Interpreting Code – Harvard Journal on Legislation

Daniel B. Listwa [*] & Adam Flaherty [**] View PDF

Abstract

In recent years, a scholarly movement has arisen focused on developing a more detailed understanding of the institutions and actors that shape the creation and propagation of statutory law. Part of that project has involved taking a closer look at the U.S. Code and drawing out from the shadows the people—the codifiers—who have a hand in creating it. Contrary to expectations, the typical role of the federal codifiers is far from ministerial. In carrying out their congressionally entrusted task of “revising” and “restating” the law, the codifiers substantially alter the statutory text: shifting placements, inserting headings, and even drafting new provisions altogether. Accordingly, this scholarship has argued, the U.S. Code—even those titles that have been enacted into positive law—must be construed in a way that is responsive to the codifiers’ role, acknowledging that certain editorial decisions cannot meaningfully (or, at least, simply) be ascribed to Congress. In sum, statutory interpretation demands a “codifier’s canon.”

And yet, despite academic advances, it appears that these insights have largely yet to penetrate the judiciary. Perhaps moving cautiously in the face of objections to the broader “process-based turn” in statutory interpretation, courts continue to overlook the significance of the codification process for accurately drawing meaning from statutory text. In this Essay, we take on this oversight, focusing on two cases from the Supreme Court’s most recent term—Dubin v. United States and Pugin v. Garland—to illustrate how an understanding of the process by which the U.S. Code is assembled is essential to faithful textualism. In particular, we show that while overlooking codification is occasionally innocuous (Dubin), in other cases (Pugin), a failure to engage meaningfully can undermine the entire edifice of an opinion’s legal reasoning. We conclude by discussing the broader implications of the Court’s failure to develop a robust framework for addressing editorial changes introduced through codification—including for difficult cases involving the future of qualified immunity that are already being raised in the lower courts.

I. Introduction

Walk into any federal judge’s chambers and you will be sure to find at least one shelf-lined wall filled with the volumes of the United States Code. Totaling over 40,000 pages in length and divided into fifty-four titles,¹ the Code is the official compendium of federal statutory law—the primary authority by which judges and lawyers access the law as it has been enacted by Congress. The Code has been a central tool for legal practitioners for nearly a century, the prism through which we have come to view our modern “republic of statutes.”² And yet, throughout that time, little thought has been given to how that prism has come to be—how the multitude of bills passed by Congress and signed by the President find their way into the Code’s bound volumes. Except in the extraordinary case in which a provision is left out of the Code (United States National Bank of Oregon v. Independent Insurance Agents of America, Inc.³U.S. Nat’l Bank of Or. v. Indep. Ins. Agents of Am., Inc., 508 U.S. 439, 448 (1993).) or new language added (Maine v. Thiboutot⁴Maine v. Thiboutot, 448 U.S. 1, 7–9 (1980).), one is hard-pressed to find acknowledgement that the Code is anything other than the law as Congress “intended.”⁵

But that has begun to change in recent years as a scholarly movement has arisen focused on developing a more nuanced understanding of the institutions that shape the creation and propagation of statutory law. This movement has, for example, revealed details about the practices of those actually tasked with drafting legislation and the congressional procedures that structure the way in which laws are made.⁶ Termed the “process-based turn” in legislative studies by now-Justice Amy Coney Barrett,⁷ this movement has broken past the “Schoolhouse Rock!” model of legislation to detail a more accurate picture of Article I’s institutions, uncovering lessons for statutory interpreters in so doing.

One portion of this broader project has involved a closer look at the codification process and the significant role that unelected officials have in shaping the law as it ultimately appears in the U.S. Code. Contrary to what one might expect, the typical role of the federal codifiers is far from ministerial. The codifiers have been statutorily entrusted with drafting “a complete compilation, restatement, and revision of the general and permanent laws of the United States,”⁸ a task that has, in practice, involved grouping and subdividing laws based on their subject matter, inserting headings, and even drafting new provisions altogether.⁹ As a result, the U.S. Code—even those titles that have been enacted into positive law—must be construed in a way that is responsive to the codifiers’ role, acknowledging that certain editorial decisions cannot meaningfully (or, at least, simply) be ascribed to Congress.

Despite the scholarly advances made, it appears that these insights have largely yet to penetrate the judiciary. Perhaps moving cautiously in the face of objections to this “new empiricism” as non-textualists,¹⁰ courts continue to overlook the significance of the codification process for accurately drawing meaning from statutory text.¹¹ This, we think, is a mistake. Cases frequently arise in which the details of codification are necessary to engage in faithful textualism, yet courts and litigants alike do not seem to notice. In this Essay, we take on this oversight, engaging in an effort to translate recent revelations regarding codification into more practical guidance for resolving difficult cases of statutory interpretation. To do so, we focus on two cases from the Supreme Court’s October Term 2022—Dubin v. United States¹²599 U.S. 110 (2023). and Pugin v. Garland¹³599 U.S. 600 (2023).—both of which involved codification-related issues that went unacknowledged by the Court.

In particular, these two cases, decided only two weeks apart, directly implicated a much-overlooked rule of construction found in Title 18—mirroring similar rules found in eleven other titles—that seems to prohibit relying on a statutory provision’s placement or caption in the Code when interpreting the statute. However, as one of us has argued, when this instruction is placed in the larger codification context, its meaning transforms. It is not a broad limitation on relying on various (important) markers of textual meaning, but rather a tailored direction—a codifier’s canon—intended to warn interpreters not to rely on particular editorial decisions made by the codifiers, as opposed to Congress.¹⁴ Applying this insight to the Court’s analyses in Dubin and Pugin, we illustrate how the Court overlooked this canon, relying heavily on captions and placements without grappling with the role played by the codifiers. As we explain, this matters not just for the resolution of the particular issues involved in those cases. By engaging directly with how interpreters should consider editorial decisions introduced by the codifiers, the Court—and the judiciary more generally—can help develop a more sophisticated form of textualism, building the tools that will shape modes of statutory interpretation in the future.

To address these arguments, this Essay proceeds in four Parts. Part II begins by discussing the idea of a codifier’s canon and contextualizing it within broader developments in how scholars have come to understand the process by which Congress’s enactments get assembled into a cohesive statutory text. Parts III and IV, respectively, then discuss the relevance of these insights to two particular cases. Part III examines a comparatively easy case, Dubin, in which the codifier’s canon helps make clear why it was appropriate for the Court to rely on the caption of 18 U.S.C. § 1028A(a)(1), “aggravated identity theft,” to narrow the provision’s seemingly broad reach. The analysis of Dubin is relatively straightforward on our approach because it involved an editorial decision that Congress made when later amending a title that it had already enacted into positive law. Part IV takes on a harder case, Pugin, which we use to illustrate how an understanding of the codification process calls into question the manner in which all of the opinions in that case invoked the set of offenses codified within Chapter 73 of Title 18, entitled “Obstruction of Justice,” to construe the scope of that same phrase as it appears in an immigration-related statute. Finally, we conclude in Part V by discussing the broader implications of the Court’s failure to develop a robust framework for addressing editorial changes introduced through codification—including for difficult cases involving the future of qualified immunity that are coming down the pike. Specifically, we consider the challenge posed by language apparently intended to abrogate common law immunity defenses that was originally included in section 1983 but inadvertently dropped during the codification process. Any account of the continuing relevance of such language requires, we argue, a deeper understanding of the relationship between the codification process and statutory interpretation.

II. Codifier’s Canon Uncovered

To understand the codifier’s canon, it helps to begin with the Supreme Court’s 2015 decision in Yates v. United States,¹⁵ a case that has already become a “legislation-course chestnut.”¹⁶ Yates presented the rather memorable question of whether a fish is a “tangible object” for purposes of the evidence destruction provision, 18 U.S.C. § 1519, of the Sarbanes-Oxley Act. A majority of the Court answered the question in the negative, with Justice Ginsburg writing a plurality opinion that emphasized section 1519’s placement and captioning in Title 18 of the U.S. Code, noting that the provision’s heading refers specifically to “records” and that the provision appeared alongside rules relating to specific kinds of—decidedly un-fishlike—evidence.¹⁷

The Yates decision garnered a great deal of attention in the months after it came down, not the least of which because of Justice Kagan’s citation to Dr. Seuss’s classic One Fish Two Fish Red Fish Blue Fish in her dissent.¹⁸ But an article by Tobias Dorsey zeroed in on an issue that other commentators had missed: it appeared that the Justices had overlooked, and indeed violated, a specific instruction from Congress barring courts from considering the headings and placement of provisions in Title 18 of the Code.¹⁹ As Dorsey noted, when Congress enacted the first federal code, the re, it included a section that read as follows:

The arrangement and classification of the several sections of the revision have been made for the purpose of a more convenient and orderly arrangement of the same, and therefore no inference or presumption of a legislative construction is to be drawn by reason of the Title, under which any particular section is placed.²⁰

Later codification efforts have included similar provisions. Most pertinently, when Congress enacted Title 18 in 1948, it included a provision—section 19 of the 1948 Act—which stated: “No inference of a legislative construction is to be drawn by reason of the chapter in Title 18, Crimes and Criminal Procedure . . . in which any particular section is placed, nor by reason of the catchlines used in such title.”²¹

This language, seemingly barring reliance on headings (that is, “catchlines”²²) and placement in the Code, is hard to locate, but you can still find it in the text of Title 18 today buried in small print in the title’s front matter.²³ And, in Dorsey’s view, Justice Ginsburg violated this congressional instruction by expressly relying on 18 U.S.C. § 1519’s caption and placement in the Code to construe the scope of the phrase “tangible object.”²⁴

Responding to Dorsey’s piece, one of us offered a different view, arguing that Title 18’s headings and placement rule—and the many similar rules found in other titles of the U.S. Code—must be understood in light of the codification process. Specifically, one of us explained that these instructions “ought not to be read as broad rejections of citing structural placement or caption,” but rather “should be understood as signaling to courts that placement and caption choices within the Code should be respected and considered when they originate in the decisions of Congress, but not when those choices are the result of intervention by the office that codifies the United States Code.”²⁵ In other words, the prohibition is a specific warning about relying on post-enactment editorial decisions made by the codifiers—hence, the “codifier’s canon.”

But why would such a warning be necessary? An overview of the codification process makes this clear. When a bill is first enacted, it is assigned a Public Law number based on the session of Congress and when it was enacted within that session.²⁶ At the end of each session of Congress, the Government Publishing Office compiles all the laws enacted within a session into a single volume known as the Statutes at Large. The Statutes at Large are “a chronological compilation of the laws exactly as they were enacted by Congress, with the same organization and content as the bill approved by Congress.”²⁷ But what the Statutes at Large possess in fidelity, they lack in readability. Because the Statutes at Large simply compile the law as it is enacted, figuring out the current form of any given federal law would require reading every volume of the Statutes at Large and piecing together original provisions with, in some cases, decades of amendments.²⁸

To address this problem, Congress has authorized the creation of the U.S. Code, a formal codification of federal statutory law. Under the present regime, the work of creating the Code falls within the ambit of the Office of the Law Revision Counsel (“OLRC”), a nonpartisan office within the U.S. House of Representatives created by Congress in 1974.²⁹ The key role of the OLRC is to create the individual subject matter titles that form the U.S. Code. This work is done essentially on a title-by-title basis, as the OLRC surveys and synthesizes the vast body of enacted federal statutes and organizes them into a coherent and accessible whole.³⁰ At first, these codified titles are referred to as “nonpositive law” titles, meaning they are not themselves the law as enacted by Congress.³¹ However, part of the OLRC’s commission is to propose draft bills that enact these codifications into law—repealing the underlying Statutes at Large in the process—and thus turning the compiled title into “positive law.” It is a slow but steady process. Today, twenty-seven of the fifty-four titles of the U.S. Code have been enacted into positive law, though those titles comprise only about twenty-five percent of the U.S. Code’s pages.³²

In preparing these proposed codifications, the OLRC is tasked with making editorial changes that clarify the law while preserving congressional intent. As the OLRC’s organic statute instructs, the OLRC should “remove ambiguities, contradictions, and other imperfections both of substance and of form,” but leave unchanged “the understood policy, intent, and purpose of the Congress in the original enactments.”³³ This is a tall order. Even the most punctilious codifier is bound to make an occasional mistake. And given how a small editorial choice can have major implications in individual cases, it is inevitable that the OLRC’s codification efforts will unintentionally alter the meaning of a given law.

This is not a new concern. The history of official codification in this country goes back at least 150 years, when, in 1866, Congress authorized President Andrew Jackson to appoint three “commissioners[] to revise, simplify, arrange, and consolidate all statutes of the United States, general and permanent in their nature.”³⁴ And the resulting Revised Statutes, enacted in 1874, were not without faults. In addition to leaving out some laws altogether,³⁵ the Revised Statutes inadvertently introduced various substantive changes—including, most remarkably, the addition of the words “and laws” to section 1983, which led the Court to extend the cause of action from constitutional rights to statutory ones in Maine v. Thiboutot.³⁶ Subsequent codification efforts, including a 1926 attempt that left out several hundred provisions of permanent law, did not fare much better.³⁷

Title 18 was enacted into positive law against the backdrop of this comedy of errors. Accordingly, it is not a surprise that Congress appended to the newly codified law a note of caution (one not unlike those included in previous codifications): do not draw inferences of legislative intent from the placement and captioning decisions made by the codifiers.³⁸ But this instruction should not be overread, as a broader understanding of the codification process makes clear. Once a title is enacted into positive law, the role of the codifiers diminishes substantially. Typically, when amending a provision in a positive law title, Congress will include specific instructions regarding codification, “directing where the amendment should be placed and how it should be captioned.”³⁹ The codifiers’ role is ministerial, carrying out the decisions made by Congress.⁴⁰

This distinction between codification decisions pre- and post-enactment into positive law has direct implications for understanding the scope of Congress’s instruction not to rely on placement and headings in the Code. Although editorial decisions made by the codifiers during codification cannot be relied on, there is no issue when placement or heading is instead the result of Congress’s direct amendment of the statute after that statute has been enacted into positive law.⁴¹

Accordingly, the general procedure for deciding whether interpretive reliance on caption and placement of a provision in the United States Code is appropriate is as follows. First, the court must ask whether the title in question has been enacted into positive law. If not, then any caption and placement decisions that differ from those reflected in the Statutes at Large cannot be relied upon—those editorial changes are not part of the law. If, however, the title has been enacted into positive law, the court must then ask whether the caption or placement originated with the codifiers during the process of enacting the title into positive law or with Congress in a subsequent amendment to an existing positive law title. Reliance on the former sort of editorial decisions is generally proscribed⁴² while reliance on the latter sort is usually permissible, subject to further complexities we explore in Parts IV and V below.

Applying this insight to Yates, there was (contrary to Dorsey’s suggestion) no violation of the codifier’s canon. The relevant provision, 18 U.S.C. § 1512, was added to Title 18 in 2002, long after positive law codification of the title in 1948, in a statute that specified each new provision’s placement and heading in the Code.⁴³ The case thus was not an instance of the Court accidentally relying on editorial changes introduced by the unelected codifiers to infer the statute’s meaning.⁴⁴

In recent years, subsequent scholarship has further developed and expanded upon the core insight of the codifier’s canon—that is, the idea that the codification process should inform interpretation.⁴⁵ Professor Jarrod Shobe, for example, has provided a much more comprehensive account of the role of the OLRC, describing, among other things, the sorts of editorial changes frequently introduced by the OLRC that go well beyond placement and headings.⁴⁶ This includes, inter alia, modifying internal references in enacted statutes (such as “this chapter” or “this Act”) to make sense in the context of the Code’s organization and terminology.⁴⁷ Because the OLRC typically changes these references to refer to organizational units of the Code, rather than the bill Congress enacted, there is a risk of the modified reference being either over- or underinclusive.⁴⁸ Shobe also described how the codification process obscures large portions of enacted text, either by relegating it to the notes section of the Code or leaving it out entirely.⁴⁹ Professor Shawn Nevers and Lecturer Julie Graves Krishnaswami have developed this point further, discussing how the Code contains tens of thousands of notes, with some containing entire statutes.⁵⁰

In an important article contextualizing the OLRC in the broader “congressional bureaucracy” of nonpartisan legislative offices, Professor Jesse Cross and Professor Abbe Gluck uncovered additional details regarding the sorts of editorial interventions made by the OLRC.⁵¹ These included the insertion of what OLRC staffers refer to as “no source” provisions, provisions that did not exist in any congressionally enacted law but rather are created by the OLRC “out of whole cloth” and added to positive law titles—typically defining new terms that are then used throughout the title, but sometimes making explicit obligations or entitlements that the codifiers believed were inferred.⁵² As Cross and Gluck noted, although these substantive changes are formally enacted by Congress, they originate in a manner that is far removed from the “core legislative process.”⁵³ Indeed, legislative history of recent codification bills as well as interviews with congressional staffers suggest that members of Congress are not even aware that the OLRC is engaged in such muscular editing.⁵⁴ More broadly, reflecting on the fact that little legislative attention is paid to the technical work of the OLRC in carrying out codification, Cross and Gluck endorse the codifier’s canon—at least as applied to structure and placement within the Code—as an appropriate “anti-deference” canon.⁵⁵

In his most recent article, Cross pushes these concepts further, describing how the codification process calls into question the very notion that there is a given “statutory text.”⁵⁶ As Cross notes, “the Supreme Court is fond of declaring that it always begins with ‘the text of the statute.’”⁵⁷ This manner of interpretation, Justice Gorsuch has explained, “honors only what’s survived bicameralism and presentment.”⁵⁸ But—and here’s Cross’s trenchant observation—“the bicameralism-and-presentment process does not produce assembled statutory texts.”⁵⁹ Instead, it produces what is, in a very real sense, an unwieldy mess of fragmented texts. The task of synthesizing those texts is a nontrivial one, and it is one that has been largely outsourced to a combination of Congress’s in-house codifiers (today, the OLRC) and third-party publishers like Westlaw and Lexis.⁶⁰ In this context, interpreters cannot just accept the synthesized texts as given. Rather, they must understand the process by which those assembled texts come to be, because to do otherwise would severely compromise their ability to draw out statutory meaning in a faithful manner. In practice, this makes an understanding of the codification process essential to statutory interpretation.

And yet, despite the importance of understanding the role of the codifiers in crafting our statutory texts, awareness of the codification process continues to elude courts and litigants. To see this illustrated, one need look no further than the Supreme Court’s most recent term. Two cases—one easy, one hard—presented prime opportunities for the Court to engage with the relevance of codification to interpretation, and in both cases, such arguments were left unaddressed.

III. The Easy Case: Dubin

Writing in dissent in Yates, Justice Kagan observed that she “[knew] of no other cases” in which the Court “relied on a title to override the law’s clear terms,” citing the “wise rule” that “the heading of a section cannot limit the plain meaning of the text.”⁶¹ Whether Justice Kagan was right at the time to suggest Yates’s reliance on section 1519’s heading was an aberration,⁶² such invocations of a provision’s caption are far from unusual today. In recent terms, statutory headings have come to play an increasingly prominent role in the Court’s statutory interpretation jurisprudence.⁶³

The Court’s 2022 term was no exception to this embrace of the statutory heading. In fact, the Court decided one case—Dubin v. United States—in a manner that put a statutory caption firmly at the center of its analysis. But, as Part II discussed, relying on captions to derive statutory meaning can be a fraught exercise, particularly when carried out without any sensitivity to either the role of the codifiers or congressional instructions regarding the propriety of relying on such types of text. In Dubin, the Court did not address these concerns, failing to engage with either. This is, in our view, a mistake—but, as we explain below, a relatively innocuous one. That is because, like Yates, Dubin exemplifies an instance in which the codifier’s canon permits the use of the provision’s caption as an interpretive resource.

The case involved David Dubin, who was convicted of health care fraud for overbilling Medicaid while working as the managing partner of a psychological services company. Dubin submitted a claim for psychological testing by a licensed psychologist when, in fact, the employee who performed the testing was only a licensed psychological associate, resulting in an overcharge of approximately $100.⁶⁴ In addition to charging healthcare fraud, 18 U.S.C. § 1347, the government also charged Dubin with “aggravated identity theft,” pursuant to 18 U.S.C. § 1028A(a)(1), which carries a mandatory two-year prison sentence.⁶⁵

Section 1028A(a)(1) applies when a defendant, “during and in relation to any [predicate offense], knowingly transfers, possesses, or uses, without lawful authority, a means of identification of another person.”⁶⁶ The predicate offenses include, among many others, healthcare fraud.⁶⁷ The Government argued at trial that section 1028A(a)(1) was “satisfied because [Dubin’s] fraudulent billing included [and thus “use[d]”] the patient’s Medicaid reimbursement number (a ‘means of identification’).”⁶⁸ The district court (reluctantly) denied the post-trial challenge to the conviction, relying on established circuit precedent, and the Fifth Circuit (first as a panel, then en banc) affirmed.⁶⁹ As the Supreme Court noted in its opinion, the Fifth Circuit’s broad reading of the statute was an outlier, as many other lower courts had read the statute in a more restrained manner.⁷⁰

In an opinion by Justice Sotomayor, the Supreme Court vacated the judgment and embraced a narrower reading of the statute.⁷¹ Specifically, instead of reading the statute as extending to any incident involving a fraudulent document that includes identifying information of another person, even if incidentally, the Court construed the statute to mean that “[a] defendant ‘uses’ another person’s means of identification ‘in relation to’ a predicate offense when this use is at the crux of what makes the conduct criminal.”⁷² The Court cited the Sixth Circuit’s test as a “helpful guide”⁷³ to understanding when the use of the person’s identity is at the “crux”⁷⁴ of the offense: “The relevant language in § 1028A(a)(1) ‘covers misrepresenting who received a certain service,’ but not ‘fraudulent claims regarding how or when a service was performed.’”⁷⁵ Applying that “heuristic”⁷⁶ to the present case illustrates why no “aggravated identity theft” occurred, as Dubin misrepresented the nature of the services provided, rather than who received those services.⁷⁷

Central to the Court’s narrow interpretation of section 1023A was the statute’s caption. The Court “[s]tart[ed] at the top, with the words Congress chose for § 1028A’s title: ‘Aggravated identity theft.’”⁷⁸ As Justice Sotomayor explained, this caption “suggests identity theft is at the core of § 1028A(a)(1).”⁷⁹ And identity theft “has a focused meaning” that supports “a reading of ‘in relation to’ where use of the means of identification is at the crux of the underlying criminality.”⁸⁰ This meaning also supports “a more targeted definition of ‘uses’” according to which “identity theft is committed when a defendant uses the means of identification itself to defraud or deceive.”⁸¹ By contrast, “[t]he Government’s broad reading,” the Court explained, “bears little resemblance to any ordinary meaning of ‘identity theft.’”⁸²

The focus on the caption in the Court’s analysis is not unexpected. In a previous generation, the Court likely would have relied heavily on the statute’s legislative history, particularly the House Report, which explicitly describes the statute as intended to “address[] the growing problem of identity theft,” defined as “crimes in which someone wrongfully obtains and uses another person’s personal data in some way that involves fraud or deception.”⁸³ Indeed, when the Court addressed a different issue involving section 1023A over a decade ago, the House Report figured prominently.⁸⁴ But times have changed. In the Court’s statutory interpretation jurisprudence, legislative history has been largely relegated to an afterthought, at best—a consequence, no doubt, of certain Justices’ unwillingness to join an opinion that explicitly relies on such evidence.⁸⁵ And in accordance with this textualist turn, the Dubin opinion mentions the House Report only in a brief footnote, prefaced with a warning that its contents were only for “[t]hose who find legislative history helpful.”⁸⁶

In its stead, more “textual” evidence of meaning has shifted to central stage, including headings. But is any given statutory heading actually part of the statutory “text” in the relevant sense? As already discussed, that the heading was included in a bill that passed through bicameralism and presentment is not enough to settle that question. Particularly in light of the version of the codifier’s canon included in Title 18, we must also ask whether the heading was added by the OLRC during the positive law codification process or added as part of a subsequent enactment by Congress. As it happens, Dubin is an instance—like Yates—in which the relevant criminal provision was added to Title 18 as an amendment to the positive law and, as a result, the heading can be confidently attributed to Congress, as opposed to the codifiers.⁸⁷

In our view, it was thus not an error for the Court to rely on the title of section 1023A in the way that it did. But that does not mean nothing was lost by the Court’s failure to acknowledge Title 18’s codifier’s canon. Not every case will be one in which the question of the appropriateness of reliance of a heading will be so easily resolved. The easy cases—like Dubin and Yates—nevertheless provide the Court with an opportunity to develop the analytical frameworks necessary for identifying which text really constitutes the relevant “statutory text,” as Cross has framed the question.⁸⁸ Indeed, it is not hard to imagine that had the Court been more deeply engaged in these questions in Dubin, it might have differently decided the harder cases implicating codification. It is to one such case we now turn.

IV. The Hard Case: Pugin

Decided just two weeks after Dubin, Pugin v. Garland⁸⁹599 U.S. 600 (2023). is a deceitfully tricky case. To the casual observer, the only unusual thing about the decision would seem to be its vote count: although a 6-3 decision on an immigration issue, the Justices’ votes do not track expected ideological lines—instead we see an unusual switch, with Justice Jackson joining the majority opinion written by Justice Kavanaugh, and Justice Gorsuch joining the more liberal Justices, Sotomayor and Kagan, in dissent. But to focus on the voting would be to overlook the more interesting issues lurking below the surface—hard questions regarding the interaction between the legislative process, codification, and statutory interpretation. To understand why, some background is necessary.

Under the Immigration and Nationality Act (“INA”), noncitizens are removable from the United States if they are convicted of an “aggravated felony.”⁹⁰ The INA defines “aggravated felony” as covering an array of federal and state crimes.⁹¹ In 1996, Congress enacted legislation—codified at 8 U.S.C. § 1101(a)(43)(S)—that expanded the definition of “aggravated felony” to include offenses “relating to obstruction of justice, perjury or subornation of perjury, or bribery of a witness, for which the term of imprisonment is at least one year.”⁹² At issue in Pugin was whether an offense “relate[s] to obstruction of justice” within the meaning of section 1101(a)(43)(S) even when the offense does not require that an investigation or proceeding be pending at the time the offensive conduct occurs.⁹³ The question arose from the immigration proceedings of two permanent residents, Jean Francois Pugin and Fernando Cordero-Garcia; Pugin was convicted of being an accessory after the fact to a felony, while Garcia was convicted of dissuading a witness from reporting a crime.⁹⁴

Under the Court’s precedent, whether a noncitizen has been convicted of an offense relating to obstruction of justice within the meaning of section 1101(a)(43)(S) depends on an application of “a categorical approach,” meaning the court must ask whether the elements of the convicted offense establish that the conviction was for an offense relating to obstruction of justice.⁹⁵ Accordingly, the question for the Court was whether the pendency of an investigation or proceeding is an element of “obstruction of justice” as Congress used the phrase when it enacted the 1996 amendments to the INA.⁹⁶ In his majority opinion, Justice Kavanaugh answered that question using an omnivorous approach, looking at such sources as dictionary definitions, state statutes, and the Model Penal Code as indicators of contemporaneous meaning—ultimately determining that there was no pendency requirement.⁹⁷ In reaching this conclusion, Justice Kavanaugh considered, as one revealing factor, the types of offenses included within Chapter 73 of Title 18 of the U.S. Code, which, the opinion notes, is “[e]ntitled ‘Obstruction of Justice.”⁹⁸ Within this chapter are two offenses that do not require a pending federal investigation or proceeding—18 U.S.C. § 1512, which covers witness tampering,⁹⁹ and 18 U.S.C. § 1519 (the provision at issue in Yates), which forbids destroying, altering, or falsifying records.¹⁰⁰ Though acknowledging that one of these provisions, section 1519, was enacted in 2002, after the 1996 amendments, Justice Kavanaugh nonetheless takes the inclusion of these offenses within the “Obstruction of Justice” chapters as indicative of the broader scope that phrase should be understood to possess.¹⁰¹ Despite the centrality of the “Obstruction of Justice” caption to the majority opinion, the Court left the questions of whether and why such interpretive reliance was permissible under Title 18’s codifier’s canon unaddressed.

This omission is even more glaring in Justice Jackson’s concurrence, according to which the textual evidence supplied by the caption is dispositive. Justice Jackson wrote separately to suggest that the arguments based on placement within the Code should be the only ones the Court ought to consider.¹⁰² The key question, Justice Jackson wrote, is what “Congress meant” when it added “obstruction of justice” to the definition of aggravated felony, not “some platonic, judicially divined meaning of Congress’s chosen words”; and, as such, the most definitive evidence is not dictionaries or state statutes, but rather evidence of how Congress itself uses the relevant term.¹⁰³ She then explains that Congress’s “description of Chapter 73 of Title 18 as concerning ‘obstruction of justice’” constitutes Congress’s “longest standing and most significant use of the phrase . . . in the Statutes at Large,” thus giving rise to the inference that when Congress amended the INA in 1996, it was likely using the offenses within that chapter as a “benchmark with respect to what qualifies as an ‘offense relating to obstruction of justice.’”¹⁰⁴ Noting, as the opinion for the Court did, that section 1517 does not require a pending proceeding, Justice Jackson concludes that this “resolve[s] the question before us”—ending with a note suggesting that “[t]he issue of whether such an approach [focused on Chapter 73] best tracks Congress’s intent can be reserved for future consideration.”¹⁰⁵ The threshold question of why the Court could rely on that caption at all in light of Title 18’s apparent proscription remained unanswered, as it did in the majority opinion.

Nor did the dissent discuss the codifier’s canon of Title 18 despite criticizing the majority’s reliance on Code-based considerations on other grounds. Specifically, Justice Sotomayor called attention to the fact that when Congress “codified chapter 73 in 1948, the chapter contained six provisions, each of which requires a connection to a pending proceeding or investigation.”¹⁰⁶ Only later did Congress add to that chapter offenses—most notably section 1512, relating to witness tampering—that did not require such a connection; and, even then, section 1512 was an outlier among the nine narrow additions to the original six offenses of Chapter 73, as the majority of these additions also contemplated pending proceedings.¹⁰⁷ Arguing that the categorical approach requires focusing only on the “heartland” of the identified offense, Justice Sotomayor relies on this evidence to support her position that, in the context of section 1101(a)(43)(S), “obstruction of justice” necessarily involves a pending proceeding.¹⁰⁸

Given that the majority, concurrence, and dissent all discuss at some length the heading of Chapter 73 of Title 18 and the placement of offenses within it, one would expect at least some mention of the legislated instruction in Title 18 regarding reliance on such evidence in the context of statutory interpretation. And yet, no mention is to be found. Blame can be at least partly laid at the feet of the parties, none of which engaged with the congressional instruction in depth.¹⁰⁹ Indeed, only the Solicitor General’s brief noted the provision at all, albeit in a short aside and without elaboration.¹¹⁰ But this raises some obvious questions: What does the instruction in Title 18 imply for the resolution of this case? And, more broadly, what would a more careful analysis of the processes that went into building the relevant statutory text suggest about the Justices’ reasoning? Answering these questions requires an account of how provisions like that in Title 18 should be construed. We thus turn to assessing the reasoning of the various opinions in Pugin in light of our account of these legislated canons.

The account offered by Dorsey takes a literal and maximal view of legislated canons like those in Title 18, taking them to universally prohibit any considerations of caption or placement, whatever their providence. As already noted, under Dorsey’s view, the answer to the first question above is easy—the Court should not have discussed the organization and caption of Chapter 73 at all.¹¹¹ Yet Dorsey’s approach is not the only option. A more nuanced look at the codification and statutory enactment history presents a different answer. The core insight of the codifier’s canon is that the initial organization and chapter titling of the original positive law version of Title 18, as enacted in 1948, reflected decisions by the codifiers (the OLRC’s predecessors)—not Congress—and thus should not be taken as indicative of Congress’s intent. This would suggest that little weight (or perhaps none at all) should be given to the fact that the original six offenses included in Chapter 73 did not include a pendency requirement.¹¹² On the other hand, the same could not be said with regard to the later additions to Chapter 73; as previously described, the placement of such subsequent additions to a previously enacted positive law title must come from Congress, not the codifiers. For example, the statute enacting section 1512—the witness tampering provision—specifies both how the provision should be numbered and that it should be added to “Chapter 73 of [T]itle 18 of the United States Code.”¹¹³ In contrast to Dorsey’s view of these canons, this account treats the providence of an editorial decision as determinative of whether courts can interpretively rely upon it.

Thus, in our view, that these subsequent additions originated with Congress, rather than the codifiers, renders their placement within Chapter 73 fair game from a statutory interpretation perspective, unhindered by the placement and captions instructions of Title 18. On the whole, this would seem to lend greater support to Justice Jackson’s argument, which focuses on the addition of section 1512, as opposed to Justice Sotomayor’s, which places relatively more emphasis on the original six offenses grouped by the codifiers under the “Obstruction of Justice” heading.

That said, we are unconvinced that these organizational decisions can bear the very substantial interpretive weight that Justice Jackson seems to contemplate. As Shobe has discussed, failure by Congress to identify where in a positive law title a new provision should go makes it likely that the new law will be relegated to notes in the U.S. Code—causing the provision to nearly disappear from the public eye.¹¹⁴ Indeed, Cross and Gluck have observed that such notes are not even visible when one pulls up a page of the Code on Westlaw or Lexis.¹¹⁵ For this reason, it might be wise to assume that Congress’s codification directions are not necessarily intended to carry extra meaning beyond that which is inferable from the main text of the enacted provision, but are instead merely efforts to maintain a convenient and accessible U.S. Code.

This latter interpretation gains support when one views the enacted version of the law bill containing section 1512. Although the heading “Obstruction of Justice” is salient when one accesses section 1512 within the Code, the phrase does not appear once in the version of the law as enacted by Congress, which instead references the relevant chapter of Title 18 only by number.¹¹⁶ Taken together with the codification history, described above, we believe this suggests that it would be erroneous to take the placement of section 1512 with Chapter 73 as a definitive guide to the scope of the phrase “obstruction of justice” as used elsewhere within the U.S. Code.

But as we previously noted, this case is particularly tricky. That is because, here, we are not asked to interpret the scope of any particular offense within Title 18, but rather the breadth of a term that appears in a wholly different context, namely an “aggravated felony” as defined by the INA to include offenses relating to “obstruction of justice.” And while the above arguments provide strong grounds against drawing inferences of congressional intent with regard to the placement of specific offenses within Chapter 73 of Title 18, one might still argue that when Congress later—that is, in 1996—used the phrase “obstruction of justice,” it did so intending to cross-reference those offenses collected under that heading within the U.S. Code.

We have already discussed the problem with inferring a longstanding congressional view from the organization of Chapter 73. But inferring an intentional cross-reference is subtly distinct from the argument that—regardless of what Congress might have thought it was doing in the past—when it amended the definition of “aggravated felony” in 1996 to include “obstruction of justice,” it intended to do so in a manner that tracked the set of offenses included in Chapter 73. But Congress made no explicit reference to Chapter 73 within its definition of aggravated felony, even though it did explicitly reference various other criminal provisions from the U.S. Code.¹¹⁷ To nonetheless read in such a reference would suggest that Congress, sub silentio, intended for the chapter to serve as a sort of constructive definition of obstruction of justice. That strikes us as unlikely, pushing the “whole code” canon beyond its appropriate limitations.¹¹⁸

Our analysis here addresses only one aspect of the majority’s reasoning in Pugin, but, given Justice Jackson’s suggestion that litigants and courts should center their arguments on what she refers to as the “Chapter 73-focused approach”¹¹⁹ in future cases, our notes of caution—informed by a nuanced look at the codification and legislative processes—are significant. But there is also a broader lesson: interpreters cannot just take the U.S. Code as the statutory text. Before one can do textualism, one needs to assemble what the relevant texts are and how they fit together. And to do that, one needs—among other things—an understanding of the codification process. That is what the codifier’s canon is all about.

V. Codification and the Future of Qualified Immunity

That the codification process is relevant to statutory interpretation is, in a sense, nothing new. As noted in Part II, in Maine v. Thiboutot,¹²⁰ a seminal case on the scope of 42 U.S.C. § 1983, a central question was the significance that should be accorded to words added—seemingly unbeknownst to Congress—by the codifiers.¹²¹ Ultimately, the Court rejected the position that the provenance of the statutory language mattered, vastly expanding the scope of Section 1983’s cause of action in the process.¹²² Thiboutot is often taught of as a fringe case, an unusual instance of a technical oversight shaping the substantive law. But the scholarly efforts over these last few years to dig deeper into the manner by which the enacted statutory text is assembled into the U.S. Code call into question such a characterization. To be sure, Thiboutot is a particularly dramatic instance of the codifiers’ intervention bearing on issues of statutory interpretation, but these types of issues are far from rare.

As this Essay has discussed, the Court encountered two such cases—Dubin and Pugin—in the last term alone, failing in each to grapple with the codification issues lurking below the surface. We think this was a mistake. Direct engagement with questions of how interpreters should weigh editorial decisions made by the codifiers might not only have impacted the result (or at least the reasoning) in Pugin but would also have allowed the Court to help build a more sophisticated textualist toolbox. And such tools are very much in need, particularly as litigants continue to respond to the federal courts’ textualist turn by developing new and creative modalities of arguably “textual” argumentation.

To take one example, lower courts have already begun grappling with the argument that Section 1983 should be read as nullifying all common-law defenses, including qualified immunity, on the basis of language originally included in the statute but then dropped, seemingly as an oversight, by the codifiers.¹²³ The Supreme Court recognized the doctrine of qualified immunity as a defense to Section 1983 damages actions more than fifty years ago in Pierson v. Ray,¹²⁴ holding that Section 1983 incorporated “the defense of good faith and probable cause, which [was] available to [police] officers in the common-law action for false arrest and imprisonment.”¹²⁵ According to the Court, had Congress intended to abrogate this sort of common law immunity with Section 1983, it would have done so explicitly, and the text of the provision lacks any such explicit abrogation.¹²⁶

This was an at least plausible reading of the text of Section 1983 available to the Pierson Court to the extent that one accepts the canon that statutes in derogation of the common law should be strictly construed. The text of the law, as codified at 42 U.S.C. § 1983, reads:

Every person who, under color of any statute, ordinance, regulation, custom, or usage, of any State or Territory or the District of Columbia, subjects, or causes to be subjected, any citizen of the United States or other person within the jurisdiction thereof to the deprivation of any rights, privileges, or immunities secured by the Constitution and laws, shall be liable to the party injured in an action at law, suit in equity, or other proper proceeding for redress.

Nothing in this language explicitly abrogates common law immunities, so the derogation canon implies that these immunities persist in actions under Section 1983.

The problem for this derogation argument is that the text of the law as codified at Section 1983 is not the text of the Civil Rights Act of 1871—that is, the text that the Reconstruction Congress enacted consistent with bicameralism and presentment. And that text, as Professor Alexander A. Reinert has uncovered, did include language explicitly abrogating common law defenses. It read,

[A]ny person who, under color of any law, statute, ordinance, regulation, custom, or usage of any State, shall subject, or cause to be subjected, any person within the jurisdiction of the United States to the deprivation of any rights, privileges, or immunities secured by the Constitution of the United States, shall, any such law, statute, ordinance, regulation, custom, or usage of the State to the contrary notwithstanding, be liable to the party injured in any action at law, suit in equity, or other proper proceeding for redress ¹²⁷

The reason that this text was never included in the codified version of Section 1983 “is the product of a decision by the first Reviser of Federal Statutes to, for unclear reasons, remove the italicized language when the first edition of the Revised Statutes of the United States was published in 1874.”¹²⁸ This was one of many codification errors in the Revised Statutes, which were supplemented and corrected gradually over time.¹²⁹ Yet the flaw in Section 1983 was neither caught nor corrected before the publication of the first United States Code in 1926, which then inherited the defective version of the law that we know today.¹³⁰

Lower courts have begun to draw attention to this potentially foundational flaw in the doctrine of qualified immunity. As Judge Don Willett recently observed in his Rogers concurrence, textualism is more dominant now than it was when the Court spelled out the qualified immunity doctrine in Pierson; the possibility that “the doctrine does not merely complement the text” that Congress enacted but “brazenly contradicts it” should thus be especially concerning to the modern Court.¹³¹ The fundamental premise of textualism is “that, under Article I, only the text voted upon by the House of Representatives and the Senate and signed by the President (or passed over a presidential veto) constitutes the law.”¹³² But what if the text that survives bicameralism and presentment is inadvertently modified by a revision effort? That is the question raised by the statutory record implicating qualified immunity.

This presents the Court with a redux of Thiboutot, raising the question of whether now, decades later, what we have learned about the codification process should lead courts to a different approach. Indeed, the codification issues raised by qualified immunity may be even thornier than those in Pugin and highlight the complexities that an understanding of the codification process introduces for faithful textualism. These are challenging questions, with answers that implicate the relationship between Congress and the judiciary. If the Court is to chart a new course from Thibotout, it must do so with these difficulties in full view.

Perhaps the Court should embrace a super-muscular version of the codifier’s canon—a “codifier’s error” canon¹³³—and read the statute as originally enacted. This would mean ignoring that Congress enacted the Revised Statutes into positive law in 1874, repealing the abrogation language in the process. Such a move might draw inspiration from the dissent in Thibotout, which emphasized that Congress directed the Reviser of Federal Statutes to not introduce substantive changes in compiling the statutes.¹³⁴ In light of this and other efforts that Congress took to expunge any substantial changes during codification, the legislature enacted the Revised Statutes with the understanding that it was not altering the substance of the law. While one might worry that a codifier’s error canon would shift Congress’s responsibility to check its own work to the courts, the saga of Section 1983 shows that even diligent legislative efforts will not unfailingly prevent the codifiers from errantly altering the law.

Whether such a canon should apply to the question of qualified immunity under Section 1983 might turn on the fact that Title 42 is not among the handful of positive law titles that include broader, if somewhat ambiguous, statutorily enacted language indicating that the codification process was not intended to make “substantive change[s]” to existing law.¹³⁵ As Cross has observed, these provisions “seem to direct interpreters to look to the underlying enactments” but “have been used inconsistently in codifications,” which “raise[s] the question of whether a negative inference is intended for titles lacking such provisions.”¹³⁶

Put differently, this raises the “question of whether the inclusio unius canon is useful” for understanding titles like Title 42 that lack a textual codifier’s canon.¹³⁷ According to Cross, Congress had “little reason . . . to believe that” it was “giving permission against the backdrop of a blanket prohibition” in enacting textual codifier’s canons.¹³⁸ There was, historically, a “default judicial practice [of] look[ing] to pre-codification enactments for controlling evidence of statutory law,” and “in the titles where Congress omitted these provisions, it consistently included them in committee reports or Reviser’s Notes and Members issued similar assurances on the chamber floor.”¹³⁹ This suggests, to Cross at least, that the inclusio unius canon’s utility might be limited in this context. Of course, for the strictest textualists on today’s Court, some of these arguments might fall flat. On the other hand, the argument that textual provisions—including rules of construction—must be understood against broader legal practices and understandings is one that has been gaining increased purchase in recent years.¹⁴⁰ In any event, understanding the significance of the presence or absence of a textual codifier’s canon seems critical to building a textualism rich enough to account for the codification process.

Similar questions arise even for those titles that do have such provisions. The language of the “codifier’s canons” found in these titles is not uniform. Some have language providing that the laws therein “may not be construed as making a substantive change” to existing law or announcing that the purpose of the title is “to restate existing law ‘without substantive change.’”¹⁴¹ Others specify that “the intent is to conform to the understood policy, intent, and purpose of Congress in the original enactments.”¹⁴² This latter language, Cross notes, echoes the statutory mandate of the OLRC, aligning the construction of codified titles with the scope of the codifiers’ statutory authority.¹⁴³ Whether and how to give effect to the variation between the codifiers canons in different titles is yet another question for a robust textualism. Perhaps the Court will interpret the provisions to uniformly proscribe reliance on substantive changes from the Statutes at Large. Alternatively, the Court might distinguish between substantial changes and structural features such as captions and placement.

Still other questions remain. The Court might grapple with whether these provisions reach codification decisions made before the creation of the U.S. Code or only those decisions made since these provisions were enacted as part of the process of compiling the Code. After all, the error in the Revised Statutes at the heart of the qualified immunity question occurred before Congress began including notes of caution in the codified law. Moreover, the Court must decide how much, if any, weight to give Congress’s apparent acquiescence to its qualified immunity decisions.¹⁴⁴

These are just a few of the many thorny questions that the codification of Section 1983 implicates, and they point to how codification can raise more questions still in other settings. But the first step toward answering them is far more mundane. It involves acknowledging the ways in which an understanding of codification is essential to a robust approach to statutory interpretation.

VI. Conclusion

We may all be textualists now,¹⁴⁵ but we do not yet all agree on how to assemble and interpret the relevant statutory texts. As scholars continue to uncover how the language with which courts work is shaped in ways that complicate the relationship between those words and the will of Congress as enacted through bicameralism and presentment, courts need pay heed if they want to be faithful to their own interpretive commitments.

As we have shown, understanding the codification process is crucial to interpreting the positive law titles compiled in the United States Code, and Congress has legislated guidance cautioning courts to be attentive to how codification might change the substance of the law. Yet, as its latest term illustrates, the Supreme Court continues to decline opportunities to address the relationship between codification and interpretation. The questions raised by the codification process are only set to multiply in quantity, complexity, and importance as scholarship advances our understanding of the law assembled in the United States Code. By acknowledging and answering these questions, the Court can build a richer and more robust textualism to guide litigants and lower courts.

[*] Daniel B. Listwa is a litigation associate at Wachtell, Lipton, Rosen & Katz and a nonresident fellow of the Center for Private Law at Yale Law School.

[**] Adam Flaherty is a member of the Yale Law School class of 2024.

Related