of this feature of discourse, see Paul Grice, Studies in the Way of Words 22–40 (1989).
  125. .Charles Freeman, A.D. 381: Heretics, Pagans, and the Dawn of the Monotheistic State 129 (2009) (“Theodosius . . . had provided the legal framework within which Christianity had been given dominance over paganism and the Nicene Creed precedence within Christianity.”); see id. at 101–03 (observing that Theodosius demanded that the sects reach agreement over “intractable philosophical problems” and by law “silenced the debate”); id. at 35 (“[I]t was not until the fourth century that the texts included in the New Testament were [finalized] as a canon . . . .”).
  126. .See Tonja Jacobi, Sonia Mittal & Barry R. Weingast, Creating a Self-Stabilizing Constitution: The Role of the Takings Clause, 109 Nw. U. L. Rev. 601, 626 (2015) (concluding that compromises historically “resolve[] the immediate issue of the crisis” and “set rules governing future policies”). See generally Barry R. Weingast, Political Stability and Civil War: Institutions, Commitment, and American Democracy (discussing the interplay of the balance rule, representative democracy, and federalism), in Analytic Narratives 148 (Robert Bates ed. 1998).
  127. .See Lopez v. Gonzales, 549 U.S. 47, 57–58 (2006) (“[T]he failure of even a single eager Assistant United States Attorney to act on the Government’s interpretation of ‘felony punishable under the [CSA]’ in the very context in which that phrase appears in the United States Code belies the Government’s claim that its interpretation is the more natural one.”).
  128. .A good example is the relatively new “elephants in mouseholes” canon, first articulated in Whitman v. Am. Trucking Associations, 531 U.S. 457 (2001). See 468 (“Congress, we have held, does not alter the fundamental details of a regulatory scheme in vague terms or ancillary provisions—it does not, one might say, hide elephants in mouseholes.”). Since Am. Trucking, the Court has referenced the “elephants in mouseholes” concept in thirteen subsequent cases. See Epic Sys. Corp. v. Lewis, 138 S. Ct. 1612, 1627 (2018); Cyan, Inc. v. Beaver County Employees Ret. Fund, 138 S. Ct. 1061, 1071 (2018); Czyzewski v. Jevic Holding Corp., 137 S. Ct. 973, 984 (2017); Puerto Rico v. Franklin California Tax-Free Tr., 136 S. Ct. 1938, 1947 (2016); Burwell v. Hobby Lobby Stores, Inc., 134 S. Ct. 2751, 2796 (2014); E.P.A. v. EME Homer City Generation, L.P., 572 U.S. 489, 528 (2014); Chamber of Commerce of U.S. v. Whiting, 563 U.S. 582, 640—41 (2011); Bilski v. Kappos, 561 U.S. 593, 645 (2010); Coeur Alaska, Inc. v. Se. Alaska Conservation Council, 557 U.S. 261, 303 (2009); Entergy Corp. v. Riverkeeper, Inc., 556 U.S. 208, 239 (2009); Ali v. Fed. Bureau of Prisons, 552 U.S. 214, 247 (2008); Metro. Life Ins. Co. v. Glenn, 554 U.S. 105, 116 (2008); Gonzalez v. Oregon, 546 U.S. 243, 267 (2006). None of these cases has characterized the “no elephants in mouseholes” principle as a “canon” per se; indeed, one called it an “English language observation.” Ali, 552 U.S. 214 at 247. Nevertheless, many scholars and lower courts consider the “elephants in mouseholes” principle to be a canon. See, e.g., Aaron-Andrew P. Bruhl, Communicating the Canons: How Lower Courts React when the Supreme Court Changes the Rules of Statutory Interpretation, 100 Minn. L. Rev. 481, 542 (2015) (“The ‘no elephants in mouseholes’ canon now occupies a secure, if limited, place in the interpretive landscape”); Jacob Loshin & Aaron Nielson, Hiding Nondelegation in Mouseholes, 62 Admin. L. Rev. 19 (2010) (discussing and evaluating the canon); Nina A. Mendelson, Change, Creation, and Unpredictability in Statutory Interpretation: Interpretive Canon Use in the Roberts Court’s First Decade, 117 Mich. L. Rev. 71, 104 (2018) (referring to the “new ‘no elephants in mouseholes’ canon”); Michael Coenen & Seth Davis, Minor Courts, Major Questions, 70 Vand. L. Rev. 777, 787 (2017); Si Min Cen v. Attorney Gen., 825 F.3d 177, 194 (3d Cir. 2016); Coal. for Responsible Regulation, Inc. v. E.P.A., No. 09-1322, 2012 WL 6621785, at *1 (D.C. Cir. Dec. 20, 2012); United States v. Franck’s Lab, Inc., 816 F. Supp. 2d 1209, 1237 (M.D. Fla. 2011), order vacated, appeal dismissed, No. 11-15350, 2012 WL 10234948 (11th Cir. Oct. 18, 2012).
  129. .See Baude & Sachs, supra note 3, at 1088–89 (discussing canons as part of the “law of interpretation”).