Harvard Law Review Harvard Law Review Harvard Law Review

Harvard Law Review Forum

The Appellate Rule of Lenity

Responding to Abbe R. Gluck & Richard A. Posner, Statutory Interpretation on the Bench: A Survey of Forty-Two Judges on the Federal Courts of Appeals

The full text of this Response, including the Appendix, may be found by clicking on the PDF link to the left.


Professor Abbe Gluck and Judge Richard Posner rightly point to the federal courts of appeals as the locus of most statutory interpretation,1×1. Abbe R. Gluck & Richard A. Posner, Statutory Interpretation on the Bench: A Survey of Forty-Two Judges on the Federal Courts of Appeals, 131 Harv. L. Rev. 1298, 1300 (2018). even though the Supreme Court usually gets the attention.2×2. For descriptions of exceptions to this rule, including an article by one author of the piece on which this Response comments, see, for example, Abbe R. Gluck, The States as Laboratories of Statutory Interpretation: Methodological Consensus and the New Modified Textualism, 119 Yale L.J. 1750 (2010); Emmett S. Collazo, Applying the Rule of Law Subjectively: How Appellate Courts Adjudicate, 4 Seton Hall Cir. Rev. 303 (2008). One major finding of their recent survey concerns appellate courts’ interpretive relationship to the Supreme Court’s use of the increasingly relevant canons of construction.3×3. Gluck & Posner, supra note 1, at 1327–34. Unsurprisingly, the rule of lenity takes on special significance.4×4. See, e.g., United States v. Santos, 553 U.S. 507, 514 (2008); United States v. Bass, 404 U.S. 336, 347–49 (1971).

Lenity’s prominence is unsurprising for a few reasons. First, as an ancient principle directing judges to construe ambiguous criminal statutes narrowly,5×5. See Antonin Scalia, A Matter of Interpretation 29 (Amy Gutmann ed., 1997) (“The rule of lenity is almost as old as the common law itself, so I suppose that is validated by sheer antiquity.” (footnote omitted)). the rule is — as the authors underline — “special.”6×6. Gluck & Posner, supra note 1, at 1331–32 (“Lenity and Avoidance Are Special.”). Of the appellate judges surveyed,

[t]hirteen judges . . . singled out lenity and/or constitutional avoidance as “actual rules” and distinguished them from the other canons, in terms of their mandatory application. They told us that those presumptions are “not canons” but rather are “substantive law.” Some judges seemed to believe these doctrines derived their special status simply because the Supreme Court said they did. . . . A few judges did not see these canons as deriving their power from the Supreme Court, but rather from the Constitution (for example, that lenity derives from the constitutional concept that federal judges cannot create crimes).7×7. Id.

It helps that most younger judges learned the legal canons in law school8×8. Id. at 1305. — where they invariably examined the rule of lenity in legislation classes like the one I teach. Additionally, Justice Scalia and Professor Bryan Garner have helped elevate the rule of lenity by including it in a set of fifty-seven recommended canons of construction in their widely read treatise on interpretation.9×9. See Antonin Scalia & Bryan A. Garner, Reading Law: The Interpretation of Legal Texts 296–302 (2012). Professor Gluck and Judge Posner report that some judges appeal to that treatise and to Professor William Eskridge’s lists of canons when interpreting statutes. See Gluck & Posner, supra note 1, at 1327–28, 1339, 1351, 1363. One judge noted: “I am a huge fan of the rule of lenity. I wish it was used even more.” Id. at 1363. For the lists of canons identified by Eskridge, sometimes with his colleagues, see William N. Eskridge, Jr., Interpreting Law: A Primer on How to Read Statutes and the Constitution 407–45 (2016); William N. Eskridge, Jr., Abbe R. Gluck & Victoria F. Nourse, Statutes, Regulation, and Interpretation app. 6 at 1091–114 (2014); and William N. Eskridge, Jr. & Philip P. Frickey, The Supreme Court, 1993 Term — Foreword: Law as Equilibrium, 108 HARV. L. REV. 26 app. at 97–108 (1994). Second, the rule of lenity carries authoritative weight. Almost one-third (13) of the 42 appellate judges surveyed considered lenity and/or constitutional avoidance “actual rules” of “mandatory application.”10×10. Gluck & Posner, supra note 1, at 1331–32. For some respondents, the Supreme Court designated these rules as binding and, for others, the Constitution did so.11×11. Id. at 1332. For further discussion of the constitutional foundations of lenity, see infra section I.B, pp. 193–95. Third, lenity is generally valid on a spectrum of interpretive approaches, regardless of ideological commitments. Most appellate judges who participated in the survey signaled a willingness to consider the canons of construction, and all judges actually used them even when they said they did not.12×12. Gluck & Posner, supra note 1, at 1343; see also id. at 1328 (“A random sampling of opinions from each of the forty-two judges we interviewed revealed all of the judges to use canons routinely, despite some telling us that they do not.”). Most explained that they consider as much information and as many canons as possible to make the most informed decision about a case of statutory construction.13×13. Id. at 1302–03. This “interpretive eclecticism”14×14. Id. at 1320; see also id. at 1313–15. suggests that appellate judges would consider lenity to be relevant to their review of challenged criminal statutes.15×15. Id. at 1344–45 (noting that six judges believe that the Supreme Court can and does dictate statutory interpretation rules, and that some judges of those who do not think the Supreme Court has the power to make statutory interpretation decisions nevertheless make exceptions for certain rules like lenity).

Given the rule’s special status, broad authority, and widespread validity, it stands to reason that lenity would feature prominently in the criminal law interpretation of appellate judges. These features of lenity combine with important features of criminal law that mark this field as different, and the rule of lenity as worthy of particular regard as interpretive precedent. Not only does criminal law entail levying the harshest of punishments on our own citizens, thus urging extreme caution, but experiences over the last half century lead to the inescapable conclusion that the criminal justice system is broken and in need of reform.16×16. See William J. Stuntz, The Collapse of American Criminal Justice (2011); Michelle Alexander, The New Jim Crow: Mass Incarceration in the Age of Colorblindness (2010). The collapse of the criminal justice system makes it imperative to devise an approach to criminal law that can correct the overcriminalization, overprosecution, and oversentencing now rampant in the United States.17×17. For excellent assessments of lenity that link the rule to the structure of the criminal justice system, see Shon Hopwood, Clarity in Criminal Law, 54 Am. Crim. L. Rev. 695, 697–700 (2017); Zachary Price, The Rule of Lenity as a Rule of Structure, 72 Fordham L. Rev. 885, 890–95 (2004). What Professor Zachary Price calls structure and links to the democratic accountability of the voter, I refer to as the “liberty interest” or “defendants’ rights” rationale for lenity that originates in a constitutional mandate, which I distinguish from the mandate for fair notice as a due process, rule-of-law principle (agreeing with Price’s categorization of the latter as structural). In my view, substantive rights to life, liberty, and equality for citizens who become criminal defendants, stand separately from due process. For further discussion of a liberty/rights rationale for lenity, see infra, notes 74, 131–133, and accompanying text Fair punishment requires corrective measures to better ensure that — alongside legislative supremacy justifications for criminal law definitions and enforcement — constitutional values of equality, liberty, and due process are also met in interpretive questions of criminal law.18×18. See Hopwood, supra note 17, at 697, 709–13. Lenity lies at the heart of interpretive questions in the criminal justice arena, the state of which adds urgency and cause for greater consider-ation of the framework and application of the rule.19×19. Id. For a thoughtful argument about the need for greater involvement of judges in interpretation of criminal law, given its distinctive features and pathological politics, to create more equitable outcomes, see Joshua Kleinfeld, Equitable Arguments and Criminal Statutes (unpublished paper) (on file with author).

To be sure, Professor Gluck and Judge Posner stress that — except in rare cases, such as Chevron deference,20×20. See Chevron U.S.A., Inc. v. Nat. Res. Def. Council, Inc., 467 U.S. 837 (1984). which every judge in their study deemed as binding21×21. Gluck & Posner, supra note 1, at 1345 (“[W]hen asked . . . about Chevron deference — the interpretive rule that courts must defer to reasonable agency interpretations of ambiguous statutes — every interviewed judge told us this rule is binding even if they disagreed with it.”). — there is “no ‘methodological stare decisis’ in the federal courts.”22×22. Id. at 1343–44. Professor Gluck and Judge Posner define “methodological stare decisis” as “the treatment of statutory interpretation rules as precedents, just as courts treat analogous decisionmaking rules such as the parol evidence rule or burden-shifting regimes.” Id. at 1344 (citing Abbe R. Gluck, Intersystemic Statutory Interpretation: Methodology as “Law” and the Erie Doctrine, 120 Yale L.J. 1898, 1970–76 (2011) [hereinafter Gluck, Intersystemic Statutory Interpretation]; cf. Abbe R. Gluck, The Federal Common Law of Statutory Interpretation: Erie for the Age of Statutes, 54 Wm. & Mary L. Rev. 753 (2013) (documenting jurisdictions that do not consider statutory interpretation rules as law); Gluck, Intersystemic Statutory Interpretation, supra (arguing that courts routinely fail to treat both state and federal statutory interpretation rules as law); Gluck, supra note 2, at 1765–66 (arguing that the Supreme Court does not treat its own methodological rules as binding on lower courts); id. at 1823–24 (contrasting the presence of methodological stare decisis in some state courts with its absence in federal courts). Yet if there were any exception to that norm, lenity would be a strong candidate. Indeed, some of the appellate judges who rejected a general notion of methodological stare decisis made exceptions for lenity among a limited set of canons that they saw as binding, based on certain “constitutional principles” or “rules” to which judges must defer.23×23. Gluck & Posner, supra note 1, at 1345.

More pointedly, there is additional reason for courts to treat the lenity regime for criminal law as interpretive precedent — that is, a lighter version of methodological stare decisis that refers to regular consider-ation of an interpretive framework that nonetheless preserves judicial discretion in applying the associated rule when certain conditions are met. Consideration of the lenity framework occurs when judges consider whether the rule of lenity is relevant, typically by determining that they are facing a challenge to the application or meaning of a criminal law statute or regulation. Judges who consider the framework may ultimately reject the rule in favor of a higher-priority canon, a claim of statutory clarity reached through reading the text and applying other interpretive tools, or a competing constitutional concern. Application of the lenity rule occurs only when judges narrowly construe a criminal or sentencing provision, typically by also citing “lenity” or the “rule of narrow construction for criminal statutes” — even if they do so as a supplementary tool of persuasion that accompanies other rationales on which the decision purportedly turns.24×24. See id. at 1302, 1330 (referring to judicial terms for such usages as “window dressing”).

The distinction between consideration of the lenity framework and application of the lenity rule is key. On it turns an accurate appraisal of lenity, which can perhaps best be analogized to the Chevron two-step framework and its accompanying rule of agency deference.25×25. See Kenneth A. Bamberger & Peter L. Strauss, Chevron’s Two Steps, 95 Va. L. Rev. 611, 611 (2009) (“The familiar two-step analysis [of Chevron] is best understood as a framework for allocating interpretive authority,” after answering a threshold question). As the most cited administrative law case,26×26. Peter M. Shane & Christopher J. Walker, Symposium, Chevron at 30: Looking Back and Looking Forward — Foreword, 83 Fordham L. Rev. 475, 475 (2014). For a chart of administrative law cases and citations, see Chris Walker, Most Cited Supreme Court Administrative Law Decisions, Yale J. on Reg.: Notice & Comment (Oct. 9, 2014), http://yalejreg.com/nc/most-cited-supreme-court-administrative-law-decisions-by-chris-walker/ [https://perma.cc/QD67-FPB9]. Chevron directs courts to uphold agency interpretations of certain regulatory statutes so long as they are reasonable.27×27. See Chevron U.S.A. Inc. v. Nat. Res. Def. Council, Inc., 467 U.S. 837, 843–44, 843 n.11 (1984). The Supreme Court has invoked its framework so consistently that it has become interpretive precedent, with methodological stare decisis effect; most commentators see it as the only legal canon to enjoy that status.28×28. See Abbe R. Gluck, What 30 Years of Chevron Teach Us About the Rest of Statutory Interpretation, 83 Fordham L. Rev. 607, 609–12 (2014). Moreover, even though Justice Thomas questioned whether agency deference precedents should, in normative terms, be “entitled to stare decisis effect,”29×29. Perez v. Mortgage Bankers Ass’n, 135 S. Ct. 1199, 1214 n.1 (2015) (Thomas, J., concurring in the judgment). most courts have ostensibly regarded the doctrine as such. Yet the consistent consideration of the agency deference framework differs from less regular application of the rule: a 2006 empirical appraisal of the 1014 cases in which the Supreme Court invoked Chevron since the decision was handed down in 1984 revealed that deference applied only 8.3% of the time.30×30. See William N. Eskridge, Jr. & Lauren E. Baer, The Continuum of Deference: Supreme Court Treatment of Agency Statutory Interpretations from Chevron to Hamdan, 96 Geo. L.J. 1083, 1089–90, 1131 (2008).

Once the framework consideration/rule application distinction is made, lenity seems to mirror the interpretive precedent of the type that Chevron exemplifies — as a matter of kind rather than degree. The claim here is not that Chevron or lenity are subject to the “super-strong presumption of correctness” stare decisis regime that the Court applies to its own statutory interpretation decisions.31×31. See William N. Eskridge, Jr., Overruling Statutory Precedents, 76 Geo. L.J. 1361, 1362 (1988). That sort of precedent is about substantive decisions, not interpretive frameworks. Nor is the claim that the number of times the Supreme Court considered the lenity framework rivals that of Chevron. Administrative agency challenges are much more numerous than criminal law statutory interpretation questions; accordingly, invocations of the Chevron framework are much more numerous as well. Instead, the claim is that the Court’s approach to lenity corresponds to that of Chevron in that the Court considers both frameworks consistently (in almost all relevant cases in criminal and administrative law challenges, respectively), and it applies the rule at a lower, but non-negligible, rate when compared to its consideration of the framework. Thus both arenas represent a type of interpretive precedent that calls on courts to consistently consider an interpretive framework for certain types of cases even when judges exercise discretion in deciding whether to apply the associated rule.

Appellate statutory criminal cases that were granted certiorari during the Roberts Court era so far have not borne out this intuition. Upon examination of these criminal law statutory interpretation cases, the Supreme Court’s approach to lenity stands in relief from the practice of appellate courts. For its part, in those cases, the Supreme Court almost always considers the lenity framework. Moreover, the Supreme Court applies the rule of lenity in about one-third of those cases, thereby modeling the discretion judges have to apply or reject the rule after consideration of the framework. By contrast, it turns out that appellate courts considered in this data set did neither. They hardly ever invoked the framework or applied the rule. This contrast may well point to a larger trend (still to be empirically examined).

If this sample of appellate court practice does indeed reflect a broader trend, this Response argues that many courts and commentators have gotten lenity wrong by equating the Supreme Court’s spotty consideration of the lenity rule with its more regular application of the lenity framework. The lower courts’ and commentators’ blunt appraisals of lenity have led to the false impression that the Supreme Court’s approach to the framework is inconsistent and weakly grounded. Because the lenity rule does not apply to a majority of criminal law statutory interpretation cases, commentators and judges tend to regard lenity as far from being interpretive precedent. Yet empirical examination of the Supreme Court’s treatment of lenity in the Roberts Court reveals fairly consistent consideration of the framework,32×32. For this Response, I reviewed all statutory interpretation cases involving criminal provisions in the Roberts Court and compared them with consideration of those same cases in the federal courts of appeals. See Appendix, pp. 211–16. In a longer paper, I review all Supreme Court cases of criminal law statutory interpretation from the beginnings of lenity jurisprudence. See Intisar A. Rabb, The Constitutional Rule of Lenity (unpublished) (on file with author). That study reveals that the Supreme Court has applied the lenity framework fairly consistently since the first major statutory construction case in United States. v. Wiltberger, 18 U.S. (5 Wheat.) 76, 95 (1820) through the latest complete Term of the Roberts Court (2016). It further shows, I argue, that the rationales for the rule have changed over time in response to prevalent structures and practices in the criminal justice system that push courts to emphasize certain constitutional values at different times. and it reveals applications of the lenity rule almost four times the rate of application of Chevron deference for the period covered in Professor William Eskridge’s and Lauren Baer’s longitudinal study of that canon.33×33. See Eskridge & Baer, supra note 30, at 1090 (contrasting the 8.3% of cases evaluating agency statutory interpretation decisions to which the Supreme Court applied Chevron deference — what I call applications of the rule, with the slight majority of those cases (53.6%) to which the Supreme Court did not apply any deference regime — what I call consideration of the framework); see also id. at 1094 (referring to the “deference regime” for Chevron as an “analytical framework” that the Court uses to evaluate agency interpretations eligible for Chevron deference); id. at 1130–31 (noting instances in which the Supreme Court, notwithstanding statutory text suggesting applicability of the Chevron framework, took “a non-deferential approach to agency interpretations of their own jurisdiction or authority” — showing specific instances of its failure to apply the rule). In other words, a review of its treatment of the framework rather than the rule reveals that, especially when compared to Chevron, the Supreme Court indeed tends to treat lenity as interpretive precedent.34×34. While it is outside the scope of this Response, these facts raise a question: why does the Court apply the lenity rule more often than it applies Chevron deference? One reason may be that the constitutional underpinnings for lenity, together with the notion that criminal law is different, create more concentrated attention on lenity for a constitutional court. This Response thus challenges the notion that the Supreme Court’s treatment of lenity justifies the stingy view of it prevalent in appellate courts and elsewhere. It also suggests a more historically based and constitutionally grounded approach to the rule.

The approach to lenity in the Supreme Court and appellate courts during the Roberts Court era illustrates the recent practice. From 2005 to 2017, the Supreme Court considered criminal law statutory interpretation questions in 47 cases. The Justices considered the lenity framework in almost all of these cases (44); it applied the lenity rule to narrow construction in about one-third of them (13 of 44 cases);35×35. Following the usage among most commentators and courts — including appellate court judges who called the rule “shorthand for [the] proposition that judges aren’t supposed to create crime,” Gluck & Posner, supra note 1, app. at 1366 (alteration in original) — I define lenity as a principle of “narrow construction of criminal statutes,” even when the Court does not use the term “lenity.” Although Justice Frankfurter was the first to name the rule lenity in Callanan v. United States, 364 U.S. 587 (1961), Chief Justice Marshall first announced the canon of narrow construction of criminal statutes long before in Wiltberger, 18 U.S. (5 Wheat.) 76. Lawrence M. Solan, Law, Language, and Lenity, 40 Wm. & Mary L. Rev. 57, 83–84, 91–94, 109 (1998). For similar usage — that is, a general identification of the canon as one of narrow construction for criminal statutes without limitation to the label lenity — see, for example, Dan M. Kahan, Lenity and Federal Common Law Crimes, 1994 Sup. Ct. Rev. 345, 345–46. Thus in coding cases, the label “lenity” refers to cases wherein the Court applied a narrow construction and the respondent won on a question involving a criminal law statute or provision. In some of these cases, the Court explicitly mentioned the lenity rule and in others merely construed the criminal statute narrowly. In the chart listing lenity cases in the Appendix, I distinguish between the two by coding the former simply as “lenity applied” and the latter as “lenity applied (narrow construction).” Even when not citing lenity by name, the latter set of opinions often carried indications that the Court considered the lenity framework — that is, the fact that it was construing a criminal statute — together with rationales for its narrow construction. See, e.g., Kellogg Brown & Root Servs., Inc. v. United States ex rel. Carter, 135 S. Ct. 1970, 1978 (2015) (“But even if there were some ambiguity in the [statute’s] use of that term, our cases instruct us to resolve that ambiguity in favor of the narrower definition. We have said that the [statute] should be ‘narrowly construed’ and ‘interpreted in favor of repose.’” (quoting Bridges v. United States, 346 U.S. 209, 216 (1953))); Moncrieffe v. Holder, 569 U.S. 184, 205 (2013) (“[W]e err on the side of underinclusiveness because ambiguity in criminal statutes referenced by the INA must be construed in the noncitizen’s favor.”); Dorsey v. United States, 567 U.S. 260, 264 (2012) (“We recognize that, because of important background principles of interpretation, we must assume that Congress did not intend those penalties to apply unless it clearly indicated to the contrary.”); Abuelhawa v. United States, 556 U.S. 816, 820 (2009) (“The traditional law is that where a statute treats one side of a bilateral transaction more leniently, adding to the penalty of the party on that side for facilitating the action by the other would upend the calibration of punishment set by the legislature . . . .”). Additionally, a case would code as an application of lenity only if the Court did not explicitly reject the rule. “Rejection” comes in the form of the majority’s explicit dismissal of the application of lenity, read broadly, or its failure to adopt or respond to a concurring or dissenting opinion’s plea to apply lenity. Accordingly, there were two instances of narrow construction that do not count as applications of lenity, because the Court explicitly dismissed the rule but narrowly construed the statute. See Fowler v. United States, 563 U.S. 668, 678 (2011) (construing the statute narrowly, but “never cit[ing] the rule of lenity,” id. at 682 (Scalia, J., concurring in the judgment)); Begay v. United States, 553 U.S. 137, 143 (2008), abrogated by Johnson v. United States, 135 S. Ct. 2551 (2015). In such instances, even though the Court rejected lenity, it still considered the framework — discernible from the explicit or implicit dismissal. Cases in which the Court rejected lenity and considered the framework are labeled as “lenity rejected (broad construction).” and it considered some of these statutory provisions on constitutional grounds of void-for-vagueness (3 cases).36×36. See Beckles v. United States, 137 S. Ct. 886 (2017); Johnson, 135 S. Ct. 2551; United States v. Stevens, 559 U.S. 460 (2010). In the lenity cases, the Court usually justified its decisions on familiar grounds of legislative supremacy and fair warning, both constitutional mandates. The Court also revived older justifications for lenity, occasionally citing the liberty interests of defendants and common law practice, the latter another term for interpretive precedent. This revival comes against the backdrop of the ailing criminal justice system, whose widespread abuses resulting in mass incarceration that disproportionately affect black men call for greater regard for defendants’ liberty interests and other constitutional rights.37×37. See supra note 16 and accompanying text. Most notable in the Roberts Court’s approach to lenity is the consistency with which it considered the lenity framework and — at 33% — the non-negligible rate at which it applied the rule, especially relative to appellate courts below (or to applications of Chevron deference).

By contrast, appellate judges who decided these same cases before they reached the Supreme Court considered the lenity framework in approximately only one-fifth of the cases (12 of 49, which were consolidated into 44) and applied the rule in 4.38×38. See infra notes 140–42. As with the Supreme Court evaluation, appellate criminal law cases were coded as lenity-eligible where they involved a challenge to a criminal statute or provision, except that here, the universe is enlarged by the fact that the Supreme Court took the case on the statutory interpretation questions, whereas the appellate courts often decided the case on other grounds — often deeming the statute clear without considering the lenity framework in light of possible ambiguities (thus coded “N/A” on lenity considerations). See Appendix, pp. 211–16. When applying the rule, the appellate courts gave the lenity rule none of the rationales used by the Supreme Court. Granted, there is no way to correlate the judges who wrote the appellate opinions reviewed in the Roberts Court lenity cases with the judges interviewed in the Gluck and Posner survey.39×39. The appellate judges who responded to the Gluck and Posner survey remain anonymous for understandable reasons. Judge Posner was on an appellate panel in only one case — a RICO question in which lenity was not mentioned and was unanimously rejected at the Supreme Court: Phoenix Bond & Indem. Co. v. Bridge, 477 F.3d 928 (7th Cir. 2007), aff’d, 553 U.S. 639 (2008). It is likewise senseless to argue that the sample of appellate cases later heard by the Supreme Court is representative of the broader universe of appellate cases. After all, given the responses to the Gluck and Posner survey signaling that at least some appellate judges consider lenity special or binding as a rule,40×40. See supra notes 6, 10, and accompanying text. we might expect to see higher rates of lenity applications in appellate cases deemed difficult. The fact that we did not may be an indication that the appellate judges in these cases thought that the interpretive questions presented to them were easier than the Supreme Court found them to be. Or, put differently, it may indicate that the Supreme Court granted certiorari only in the difficult cases as it defined it, perhaps sometimes to settle a longstanding dispute that only the highest court could resolve.41×41. For example, the Court twice heard challenges to constructions of the term “violent felony” in the Armed Career Criminal Act, 18 U.S.C. § 924(e)(1) (2012), imposing mandatory minimum sentences for possession of firearms by a felon — once in favor of the Government and rejecting lenity over Justice Scalia’s objection, James v. United States, 550 U.S. 192 (2007), and once in favor of the petitioner and again rejecting lenity over Justice Scalia’s objection in his concurrence in the judgment, Begay, 553 U.S. 137. Justice Scalia upped the ante and eventually got his way when he authored the opinion in Johnson v. United States that ruled the enhanced penalty provision void for vagueness, thereby overturning longstanding precedent and the Court’s more recent lenity decisions, along with several appellate cases decided on the basis of the old norm. 135 S. Ct. 2551. For further discussion of subsequent jurisprudence on related matters, see infra note 56. Finally, it may indicate a divergence between interpretive theory, where the Gluck and Posner survey shows that judges regarded lenity to be special, and interpretive practice on appellate courts, where this review shows that judges did not apply lenity as a special rule — a reversal of Professor Gluck and Judge Posner’s observation that all judges routinely used canons even when some said they did not.42×42. See supra note 12 and accompanying text. Despite these open questions and the fact that the judges’ survey responses and written opinions do not entirely line up, this comparison of the appellate court and Supreme Court treatments of criminal law statutory interpretation cases that were granted certiorari in the Roberts Court is useful because it offers rich insight into attitudes toward lenity among a significant set of federal judges. The comparison also sheds light on the utility of disaggregating lenity in terms of judicial consideration of the framework versus applications of the rule, and the extent to which judges do or should consider lenity to be interpretive precedent. This Response explores those questions.

Part I outlines the lenity framework and the constitutional bases for the rule. Part II discusses the widespread Roberts Court consideration of the framework and its differing approaches to applications of the rule. Part III discusses the appellate courts’ much more limited application of both. The Response concludes with reflections about the significance of the framework-rule distinction in lenity jurisprudence, and argues that this more informed understanding can offer courts a more historically grounded and constitutionally mandated approach to the rule of lenity.

I. The Lenity Framework

Lenity operates as both an interpretive framework and a constitutional rule. This Part defines the lenity framework, and shows how the Court’s choice of one or two constitutional bases for lenity in lieu of all three often drives the application of the rule in ways that correspond to the test that the Court laid out in Chevron — with ambiguity at Step Zero being a trigger for application of the rule at Steps One or Two. This Part then considers the application of the rule — determining whether lenity enters into the interpretive process early or virtually not at all: lenity first or lenity last. The Court’s conflicting approaches to lenity, now coupled with the dire straits of the criminal justice system, demand resolution. Fortunately, the Court’s own historical practice provides one. A close examination of past judicial practice reveals early and deliberate emphasis on lenity first, which is now experiencing a late-onset displacement that is decidedly at odds with the full range of lenity’s constitutional underpinnings. When combined with the Court’s regular consideration of the framework, the lenity-first approach best follows historical precedent and constitutional mandates. Moreover, this combined approach allows judges to provide an institutional check on the political excesses that permit unclear laws, prosecutorial overreach, and infringements on liberty that have resulted in the discriminatory mass incarceration, overcriminalization, and overpunishment that characterize the American criminal justice system today.

A. The Lenity Framework Defined

Determining whether to construe criminal statutes with lenity is a two-step process that, like Chevron, involves a threshold inquiry as to whether ambiguity exists, importantly, in the relevant area of law — administrative for Chevron, criminal for lenity.43×43. See Chevron U.S.A. Inc. v. Nat. Res. Def. Council, Inc., 467 U.S. 837 (1984); see also Cass R. Sunstein, Chevron Step Zero, 92 Va. L. Rev. 187, 190–91 (2006) (defining Step Zero as “the initial inquiry into whether the Chevron framework applies at all”; Step One as “whether Congress ‘has directly spoken to the precise question at issue,’ . . . [meaning] whether Congress’ intent is ‘clear’ and ‘unambiguously expressed’”; and Step Two as asking “whether the agency’s interpretation is ‘permissible,’ which is to say reasonable in light of the underlying law” and therefore deserving of deference (quoting Chevron, 467 U.S. at 842–43)). At Step Zero of the lenity framework, courts make an initial determination whether they must construe a challenged provision of a criminal statute or a civil statute with criminal consequences.44×44. Compare Price, supra note 17, at 889–95, with Zachary Price, The Court After Scalia: The Rule of Lenity, SCOTUSblog (Sept. 2, 2016, 2:14pm), http://www.scotusblog.com/2016/09/the-court-after-scalia-scalia-and-the-rule-of-lenity [https://perma.cc/R9BD-T58J]. If not, lenity is not triggered and the plain import of the statute prevails45×45. See, e.g., United States v. Turkette, 452 U.S. 576, 580 (1981) (“If the statutory language is unambiguous, in the absence of ‘a clearly expressed legislative intent to the contrary, that language must ordinarily be regarded as conclusive.’” (quoting Consumer Prod. Safety Comm’n v. GTE Sylvania, Inc., 447 U.S. 102, 108 (1980)). — typically in a way that disfavors the defendant challenging the applicability of the statute in the first place.46×46. See Hopwood, supra note 17, at 738. If so, courts are to identify all plausible interpretations based on the usual interpretive considerations (Step One).47×47. See, e.g., Yates v. United States, 135 S. Ct. 1074, 1088 (2015) (“[I]f our recourse to traditional tools of statutory construction leaves any doubt about the meaning of [the challenged term], . . . we would invoke the rule that ‘ambiguity concerning the ambit of criminal statutes should be resolved in favor of lenity.’” (quoting Cleveland v. United States, 531 U.S. 12, 25 (2000))); Burrage v. United States, 134 S. Ct. 881, 891 (2014) (“Especially in the interpretation of a criminal statute subject to the rule of lenity, we cannot give the text a meaning that is different from its ordinary, accepted meaning, and that disfavors the defendant.” (citation omitted)); United States v. Santos, 553 U.S. 507, 514 (2008) (“From the face of the statute, there is no more reason to think that ‘proceeds’ means ‘receipts’ than there is to think that ‘proceeds’ means ‘profits.’ Under a long line of our decisions, the tie must go to the defendant. The rule of lenity requires ambiguous criminal laws to be interpreted in favor of the defendants subjected to them.”). If the statutory construction yields more than one plausible reading, lenity would apply to require selection of the narrowest such reading (Step Two).48×48. See United States v. Thompson/Ctr. Arms Co., 504 U.S. 505, 517–18 (1992) (“After applying the ordinary rules of statutory construction, then, we are left with an ambiguous statute. . . . It is proper, therefore, to apply the rule of lenity and resolve the ambiguity in [the defendant’s] favor.”); cf. Price, supra note 17, at 889–96.

The above formulation of the test may be called lenity first. It represents early versions of the rule, as advanced by Chief Justice Marshall and Justice Holmes.49×49. See United States v. Wiltberger, 18 U.S. (5 Wheat.) 76, 95 (1820); McBoyle v. United States, 283 U.S. 25, 27 (1931). Applying it, courts consider the lenity framework relevant from the outset for all criminal statutes. If they make an initial determination of ambiguity, they narrow constructions where, for example, the text was written broadly but the legislative history suggests a narrower purpose for a particular provision at issue, or where the text was written narrowly but the political context suggests a broader purpose.50×50. Cf. Price, supra note 17, at 889–96. This approach was common on the Supreme Court when lenity was first incorporated into American law, and continued to be in use until relatively recently.51×51. See, e.g., Bifulco v. United States, 447 U.S. 381, 387 (1980) (noting that lenity informs the construction of challenged criminal statutes).

A contrasting approach may be called lenity last, and it is the more popular approach in today’s Court. Here, too, the inquiry begins with a question at Step Zero as to whether the lenity framework applies: Is there a criminal statute challenged for ambiguous construction? To determine the answer, courts pursue a robust Step One — identifying all plausible interpretations through their preferred methods of statutory construction.52×52. On tools used in pragmatic interpretation, see William N. Eskridge, Jr. & Philip P. Frickey, Statutory Interpretation as Practical Reasoning, 42 STAN. L. REV. 321 (1990) (describing a “practical reasoning approach” as an alternative to textualism). In another piece, Eskridge succinctly lays out his alternate (“dynamic”) approach to textualism. William N. Eskridge, Jr., Textualism, the Unknown Ideal?, 96 Mich. L. Rev. 1509 (1998) (reviewing Antonin Scalia, A Matter of Interpretation (Amy Gutmann ed. 1997)) (contrasting dynamic statutory interpretation with textualism to include legislative purpose, which sometimes requires judges to consider legislative history and other extratextual factors). Courts may stop the inquiry there and, at Step Two, apply what they deem to be the most plausible or fair reading — including broad constructions based on a purpose gleaned from the overall policy or legislative history that they read as informing the text.53×53. See, e.g., Moskal v. United States, 498 U.S. 103, 110 (1990) (referring to a statute’s general purpose to reject a narrow construction of its text); Dixson v. United States, 465 U.S. 482, 496 (1984) (using legislative history to reject a narrow construction of the text); see also Abbott v. United States, 562 U.S. 8, 28 n.9 (2010) (“‘[T]he touchstone of the rule of lenity is statutory ambiguity.’ . . . ‘[A]fter consulting traditional canons of statutory construction,’ we are persuaded that none remains here. . . . Although the clause might have been more meticulously drafted, the ‘grammatical possibility’ of a defendant’s interpretation does not command a resort to the rule of lenity if the interpretation proffered by the defendant reflects ‘an implausible reading of the congressional purpose.’” (alterations in original) (citations omitted)). On this approach, courts only resort to lenity if, after considering every other interpretive tool available, they still cannot make sense of the statute and “grievous ambiguity” persists.54×54. See, e.g., Ocasio v. United States, 136 S. Ct. 1423, 1434 n.8 (2016) (“[Lenity] applies only when a criminal statute contains a ‘grievous ambiguity or uncertainty,’ and ‘only if, after seizing everything from which aid can be derived,’ the Court ‘can make no more than a guess as to what Congress intended.’” (quoting Muscarello v. United States, 524 U.S. 125, 138–39 (1998))); Barber v. Thomas, 560 U.S. 474, 488 (2010) (“[T]he rule of lenity only applies if, after considering text, structure, history, and purpose, there remains a ‘grievous ambiguity or uncertainty in the statute.’” (also quoting Muscarello, 524 U.S. at 139)). Only on the rare occasion when they find such grievous ambiguity would courts narrow the construction at Step Two. On this approach, once “a sort of ‘junior version of the vagueness doctrine,’”55×55. United States v. Lanier, 520 U.S. 259, 266 (1997) (quoting Herbert L. Packer, The Limits of the Criminal Sanction 95 (1968)); see also John Calvin Jeffries, Jr., Legality, Vagueness, and the Construction of Penal Statutes, 71 Va. L. Rev. 189, 199–200 (1985). lenity on these terms merges into vagueness, with neither doctrine applying unless there is no fair or consistent application of the challenged provision over time.56×56. Indeed, as noted above, supra note 41, the Court has long grappled with whether lenity or vagueness applies to challenges of “violent felony” convictions from the residual clause of the Armed Career Criminal Act of 1984 (ACCA). The Court finally declared the phrase to be unconstitutionally vague. See Johnson v. United States, 135 S. Ct. 2551 (2015); see also Welch v. United States, 136 S. Ct. 1257 (2016) (applying Johnson retroactively). But that declaration came only after judgments about individual applications of the term to particular prior convictions resulted in a series of 5–4 split decisions that illustrated the Court’s apparent inability to clarify the term with a lenity-last approach. Compare Begay v. United States, 553 U.S. 137 (2008) (rejecting lenity but applying a narrow construction), with James v. United States, 550 U.S. 192 (2007) (rejecting lenity and applying a broad construction). Further muddying the line between lenity and vagueness (and retroactivity), the Court then rejected a vagueness challenge to the Federal Sentencing Guidelines despite the fact that the “crime of violence” provision at issue was nearly identical to the “violent felony” provision in the ACCA just invalidated in Johnson the year before. See Beckles v. United States, 137 S. Ct. 886 (2017). For further analysis, see Hopwood, supra note 17, at 708–09, 723–24. This stingy approach puts lenity “dead last,” replacing the Court’s once-dominant lenity-first approach.57×57. Price, supra note 17, at 891, 891–99 (pointing to a displacement of lenity in the 1990s) (citing, inter alia, Moskal, 498 U.S. at 108); cf. United States v. R.L.C., 503 U.S. 291 (1992) (using legislative history to trump lenity); United States v. Bass, 404 U.S. 336, 347 (1971) (concluding that lenity applies only after judicial consideration of text, context, and legislative history fails to clarify the text).

Justice Scalia sought to modify both of these approaches with a textualist rule of lenity — one that substantively took lenity back to its origins as conceived by Chief Justice Marshall.58×58. See Solan, supra note 35, at 108–15 (describing Justice Scalia’s emphasis on lenity as a return back to the Marshallian view of it, and a departure from Justice Frankfurter’s narrow conception). Justice Scalia’s version advocates considering lenity as a part of an overarching framework for approaching challenges to criminal statutes (Step Zero), but cutting off the Step One inquiry to exclude all but textual tools of interpretation — including statutory texts, some canons of construction, and the common law — from the determination of whether plausible readings exist. This formulation would counsel applying a narrow construction at Step Two based on a smaller set of plausible interpretations from the lenity-first approach. The idea is that a clearer and more limited set of interpretive tools used at Step One would better guide judges in making lenity determinations. It would spare them from having to simply conclude that a statute could have multiple meanings to trigger application of the lenity rule, on the lenity-first approach (even if far from a defendant’s understanding of the law);59×59. See, e.g., McBoyle v. United States, 283 U.S. 25, 27 (1931). or from concluding that a statute was either clear enough that lenity should not apply or else so hopelessly vague that it should, on the lenity-last approach (also without regard for a defen-dant’s asserted understanding of the law).60×60. E.g., Ocasio v. United States, 136 S. Ct. 1423, 1434 (2016); see also Scalia, supra note 5, at 29; Price, supra note 17, at 890–91. See generally Scalia & Garner, supra note 9. With this modified approach to lenity, Justice Scalia became lenity’s most unexpectedly ardent advocate.61×61. Sarah Newland, Note, The Mercy of Scalia: Statutory Construction and the Rule of Lenity, 29 Harv. C.R.-C.L. L. Rev. 197, 210–11 (1994) (reporting on Justice Scalia’s first eight years on the Court). As had earlier Justices, he argued that the rule was the best means of respecting three important values: legislative supremacy, fair warning (with prosecutorial restraint), and defendants’ rights — all of which bolstered norms of judicial restraint and interpretive consistency in criminal law.62×62. See Solan, supra note 35, at 108–15 (comparing Justice Scalia’s approach to lenity as closer to that of Chief Justice Marshall and in contrast from the more restrictive vision of it promoted by Justice Frankfurter and the majority of the Courts on which he served). Even though Justice Scalia did not always write the opinion (for example, he wrote a lenity majority in only three cases of the dozen or so applications of lenity on the Rehnquist and Roberts Courts respectively), he did frequently join lenity majorities, and he advocated for lenity over legislative history, among other factors, in his concurrences and dissents to opinions with criminal law constructions.

These various tests fall within what I call the “lenity framework”: consideration of possible ambiguity in all criminal law statutory interpretation cases as a separate prerequisite from the inquiry of whether to apply the rule. Like the Chevron two-step test (with its own “Step Zero”),63×63. Cf. Sunstein, supra note 43, at 191 (referring to “the initial inquiry into whether the Chevron framework applies at all”). the lenity framework is a meta-rule meant to direct judges on the appropriate use of lenity in a hierarchy of various interpretive tools.64×64. See Price, supra note 17, at 889. Differences about the place of lenity in that hierarchy, and thus differences in approaches to the test, reflect disputes about the appropriate judicial role in statutory interpretation.

Essentially, the main approaches to lenity — lenity first versus lenity last — map onto the general distinction between presumptions and tiebreakers. The conventional wisdom about substantive legal canons is that judges who use them do so in three core ways: as presumptions, tiebreakers, or clear statement rules.65×65. See William N. Eskridge, Jr., Philip P. Frickey & Elizabeth Garrett, Cases and Materials on Legislation 850–51 (3d ed. 2001). Lenity first functions as a “front-end presumption effectively shaping the interpretive process” when judges use it to adopt a (plausible) narrow construction of a criminal statute.66×66. James J. Brudney, Canon Shortfalls and the Virtues of Political Branch Interpretive Assets, 98 Calif. L. Rev. 1199, 1208 (2010) (citing Muscarello v. United States, 524 U.S. 125, 138–39 (1998), as an example of the Court’s use of lenity as a tiebreaker in the majority opinion, and id. at 148–50 (Ginsburg, J., dissenting), as an example of a dissenting opinion coding lenity as a front-end presumption). Lenity last works more like a tiebreaker when judges invoke it only “at the back end of that process” to adopt a narrow construction only when there is no other means of resolving the meaning of a disputed provision of a criminal statute.67×67. Id. Justice Scalia’s textualist rule of lenity works more like a clear statement rule when, at the slightest hint of ambiguity, he advocates for lenity as a way of requiring Congress to clearly speak to the issue at hand in lieu of declaring criminal liability.68×68. Philip P. Frickey, Interpretive-Regime Change, 38 Loy. L. Rev. 1971, 1989 (2005). Professor Shon Hopwood has recently argued that judges should adopt a clear statement rule of lenity to more broadly address the ills of the criminal justice system. Hopwood, supra note 17, at 697, 709–13. While sympathetic to his argument, this Response advocates a coupling of the Court’s existing regular consideration of the lenity framework with renewed emphasis on its early constitutionally informed lenity-first presumption — the latter of which can be overcome only by a persuasive contrary interpretation or overriding constitutional value. This schema better preserves judicial discretion, which judges jealously guard and for which they typically reject the idea of interpretive precedent, while also providing constitutional guidance for more historically grounded and institutionally sound exercises of that discretion on questions of criminal law today.

The “canon drift” characterizing lenity jurisprudence — with shifts in meaning over time — is a regular feature of substantive canons that often precludes judges from using them in neutral or predictable ways.69×69. See Brudney, supra note 66, at 1209 (noting that “canon drift poses a threat to predictability” and that “more dramatic shifts in canon meaning give rise to additional concerns about neutrality” (emphasis omitted)). This drift matters in the criminal justice arena because the Justices’ disagreement about approaches to lenity reflects and affects their approach to the criminal justice system at a crucial juncture that lacks predict-ability or fairness. Resolution likely requires looking beyond the fact of the drift to interrogate larger questions of the Court’s institutional and constitutional role in this arena.70×70. Frickey, supra note 68, at 1981 (“Any interpretive regime worthy of respect should provide transparency. The basic idea is that, from the cluster of plausible but somewhat inconsistent approaches, the highest appellate court should develop a predictable and settled methodology.”). Accordingly, the next section details the constitutional rationales that Justices variously use to inform the Court’s varied approaches to lenity.

B. Lenity as a Constitutional Rule

Judges regularly appeal to lenity’s constitutional foundations. When the Supreme Court interpreted the first U.S. federal criminal statute, the Crimes Act of 1790,71×71. For analysis of the question raised about this statute, see Price, supra note 17, at 897–98. As Price notes, this question followed the abrogation of federal common law crimes in United States v. Hudson & Goodwin, 11 U.S. (7 Cranch) 32, 34 (1812). But see Dan M. Kahan, Is Chevron Relevant to Federal Criminal Law?, 110 Harv. L. Rev. 469, 471–79 (1996) (arguing that the federal system never really abolished common law crimes and instead permits Congress to delegate criminal lawmaking to courts); Kahan, supra note 35, at 347 (arguing the same). Chief Justice Marshall in United States v. Wiltberger72×72. 18 U.S. (5 Wheat.) 76 (1820). formally incorporated lenity into American law, calling it a rule of “strict construction.”73×73. Id. at 90. Several Courts had contemplated strict construction before, but Wiltberger was the first in an explicitly criminal context to advance rationales that would become central to the application of lenity. See, e.g., United States v. Sheldon, 15 U.S. (2 Wheat.) 119, 121 (1817); Arnold v. United States, 13 U.S. (9 Cranch) 104, 119–20 (1815); United States v. Fisher, 6 U.S. (2 Cranch) 358, 386 (1805). He justified the rule on grounds of constitutional rights (liberty) and structure (legislative supremacy):

The rule that penal laws are to be construed strictly is perhaps not much less old than construction itself. It is founded on the tenderness of the law for the rights of individuals; and on the plain principle that the power of punishment is vested in the legislative, not in the judicial department.74×74. Wiltberger, 18 U.S. at 95 (emphasis added).

To these principles, Justice Holmes added a third constitutional principle a century later: a due process requirement that “a fair warning should be given to the world in language that the common world will understand, of what the law intends to do if a certain line is passed.”75×75. McBoyle v. United States, 283 U.S. 25, 27 (1931). For a discussion of the due process requirements of fair warning, see Trevor W. Morrison, Fair Warning and the Retroactive Judicial Expansion of Federal Criminal Statutes, 74 S. Cal. L. Rev. 455 (2001).

Since, the tendency has been to narrow the ambit of lenity’s constitutional foundations. Courts tend to ignore the first rationale, of defendants’ rights.76×76. Only the Warren Court consistently used the defendants’ rights justification for lenity. See Rabb, supra note 32. And most courts and commentators conceive of lenity as a doctrine lodged only on the constitutional values of legislative supremacy and/or fair warning.77×77. See supra notes 32–42 and 62–74 and accompanying text; see also, e.g., Einer Elhauge, Statutory Default Rules 168 (2008) (contemplating the legislative supremacy functions of the rule of lenity as a “preference-eliciting default rule”); Scalia, supra note 5, at 9–29 (naming textualism and canons like lenity as doctrines of legislative supremacy); Jeffries, supra note 55, at 199–200 (discussing legislative supremacy); Kahan, supra note 35, at 350 (describing how the rule of lenity promotes legislative supremacy through a nondelegation doctrine); Anita S. Krishnakumar, Longstanding Agency Interpretations, 83 Fordham L. Rev. 1823, 1868 n.208 (2015) (describing the rule of lenity as a “due-process based canon”); Paul H. Robinson, Fair Notice and Fair Adjudication: Two Kinds of Legality, 154 U. Pa. L. Rev. 335, 340–41 (2005) (discussing rationales for limitations on common law crimes that often double as justifications for lenity: notice, compliance, legislative supremacy, uniformity, and prosecutorial constraint); Solan, supra note 35, at 108–15; Phillip M. Spector, The Sentencing Rule of Lenity, 33 U. Tol. L. Rev. 511, 534–56 (2002) (naming as “[q]uasi-[c]onstitutional” rationales: separation of powers as principles of nondelegation and legislative supremacy, and due process as principles of due notice and against arbitrariness); Newland, supra note 61, at 201; Note, The New Rule of Lenity, 119 Harv. L. Rev. 2420, 2424 (2006) (explaining how traditional rationales for use of the rule of lenity range from legislative supremacy as an element of separation of powers to fair notice as an element of the rule of law). For scholarly commentary acknowledging lenity’s individual rights foundations, see Kahan, supra note 35, at 345–46. For additional rationales, see Price, supra note 17, at 910–19 (advancing a new justification for a “strong” rule of lenity as a tool for the judiciary to “respon[d] to the defects of the criminal lawmaking process [in a way] that seeks to enhance popular control of criminal justice, rather than remove it,” id. at 915, as a “disclosure requirement” for legislators, id. at 916, and as an “electoral constraint[] on executive officials,” id. at 918). Some appellate judges responding to Professor Gluck and Judge Posner’s survey deemed lenity to be a binding rule justified on the basis of legislative supremacy alone.78×78. Gluck & Posner, supra note 1, at 1331–32 (“A few judges [saw] these canons as deriving their power from . . . the Constitution (for example, that lenity derives from the constitutional concept that federal judges cannot create crimes).”).

The Supreme Court’s turn away from the full ambit of lenity’s original constitutional foundations fueled its transformation from lenity first to lenity last.79×79. See Solan, supra note 35, at 102–15 (describing Justice Frankfurter’s naming and narrowing of the rule from its early application and use). No doubt noting this trend and that the dual rationales typically used to support his textualist rule of lenity failed to hold sway with the majority of the Court, Justice Scalia advocated in United States v. Santos80×80. 553 U.S. 507 (2008). for a return to an originalist conception: applications of lenity founded on structural constitutional requirements of legislative supremacy and fair notice as well as on rights-based constitutional mandates protecting liberty and defendants’ rights.81×81. Id. at 514–15.

It remains to be seen whether the Court will reconsider the full range of the rule’s constitutional foundations. Without reconsideration of its past precedents and its constitutional mandate, a turn seems unlikely, as the Court reverted to lenity last on the dual rationales of legislative supremacy and fair warning as recently as 2015.82×82. See, e.g., Yates v. United States, 135 S. Ct. 1074 (2015); Reynolds v. United States, 565 U.S. 432 (2012). Even Justice Scalia scaled back his full-throated defense of lenity in his majority opinions applying it, for instance, in Burrage v. United States, 134 S. Ct. 881, 891 (2014), although he scaled up in other opinions, including more characteristically full-throated dissents, for instance, in Abramski v. United States, 134 S. Ct. 2259, 2275 (2014) (Scalia, J., dissenting).

By shedding light on past precedent and constitutional mandates for lenity, the aim here is to highlight the imperative for that reconsideration of lenity, with respect to the liberty and defendants’ rights rationale that first animated the rule. That rationale, which supported a lenity-first approach, mandates broader application of the rule. That approach would in turn enable a judicial corrective to institutional imbalances that have resulted in the excessive legislative ambiguity and prosecutorial discretion in federal criminal law that underlie our current landscape of overcriminalization and overpunishment.83×83. See Hopwood, supra note 17, at 724–34. In short, the rule gains authority when grounded in the full range of constitutional principles that first animated it; and the consideration of the framework with the lenity-first approach best accords with early practice and with constitutional mandates to guard not only the values of legislative supremacy and fair warning, but also in the current criminal law climate, the liberty interests and defendants’ rights mandates that are more urgently relevant today.

II. Roberts’s Rules of Lenity

In the Roberts Court’s criminal law statutory construction cases between 2005 and 2017, lenity was potentially at play in 45 cases. The Justices mentioned the rule in almost all of these cases, and they either applied lenity by name or otherwise construed statutes narrowly in 13 cases. They usually justified their decisions on familiar grounds of legislative supremacy and fair notice. But they also revisited the substantive rights justification for lenity and signaled that they viewed the rule as interpretive precedent.

Appeals to the lenity framework were generally not limited by ideology. Almost every single Justice on the Roberts Court has considered the lenity framework when addressing criminal law questions of statutory interpretation.84×84. Justice Gorsuch, who joined the Court in April 2017, is not counted here as he did not take part in any of these decisions. Justice Day O’Connor also did not take part in lenity decisions in her one year on the Roberts Court, but she applied it several times in her twenty-four-year tenure on the Rehnquist and Burger Courts before Chief Justice Roberts became Chief in 2005. For lenity applications in opinions she authored, see the majority opinions in United States v. Kozminski, 487 U.S. 931 (1988), and Tanner v. United States, 483 U.S. 107 (1987), and the dissent in Dixson v. United States, 465 U.S. 482, 501–02, 506 (O’Connor, J., dissenting). For lenity rejections in majority opinions she authored, see United States v. Shabani, 513 U.S. 10 (1994); Beecham v. United States, 511 U.S. 368 (1994); Smith v. United States, 508 U.S. 223 (1993); Regents of University of California v. Public Employment Relations Board, 485 U.S. 589 (1988); United States v. Albertini, 472 U.S. 675 (1985); and McElroy v. United States, 455 U.S. 642 (1982). Most Justices — that is, of the thirteen who have served on the Roberts Court — have been on both sides of the lenity equation.85×85. This fact alone does not mean that ideology or interpretive philosophy plays no role in the propensity for applying or rejecting lenity — analysis of which would require a closer study of each Justice. For current purposes, it is enough to note that most Justices have both applied and rejected the rule.

There is one exception. Justice Anthony Kennedy has tended to reject lenity applications or sign onto only majority opinions that did not name the rule when narrowly construing criminal statutes. He has never authored an opinion applying lenity, and he has instead authored two opinions rejecting it or dissenting from its application.86×86. Maracich v. Spears, 570 U.S. 48 (2013); Negusie v. Holder, 555 U.S. 511 (2009). To be sure, this pattern is not conclusive evidence of a categorical stance against lenity as opposed to being a coincidental byproduct of opinions to which he was assigned. After all, Justice Kennedy has signed onto a few majority opinions applying lenity.87×87. See Burrage, 134 S. Ct. 881 (a unanimous decision authored by Justice Scalia); Reynolds, 565 U.S. 432 (a 7–2 opinion, authored by Justice Breyer); Carachuri-Rosendo v. Holder, 560 U.S. 563 (2010) (a unanimous decision authored by Justice Stevens). Yet those opinions were near-unanimous and, more often than not, he signed onto majority opinions that narrowly construed an ambiguous criminal statute without naming the lenity rule.88×88. Esquivel-Quintana v. Sessions, 137 S. Ct. 1562 (2017) (8–0, Thomas); McDonnell v. United States, 136 S. Ct. 2355 (2016) (9–0, Roberts); Kellogg Brown & Root Servs., Inc. v. United States ex rel. Carter, 135 S. Ct. 1970 (2015) (9–0, Alito); Bond v. United States, 134 S. Ct. 2077 (2014) (9–0, Roberts); Burrage, 134 S. Ct. 881 (9–0, Scalia); Sekhar v. United States, 570 U.S. 729 (2013) (9–0, Scalia); Moncrieffe v. Holder, 569 U.S. 184 (2013) (7–2, Sotomayor); Dorsey v. United States, 567 U.S. 260 (2012) (5–4, Breyer); Bloate v. United States, 559 U.S. 196 (2010) (7–2, Thomas); Abuelhawa v. United States, 556 U.S. 816 (2009) (9–0, Souter); Begay v. United States, 553 U.S. 137 (2008) (5–4, Breyer), abrogated by Johnson v. United States, 135 S. Ct. 2551 (2015). At the very least, this pattern signals that he has not viewed lenity as always key.

Justice Kagan at first seems to follow a similar trajectory, putting into question her acceptance of the lenity framework or the rule. She has authored no majority opinion applying lenity and tends to join majority opinions with narrow constructions of criminal statutes only when they do not mention lenity by name;89×89. McDonnell, 136 S. Ct. 2355, Kellogg Brown & Root Servs., Inc., 135 S. Ct. 1970; Bond, 134 S. Ct. 2077; Burrage, 134 S. Ct. 881; Sekhar, 570 U.S. 729; Moncrieffe, 569 U.S. 184; Dorsey, 567 U.S. 260. A possible exception is Reynolds, but in that case the Court narrowly construed the statute, and lenity only appears in a quotation cited from an earlier case, United States v. Lanier, 520 U.S. 259 (1997). Reynolds, 565 U.S. at 442. she has rejected lenity in majority opinions to which it might have applied;90×90. See Voisine v. United States, 136 S. Ct. 2272 (2016) (6–2); Loughrin v. United States, 134 S. Ct. 2384 (2014) (9–0); Abramski v. United States, 134 S. Ct. 2259 (2014) (5–4) (not by name); Holder v. Martinez Gutierrez, 566 U.S. 583 (2012) (9–0). and she once wrote a dissent specifically to reject the application of lenity to a highly contested case.91×91. See Yates v. United States, 135 S. Ct. 1074, 1090–1101 (2015) (Kagan, J., dissenting). But she dispelled any questions about her support for lenity when, in her dissent to Lockhart v. United States,92×92. 136 S. Ct. 958 (2016). she took up Justice Scalia’s position not long after his death.93×93. Id. at 969 (Kagan, J., dissenting). There, she complained that:

This Court has a rule for how to resolve genuine ambiguity in criminal statutes: in favor of the criminal defendant. As the majority puts the point, the rule of lenity insists that courts side with the defendant “when the ordinary canons of statutory construction have revealed no satisfactory construction.” At the very least, that principle should tip the scales in Lockhart’s favor, because nothing the majority has said shows that the modifying clause . . . unambiguously applies . . . .94×94. Id. at 977 (citations omitted).

In other words, she questioned the majority’s failure to consider the lenity framework even if they were to adopt a lenity-last approach.95×95. See id. at 968 (“Finally, Lockhart asks us to apply the rule of lenity. We have used the lenity principle to resolve ambiguity in favor of the defendant only ‘at the end of the process of construing what Congress has expressed’ when the ordinary canons of statutory construction have revealed no satisfactory construction. . . . But the arguable availability of multiple, divergent principles of statutory construction cannot automatically trigger the rule of lenity. . . . We will not apply the rule of lenity to override a sensible grammatical principle buttressed by the statute’s text and structure.” (citation omitted)). For Justice Kagan, the lenity framework should have applied, and the defendant should have prevailed even on a lenity-last approach.

In the end, Justice Kennedy is the lone outlier in his apparent opposition to lenity. If he is indeed opposed to the rule, his opposition has not affected the Court’s regular consideration of the lenity framework. To the contrary, lenity was considered in almost every case by almost every Justice, typically for one of the traditional constitutional ratio-nales. Taken in the aggregate, the Court has been consistent in considering the lenity framework in its construction of ambiguous criminal statutes, and thus successful at forming another interpretive precedent.

A. The Lenity Framework Applied

The Supreme Court considered the lenity framework in almost all of its criminal law statutory construction cases and applied the lenity rule in a relatively significant number of those cases (though not a majority). Specifically, the Court applied narrow constructions to challenged provisions of criminal statutes in about one-third of the cases (13 of 44) considered by the Roberts Court so far. In half (7 of 13), the Court invoked lenity by name, and typically justified it by appeal to the now-familiar rationales of legislative supremacy and sometimes fair notice.96×96. See McDonnell v. United States, 136 S. Ct. 2355, 2373 (2016); Yates, 135 S. Ct. at 1088; Burrage v. United States, 134 S. Ct. 881, 891 (2014); Reynolds v. United States, 565 U.S. 432, 442 (2012); Skilling v. United States, 561 U.S. 358, 410–11 (2010); Carachuri-Rosendo v. Holder, 560 U.S. 563, 574–75, 581 (2010); United States v. Santos, 553 U.S. 507, 514 (2008) (plurality opinion). In some of these cases, the Court also mentioned constitutional avoidance as a secondary reason. See, e.g., McDonnell, 136 S. Ct. at 2373; Skilling, 561 U.S. at 408–09. In Carachuri-Rosendo and Moncrieffe, the Court applied the rule — by name in the former case and not by name in the latter — as interpretive precedent governing immigration statutes with penal consequences. See Carachuri-Rosendo, 560 U.S. at 581; Moncrieffe v. Holder, 569 U.S. 184, 205 (2013). As discussed below, Justice Scalia also justified the two lenity decisions he authored in terms of defendants’ rights. See infra notes 130–132 and accompanying text. In the other half, the Court applied narrow constructions without mentioning lenity by name. Sometimes it did so on grounds of interpretive precedent — that criminal statutes or “crimmigration” provisions not directly confronting an agency interpretation required narrow construction.97×97. On “crimmigration,” see generally Juliet Stumpf, The Crimmigration Crisis: Immigrants, Crime, and Sovereign Power, 56 Am. U. L. Rev. 367, 376 (2006) (coining the term). For excellent recent scholarship on the field, see, for example, Jennifer M. Chacón, Criminalizing Immigration, in Acad. for Justice, 1 Reforming Criminal Justice: Introduction and Criminalization 205 (Erik Luna ed., 2017). At other times the Court did so on grounds of “plain meaning,” the fair reading arrived at after applying a limited set of interpretive tools, or the need for constitutional avoidance of federalism questions raised with broad applications of a federal criminal statute.98×98. See Esquivel-Quintana v. Sessions, 137 S. Ct. 1562, 1572 (2017) (plain meaning); McDonnell, 136 S. Ct. at 2373 (fair notice, federalism/constitutional avoidance); Kellogg Brown & Root Servs., Inc. v. United States, 135 S. Ct. 1970, 1978 (2015) (interpretive precedent for criminal statutes); Bond v. United States, 134 S. Ct. 2077, 2090 (2014) (fair reading, federalism); Moncrieffe, 569 U.S. at 205 (interpretive precedent for crimmigration statutes); id. at 210–11 (Alito, J., dissenting) (same); Dorsey v. United States, 567 U.S. 260, 281 (2012) (fair reading, police preferences); Bloate v. United States, 559 U.S. 196, 206–07 (2010) (plain meaning); Abuelhawa v. United States, 556 U.S. 816, 820 (2009) (ordinary usage over plain meaning). Notably, federalism concerns show up twice as reasons to apply lenity. Bond famously raised federalism above lenity in the course of arguing for fair reading and applying lenity last, but nevertheless construing the statute narrowly. 134 S. Ct. at 2090. McDonnell followed suit. See 136 S. Ct. at 2373.

On the flip side, the Roberts Court rejected lenity in favor of broad constructions in 31 of its 44 lenity-eligible cases. In the vast majority of these cases (all but 6), the Court asserted that its readings were dictated by the texts’ “plain meaning[s]” or the meanings arrived at after prioritizing an expansive array of interpretive tools — that is, lenity last. Where the Court rejected lenity, its authoring Justices typically did not cite constitutional reasons for doing so.99×99. See, e.g., Shaw v. United States, 137 S. Ct. 462, 469 (2016) (“The statute is clear enough that we need not rely on the rule of lenity.”); Salman v. United States, 137 S. Ct. 420, 429 (2016) (“We also reject Salman’s appeal to the rule of lenity, as he has shown ‘no grievous ambiguity or uncertainty that would trigger the rule’s application.’” (quoting Barber v. Thomas, 560 U.S. 474, 492 (2010))); Lockhart, 136 S. Ct. at 968 (“We have used the lenity principle to resolve ambiguity in favor of the defendant only ‘at the end of the process of construing what Congress has expressed’ when the ordinary canons of statutory construction have revealed no satisfactory construction. That is not the case here.” (quoting Callanan v. United States, 364 U.S. 587, 596 (1961))); Robers v. United States, 134 S. Ct. 1854, 1859 (2014) (“[T]he rule of lenity applies only if, after using the usual tools of statutory construction, we are left with a ‘grievous ambiguity or uncertainty in the statute.’ Having come to the end of our analysis, we are left with no such ambiguity or uncertainty here.” (quoting Muscarello v. United States, 524 U.S. 125, 139 (1998))); Kawashima v. Holder, 565 U.S. 478, 489 (2012) (“It is true that we have, in the past, construed ambiguities in deportation statutes in the alien’s favor. We think the application of the present statute clear enough that resort to the rule of lenity is not warranted.” (citation omitted)); Kasten v. Saint-Gobain Performance Plastics Corp., 563 U.S. 1, 16 (2011) (“But after engaging in traditional methods of statutory interpretation, we cannot find that the statute remains sufficiently ambiguous to warrant application of the rule of lenity here.”); United States v. Hayes, 555 U.S. 415, 429 (2009) (“We apply the rule ‘only when, after consulting traditional canons of statutory construction, we are left with an ambiguous statute.’ Section 921(a)(33)(A)’s definition of ‘misdemeanor crime of domestic violence,’ we acknowledge, is not a model of the careful drafter’s art. But neither is it ‘grievous[ly] ambigu[ous].’ The text, context, purpose, and what little there is of drafting history all point in the same direction: Congress defined ‘misdemeanor crime of domestic violence’ to include an offense ‘committed by’ a person who had a specified domestic relationship with the victim, whether or not the misdemeanor statute itself designates the domestic relationship as an element of the crime.” (citations omitted) (alterations in original)). Instead, they construed the statute’s meaning to be plain — despite concurrences or dissents raising alternative plausible meanings and sometimes arguing for lenity.100×100. See, e.g., Ocasio v. United States, 136 S. Ct. 1423, 1434 (2016) (6–3, Alito); Abramski v. United States, 134 S. Ct. 2259, 2272 n.10 (2014) (5–4, Kagan); Dolan v. United States, 560 U.S. 605, 621 (2010) (5–4, Breyer); James v. United States, 550 U.S. 192, 198 (2007) (5–4, Alito).

The six cases in which the Court rejected lenity for other substantive reasons involved subject-specific areas of immigration or organized crime, two areas to which the Court has concluded categorically that lenity does not apply.101×101. But see Nina L. Mendelson, Change, Creation, and Unpredictability in Statutory Interpretation: Interpretive Canon Use in the Roberts Court’s First Decade, 117 Mich. L. Rev. (forthcoming 2018) (manuscript at 42) (on file with the Harvard Law School Library) (arguing that the Roberts Court seemed to robustly apply lenity to the immigration setting in its first ten years, because “ambiguity in criminal statutes referenced by the [Immigration and Nationality Act] must be construed in the noncitizen’s favor,” id. (quoting Moncrieffe, 133 S. Ct. at 1693) (alterations in original)). In the immigration context, the Court reasoned that the rationales behind Chevron deference outweighed applications of lenity,102×102. Whitman v. United States, 135 S. Ct. 352, 353 (2014) (Scalia, J., concurring with denial of certiorari); Holder v. Martinez Gutierrez, 566 U.S. 583 (2012); Negusie v. Holder, 555 U.S. 511, 518 (2009). reversing its previous stance to the contrary.103×103. For analysis of the conflict between Chevron and lenity for immigration cases in the Supreme Court and in appellate courts just prior to the Roberts Court era, see Brian G. Slocum, The Immigration Rule of Lenity and Chevron Deference, 17 Geo. Immigration L.J. 515, 539–43 (2003). For organized crime, the Court has long determined that the plain meaning of the Racketeer Influenced and Corrupt Organizations Act’s (RICO) unique liberal construction clause excludes lenity.104×104. Boyle v. United States, 556 U.S. 938, 950 (2009); Begay v. United States, 553 U.S. 137, 143 (2008), abrogated by Johnson v. United States, 135 S. Ct. 2551, 2557 (2015). The Racketeer Influenced and Corrupt Organizations Act, 18 U.S.C. §§ 1961–1968 (2012), was passed as a part of the Organized Crime Control Act of 1970, Pub. L. No. 91-452, 84 Stat. 922. The Organized Crime Control Act states that “the provisions of this title shall be liberally construed to effectuate its remedial purposes.” Id. § 904(a). By contrast, many appellate judges have narrowed constructions of that statute in racke-teering cases.105×105. See, e.g., United States v. Turkette, 632 F.2d 896, 906 (1st Cir. 1980), rev’d, 452 U.S. 576 (1981); see also Ellsworth A. Van Graafeiland, RICO and the Rule of Lenity, 9 N. Ill. L. Rev. 331, 332–38 (1989) (collecting cases); Craig W. Palm, Note, RICO and the Liberal Construction Clause, 66 Cornell L. Rev. 167, 169 n.9 (1980). For instances in which other judges liberally construed such statutes, see, for example, United States v. Whitehead, 618 F.2d 523, 525 n.1 (4th Cir. 1980) and cases cited in Palm, supra, at 168 n.8. And at least one appellate judge has advocated for lenity over severity in such cases, on grounds that ambiguities abound in RICO that trigger constitutional commitments to the legislative supremacy and due process requirements undergirding lenity.106×106. Van Graafeiland, supra note 105, at 340–44. The author was a judge on the United States Court of Appeals for the Second Circuit from 1974 until 2004. But this same judge acknowledged that he was bound to follow the Supreme Court’s explicit RICO jurisprudence to the contrary.107×107. Id. at 333. In short, the cases where lenity lost generally fell into two categories wherein the Court explicitly determined that a greater-priority canon, Chevron deference, or an anti-lenity clause of broad construction was in the plain text of the statute. These cases are thus exceptions that in some measure prove the rule.

The extent to which appellate judges diverge from their usual stingy applications of lenity in these same cases suggests that they substantively disagree with the Supreme Court’s hierarchy of canons and decisions about plain meaning. The judge above noted as much even as he felt bound to follow the high court rulings on RICO; and the appellate judge in Sessions v. Esquivel-Quintana108×108. Esquivel-Quintana v. Lynch, 810 F.3d 1019 (6th Cir. 2016), rev’d, 137 S. Ct. 1562 (2017). noted as much too, even though he felt bound to follow the high court’s rulings elevating Chevron above lenity in crimmigration contexts.109×109. See infra Part III.B, pp. 206–09. The divergence further suggests that the precise contours of the lenity framework are not well understood — a lack this Response aims to remedy.

The constitutional rationales behind the Court’s current treatment of lenity are revealing. When applying lenity, the Court often relies on legislative supremacy, sometimes in the guise of plain meaning, as noted above, or a clear statement rule.110×110. E.g., Burrage v. United States, 134 S. Ct. 881, 891 (2014) (“The language Congress enacted requires death to ‘result from’ use of the unlawfully distributed drug, not from a combination of factors to which drug use merely contributed. Congress could have written [the provision] to impose a mandatory minimum when the underlying crime ‘contributes to’ death or serious bodily injury, or adopted a modified causation test tailored to cases involving concurrent causes, as five States have done. It chose instead to use language that imports but-for causality. Especially in the interpretation of a criminal statute subject to the rule of lenity, we cannot give the text a meaning that is different from its ordinary, accepted meaning, and that disfavors the defendant.” (citations omitted)); Skilling v. United States, 561 U.S. 358, 404 (2010) (“[T]o preserve what Congress certainly intended the statute to cover, we pare that body of precedent down to its core.”); id. at 411 (“[W]e resist the Government’s less constrained construction absent Congress’ clear instruction otherwise.”). It less often relies on fair warning.111×111. See, e.g., McDonnell v. United States, 136 S. Ct. 2355, 2373 (2016) (“[U]nder the Government’s interpretation, the term ‘official act’ is not defined ‘with sufficient definiteness that ordinary people can understand what conduct is prohibited,’ or ‘in a manner that does not encourage arbitrary and discriminatory enforcement.’ Under the ‘standardless sweep’ of the Government’s reading, public officials could be subject to prosecution, without fair notice, for the most prosaic interactions. ‘Invoking so shapeless a provision to condemn someone to prison for up to 15 years raises the serious concern that the provision ‘does not comport with the Constitution’s guarantee of due process.’” (citations omitted)); see also Yates v. United States, 135 S. Ct. 1074 (2015); Reynolds v. United States, 565 U.S. 432 (2012). And it hardly ever relies on defendants’ rights, as discussed below. When rejecting lenity, the Court typically announces its results on grounds of plain meaning, as another implicit invocation of legislative supremacy, and it often displays lenity-last reasoning.112×112. See, e.g., Shaw v. United States, 137 S. Ct. 462, 465 (2016). The Court thereby oscillates between an implicit legislative supremacy rationale for the former and an assertion of its discretion to interpret laws for the latter. That the Court appeals to legislative supremacy to both apply and reject lenity is consistent with a longstanding practice of privileging that value as a strong constitutional norm in criminal law.113×113. See William N. Eskridge, Jr., Spinning Legislative Supremacy, 78 GEO. L.J. 319 (1989). But see Kahan, supra note 35, at 348–89. Thus even when the Court has invoked plain meaning but rejected lenity, it was rejecting the rule while still respecting the rationale of legislative supremacy. Put differently, legislative supremacy has been used to both support and undermine lenity applications. That part is logical, because the Court is meaningfully deciding cases and disagreements over a solid constitutional foundation for lenity. The problem with all of this lenity jurisprudence, though, lies in the fact that, when the Court rejects lenity on legislative supremacy grounds alone, it ignores the other key constitutional mandates to protect a defendant’s liberty and other substantive rights.

B. United States v. Santos (2008)

Perhaps realizing the pattern, and that the two sides of lenity were talking past each other in ways at odds with the full range of its constitutional foundations, Justice Scalia sought to clarify the rule in his first prevailing lenity opinion after Chief Justice Roberts’s elevation to the bench. In United States v. Santos, he considered both lenity’s origins and its constitutional rationales.114×114. United States v. Santos, 553 U.S. 507 (2008) (plurality opinion).

Efrain Santos headed an illegal lottery operation in Indiana, and Benedicto Diaz helped process the payments.115×115. Id. at 509. For almost two decades, a group of runners gained commissions from bets they gathered, and portions of the remaining money from those bets were used to pay salaries for Diaz and other collectors, as well as to pay winning gamblers.116×116. Id. Based on these payments and on Diaz’s guilty plea to conspiracy to launder money, a federal district court convicted Santos and Diaz of violating the federal money laundering statute,117×117. Id. at 510 (reporting the results of the jury conviction and sentencing in the trial court). which prohibits the use of the “proceeds” of criminal activities for various purposes, including engaging in, and conspiring to engage in, transactions intended to promote the carrying on of unlawful activity.118×118. See 18 U.S.C. §§ 1956(a)(1)(A)(i), (h) (2012).

The Seventh Circuit affirmed the convictions.119×119. United States v. Santos, 218 F.3d 784, 799 (7th Cir. 2000). On collateral review, the District Court concluded that the defendants could benefit from what was then a new Seventh Circuit ruling — decided after Santos’s direct appeal — that the word “proceeds” in the federal money laundering statute applies only to transactions involving criminal profits, not criminal receipts.120×120. United States v. Santos, 342 F. Supp. 2d 781, 794–99 (N.D. Ind. 2004) (discussing United States v. Scialabba, 282 F.3d 475 (7th Cir. 2002)). Because it found no evidence that the money laundering convictions involved lottery profits, the Court vacated the money laundering convictions.121×121. Id. at 799. The Seventh Circuit again affirmed, this time upholding the reversal of the convictions.122×122. Santos v. United States, 461 F.3d 886, 891 (7th Cir. 2006).

The Government then asked the Supreme Court to review the question. The Solicitor General asked the Court to determine whether the term “proceeds” from the federal money laundering statute, criminalizing transactions designed to promote criminal activity, means only “profits” or also “receipts.”123×123. United States v. Santos, 553 U.S. 507, 509 (2008) (plurality opinion).

In a plurality opinion, the Supreme Court agreed with the Seventh Circuit on a narrow reading of the term.124×124. Id. at 524. If use of the prohibited “proceeds” referred unambiguously to profits and receipts, the convictions would have been upheld.125×125. Id. at 515. But if the term plausibly referred only to profits, the convictions could not be sustained.126×126. Id. In this case, the Court chose the narrower of the two readings: “proceeds,” which would refer only to profits, not receipts, for a prosecution involving a standalone illegal gambling operation.127×127. Id. More generally, the Court ultimately concluded that the statutory term “proceeds” was ambiguous, thus triggering the rule of lenity.128×128. Id. at 514.

Justices on all sides strongly debated the relevant applications and rationales for the lenity rule. At Step One, Justice Scalia noted the “inherent ambiguity” of the word “proceeds,” commenting that Congress had defined the term differently in various criminal provisions, sometimes to mean profits and sometimes to mean receipts.129×129. Id. at 511–12. Also noting that the statute cohered if “proceeds” were read to carry either meaning, he concluded at Step Two that lenity must break the tie.130×130. Id. at 513–14 (“Under either of the word’s ordinary definitions, all provisions of the federal money-laundering statute are coherent; no provisions are redundant; and the statute is not rendered utterly absurd. From the face of the statute, there is no more reason to think that ‘proceeds’ means ‘receipts’ than there is to think that ‘proceeds’ means ‘profits.’ Under a long line of our decisions, the tie must go to the defendant. The rule of lenity requires ambiguous criminal laws to be interpreted in favor of the defendants subjected to them.” (citations omitted)). Significantly, in explaining why, Justice Scalia justified the rule on all three of the rule’s constitutional rationales:

This venerable rule not only vindicates the fundamental principle that no citizen should be held accountable for a violation of a statute whose commands are uncertain [the due process / fair notice rationale], or subjected to punishment that is not clearly prescribed [the liberty / defendants’ rights rationale]. It also places the weight of inertia upon the party that can best induce Congress to speak more clearly and keeps courts from making criminal law in Congress’s stead [the nondelegation / legislative supremacy rationale]. Because the “profits” definition of “proceeds” is always more defendant-friendly than the “receipts” definition, the rule of lenity dictates that it should be adopted.131×131. Id. at 514.

In other words, Justice Scalia not only regarded lenity as a venerable rule sometimes justified by “sheer antiquity,”132×132. Scalia, supra note 5, at 29. Justice Scalia’s language in Santos, “venerable rule,” echoes the “sheer antiquity” rationale he suggested in his book. See 553 U.S. at 514. but he also insisted that constitutional principles of legislative supremacy in criminal lawmaking, due-process fair warning, and substantive rights to life and liberty also require application of the rule. He underscored the interest of defendants’ rights in criminal proceedings this way:

The Government also argues for the “receipts” interpretation because — quite frankly — it is easier to prosecute. . . . Essentially, the Government asks us to resolve the statutory ambiguity in light of Congress’s presumptive intent to facilitate money-laundering prosecutions. That position turns the rule of lenity upside down. We interpret ambiguous criminal statutes in favor of defendants, not prosecutors.133×133. Santos, 553 U.S. at 519.

In a concurrence, Justice Stevens questioned the application of lenity, disagreeing with the textualist reading that Justice Scalia had presented. That is, Justice Stevens would not have stopped the inquiry at consideration of the text and other provisions of criminal statutes to define the term, as did Justice Scalia. He would, and did, also consult legislative history. Finding nothing in the legislative history, Justice Stevens agreed with a narrow construction in this case, but reserved space for rejecting lenity where the legislative history clarified the import of a criminal statute.134×134. Id. at 528 (Stevens, J., concurring) (noting that the plurality’s “conclusion dovetails with what common sense and the rule of lenity would require” given the “lack of legislative history speaking to the definition of ‘proceeds’” and “the perverse result that would obtain” were lenity not applied). In short, he maintained, ambiguous criminal statutes do not always require narrow construction; instead, they amount to a delegation of interpretive authority to the courts to give meaning to the statutes — whether by assessing statutory purpose in the legislative history or through other means.135×135. Id. at 524 (“When Congress fails to define potentially ambiguous statutory terms, it effectively delegates to federal judges the task of filling gaps in a statute.” (citing Comm’r v. Fink, 483 U.S. 89, 104 (1987) (Stevens, J., dissenting)) (“In the process of legislating it is inevitable that Congress will leave open spaces in the law that the courts are implicitly authorized to fill.”)).

Justice Stevens’s statement was controversial and highlighted the dispute between lenity proponents and critics based on the interpretive scope afforded by the constitutional rationale adopted. In one way, Justice Stevens had simply made a robust argument for a lenity-last approach to criminal law. But in another way, his statement was shocking, as his call for giving more leeway to legislative history might be read as explicitly favoring judicial authority, rather than legislative authority, to clearly define criminal law, as Professor Dan Kahan has suggested courts regularly do, implicitly.136×136. See Kahan, supra note 35, at 347. On this understanding of Justice Stevens’s argument, the creation of statutory ambiguity is an attempt by the legislature to impose legislative direction without sufficient clarity — meaning that judges are “implicitly authorized” to specify the statute’s meaning. In response, Justice Scalia attempted to clarify the line between interpretation and law making:

Justice Stevens’ position is original with him; neither the United States nor any amicus suggested it; it has no precedent in our cases. . . . Our obligation to maintain the consistent meaning of words in statutory text does not disappear when the rule of lenity is involved. . . . If anything, the rule of lenity is an additional reason to remain consistent, lest those subject to the criminal law be misled. And even if, as Justice Stevens contends, statutory ambiguity “effectively” licenses us to write a brand-new law, we cannot accept that power in a criminal case, where the law must be written by Congress.137×137. Santos, 553 U.S. at 522–23.

Santos is important for the tripartite constitutional foundations it insists upon for lenity. Namely, it draws upon the original concepts animating lenity in the founding era — defendants’ rights alongside legislative supremacy and fair warning; and it calls on courts to help police the institutional boundaries required to maintain each. This treatment of lenity contrasts sharply with the appellate courts’ treatment of lenity.

III. The Appellate Rule of Lenity

Finally, we come to the courts of appeals’ applications of lenity, where the analysis will be necessarily shorter. Here, I examined all of the appellate cases that gave rise to the Roberts Court lenity cases, to see when, how, and why appellate judges considered the lenity framework, and whether they applied or rejected the rule. This exercise provides a concrete test of the extent to which these courts accord with the Supreme Court’s lenity jurisprudence.

A. The Lenity Framework Denied

Appellate judges considered the lenity framework in a dramatically smaller set of cases than did the Supreme Court. The Supreme Court considered the lenity framework in almost of all lenity-eligible cases (44 of 47 cases, with a vagueness analysis — the “senior” version of lenity — applying to the remaining 3), and applied the rule in 13 cases.138×138. See supra Part II, pp. 195–205. By contrast, appellate courts considered the framework in only some 12 of 53 lenity-eligible cases (22%),139×139. Esquivel-Quintana v. Lynch, 810 F.3d 1019 (6th Cir. 2016); United States v. McDonnell, 792 F.3d 478 (4th Cir. 2015); United States v. Lockhart, 749 F.3d 148 (2d Cir. 2014); United States v. Ocasio, 750 F.3d 399 (4th Cir. 2014); United States v. Yates, 733 F.3d 1059 (11th Cir. 2013); United States v. Armstrong, 706 F.3d 1 (1st Cir. 2013) (consolidated with United States v. Voisine, 495 F. App’x 101 (1st Cir. 2013), aff’d 136 S. Ct. 2272 (2016)); United States v. Castleman, 695 F.3d 582 (6th Cir. 2012); see also United States v. Voisine, 778 F.3d 176 (1st Cir. 2015); Carachuri-Rosendo v. Holder, 570 F.3d 263 (5th Cir. 2009); United States v. Hayes, 482 F.3d 749 (4th Cir. 2007); United States v. Begay, 470 F.3d 964 (10th Cir. 2006); United States v. Rodriguez, 464 F.3d 1072 (9th Cir. 2006); United States v. Santos, 461 F.3d 886 (7th Cir. 2006). and they applied it in only 4 (7%).140×140. Castleman, 695 F.3d 582; Hayes, 482 F.3d 749; Rodriguez, 464 F.3d 1072; Santos, 461 F.3d 886. To be sure, determining whether this sample is representative of the appellate docket more generally would require a larger survey. Perhaps one might expect, on one hand, that the cases granted certiorari were hard cases, in which the Supreme Court might have expected a higher rate of consideration of the lenity framework and application of the rule and thus sought to correct the appellate courts where the Justices did not find either. On the other hand, it could be that the cases were not that hard and that the Supreme Court applied the lenity framework to show how they could be resolved, applying the rule to correct overcriminalization through judicial interpretation. In both scenarios — whether as a corrective measure or as guidance — the Court demonstrated the more general point that the lenity framework is appropriate for statutory questions of criminal law, helping to shore up the extent to which it treats or models approaches to the rule as interpretive precedent.

Astonishingly, the reasons behind appellate court applications of lenity accorded with none of the rationales advanced at the Supreme Court. One panel applied lenity, in Santos v. United States, based on interpret-ive precedent in its own Circuit (which the Supreme Court affirmed for reasons that instead combined the three constitutional rationales above).141×141. Santos, 461 F.3d 886, aff’d, 553 U.S. 507. Justice Scalia authored the plurality opinion. The other panel that applied the rule in United States v. Hayes treated lenity like a residual savings clause, perhaps hedging against the Supreme Court finding its construction of the statute flawed (it did, and reversed, rejecting lenity).142×142. Hayes, 482 F.3d at 749, rev’d, 555 U.S. 415. Justice Ginsburg authored the 7–2 opinion.

B. Sessions v. Esquivel-Quintana (6th Cir. 2016)

Interestingly, though, where appellate courts explicitly rejected lenity, they often did so citing recent Supreme Court precedent on the rule, or at least their understandings of it. Sessions v. Esquivel-Quintana is illustrative. In that case, the United States Court of Appeals for the Sixth Circuit was tasked with deciding whether a conviction under a state statute criminalizing consensual sexual intercourse between a twenty-one-year-old and a seventeen-year-old qualifies as an “aggravated felony” under the Immigration and Nationality Act (INA), which permits the Attorney General to deport ‘‘[a]ny alien who is convicted of an aggravated felony.”143×143. 8 U.S.C. § 1227(a)(2)(A)(iii) (2012). One of the many crimes that constitutes an aggravated felony under the INA is ‘‘sexual abuse of a minor,’’ which the statute does not expressly define. Id. § 1101(a)(43)(A). The twenty-one-year-old in question was Mexican native and citizen Mr. Esquivel-Quintana, who sought review of a Board of Immigration Appeals decision ordering his removal on a finding that his California conviction for unlawful sexual intercourse with a minor was an aggravated felony under the INA. The Sixth Circuit concluded that it could not overturn the removal order.144×144. Esquivel-Quintana, 810 F.3d at 1021.

Even though the Supreme Court ultimately disagreed with it, the Sixth Circuit’s reasoning in Esquivel-Quintana is important for the extent to which it treated lenity as interpretive precedent, and to which the judicial panel sought to take guidance from the Supreme Court. When Esquivel-Quintana raised the issue of lenity, arguing for a narrow application of “sexual abuse of a minor” in his favor, the Circuit Court first acknowledged the general applicability of the lenity framework, in line with Supreme Court lenity jurisprudence: “According to the rule of lenity, when a criminal statute is ambiguous, that ambiguity must be resolved in the defendant’s favor.”145×145. Id. at 1023. But the Circuit Court doubted whether lenity applies in contexts where, as there, a civil statute carries criminal penalties. Somewhat at odds with regular Supreme Court pre-cedent on this question, which yields a pretty unequivocal “yes,”146×146. In addition to the cases cited below, see Moncrieffe v. Holder, 569 U.S. 184, 205 (2013) (applying narrow construction) (“[W]e err on the side of underinclusiveness because ambiguity in criminal statutes referenced by the INA must be construed in the noncitizen’s favor.”). See also United States v. Thompson/Ctr. Arms Co., 504 U.S. 505, 517–18 (1992) (“The key to resolving the ambiguity lies in recognizing that although it is a tax statute that we construe now in a civil setting, the NFA has criminal applications that carry no additional requirement of willfulness.”); Rewis v. United States, 401 U.S. 808, 812 (1971) (Travel Act); Toussie v. United States, 397 U.S. 112, 122 (1970) (federal draft law); United States v. Laub, 385 U.S. 475, 480 (1967) (INA and the Passport Act of 1926); United States v. Bos. & Me. R.R., 380 U.S. 157, 160 (1965) (Clayton Act on antitrust); United States v. Mersky, 361 U.S. 431, 440 (1960) (Tariff Act of 1930 and Criminal Appeals Act); Comm’r v. Acker, 361 U.S. 87, 91–92 (1959) (Internal Revenue Code and Treasury Regulation); Bonetti v. Rogers, 356 U.S. 691, 699 (1958). For an appellate construction that places lenity ahead of Chevron, see, for instance, WEC Carolina Energy Sols. LLC v. Miller, 687 F.3d 199, 204 (4th Cir. 2012) (“Where . . . our analysis involves a statute whose provisions have both civil and criminal application, our task merits special attention because our interpretation applies uniformly in both contexts. Thus, we follow ‘the canon of strict construction of criminal statutes, or rule of lenity.’” (quoting United States v. Lanier, 520 U.S. 259, 266 (1997)). the Circuit Court looked to a recent denial of certiorari memorandum and to its own precedent to arrive at an equivocal view that it did not apply. The Circuit Court suggested that only “[a]n increasingly emergent view asserts that the rule of lenity ought to apply in civil cases involving statutes that have both civil and criminal applications.”147×147. Esquivel-Quintana, 810 F.3d at 1023 (citing Whitman v. United States, 135 S. Ct. 352, 352–54 (Scalia, J., concurring with denial of certiorari); see also Carachuri-Rosendo v. Holder, 560 U.S. 563, 581 (2010); Leocal v. Ashcroft, 543 U.S. 1, 11 n.8 (2004); Carter v. Welles-Bowen Realty, Inc., 736 F.3d 722, 729–36 (6th Cir. 2013) (Sutton, J., concurring) (arguing that the rule of lenity should be applied before Chevron deference in cases where agency interpretations implicate criminal penalties)). The reason, the Circuit Court concluded, is that lenity meets the fair notice and legislative supremacy requirements that must apply when there are penal consequences for violation of a law.148×148. Esquivel-Quintana, 810 F.3d at 1023 (“[A]mbiguous statutes must be construed in favor of defendants under the rule of lenity. The rule of lenity ensures that the public has adequate notice of what conduct is criminalized, and preserves the separation of powers by ensuring that legislatures, not executive officers, define crimes. Taken together, these two principles lead to the conclusion that the rule of lenity should apply in civil cases involving ambiguous statutes with criminal applications.”). It did not consider the defendants’ rights/liberty rationales. It further mounted a spirited offense against privileging Chevron deference over lenity — arguing that the twin fair notice and nondelegation constitutional rationales for lenity are meant to prevent both courts from criminalizing otherwise permissible acts and agencies from creating “new crimes at will.”149×149. Id. at 1023–24 (“Giving deference to agency interpretations of ambiguous laws with criminal applications would allow agencies to ‘create (and uncreate) new crimes at will, so long as they do not roam beyond ambiguities that the laws contain.’ Writing criminal laws is the legislature’s prerogative, not the executive’s. Furthermore, deferring to agency interpretations of criminal laws violates the principle that ‘criminal laws are for courts, not for the Government, to construe.’ Left unchecked, deference to agency interpretations of laws with criminal applications threatens a complete undermining of the Constitution’s separation of powers.” (quoting, respectively, Whitman, 135 S. Ct. at 353 (Scalia, J., concurring with denial of certiorari), and Abramski v. United States, 134 S. Ct. 2259, 2274 (2014)).

But in the end, the Circuit Court concluded that its hands were tied. Noting accurately that the Supreme Court’s recent lenity decisions in the crimmigration context placed Chevron above lenity when the two canons faced off in a head-on duel,150×150. But see Moncrieffe, 569 U.S. at 205 (applying a narrow construction in favor of a noncitizen who was convicted of an “aggravated felony” and deemed deportable for marijuana possession, and noting that “we err on the side of underinclusiveness because ambiguity in criminal statutes referenced by the INA must be construed in the noncitizen’s favor”) (citations omitted). the Circuit Court followed suit reluctantly and rejected lenity in this case.151×151. Esquivel-Quintana, 810 F.3d at 1024 (“[W]hile this view is increasing in prominence, the Supreme Court has not made it the law. To the contrary, the Court has reached the opposite conclusion.”). Moreover, it presented itself as bound to doing so as part of a lenity jurisprudence that extended to the administrative context more generally — dating back at least to Babbitt v. Sweet Home Chapter of Communities for a Great Oregon,152×152. 515 U.S. 687 (1995). the 1995 case construing the Endangered Species Act.153×153. Esquivel-Quintana, 810 F.3d at 1024 (citing Babbitt, 515 U.S. at 704 n.18). In Babbitt, the Court rejected the rule of lenity in these terms: “We have never suggested that the rule of lenity should provide the standard for reviewing facial challenges to administrative regulations whenever the governing statute authorizes criminal enforcement.” 515 U.S. at 704 n.18. The Sixth Circuit in Esquivel-Quintana further suggested that the Supreme Court’s indications that the rule of lenity should apply came only in dicta. 810 F.3d at 1024 (citing Leocal v. Ashcroft, 543 U.S. 1, 8–13 (2004); Kasten v. Saint-Gobain Performance Plastics Corp., 563 U.S. 1, 1 (2011); Maracich v. Spears, 133 S. Ct. 2191, 2209 (2013)). The Circuit Court also adopted the faithful agent metaphor, except that rather than envisioning a congressional principal as is typical, it named a Supreme Court principal: “As an ‘inferior’ court, our job is to adhere faithfully to the Supreme Court’s precedents.”154×154. Esquivel-Quintana, 810 F.3d at 1024. The Circuit Court arguably got its appraisal of the practice wrong. But what is notable is that this appellate court regarded and attempted to treat the Supreme Court consideration of lenity as interpretive precedent.


Professor Gluck and Judge Posner’s focus on the federal courts of appeals has appropriately shifted our sights onto statutory interpretation in the arena where it happens most. Their work provides insight into the larger questions of statutory interpretation, and digging even deeper into the way judges approach particular canons is an important next step to understanding fully the interpretive work performed by appellate courts.

Yet, in considering the relationship between appellate courts and the Supreme Court, lenity does not fare well. The Court’s fairly consistent application of the lenity framework for ambiguous criminal statutes arguably mimics its consistent application of Chevron’s framework for agency deference for ambiguous regulatory statutes. The problem is that the Justices do not always agree on the point at which the lenity rule should trigger narrow construction, nor do they always rely on the same constitutional rationales behind lenity when considering its early application (lenity first) against countervailing canons or other constitutional considerations (lenity last). Despite the consistent application of the lenity framework, most commentators and appellate court judges have not treated it as interpretive precedent. To wit: in the appellate courts surveyed whose criminal law opinions were granted certiorari, appellate judges rarely invoked the framework and applied the rule even less.

These features of current lenity jurisprudence reflect an approach to the rule that is at odds with the historical practice and constitutional mandates for the rule. Close examination of each, as provided here, reveals an imperative for regular consideration of the framework informed by the full range of constitutional mandates driving the rule: legislative supremacy, fair notice, and liberty interests by which the court ensures the Constitution’s structural integrity and individual rights. Exercise of that judicial role has never been more important than now, when the tenor of discriminatory and pathological politics has combined with a lack of legislative clarity and prosecutorial restraint to result in a broken criminal justice system in dire need of interpretive attention, among other checks and balances.

Similarly, the appellate courts’ grudging approach to lenity is at odds with that of the Supreme Court. Appellate courts’ frequent failure to even consider the lenity framework confirms the notion gestured at in Professor Gluck and Judge Posner’s Article: that appellate courts often deviate significantly from the Supreme Court in terms of methodological approaches to statutory interpretation. If close scrutiny of the Roberts Court criminal law statutory interpretation cases is a guide, the Supreme Court’s consistent consideration of the lenity framework suggests that some modifications are in order to the common notion of lenity, now prevalent among commentators and represented on appellate courts, as inconsistent and weakly anchored. These modifications would commend increased consideration of the lenity framework in criminal law statutory interpretation cases on appellate courts to better match Supreme Court treatment. Fuller consideration of the constitutional mandates for lenity would also command increased application of the lenity-first approach to the rule. Taken together, this conception of lenity as an interpretive precedent that takes important criminal law questions seriously would give full effect to the constitutional mandates directing judges to help ensure legislative supremacy, fair notice, and liberty in the criminal process, particularly in these times.

A modified rule of lenity that asks judges always to consider the lenity framework for criminal law statutory interpretation challenges and to use their discretion in applying the rule is not only desirable, but it is arguably required if courts are to adhere to lenity’s historical and constitutional foundations. It would better accord with Supreme Court signals and appellate-court sometimes-recognition of the “special” status of lenity in criminal law interpretation without imposing undue constraints on the judicial task of deciding questions of law at the appellate level. And it would guide judges to respond appropriately to the increasingly urgent need for more cautionary approaches to criminal law statutory interpretation in view of the collapse of the American criminal justice system.

* Professor of Law, Harvard Law School; Professor of History, Harvard University; Susan S. & Kenneth L. Wallach Professor, Radcliffe Institute for Advanced Study at Harvard. This paper benefited from helpful comments and feedback from Ryan Doerfler, Abbe Gluck, Joshua Kleinfeld, Nina Mendelson, Zachary Price, and Jane Schacter. Michael Decker, Marilyn Robb, and Samuel Rubinstein provided excellent research assistance. The author thanks all of them as well as the editors of the Harvard Law Review for their scholarly enthusiasm, as-tute comments, and careful editing.

Recommended Citation: Intisar A. Rabb, Response, The Appellate Rule of Lenity, 131 Harv. L. Rev. F. 179 (2018), https://harvardlawreview.org/2018/06/the-appellate-rule-of-lenity/.