Reading Law with Linguistics: The Statutory Interpretation of Artifact Nouns

Brandon Waldon[*]

Cleo Condoravdi[**]

James Pustejovsky[***]

Nathan Schneider[****]

Kevin Tobia[†]

Abstract

Is an airplane a “vehicle”? Is a floating home a “vessel”? Is an unassembled gun a “firearm”? Such questions about “artifact nouns”—nouns that describe human-created entities—are fodder for legal philosophy. They are also common statutory interpretation issues, which today’s textualist courts resolve with linguistic analysis. We propose that textualist courts complement familiar tools, like dictionaries, with insights from linguistics.

We examine as a case study Garland v. VanDerStok, which the Supreme Court will soon decide. It concerns “gun parts kits,” firearm parts that can become operable firearms through combination or part finishing. These kits have been used in several mass shootings, and the case concerns whether such a kit is a “firearm” subject to regulation under the 1968 Gun Control Act. To analyze the statute’s meaning, we apply insights from linguistic theory, new data from language usage, and a survey study of ordinary Americans. This evidence supports that the gun parts kits identified by the government fit within the statutory meaning of “firearm.”

The Article’s case study in the legal interpretation of artifact nouns also carries broader implications. We develop lessons for the practice of legal interpretation, statutory interpretation theory, and broader debates in legal philosophy.

I. Prefatory Note

We wrote this Article and an associated amicus brief in Garland v. VanDerStok¹Brief for Professors and Scholars of Linguistics and Law as Amici Curiae in Support of Petitioners at 15, Garland v. VanDerStok, 144 S. Ct. 1390 (2024) (No. 23-852) [hereinafter Linguistics Brief]. over the summer and fall of 2024. The Supreme Court heard the oral argument on October 8, 2024. The case remained undecided as this Article entered the final stages of publication with the Harvard Journal of Legislation in early 2025. On March 26, 2025, the Supreme Court decided the case: Bondi v. VanDerStok.² We were pleased that seven Justices applied a commendable linguistic analysis, and that the majority referenced our amicus brief,³ upon which this Article expands.

The publication timeline did not allow time to substantively update this Article to address the Court’s analysis. As such, we ask readers to consider the pre-decision window as the reference context for our Article. In keeping with that context, we have made no changes to the Article in light of the decision (including even minor changes to reflect the final caption, “Bondi v. VanDerStok”). We are grateful to the Harvard Journal on Legislation for allowing us this introductory note, the remainder of which briefly addresses the relationship between some of this Article’s arguments and the Court’s opinion.

In what follows in this Prefatory Note, we assume that the reader is already acquainted with the relevant background of Bondi v. VanDerStok and with our linguistic analysis as presented in this Article and in our amicus brief.

The Court’s opinion explicitly identifies weapon, frame, and receiver with linguistic terminology: these nouns are “artifact nouns” denoting entities of human creation and/or assimilation.⁴ The Court characterizes its task as determining whether “Congress… use[d] artifact nouns to reach incomplete objects” in the statutory definition of firearm, another artifact noun.⁵ The Court recognizes a key aspect of artifact noun meaning: artifact nouns are distinguished by the potential to exhibit human-intended functions.⁶ Thus, “everyday speakers sometimes use artifact nouns to refer to unfinished objects—at least when their intended function is clear.”⁷

Nevertheless, just because artifact nouns can be used in a way that emphasizes their function does not mean they should always be read in this way. The Court recognizes, rather, that context is crucial for determining the relevance of design considerations and functional potential in the Gun Control Act of 1968 (“GCA”) definition of firearm.⁸ Some contexts may exclusively center entities’ physical form or physical constitution. The GCA definition of firearm does not provide such a context: rather, “Congress explained that a ‘weapon’…qualifies for regulation if it is either ‘designed’ to accomplish that function or ‘capable of being ‘readily…converted’ to do so.’”⁹

The Court acknowledges that the context supplied by the GCA definition is by no means unusual in this respect: ordinary speakers routinely deploy terms such as weapon to refer to incomplete, unfinished, or disassembled objects. “Imagine a rifle disassembled for storage, transport, or cleaning…as a matter of every day speech, that rifle is a weapon, whether disassembled or combat ready.”¹⁰ More generally, the Court finds it relevant to consider how ordinary Americans and weapons manufacturers use language to refer to gun parts kits. The Court discusses one such kit, Polymer80’s “Buy Build Shoot” kit, at length. For the Court, this kit’s “intended function as instrument of combat is obvious. Really, the kit’s name says it all: ‘Buy Build Shoot.’”¹¹

Moreover, the Court recognizes that artifact nouns encode an inherently vague and context-dependent notion of functional potential.¹² “[A]t some point a kit may be so incomplete or cumbersome to assemble that it can no longer fairly be described as a ‘weapon’ capable of ‘read[y]… conver[sion]’ into a working firearm.”¹³ The language of the GCA definition makes clear that some gun parts kits meet the threshold of potential, on any plausible specification of that threshold. “If Congress had wanted to regulate only operable firearms, it could have simply addressed ‘weapons’ that can ‘expel a projectile by the action of an explosive.’ But Congress didn’t stop there.”¹⁴ Thus, “the ATF’s rule [at CFR §478.11] is not facially inconsistent with the GCA,”¹⁵ contra the Respondents’ position.

The Court also employs insights from linguistic theory in its analysis of frame and receiver. In addition to recognizing that these terms are artifact nouns,¹⁶ the Court recognizes that subsection (B) of the statutory definition contains an anaphoric expression any such weapon that “refers us back to weapons encompassed by subsection (A).”¹⁷ Subsection (A) in turn “encompasses some things that are not yet fit for effective use in combat, including…certain weapon parts kits” containing unfinished frames/receivers.¹⁸ For us and for the Court, this relationship between subsections (A) and (B) undermines the Respondents’ “complete-items-only reading of subsection (B).”¹⁹

The Court’s interpretation of frame and receiver is further supported by the following principle of linguistic interpretation: ordinarily, “we expect context to clarify language in a consistent manner.”²⁰ Having established that weapon in (B) refers back to weapons “not yet fit for effective use,”²¹ the Court contends that (B) should be read such that the artifact nouns frame and receiver similarly reach unfinished objects: “the fact that Congress used one artifact noun (‘weapon’) in subsection (B) to reach some unfinished articles suggests it used two other artifact nouns (‘frame’ and ‘receiver’) in the same way in the same provision.”²²

The Court’s analysis stands in tension with an alternative, which proceeds from the observation that Congress explicitly contemplates weapons “readily converted” to exhibit particular functions in (A) but does not include similar language in (B). This alternative analysis takes this omission to be a meaningful variation: whereas (A) may reach unfinished or incomplete weapons, (B) extends only to complete frames and receivers.²³

However, the Court’s analysis provides a ready explanation for this omission. There are sufficient clues from the surrounding context of (B)—including the anaphoric expression any such weapon linking back to (A)—to support the government’s interpretation of frame and receiver in that subsection.²⁴ Indeed, “reading subsection (B) in light of subsection (A) does more to undermine than to advance the plaintiffs’ cause.”²⁵

In VanDerStok, the Court leveraged analytical insights from linguistic theory to navigate a contested matter of statutory interpretation. In our Article, we argue that these insights have implications not only for VanDerStok, but for future hard cases involving the interpretation of artifact nouns. We also claim that our analysis bears on theories of legal interpretation more generally.

II. Introduction

In 2019, a sixteen-year-old shooter in California killed two students and injured several others.²⁶ This shooter’s gun had been assembled from a “gun parts kit,” an unassembled collection of firearm parts. A sixteen-year-old is not legally eligible to buy a firearm.²⁷ But the sale of “gun parts kits” has made it easier to avoid age and background requirements; since 2019, there have been at least 150 shootings involving these guns, many of which involve teen shooters.²⁸ “Gun parts kits” are typically unserialized, making them difficult to trace, and they are sold from many online retailers. Internet tutorials demonstrate how to convert parts kits into an AR-15 assault rifle in as little as thirty minutes.²⁹ Federal law imposes requirements concerning age minimums, licensing, and background checks for the purchase of “firearms.”³⁰ However, the Fifth Circuit recently held that “gun parts kits” are not firearms within the definition of these federal laws.³¹

The Supreme Court will decide that decision’s appeal in 2025. Garland v. VanDerStok³² asks whether the Gun Control Act of 1968 (“GCA”), allowing government regulation of “firearms,”³³ includes “gun parts kits” or “ghost guns.”³⁴

The Court’s decision will turn on its application of textualism,³⁵ an interpretive theory that gives statutory text the meaning it would communicate to an ordinary person, “reasonable person,”³⁶ or “ordinary . . . speaker.”³⁷ Although the Court’s textualists have some theoretical disagreements,³⁸ the Justices are “all textualists”³⁹ in the sense that the debate will start with the linguistic meaning of the statutory definition, which implicates the meaning of terms like firearm and weapon.⁴⁰ The textualist Court often relies on dictionary definitions and intuitive hypotheticals.⁴¹ This Article argues that textualist courts should also consider analytical insights from the field of linguistics. In making this recommendation, the Article builds on prior calls to “triangulate” ordinary meaning,⁴² which emphasize the use of interpretive tools like analysis of naturally occurring language and survey data about how ordinary Americans understand language.⁴³

The Article considers VanDerStok and the regulation of “gun parts kits” as an important practical example. These firearms are sold without serial numbers and allow minors and other parties to obtain weapons that would normally be prohibited, with no paperwork or background check. Each year, law enforcement recovers thousands of these weapons.

At the same time, our linguistic analysis applies more broadly. Firearm is an artifact noun, denoting entities of human-made creation. Legal interpretation often implicates the meaning of these nouns, asking questions like: Is an airplane a vehicle, and is a floating home a vessel?⁴⁴ Our case study’s analysis also illustrates lessons for other interpretive disputes, statutory textualism, and legal philosophy.

Part III of the Article provides brief background on VanDerStok and traditional textualist tools like dictionary definitions and intuitive examples. Part IV introduces and applies relevant research from linguistics to VanDerStok. First, linguistic theory suggests that the interpretation of artifact nouns (like weapon, firearm, frame, receiver, bicycle, or table) depends critically on context. We explain how context indicates the relevant facets of meaning of these artifact nouns, including facets related to the noun’s potential function or manner of creation. Next, we argue that the statutory context of the GCA strongly emphasizes the functional and creational facets of firearm and frame or receiver over other facets.

In Part V, we supplement our theoretical argument with two types of new empirical data. First, we report examples of naturally occurring language in which people describe firearm parts kits as “weapons” and “firearms.” This indicates that, in many contexts, parts kits fit comfortably within the ordinary meaning of firearm and weapon. Next, we present an original experimental study of ordinary Americans. The survey has two primary implications. First, most (65%) of our sample understood parts kits as firearms, even with no further context. Second, when participants had context mirroring the statutory definition,⁴⁵ they agreed even more (73%) that parts kits are firearms. Part VI of the Article extends our linguistic analysis to the second question in VanDerStok, concerning the interpretation of frame and receiver.

Part VII elaborates the implications. First, the Article has implications for legal practice. This includes a recommendation for how the Supreme Court should analyze VanDerStok. Insofar as the Court seeks to resolve the case through textualism, the Court should not rely solely on dictionary definitions and arbitrary hypotheticals about word meaning.⁴⁶ Instead, it should look to the rich context provided by the statutory definition, which emphasizes an object’s potential functionality and design as a firearm. We additionally discuss recommendations for other cases involving the interpretation of artifact nouns. The Article’s case study also carries broader implications for textualist theory, which we elaborate in Section VII.B, and for longstanding legal philosophical debates about text and purpose, which we discuss in Section VII.C.

III. VanDerStok and Traditional Textualist tools

A. Background on Garland v. VanDerStok

The Gun Control Act of 1968 (“GCA”), codified at 18 U.S.C. § 921, defines firearm as:

(A) any weapon (including a starter gun) which will or is designed to or may readily be converted to expel a projectile by the action of an explosive;

(B) the frame or receiver of any such weapon;

(D) any destructive device.⁴⁷

Congress delegated authority to administer the GCA to the Attorney General,⁴⁸ who delegated that authority to the Director of the Bureau of Alcohol, Tobacco, Firearms, and Explosives (“ATF”).⁴⁹ In 2022, the ATF promulgated a rule clarifying its interpretation of the definition.⁵⁰ The ATF interpreted firearm to include “a weapon parts kit that is designed to or may readily be completed, assembled, restored, or otherwise converted to expel a projectile by the action of an explosive.”⁵¹ The rule also explained that the agency interprets frame or receiver⁵²27 C.F.R. § 478.12(a) provides the following definitions: “(1) The term ‘frame’ means the part of a handgun, or variants thereof, that provides housing or a structure for the component (i.e., sear or equivalent) designed to hold back the hammer, striker, bolt, or similar primary energized component prior to initiation of the firing sequence, even if pins or other attachments are required to connect such component (i.e., sear or equivalent) to the housing or structure. (2) The term ‘receiver’ means the part of a rifle, shotgun, or projectile weapon other than a handgun, or variants thereof, that provides housing or a structure for the primary component designed to block or seal the breech prior to initiation of the firing sequence (i.e., bolt, breechblock, or equivalent), even if pins or other attachments are required to connect such component to the housing or structure.” to include certain partially complete, disassembled, or nonfunctional frames and receivers, including a parts kit: “The terms ‘frame’ and ‘receiver’ shall include a partially complete, disassembled, or nonfunctional frame or receiver, including a frame or receiver parts kit, that is designed to or may readily be completed, assembled, restored, or otherwise converted to function as a frame or receiver.”⁵³

Jennifer VanDerStok and a set of individual, manufacturer, and organizational plaintiffs who sought to utilize, purchase, or sell gun parts kits (hereinafter “VanDerStok”) challenged the agency’s interpretation in a Texas district court, which held that the ATF exceeded its statutory authority and vacated the final rule nationwide.⁵⁴ The Fifth Circuit upheld the decision,⁵⁵ and the Supreme Court granted certiorari.⁵⁶ In the October 2024 Term, the Supreme Court will decide two questions:

1. Whether “a weapon parts kit that is designed to or may readily be completed, assembled, restored, or otherwise converted to expel a projectile by the action of an explosive,” 27 C.F.R. 478.11, is a “firearm” regulated by the Act.

2. Whether “a partially complete, disassembled, or nonfunctional frame or receiver” that is “designed to or may readily be completed, assembled, restored, or otherwise converted to function as a frame or receiver,” 27 C.F.R. 478.12(c), is a “frame or receiver” regulated by the Act.⁵⁷

VanDerStok will have important consequences for the regulation of firearms. Although some states have passed laws regulating “gun parts kits,”⁵⁸ most have not. And although some significant “gun parts kits” manufacturers have recently shut down,⁵⁹ many others are operating.⁶⁰ Tens of thousands of “gun parts kits” are recovered by law enforcement each year, and VanDerStok will determine the efficacy of existing federal regulation of these kits.

B. Textualism’s Traditional Tools

The questions presented in VanDerStok concern statutory interpretation: Do the weapons parts kits, frames, and receivers identified by the ATF fall within the statutory definitions of firearm, frame, and receiver? For a textualist Supreme Court, the case turns on the statute’s linguistic meaning, including what it would communicate to an ordinary reader.

The textualist Court often relies on dictionary definitions and intuitive hypotheticals.⁶¹ These traditional tools infuse both the VanDerStok lower court opinions⁶² and Supreme Court briefs.⁶³ They appeal to dictionary definitions of firearm and colorful hypotheticals about ordinary language: When IKEA sells a “table parts kit” that must be assembled by the purchaser into a table, isn’t this parts kit a table?⁶⁴

Definitions and hypotheticals can assist in clarifying how an ordinary speaker of English understands language, but these tools must be carefully employed. Dictionary definitions can be cherry-picked.⁶⁵ Intuitive hypotheticals can also be easily cherry-picked, and they may not accurately reflect ordinary language.⁶⁶ Moreover, “[w]hen Justices—elite lawyers—debate how ‘ordinary people’ talk, there is a serious risk that their renderings will speak with an upper-class, judicially-inflected accent.”⁶⁷

IV. Analytical Insights from Research in Linguistics

This article augments the traditional textualist toolkit with formal linguistic theory and empirical linguistic methods. We show that out of context, artifact nouns such as firearm are underdetermined as a matter of their ordinary linguistic meaning. Specifically, the interpretation of firearm depends everywhere and always on the context in which that term appears. Though dictionary definitions may characterize the interpretation of firearm in frequently encountered contexts, such definitions cannot possibly anticipate the rich variety of contexts in which the word is used or how those contexts shape interpretation. We review empirical evidence that supports our theoretical perspective, which is a mainstream one in linguistics.

Linguistic theory also provides guidance on how an artifact noun’s meaning is resolved as a function of context. This allows us to ascertain when a statutory definition, e.g., the definition of firearm in the GCA, departs from ordinary meaning, as has been debated among the litigants in VanDerStok. We complement this theoretical exercise with two empirical contributions: naturally occurring evidence of ordinary linguistic usage and experimental survey data on ordinary linguistic interpretation.

A. The Ordinary Meaning of Artifact Nouns

To ascertain the ordinary meaning of firearm, we look to linguistic theory, which has extensively studied the linguistic category to which firearm belongs: artifact nouns (e.g., firearm, weapon, table, and bicycle). Artifact nouns “denote[] entities of human invention and/or entities that, through some assimilative procedure, come to serve some human-intended function.”⁶⁸ These nouns differ in this respect from so-called natural kind-denoting nouns (e.g., egg, dog, and water). Linguists have long recognized that as a matter of literal semantic meaning, an artifact noun can be characterized by basic attributes that pertain to members of the category denoted by the noun. A widely cited analytical framework was developed by Professor Pustejovsky, who associates words of the English lexicon with a semantic structure representing the essential facets (‘qualia roles’) of artifact noun meaning:

(1) The Constitutive facet: the relationship between an object and its constituents, or proper parts;

(2) The Formal facet: that which distinguishes the object in the larger domain;

(3) The Potential (‘telic’) facet: the potential of an object to fulfill some function or purpose;

(4) The Design (‘agentive’) facet: factors involved in the origin or ‘bringing about’ of an object.⁶⁹

In a given context, only select facets of nominal meaning may be relevant for the purposes of linguistic interpretation. For example, consider sentence (1) below. This sentence is underspecified as to the action that Noel undertakes. On one interpretation, (1) implies that Noel initiated the process of reading a novel; on another interpretation, (1) implies that Noel initiated the process of writing a novel:

Noel began a novel.

Linguistic theory illuminates the source of this underspecificity. On the ‘begin-to-read’ interpretation, the relevant facet of the noun novel is its ‘Potential’ facet: here, the novel realizes its potential to fulfill its intended function (i.e., to be read). On the ‘begin-to-write’ interpretation, the relevant facet is the ‘Design’ facet, i.e., the property of being brought into existence through writing. Context is crucial for identifying the relevant facet(s) and hence the intended meaning of the sentence. Out of context, (1) does not have a stable interpretation.⁷⁰

According to one naïve hypothesis of linguistic meaning, contexts that center an artifact noun’s Potential facet are ones in which members of the category realize their intended function. On this hypothesis, the set of vehicles in such a context consists exclusively of objects that are (currently) engaged in the action of transporting people or things. Clearly, this hypothesis is untenable: it predicts that parked cars can never belong to the set of vehicles (or even cars). A more plausible hypothesis states that category membership in such contexts is determined by functional potential that may or may not be realized in context.⁷¹ The parked car is a vehicle in part because it is ready to be used for transportation (once one turns the ignition key and starts driving). The Potential facet attends to potential capabilities, rather than exhibited behaviors.

The line between ‘capable’ and ‘incapable’ vis-à-vis function is vague and heavily dependent on context and the artifact noun under consideration. The government’s bicycle hypothetical in VanDerStok supports this point:

A bicycle is still a bicycle even if it lacks pedals, a chain, or some other component needed to render it complete or allow it to function. So too if the bicycle is shipped with plastic guards attached to the gears or brakes that must be removed before operation, or with a seat tube that the user must cut to length before installing. No one would deny that a company selling and shipping products in any of those conditions was engaged in selling ‘bicycles.’⁷²

This hypothetical considers objects that, with some manipulation, can fulfill the canonical intended purpose of bicycles. The relevant context is one in which “a company [is] selling and shipping products” to buyers who fully expect to put in some elbow grease before they can ride the thing they have purchased. Form and constitution—including the physical shape of the object and the relative position of constituent parts in physical space—are secondary considerations; what is important is that the buyer receives a product that is capable of being ridden, given an appropriate amount of effort. If the company ships a hunk of aluminum and other raw materials sufficient to produce a bicycle, the buyer can justifiably pursue a refund on the grounds that he did not receive a bicycle. Of course, this is a clear case on a continuum of possibilities; there is no crisp line between ‘appropriate’ and ‘inappropriate’ effort.

In sum, this theoretical perspective takes the ordinary meaning of artifact nouns like weapon, firearm, and table to be inherently underspecified. There is no complete interpretation of these artifact nouns without context. In some contexts, a firearm must be completed; in others, firearm or gun may have an interpretation that includes incomplete members. Moreover, context is crucial to determine both (1) whether, in context, an artifact noun refers to entities that share common functional properties and (2) how to characterize that common functional ‘core.’ Below, we apply this linguistic framework to analyze how statutory context resolves the interpretation of firearm in the GCA.

B. Context’s Role in Interpreting “Firearm”

For the textualist Supreme Court, VanDerStok turns on the meaning of firearm in the GCA. This section addresses this linguistic question in light of the statutory context, which has been emphasized by textualists from Scalia⁷³ to Barrett.⁷⁴ We restrict our attention to a contested subpart of the statutory definition, section 921(a)(3)(A) [hereinafter “(A)”], the part of the statutory definition from which the ATF derives its authority to regulate firearm parts kits. Below, we argue that (A) clearly extends beyond the narrowest category of completed, operable weapons. We then argue that (A) does so in a manner consistent with the ordinary meaning of firearm.

1. Analysis of the GCA Definition

When a statute “includes an explicit definition,” the Court “must follow that definition, even if it varies from a term’s ordinary meaning.”⁷⁵ While the meaning of firearm in the GCA may be informed by contemplating the ordinary meaning of that term, it is crucial to attend to the statutory context to discern the contextual meaning that firearm has within the Act. We might not call a firearm silencer or a bomb kit a firearm in ordinary language. But the statutory definition of firearm explicitly includes a wide range of objects, from firearm frames and receivers,⁷⁶ to any combination of parts designed and intended for use in assembling a firearm silencer,⁷⁷ to bombs and bomb-making kits.⁷⁸

In VanDerStok, a locus of disagreement is part (A) of the GCA definition, which defines firearm as “any weapon (including a starter gun) which will or is designed to or may readily be converted to expel a projectile by the action of an explosive.”⁷⁹ The Respondents in the case sometimes suggest that the only “actual” firearms are completed ones, and that the statute covers only such firearms: “It covers only an actual firearm…it does not cover mere parts or kits of parts that might be manufactured into one.”⁸⁰

The plain language of (A) undermines this analysis. (A) communicates three relevant categories of firearms, comprising a list of items interpreted disjunctively:

(1) “any weapon…which will…expel a projectile by the action of an explosive”;

(2) “any weapon…which…is designed to…expel a projectile by the action of an explosive”; or

(3) “any weapon…which…may readily be converted to expel a projectile by the action of an explosive.”⁸¹

To read (A) to include completed, assembled, currently functional firearms gives effect to the first disjunct (1). But to read (A) to include only such entities renders (2) and (3) superfluous, and “courts should avoid treating any statutory language as surplusage.”⁸²

Thus, the statutory context of (A) emphasizes that firearm extends to weapons beyond fully completed, functional firearms. The text clearly communicates that firearm includes some weapons beyond those that “will . . . expel” a projectile by the action of an explosive; it also includes (2) weapons designed to or that (3) may readily be converted to do so.⁸³

Though we expect the statutory definition of firearm to take precedence over its ordinary meaning in VanDerStok (if the two diverge), the latter notion nevertheless features prominently in that case: does (A) embellish the ordinary meaning of firearm, or does it explicate an interpretation consistent with that term’s ordinary meaning? We turn to that question below.

2. A Linguistic Perspective on the Statutory Definition

The litigants in VanDerStok no doubt recognize the primacy of the statutory definition of firearm for the purposes of the GCA (and that Act’s implementation by the ATF). However, VanDerStok claims that (A) is “an expansion of the ordinary usage” of firearm because it covers weapons “that could readily be converted” to expel projectiles.⁸⁴ The Government disagrees, claiming that (A) explicates the term’s ordinary meaning rather than embellishing it (it is an “‘explicit’” definition of the ordinary meaning).⁸⁵

This dispute over ordinary meaning is consequential. First, as discussed in Part IV, the Government’s ‘ordinary meaning’ analysis of firearm is important for its analysis of the phrase frame or receiver as it appears in section 921(a)(3)(A) [hereinafter “(B)”]. Our analysis of firearm’s ordinary meaning will similarly inform our own analysis of frame or receiver. Another reason to consider this question is that while statutory definitions take priority by default, “[s]ometimes a definition itself contains a term that is not clear.”⁸⁶ In such cases, “the usual criteria of interpretation . . . are brought to bear,”⁸⁷ and “[f]ar and away the most important of those is the contextual factor of the word actually being defined.”⁸⁸ In other words, judges should look to a term’s ordinary meaning and surrounding context when specialized statutory definitions are unclear. More broadly, “the meaning of the definition is almost always closely related to the ordinary meaning of the word being defined.”⁸⁹

As far as (A) is concerned, there is no conflict between the ordinary meaning of firearm and its statutory definition. Rather, (A) contains ample linguistic context to specify the noun’s ordinary meaning. (A) identifies the relevant facets of meaning for the purposes of interpretation: there is a clear and specific elaboration that firearm’s Design facet (“designed to”) and Potential facet (“may readily be converted to”) are essential.⁹⁰

The Potential facet encodes a notion of potential capability that is inherently vague and context dependent. (A) faithfully reflects this inherently vague dimension of artifact noun meaning: the definition explicitly identifies objects that “may readily be converted to expel a projectile by the action of an explosive.”⁹¹ With “converted,” we see Congress contemplating scenarios in which the user must perform a non-trivial action to get the object to perform its intended function, “to expel a projectile by the action of an explosive.”⁹² Nevertheless, the line between readily converted and its negation is inherently fuzzy. In this regard, the definition takes pains not to narrow the reach of the term firearm relative to its ordinary meaning. That is, the definition explicates a property of artifact noun interpretation that constitutes part of the implicit linguistic knowledge of competent ordinary speakers, who recognize that an object can be called a firearm even when it does not presently realize its intended function.

Finally, (A) covers “any weapon” that exhibits the design and potential characteristics.⁹³ Thus, to determine whether (A) is faithful to the ordinary meaning of firearm, we must understand the contribution of mentioning “weapon” in the definition.⁹⁴ Because weapon, an artifact noun, is not defined elsewhere in the statute, we consider its ordinary meaning.

The relationship between firearm and weapon is well-characterized in both linguistic and psychological research, which draws a distinction between ‘basic-level’ categories, the most typical names we assign to everyday things (e.g., those categories denoted by tree, book, table), and superordinate-level categories, which reside at a higher level of abstraction (e.g., organism and furniture). Superordinate terms group together categories denoted by basic-level terms: for example, the superordinate category mammal groups together the categories denoted by dog and cat.

Firearm and weapon stand in exactly this basic-superordinate relationship. (A) grounds the definition of firearm in its superordinate category, just as the category denoted by firearm constitutes part of the category denoted by weapon as a matter of ordinary linguistic meaning.

Like basic-level artifact nouns, superordinates tend to identify entities that have common potential-related properties; however, there may be few (if any) common perceptual attributes that characterize a superordinate category.⁹⁵ For example, the superordinate noun vehicle identifies entities of many shapes, sizes, and descriptions; what makes something a vehicle is the potential to transport people or things.

As a matter of ordinary meaning, weapon contemplates entities with common functional potential. Moreover, like firearm, weapon may denote entities that are far from operable. Thus, weapon harmonizes with the rest of (A), which explicitly emphasizes the role of potential while downplaying the relevance of physical characteristics.

In sum, there are multiple sources of linguistic context within (A) that point to the relevance of the Design and Potential facets of firearm while de-emphasizing considerations of physical form and/or constitution. This interpretation is routinely available for artifact nouns, and the Government offers compelling evidence that supports this generalization regarding entities that require further assembly to become functional: for example, “[e]very speaker of English would recognize that a tax on sales of ‘bookshelves’ applies to IKEA when it sells boxes of parts and the tools and instructions for assembling them into bookshelves.”⁹⁶ The same is true of entities that require finishing, with additional tools sold separately, to become functional. For example, Ticonderoga’s “Classic Yellow Wood-Cased Pencils” are sold unsharpened.⁹⁷ They can only actually function (write) after the purchase of a sharpener and an additional finishing step: sharpening the pencil. Nevertheless, buyers and sellers recognize that these are clearly pencils. Similarly, customizable belts and watches are sold without holes. Users must hole punch the belts and watches with additional tools to complete them. Nevertheless, the objects as sold are belts and watches.

Thus, from a linguistic perspective, Congress’s definition of firearm in (A) is best understood not as an instruction to disregard or embellish the noun’s ordinary meaning but as an attempt to provide sufficient context to resolve an indeterminacy that is inherent to its ordinary meaning. Because the resolution depends on context, we expect that some contexts may suggest a different resolution. For example, when we call a child’s toy gun his gun, the relevant facets are the Formal and Constitutive facets: the object has certain physical properties that make it a suitable object for make-believe play. (A) is not such a context; the language of the definition points to an interpretation that centers the object’s functional capabilities and the circumstances of its creation.

Notably, the Government and Respondents in VanDerStok appear to dispute whether the ATF’s interpretation of the GCA departs from the ordinary meaning of ‘firearm’:

[A] covered firearm parts kit qualifies as a firearm as a matter of ordinary usage . . . An ordinary speaker of English would recognize that a company in the business of selling kits that can be assembled into firearms in minutes—and that are designed, marketed, and used for that express purpose—is in the business of selling firearms.⁹⁸

The district court correctly held that [the ATF] . . . extend[s] the definitions of “firearm” and “frame or receiver” in federal law beyond any reasonable understanding of those terms.⁹⁹

As mentioned above, we expect the statutory definition of firearm to be privileged over considerations of ‘ordinary usage’ or ‘reasonable,’ ‘commonsense’ understanding—insofar as the statutory definition diverges from ordinary meaning. Nevertheless, the above passages suggest that some parties believe that the Court’s interpretation of the statutory definition should also cohere with ordinary meaning and ‘commonsense’ linguistic intuition. The analysis presented in this section helps clarify this issue: to the extent that the ATF’s regulatory authority over firearm parts kits is grounded in (A), its authority is also grounded in the ordinary meaning of firearm, since (A)’s definition is fundamentally underscoring the ordinary meaning of firearm in context. In the next section, we present empirical linguistic data that speaks in favor of the ATF’s interpretation of (A) and provides further evidence that this interpretation coheres with ordinary meaning considerations.

V. Empirical Data on the Meaning of “Firearm”

In this Part, we supplement Part IV’s theoretical linguistic analysis with two sources of empirical evidence: naturally occurring linguistic data, which demonstrates that ordinary speakers can felicitously refer to a parts kit as a firearm, weapon, gun, rifle, and/or assault rifle (AR), and experimental survey data, which demonstrates that ordinary speakers interpret firearm to include parts kits in the context of definitions akin to (A).

A. New Data on Ordinary Language Use

In assessing whether firearm parts kits fall within the ordinary meaning of firearm and weapon, actual linguistic usage is instructive. The Court regularly considers hypothetical linguistic examples to assess the ordinary meaning of statutes, particularly in criminal contexts.¹⁰⁰ Members of the Court have also recommended considering patterns of actual language use through corpus linguistics.¹⁰¹

Ordinary usage shows that firearm parts kits fall within the ordinary meaning of both firearm and weapon in many contexts—especially contexts similar to § 921. Below, we present several examples from online product advertisements and reviews, with the relevant term bolded for emphasis. The Appendix has an extensive set of examples.

First, grammatical evidence links the label “firearm” to a parts kit:

● “Introducing the AR-40 4.5″ MOD1 Billet Upper Receiver Pistol Build Kit, a powerful and compact firearm designed to deliver outstanding performance in the dynamic world of pistol builds.”¹⁰²

● “Tiger Rock AR-15 Enhanced Flat Dark Earth Rifle Kit with a 10” Handguard, a precision-engineered firearm designed for optimal performance and durability.”¹⁰³

● “Tiger Rock AR-15 Enhanced Robins Egg Blue Pistol Kit – a compact and powerful firearm designed for performance and style.”¹⁰⁴

● “Discover the iconic 16” Burnt Bronze Rifle Kit 5.56 from House Keymod, a stylish and versatile firearm made in the USA.”¹⁰⁵

● “Invest in excellence with the Tiger Rock AR-15 Burnt Bronze 5” Complete Pistol Kit – a versatile, reliable, and aesthetically pleasing firearm that stands out in both performance and style.”¹⁰⁶

The construction used above (“the…Kit, a powerful and compact firearm”),¹⁰⁷ with a comma or hyphen separating the full product name and a descriptive phrase, conveys that the speaker believes that the kit is a firearm. Specifically, this is an illustration of an “ascriptive [noun phrase] supplement[].”¹⁰⁸ For example, the first part of the sentence “Kim Jones, a quite outstanding student, won a scholarship to MIT,” is equivalent in meaning to “Kim Jones is a quite outstanding student.”¹⁰⁹

Second, many advertisements name the kit and then immediately refer to a “weapon” or “rifle.” For example:

● “Looking for a little more firepower in your life? Say hello to the 16” Flat Dark Earth Rifle Kit 5.56 with 12” Keymod. This powerful rifle comes equipped with an M4 Feed Ramp, a 1×7 barrel twist, and a 1/2×28 thread, making it perfect for taking down even the most tough targets.”¹¹⁰

Others refer to parts kits themselves as weapons:

● “When you need the best AR-15 rifle available, look no further than this Blue Titanium 16” Rifle Kit 5.56 12” House M-LOK. Designed and manufactured with an obsessive attention to detail, this rugged and dependable weapon is perfect for the professional gun owner or enthusiast.”¹¹¹

● “Outstanding! A great value and a great weapon! Assembles pretty easy, I would recommend it for anyone who is interested in making their first build.”¹¹²

Customer reviews also tightly tether references to the unassembled firearm parts kits and a “gun” or “rifle.”¹¹³

● “Since I bought this kit I have bought several more AR’s. This one by far is still my go to range rifle as it nails at 200 yards.”¹¹⁴

● “The kit came in as advertised. Great inexpensive rifle.”¹¹⁵

● “Purchased this rifle. The assembly was a learning experience since this was my first build.”¹¹⁶

The purchased entity is an uncompleted firearm kit, not a completed firearm—so it cannot be the case that “rifle” in these sentences refers only to completed firearms. Rather, the writers’ use of the demonstrative word “this” to modify “rifle,” near references to the kit, shows that the customers refer to both the kit and the firearm assembled from the kit as one holistic product.¹¹⁷

The preceding examples offer ample evidence that sellers and consumers readily deploy firearm and similar nouns to refer to a product over its lifespan, or to the product that is purchased (as a kit). These examples demonstrate that ordinary speakers do not confine firearm to just completed weapons or those extremely close to operability as a matter of ordinary meaning.

B. New Data from a Survey Experiment

Supreme Court justices have also recognized the usefulness of survey data to textualist analysis.¹¹⁸ Survey data are a useful complement to traditional textualist sources of linguistic evidence, especially insofar as the Court seeks to understand how “the ordinary English speaker . . . would understand the words of the statute” or how the “reasonable person” uses words. In an experiment, we observe that ordinary speakers readily construe firearm to include parts kits. This observed decision by consensus among ordinary speakers is particularly strong in contexts akin to (A), which clearly centers firearm’s design and potential facets.

1. Methods

Participants: We recruited 1250 participants from Prolific, a widely used online crowdsourcing platform. We recruited a U.S.-based sample, with a minimum 99% approval rating, 50% men and 50% women, and 50% Republican–identified, 50% Democrat–identified, <1% Independent, Other or No Preference. We preregistered three comprehension check questions and one CAPTCHA (to screen out bots). There were 988 participants who completed all checks correctly and were included in the analysis (50% male, 49% female, 1% non-binary; 49% Republican, 47% Democrat, 3% Independent, <1% Other or No Preference). Participants were paid $1.00, and median completion time was 3 minutes and 53 seconds, for a median compensation rate of $15.44/hour.¹¹⁹

Materials: Experiment trials are schematized in Figure 1. Participants were asked to read a short passage which described a disagreement between two parties. The locus of the disagreement was the interpretation of some artifact noun, which was defined within the context of a law (in “legal”-domain trials) or a company rule (in “ordinary”-domain trials).

In all trials, the definition specified that members of the relevant nominal category are also members of a corresponding superordinate category. (E.g., for X to meet the definition of firearm, X must also be a weapon). However, these definitions were further elaborated in ways that varied across experimental conditions. In trials featuring a “restricted” definition, the law/rule explicitly contemplated entities with the immediate potential to exhibit a canonical function. In trials featuring a “full” definition, the law/rule additionally contemplated the Design and Potential of such entities. These definitions were designed to closely track the statutory definition of firearm in (A), modulo relevant manipulations. There were 12 trial types total: 3 Noun conditions [firearm, table, or bicycle] × 2 Domain conditions [legal or ordinary] × 2 Definition conditions [restricted or full].

Participants were then asked to provide their judgment as to the permissibility of an [AN] parts kit, where [AN] was the artifact noun defined in the law/rule. Participants were told that these kits can be purchased online, that creating a functional [AN] requires combining the parts, and that most people could combine the parts in one or two hours with the right tools.¹²⁰ Participants provided a binary “Yes”/ “No” judgment to the following question: “In your view, does the rule’s definition of “[AN]” include [AN] parts kits?”

Procedure: Each participant completed one trial of a randomly assigned type. Before the trial, participants completed two tasks which served as exclusion criteria: an attention task, designed to ensure that participants were reading experiment instructions in full, and a simple reasoning problem task. After the trial, participants completed a comprehension task, in which they were asked to recall how the artifact noun of interest was defined in the experimental trial. The two selection options corresponded to the artifact noun’s two associated Definition conditions. This task served as a third exclusion criterion. Data from participants who met at least one criterion were excluded from analysis.

Figure 1a. Experimental materials, legal condition

Figure 1b. Experimental materials, ordinary condition

Figure 2. Proportions of ‘yes’ response in Experiment 1. Error bars represent 95% binomial confidence intervals (as computed with the Hmisc function in R).

2. Results

In total, 262 participants (≈21%) met at least one exclusion criterion or exited the study early, leaving data from 988 participants for analysis.¹²¹ Figure 2 visualizes participant responses as a function of Definition, faceted by Noun. A majority of participants (69%; 95% CI = [63%, 74%]) considered firearm parts kits to be firearms. This trend held in both the legal context (64%; 95% CI = [56%, 71%]) and the ordinary context (74%; 95% CI = [67%, 80%]), regardless of whether firearm received a restricted definition (65%; 95% CI = [58%, 72%]) or full definition (73%; 95% CI = [65%, 79%]).

To investigate whether and how participant responses varied as a function of Noun (reference level firearm), Domain (reference level “legal”), and Definition (reference level “restricted”), we conducted a binary logistic regression predicting log odds of “Yes” response from fixed effects of these three variables and all possible interactions. None of the interaction terms were significant predictors of the outcome variable, so we dropped these terms from the model and conducted a second regression with only the additive fixed effects of the three variables. Relative to firearm trials, the likelihood of “Yes” response was significantly lower on bicycle trials (β = -0.7073, SE = 0.1637, z = -4.319, p = 1.56e-05) but not on table trials (β = -0.1626, SE = 0.1671, z = -0.973, p = 0.330611). Moreover, we found no evidence of an effect of Domain (β = 0.1096, SE = 0.1338, z = 0.819, p = 0.412656); i.e., there was no evidence that the likelihood of “Yes” response was either higher or lower in the “legal” condition as compared to the “ordinary” condition. However, there was evidence of significant positive effect of Definition (β = 0.5136, SE 0.1359, z = 3.780, p = 0.000157); i.e., there was evidence that the likelihood of “Yes” response was higher when participants were provided with a ‘full’ definition of the relevant noun.

3. Discussion

Across multiple tested contexts, a majority of ordinary Americans evaluated firearm parts kits as members of the category denoted by firearm. Note that this occurred without any further language emphasizing the relevance of design or potential function, such as the language of (A) (“designed to” and “may readily be converted to”). We found that when such language was provided, an even greater proportion of participants considered firearm parts kits to be firearms. That is, ordinary readers recognize that a clear statement including weapons that are “designed to” or “may readily be converted to” function as a firearm indicates that parts kits are included.

This result is fully compatible with the theoretical framework presented in Part IV. Without definitional language that explicitly identifies the relevant facets of artifact noun meaning, ordinary speakers attempt to infer those relevant facets under considerable uncertainty. We predict this uncertainty to be reflected as population-level variation in interpretive judgments, which is what we observe in the “restricted” condition of the experiment. When the context includes language which explicitly identifies the Design and Potential facets of firearm, the uncertainty is resolved.¹²² As predicted, we see greater levels of population-level agreement in the expected direction, with a greater proportion of speakers construing firearm to include parts kits.

These results also carry methodological implications. Not all artifact nouns (or artifact noun parts kits) were treated equally: most participants reported that the term firearm includes firearm parts kits, but a significantly smaller proportion of participants judged bicycle to include bicycle parts kits. In VanDerStok, the Government, Respondents, and lower courts have all attempted to elucidate the meaning of firearm and firearm parts kit by considering ‘analogous’ linguistic expressions. (Recall, e.g., the Government’s “IKEA bookshelf” hypothetical.)¹²³ Our results suggest that there are limits to this approach: it may obscure substantial divergences in how ordinary Americans would interpret the words or phrases under comparison. This cautionary note applies even when making analogies between highly related linguistic expressions. (For example, bicycle and firearm are both alike in that they are both basic-level artifact terms.)

VI. The Meaning of “Frame” and “Receiver”

Beyond the interpretation of firearm, VanDerStok implicates the interpretation of frame or receiver as that phrase appears in (B), the second disjunct of the GCA’s statutory definition of firearm. The Government contends that (B) includes frame or receiver parts kits that are designed to or may readily be converted to function as a frame or receiver. The respondents claim that “[i]f an item potentially could be made into a frame or receiver but is not a[n assembled] frame or receiver that is insufficient under the Act’s plain text.”¹²⁴ That is, the respondents contend that (B) extends only to fully-assembled, fully-functional frames and receivers.

A. Linguistic Theory

As a reminder, recall the language of 18 U.S.C. § 921(a)(3):

The term “firearm” means

(A) any weapon (including a starter gun) which will or is designed to or may readily be converted to expel a projectile by the action of an explosive;

(B) the frame or receiver of any such weapon;

Like firearm, both frame and receiver are artifact nouns whose linguistic meanings are made more precise by context. We argued that (A) underscores that the contextual meaning of firearm includes the ATF-identified gun parts kits, by foregrounding the Potential (“expel projectiles . . .”) and Design (“designed”) facets of the noun’s ordinary linguistic meaning and backgrounding the Formal and Constitutive facets (no mention of size, shape, trigger, etc.). Notably, however, Congress does not similarly underscore the meaning of frame or receiver in (B) with similar language (e.g., “any object which will or is designed to or may readily be converted to function as a firearm frame or receiver”).

The Government argues that the inclusion of (A)’s “designed” and “readily be converted” language merely serves to ground the statutory definition of firearm in a familiar, ordinary sense of that term;¹²⁵ and in (B), frame or receiver is similarly interpreted according to its ordinary meaning, which for the Government “include[s] a partially completed frame or receiver that can readily be made functional.”¹²⁶

In contrast, the lower courts and respondents take the omission of such language in (B) to be deliberate and meaningful, and they conclude on that basis that frame or receiver extends only to fully assembled and functional frames and receivers. However, proper understanding of artifact nouns reveals that this conclusion does not follow. (B) not only omits the “designed” and “readily be converted” language found in (A) but also omits language analogous to (A)’s “will . . . expel a projectile by the action of an explosive.” (B) could have been written to explicitly identify (only) fully assembled, fully functional frames or receivers, e.g., ‘a frame that will hold the essential mechanism’ (“will” to express purpose-readiness) of ‘any such weapon’ defined in (A). But (B)’s language varies from (A)’s in this respect. As we argue below, we do not think these strong inferences should be drawn from the “variation” between (A) and (B). But even for those who propose that this variation is meaningful, it does not follow that the variation cuts in one direction. (B) has neither the explicit mirroring language that would favor the government (“any object which will or is designed to or may readily be converted to function as a firearm frame or receiver”) nor the explicit mirroring language that would favor the respondents (“any object which will function as a firearm frame or receiver”).

1. The Context of Section 921(a)(3)(B)

How, then, does context contribute to the meaning of frame or receiver? To answer this question, we first note that language users routinely leverage contextual information to inform a word’s contextual meaning. Recall from Section IV.A that verb + artifact noun combinations may leave unspecified the nature of the activity. (2) shows that when multiple under-specifications are present, there is pressure to specify them similarly:

(2) Last week, Noel started a novel, and Liz finished a nonfiction book.

In (2), the nouns novel and nonfiction book have a possible ‘Potential’–oriented interpretation: Noel and Liz are reading (realizing the object’s intended function). Both nouns also have a possible ‘Design’ interpretation: Noel and Liz are writing (creating) the object. Where available, interpreters tend to apply contextual information broadly to specify nouns consistently. This is demonstrated by (3) and (4), in which the first sentence provides a context clue for how to resolve the indeterminacy in (start a) novel. Because of the strong default expectation of similar specification, this context clue may also serve to determine an interpretation of (finished a) nonfiction book.

Noel is a prolific writer. Last week, he started a new novel. A few days later, his friend Liz finished a nonfiction book.

Noel is an avid reader. Last week, he started a new novel. A few days later, his friend Liz finished a nonfiction book.

The context of (3) clarifies that Liz is writing, and the context of (4) clarifies that Liz is reading. The same kind of contextual inferences clarify section 921(a)(3). In (A), Congress furnishes the reader with the context sufficient to identify the relevant facet of firearm’s ordinary meaning for the purposes of the definition. Frame or receiver does not receive a similar elaboration in (B), but because frame or receiver is underspecified in the same way that firearm is, we can readily account for this omission by appealing to the same principle of interpretive consistency exemplified in (2), (3), and (4). Section 921(a)(3)(A) contemplates gun parts kits, including those with 80% frames or receivers, as we have argued above in Part IV; in context, section 921(a)(3)(B)’s “frames or receivers” are just those same frames or receivers. Even though 80% frames or receivers are not immediately ready to function, they are regularly included as part of a kit that may readily be converted into a firearm, and therefore the frame or receiver of any such weapon contemplated in (A).

The remainder of section 921(a)(3) bolsters this conclusion. Part (C) defines “any firearm muffler or firearm silencer” as a firearm; (D) further extends this definition to include “any destructive device.” Both (C) and (D) ‘bottom out’ in statutory definitions which, like (A), foreground the Potential and Design facets over and above Formal or Constitutive facets:

The terms “firearm silencer” and “firearm muffler” . . . includ[e] any combination of parts, designed or redesigned, and intended for use in assembling or fabricating a firearm silencer or firearm muffler.”¹²⁷

The term “destructive device” means . . . (B) any type of weapon . . . which will, or which may be readily converted to, expel a projectile by the action of an explosive or other propellant . . . and (C) any combination of parts either designed or intended for use in converting any device into any destructive device . . . and from which a destructive device may be readily assembled.¹²⁸

(C) and (D) thus provide further contextual information which helps to resolve the indeterminacy of frame or receiver in (B). In sum, neither context nor considerations of ordinary linguistic meaning suggest that frame or receiver extends exclusively to fully assembled frames and receivers in (B).

This analysis contrasts starkly with one presented in the lower court opinions, which apply a familiar heuristic of legal interpretation and achieve the opposite interpretive result. The lower courts’ preferred heuristic—sometimes called the Presumption of Consistent Usage canon of statutory construction—states that “[a] word or phrase is presumed to bear the same meaning throughout a text; a material variation in terms suggests a variation in meaning.”¹²⁹ From the Meaningful Variation corollary to this canon, we can derive an ‘omitted terms, negative implication’ principle of the kind invoked by the lower courts.¹³⁰

By contrast, the argument developed in this section is evocative of another heuristic, the Associated Words canon (noscitur a sociis), which states that “[a]ssociated words bear on one another’s meaning.”¹³¹ Scalia and Garner elaborate: “When several nouns or verbs or adjectives or adverbs . . . are associated in a context suggesting that the words have something in common, they should be assigned a permissible meaning that makes them similar.”¹³² The Associated Words and Consistent Usage (Meaningful Variation) canons appear to offer contradictory guidance when it comes to the meaning of frame or receiver in (B). How can we be confident that Associated Words takes “precedence” over Meaningful Variation in this case?

According to Scalia and Garner, “[t]he [Associated Words] canon has tremendous value in a broad array of cases”;¹³³ by contrast, “[b]ecause it is so often disregarded, [the Consistent Usage] canon is particularly defeasible by context . . . [it] can hardly be said to apply across the whole corpus juris.”¹³⁴ The Consistent Usage canon “more than most other canons . . . is not often achieved.”¹³⁵ Thus, say Scalia and Garner, as a general matter the Associated Words canon is more broadly applicable than Consistent Usage reasoning.¹³⁶

2. The Role of Anaphora in Section 921(a)(3)

The specific language and context of section 921 further confirm the conclusion recommended by the Associated Words canon. There are clear linguistic indications that count against inferring any “meaningful” variation from differences between (A) and (B). Recall that section 921 defines firearm to include:

(A) any weapon (including a starter gun) which will or is designed to or may readily be converted to expel a projectile by the action of an explosive;

(B) the frame or receiver of any such weapon.¹³⁷

The phrase “any such weapon” in (B) refers back to “weapon” as defined in (A). Linguists call this relationship “anaphora.” This is “a relation between two linguistic elements, wherein the interpretation of one (called an anaphor) is in some way determined by the interpretation of the other (called an antecedent).”¹³⁸

This clear reference from Congress, using “any such weapon” in (B) to refer back to “weapon” in (A), is essential elaboration of the contextual meaning of “frame” and “receiver.” It indicates a close connection between (A) and (B). As Parts I and II of this Article explain, (A) describes weapons that will, are designed to, or may readily be converted to expel a projectile by means of an explosive. Section (B), then, describes frames and receivers of “such weapons” that will, are designed to, or may readily be converted to expel a projectile by means of an explosive.

Anaphora’s contribution to this contextual meaning clarifies why “receiver” in (B) would not include, for example, the completed metal receiver of a toy gun.¹³⁹ Literally, such a receiver is a completed receiver, but there is no statutory context to indicate that “completion” in the constitutive sense is the relevant facet of meaning. Overreliance on abstract dictionary definitions could imply the same bizarre conclusion.¹⁴⁰ But the completed receiver of a toy gun—or the completed receiver of a metal model gun—is not a receiver in this context: Those are not receivers of “any such weapon,” as defined in (A). The text says nothing about whether the relevant receivers should be restricted to only completed ones; to the contrary, it emphasizes that the relevant receivers are ones compatible with the “weapons” described in (A).

Recognizing the contextual meaning of (B) underscores that it includes an “incomplete” 80% receiver that is designed to or could be converted in one hour into the essential firing mechanism of an AR-15. Such receivers are frequently sold as parts of weapons that are firearms under (A), as we argued in Parts I–III.¹⁴¹ Those (unfinished) receivers are receivers of any such weapon as defined in (A).

The presence of the anaphoric construction in (B) (“any such weapon”)—coupled with the linguistic context provided by (A) (elaborating firearm as “any weapon” that is designed to or may readily be converted to expel a projectile), which contains the antecedent—clarifies why the absence of similar language in (B) does not point to such a restrictive construal of frame or receiver. These two linguistic phenomena are sufficient to determine the interpretation of frame or receiver in context. For this reason, it would have been superfluous for Congress to add overt ‘designed to or may readily be converted to’ language directly into (B). The absence of such language in (B) is further evidence that the interpretation of frame or receiver is resolved by considering the context of (A), which in turn suggests that frame or receiver extends beyond fully completed frames and receivers in context. The textual indications from section 921 strongly support applying the Associated Words canon.

In contrast, there are none of the indications one would expect to support the Meaningful Variation canon. That canon carries the most weight when there is clear linguistic evidence that text was added in one place and not the other in order to draw a contrast.¹⁴² That is not the case here. Consider a hypothetical alternative where instead of (B), there was (B′), coupled with (A) as follows:

(A) any weapon (including a starter gun) which will or is designed to or may readily be converted to expel a projectile by the action of an explosive;

(B′) any frame or receiver of a weapon which will expel a projectile by the action of an explosive.

In this hypothetical, the parallel language (“which will expel”) and absence of any connection between the two clauses (i.e., no anaphoric “such”) provide a stronger basis for reading the absence of “is designed to or may readily be converted to” from (B′) as deliberate and meaningful.

The actual statutory text, however, is not a good candidate for such a Meaningful Variation inference. It does not use two different terms (e.g., land vs. real estate, completed frame vs. frame), nor does it include any language that indicates a contrast between (A) and (B). It does the opposite: It explicitly binds the two clauses together (with “any such weapon”) such that they are best read as sharing a context in which both Design and Potential facets of meaning are salient.

B. New Data on Ordinary Language Usage

The preceding linguistic analysis may seem complex, and the contribution of context here is easily overlooked. But the preceding analysis is confirmed by ordinary examples and common sense.

1. Unfinished frames as frames and unfinished receivers as receivers

Our analysis of frame or receiver in (B) extends to usages of frame and receiver in ordinary contexts. As the government’s reply brief in VanDerStok notes, sellers (including Respondents in VanDerStok) refer to 80% frames as simply frames.¹⁴³ Ordinary people also refer to incomplete receivers as receivers:

● “What additional parts do I need to assemble a complete rifle or pistol using this receiver?”¹⁴⁴

People also regularly refer to unassembled packages that include an unfinished 80% receiver as a firearm and rifle.

● “Get ready to rock and roll with the brand new 16” AR-15 Rifle Kit with 15” Slim Keymod and 80% Lower Receiver – Burnt Bronze! Not only does it look great, thanks to its stylish burnt bronze color, but it’s also the perfect firearm for any shooter.”¹⁴⁵

● “Introducing the 16” AR-15 Rifle Kit with 15” Slim Keymod with 80% Lower Receiver – Titanium Blue; the perfect addition to your collection! This rifle has many of the features you’d expect from a higher-end model, like a .223/5.56 M4 Feed Ramp and 1:7-barrel twist.”¹⁴⁶

An unfinished 80% frame or receiver falls within the ordinary meanings of frame and receiver, respectively. As such, it is sensible for a statute to simply employ these terms with this meaning (inclusive of, e.g., 80% receivers), with no additional language to underscore this meaning.

2. Analogous ordinary examples do not support an inference from “meaningful variation” or “lack of similar language.”

Section VI.A’s linguistic analysis of artifact nouns extends beyond firearm frames and receivers. Here we present five ordinary examples that mirror the structure of section 921(a)(3). Imagine, as a simple example, that a high school prohibits students from bringing cigarettes to school. The rule defines cigarettes as follows.

Example 1: Cigarettes

The term “cigarette” means

(A) any object which will or is designed to or may readily be converted to smoke tobacco;

(B) the tobacco tube of any such object.

(A)’s express definition prohibits finished cigarettes as well as an unassembled and unfinished “cigarette kit,” complete with tobacco, rolling papers that must be unboxed and torn out of their holder, and a filter. (B) does not explicitly define “tobacco tube” of “any such object.” Now imagine that a student entered the school with (only) unopened/unboxed rolling papers and claimed that these do not fall under (B) because they are only an “unfinished” and “unassembled” tobacco tube and (B) does not have language mirroring (A), like “any object which will or is designed to or may readily be converted to be used as a tobacco tube.” The school principal would rightly reject this argument as an unsuccessful effort to evade the rule’s meaning. The absence of an explicit definition in (B) is not meaningful in the sense that it implies that (B) is artificially limited to only 100% completed tobacco tubes. The language “any such object” in (B) underscores that the contextual meaning of “tobacco tube” is not artificially restricted in this way: (B) explicitly includes those tobacco tubes that are part of a kit falling under (A).

Before turning to the next example, note that the anaphora (“any such object”) also underscores the contextual meaning of tobacco tube by emphasizing the exclusion of entities that might fall under the meaning. A candy cigarette tube (that is solid and cannot hold tobacco) does not fall under (B). This reasoning also answers the Fifth Circuit’s suggestion that the government’s reading of section 921(a)(3)(B) might extend to objects that “look like frames or receivers.”¹⁴⁷ The court worries that:

One could make a cake that looks like a hamburger, just as one could make a cake that looks like a gun frame or receiver. One is ‘clearly identifiable’ as a hamburger, just as the other is ‘clearly identifiable’ as a gun part. But that does not make the former taste like a Big Mac, just as it does not make the latter covered by the GCA.¹⁴⁸

The court is correct that section 921(a)(3)(B) does not include a cake that looks like a firearm receiver, and that this would be an absurd conclusion. But the court is incorrect that such an absurd conclusion must follow from an interpretation that extends section 921(a)(3)(B) beyond fully completed receivers. In Example 1 above, tobacco tube extends to unfinished and unassembled tubes but not toy tubes or cake tubes. This is underscored by the statutory context, indicating the relevant class of tubes with “any such tube.”

As a second example, suppose that a middle school prohibits children from bringing board games to school. The rule defines board games as follows.

Example 2: Board games

The term “board game” means

(A) any object which will or is designed to or may readily be converted to be played as a game by placing or moving pieces on a board;

(B) the board of any such object.

Suppose a student enters the school with a brand-new cardboard sheet of tiled board pieces from the board of the game Settlers of Catan. To convert these tiles into a functional board, the pieces must be “punched out” of the cardboard sheet that holds them and assembled, but the student has clearly brought to school a board within the meaning of (B). The same would be true of a student who brought unfinished and unassembled boards of Guess Who? or Mousetrap. There is no Meaningful Variation inference to draw from the explicit definition in (A) to an artificial restriction of the ordinary meaning of (B)’s terms. The best reading in context, underscored by the anaphora (“any such object”) connecting (A) and (B), is that (B) includes such an unfinished and unassembled board.

We are cautious about drawing too much from ordinary examples for legal interpretation. But these examples show that there is nothing unusual about the contextual argument from Section VI.A. There is no Meaningful Variation inference that must be drawn from a construction like section 921(a)(3)’s.¹⁴⁹

The real examples from “gun parts kits” advertisements and customer reviewers also illustrate that the Design and Potential function of artifact nouns, like unfinished frames and receivers, are often contextually relevant. Moreover, these examples illustrate that it is a mistake to assume that the “variation” in descriptions of artifact nouns in sections 921(a)(3)(A) and 921(a)(3)(B) is a sound basis of contextual restriction of those nouns’ meaning. The opposite inference is often contextually warranted: artifact nouns described in slightly different terms should be construed consistently, not artificially inconsistently. Finally, the construction in section 921(a)(3) underscores this opposite contextual inference: The anaphoric “any such weapon” explicitly unifies the context of sections 921(a)(3)(A) and 921(a)(3)(B).

VII. Implications

This Article’s analyses have implications for legal interpretive practice. These include (1) a recommendation for deciding VanDerStok, and (2) lessons for other interpretive disputes involving artifact nouns.

The case study has broader implications for statutory interpretation theory. Today’s textualists rely heavily on what we call a “word sense disambiguation” (“WSD”) paradigm. The analysis here illustrates the limits of that approach. This also has implications for the increasingly important determination of textual clarity versus ambiguity.

Most broadly, this study advances longstanding discussions in legal philosophy. First, vehicle is an artifact noun, and we turn to the classic Hart-Fuller debate about the hypothetical “no vehicles in the park” rule. That debate has been explored thoroughly,¹⁵⁰ but the Article’s linguistic analysis provides new support for Fuller’s (sometimes disparaged) view of language and the relationship between text and purpose. Second, the Article’s case study illustrates the benefits to legal interpretation of employing intensional rather than extensional definition.¹⁵¹

A. Legal Interpretive Practice

1. Implications for Garland v. VanDerStok

The linguistic theories and data presented in this Article bear directly on how to interpret the terms firearm, frame, and receiver in the GCA. These interpretive questions are at the heart of VanDerStok. Our arguments support the claim that the GCA grants the ATF regulatory authority over the firearm parts kits and unfinished frames and receivers identified by the government.

Begin with ordinary meaning. The linguistic theory explained here clarifies that artifact nouns (like firearm, weapon, frame, receiver, table, and bicycle) have context-sensitive meanings, which are heavily influenced by the entity’s design and potential function. In many contexts, artifact nouns include members missing parts (a book missing a page is a book), members that are unassembled (an IKEA table is a table), and members with unfinished parts that can only be completed by applying additional tools (an unsharpened pencil is a pencil).

This linguistic theory is supported by a survey experiment of a sample of ordinary Americans. The majority of participants evaluated a firearm parts kit as a firearm without any further context. Ordinary linguistic usage further supports this finding. Moreover, in online consumer reviews of weapon parts kits, people regularly refer to such a product as a firearm, weapon, gun, rifle, and AR (“assault rifle”).

The statutory context further underscores this ordinary meaning of firearm and weapon. Section 921(a)(3)(A) explicitly defines firearm with reference to design (“designed to”) and potential function (“may readily be converted to”) as key facets of meaning of the firearm-weapons contemplated by the statute. Our empirical results demonstrate that ordinary readers are sensitive to these contextual cues. When survey participants were presented with this statutory language (a firearm is “any weapon which will or is designed to or may readily be converted to…”), their categorization of weapon parts kits as firearms increased to 73%.

Frame and receiver are also artifact nouns, and thus they are also context-sensitive and heavily influenced by design and functional potential. In many contexts, an “80% receiver,” which can be converted into a fully functional receiver in a few hours, fits comfortably within the ordinary meaning of receiver. Ordinary language confirms this: Firearms manufacturers and consumers refer to such unfinished frames and receivers as simply a “frame” and “receiver.”

Moreover, the statutory context underscores this meaning. Section 921(a)(3)(A) contains an explicit disjunctive definition of the ordinary meaning of firearm, while section 921(a)(3)(B) does not include such a definition. However, it is a mistake to assume that this construction implies an artificial restriction of the ordinary meaning of frame and receiver in 921(a)(3)(B). Moreover, section 921(a)(3)(B) contemplates the “frame” and “receiver” of “any such weapon,” as defined in Section 921(a)(3)(A). In linguistics terminology, this language establishes an anaphoric relation, clarifying that the relevant class of frames and receivers are the ones for the weapons contemplated by section 921(a)(3)(A).

2. Implications for Artifact Noun Disputes

This Article has focused on the artifact nouns firearm, weapon, frame and receiver. But litigation implicates many other artifact nouns. Since 2010, the Supreme Court has examined the meaning of artifact nouns including administration, certified mail, document, employee, jail, money, prison, and public accommodation.¹⁵² Artifact nouns also arise regularly in contract litigation,¹⁵³ as the next sub-section’s example illustrates. And there is dispute about artifact nouns in common law adjudication.¹⁵⁴ The analysis here provides a template for linguistic analysis of these terms.

Consider, for example, the meaning of vessel. The meaning of this term was the central issue in both Stewart v. Dutra Construction Co., which the Supreme Court decided in 2005,¹⁵⁵ and Lozman v. City of Riviera Beach, which the Supreme Court decided in 2013.¹⁵⁶ The Rules of Construction Act, codified at 1 U.S.C. §§ 1–8, defines vessel:

The word “vessel” includes every description of watercraft or other artificial contrivance used, or capable of being used, as a means of transportation on water.¹⁵⁷

Begin with Stewart. As part of the City of Boston’s “Big Dig” construction project, it employed Dutra Construction Company. Dutra owned and used the “Super Scoop,” a massive floating platform with a clamshell bucket that removes silt from the ocean floor. Dutra hired Willard Stewart to monitor the Super Scoop’s mechanical systems, and Stewart was injured in an accident on the Super Scoop. Stewart sued under the Jones Act, alleging Dutra’s negligence, and under the Longshore and Harbor Workers’ Compensation Act (LHWCA), which allows employees to sue a “vessel” owner for an injury caused by that owner’s negligence. Ultimately, one question reached the Supreme Court: Is the Super Scoop a vessel for the purposes of the LHWCA?

At the time of the accident, the Super Scoop was not moving in the water. So, an interpreter might be inclined to think that it is not a vessel—vessels are only watercrafts that are actively transporting. But Justice Thomas’s majority came to the opposite conclusion, that the stationary Super Scoop is a vessel:

A ship long lodged in a drydock or shipyard can again be put to sea, no less than one permanently moored to shore or the ocean floor can be cut loose and made to sail. The question remains in all cases whether the watercraft’s use “as a means of transportation on water” is a practical possibility or merely theoretical one.¹⁵⁸

Here, Thomas’s majority reads 1 U.S.C. § 3’s “capable of being used as a means of transportation on water” not to mean literally being used or extremely close to being used, but rather practically capable of being used. There is, of course, inevitable fuzziness around the category boundaries of “practically capable.” But Justice Thomas, rightly, did not take that fuzziness to imply that any object could be a vessel within the meaning of 1 U.S.C. § 3.

This Article’s discussion of the linguistics of artifact nouns clarifies both why Justice Thomas’s interpretation is consistent with the meaning of vessel and a sensible reading in context. The meaning of artifact nouns is not necessarily limited to currently operable members. Firearm often includes firearms without bullets or magazines; table often includes unassembled IKEA tables; and so on. Similarly, vessel need not be restricted to vessels that are currently transporting entities on water, or even vessels that are extremely close to doing so. Vessels that have a “practical possibility” of transporting on water are also vessels.

The statutory definition of vessel underscores this ordinary meaning. Like the explicit definition of firearm in the GCA, the definition of vessel includes a disjunction: “The word ‘vessel’ includes every description of watercraft or other artificial contrivance used, or capable of being used, as a means of transportation on water.”¹⁵⁹ This context underscores that vessel extends beyond only watercrafts “used.”

As a second vessel example, consider Lozman v. City of Riviera Beach.¹⁶⁰ In 2002, Fane Lozman bought a floating, two-story structure that was about fifty-seven feet long. Lozman lived there for three years, until a hurricane struck, when Lozman moved the structure to the Riviera Beach marina and lived there for a year. About two years later, the Riviera Beach city brought an in rem action against the structure. The City argued that Lozman’s structure was a vessel, and thus that maritime liens attach to the vessel; Lozman argued that his structure was not a vessel.

The district court held that Lozman’s floating home was a vessel under the Rules of Construction Act, which defines a vessel as including “every description of watercraft or other artificial contrivance used, or capable of being used, as a means of transportation on water.”¹⁶¹ This is a broad definition, which would seemingly include a floating mobile home, capable of moving on water (and thus transporting on water). The Eleventh Circuit agreed, holding that Lozman’s structure was a vessel.¹⁶²

However, the Supreme Court reversed. Justice Breyer rejected the proposal that a vessel is “anything that floats.” Instead, Justice Breyer’s majority held that a vessel is something that a reasonable observer would consider designed for transportation on water; as such, Lozman’s floating home was not a vessel. Justice Breyer’s opinion highlighted the importance of design to artifact nouns:

Not every floating structure is a “vessel.” To state the obvious, a wooden washtub, a plastic dishpan, a swimming platform on pontoons, a large fishing net, a door taken off its hinges, or Pinocchio (when inside the whale) are not “vessels,” even if they are “artificial contrivance[s]” capable of floating, moving under tow, and incidentally carrying even a fair-sized item or two when they do so . . . . Consequently, in our view a structure does not fall within the scope of this statutory phrase unless a reasonable observer, looking to the home’s physical characteristics and activities, would consider it designed to a practical degree for carrying people or things over water . . . .

The home has no other feature that might suggest a design to transport over water anything other than its own furnishings and related personal effects. In a word, we can find nothing about the home that could lead a reasonable observer to consider it designed to a practical degree for “transportation on water.”¹⁶³

Justice Breyer’s opinion also emphasized that actual use would be relevant. Lozman’s home was towed from location to location, and was rarely (if ever) used for transportation. Justice Breyer writes that “[t]his is far too little actual ‘use’ to bring the floating home within the terms of the statute,”¹⁶⁴ implicitly acknowledging that more robust actual use could bring the home within the meaning of the statute.

This Article’s analysis of artifact nouns illuminates the analysis and discussion in these two cases. In Stewart, the Court held that the Super Scoop was a vessel: The term also included vessels that have a practical possibility of transporting on water.¹⁶⁵ In linguistic terms, this is to emphasize the Potential (telic) facet. In Lozman, the Court held that Lozman’s floating home was not a vessel: The term did not apply to all floating structures, though it did apply to those that are designed to transport.¹⁶⁶ This emphasizes the Design (agentive) facet.

This linguistic analysis also clarifies the source of seeming contradiction between Stewart and Lozman. If the Super Scoop was a vessel, surely a floating mobile home was also one? Both can transport people and goods on water, though neither’s primary purpose is to do so. Part of what explains the difference between the opinions is that Justice Thomas emphasizes the Potential (telic) facet (what can practically operate as a vessel) while Justice Breyer emphasizes the Design (agentive) facet (what is designed to be a vessel).

In these cases, and others concerning artifact nouns,¹⁶⁷ a proper understanding of artifact nouns illuminates legal decision making.

B. Statutory Interpretation Theory

The case study presented here also has broader implications for statutory interpretation. First, the Article’s analysis illustrates the limitations of “word sense disambiguation” (“WSD”). WSD is a common approach in current textualist discourse and the foundation of contemporary legal corpus linguistics; but, we argue, this approach has important limitations. Second, the case study illuminates current interpretive debate about textual clarity and ambiguity.

1. Beyond Word Sense Disambiguation

Contemporary textualist theory emphasizes word sense, often asking: Does the statutory term express sense A or sense B? Important distinctions, like ordinary versus technical meaning, are framed as ones about competing word senses.¹⁶⁸ Traditional tools, like dictionaries, are often employed to distinguish between competing word senses.¹⁶⁹ New tools, like corpus linguistics, are also framed in relation to sense differentiation.¹⁷⁰

A broader lesson of this Article’s analysis is that textualist interpreters should think beyond what we call the WSD paradigm. This paradigm views interpretation through the lens of selecting the right “sense” of a contested word, among various competing senses.

Some textualists might be tempted to analyze VanDerStok as a WSD problem. This analysis would begin by positing two separate senses of firearm: a narrower “completed sense” and a broader “completed-or-uncompleted sense.” Next it would analyze which of those senses is the ordinary meaning. This might involve quantifying which sense is more frequent in ordinary language.¹⁷¹

Such a frequency approach is problematic for many reasons,¹⁷² and it would be unhelpful to conduct a corpus linguistic analysis reporting that most ordinary uses of firearm or weapon refer to completed or uncompleted entities. The question in VanDerStok is about the contextual meaning of a statutory definition, not the meanings that firearm or weapon most often take in literature or on the internet. Just because most (or even all) uses of table in a corpus refer to completed tables does not imply that an unassembled IKEA table is outside of the ordinary meaning of that term in all contexts. And even if most uses of pencil in a corpus refer to finished (i.e., sharpened) pencils, this does not imply that an unfinished (i.e., unsharpened) pencil must fall outside of the term’s ordinary meaning.

This Article’s analysis also challenges textualism’s familiar WSD paradigm in a more fundamental way. A term like bank has separate senses: (1) land beside a body of water; (2) a financial institution.¹⁷³ But there is no basis to assert that an artifact noun like firearm has a “completed sense” and an “uncompleted sense.” This distinction applies to many artifact nouns: firearm, weapon, table, chair, house, hammer, etc. A parsimonious explanation of this phenomenon (and the one arrived at by dictionary-makers¹⁷⁴) is that (non-)completion simply does not warrant specification as part of the listed senses of these nouns. Firearm or table does not usually have a distinct “complete” versus “uncompleted” sense listed in a dictionary; instead, as discussed in Section IV.A, it has an underspecified sense, the complete meaning of which depends on context.¹⁷⁵

This one-sense vs. multiple-sense distinction matters for textualist theory.¹⁷⁶ On the multiple-sense view, the interpretive debate may center on whether Congress has clearly enough replaced the “ordinary sense” of firearm with a technical sense. But on the one-sense view of artifact nouns, Congress is not “overriding” ordinary meaning in definitions like the GCA’s; with these artifact nouns, context is necessary to specify what is relevant. This is what occurs in part (A) of the statutory definition.¹⁷⁷ This provides a different reason from those mentioned above to reject a corpus linguistic study of the “common” use of firearm in ordinary language. On the one-sense view, every example demonstrates the same ordinary meaning of firearm, an ordinary meaning that must be further specified by context.

Textualists should not abandon word sense disambiguation, but they should grapple with the existence of other linguistic phenomena besides lexical ambiguity. Current textualism often operates with the hammer of WSD, viewing many interpretive problems as lexically ambiguous nails. But not all linguistic indeterminacy is the product of lexical ambiguity. Treating most linguistic indeterminacy as lexical ambiguity likely contributes to interpretive overreliance on the choice among dictionary definitions. Some definitions, like those of firearm, vessel, or vehicle, explain one sense that is context-dependent. The approach advanced in this Article illustrates how textualism can make progress in interpreting one sense of an artifact noun, in context.

2. Clarity vs. Ambiguity

Consider one more implication of the overreliance on the WSD paradigm. Begin by noting that textualists often use the word “ambiguity” to refer to any linguistic indeterminacy (including, for example, indeterminacy caused by vagueness, polysemy, or underspecification). For textualists, “ambiguity” is the opposite of clarity. For linguists, lexical ambiguity is simply one type of linguistic indeterminacy, and its proper resolution may or may not be clear from context.¹⁷⁸ Even when context serves to resolve the lexical ambiguities of a text, there may be other unresolved indeterminacies that serve to make the text unclear.

For this reason, conflating linguistic ambiguity with textual indeterminacy can lead textualists astray. The typical textualist approach to clarity vs. ambiguity is to draw a line between competing word senses: If the statutory term could express sense A or sense B, we will treat it as “clear” so long as we’re 90% (or 65% or 55%) certain it expresses A; otherwise, it is ambiguous between A and B.¹⁷⁹ When faced with a true case of lexical ambiguity (in context, does bank refer to a financial institution or to a river bank?), this is a sensible paradigm. But when faced with other sources of indeterminacy, this paradigm is less useful.

In contrast, Justice Kavanaugh has proposed a “best reading” standard: Judges should determine whether reading A or reading B is the better reading.¹⁸⁰ This idea animates Loper Bright Enterprises v. Raimondo in which the Court overruled Chevron deference to agencies adopting reasonable interpretations of “ambiguous” statutes and replaced it with a command for judges to implement the statute’s best reading.¹⁸¹ For cases like VanDerStok, this “best reading” paradigm is a better fit than a paradigm that seeks to adjudicate between two senses of a term. To the question, “Are terms like firearm, weapon, and frame ambiguous between two senses in the GCA?,” there is an easy answer: No. The relevant question is about the contours of the senses of these terms that the GCA expresses: Does the statutory context indicate that the Formal, Potential, and/or Design facets are most relevant? In Justice Kavanaugh’s terms, the linguistic question is one that should concern the “best reading” of an entirely unambiguous sense of firearm, weapon, frame, and receiver.

Attending to this distinction has implications for the rule of lenity and the degree of indeterminacy required to trigger it.¹⁸² The rule of lenity holds that ambiguities in criminal statutes should be resolved by construing the statute favorably to the defendant. Justice Gorsuch’s trigger for lenity occurs when traditional interpretive tools yield no clear answer.¹⁸³ Justice Kavanaugh’s trigger is a grievous ambiguity,¹⁸⁴ i.e., an unresolvable indeterminacy for which there is no “best” reading. Understanding the linguistic issue in VanDerStok to involve artifact nouns clarifies that there is no grievous ambiguity in Justice Kavanaugh’s sense: There is rich statutory context that indicates which facets of meaning (i.e., the Potential and Design facets) are relevant.

C. Legal Philosophy

This Article’s analysis also speaks to broader philosophical issues. We discuss implications for (1) the famous “no vehicles in the park” debate and (2) the distinction between intensional and extensional definition.

1. The Hart-Fuller Debate: No Vehicles in the Park

Recall the famous debate between philosophers H.L.A. Hart and Lon Fuller.¹⁸⁵ Hart argued that a rule’s text resolves most cases, while Fuller argued that decisions always require looking to a rule’s purpose.¹⁸⁶ This debate introduced one of the most famous and well-known hypotheticals: Imagine a rule stating that “no vehicles may enter the park.” Hart insisted that this rule gives rise to debate about edge cases (e.g., a bicycle or a drone) but that the easy cases are resolved by the text alone: The rule prohibits cars and trucks.

Fuller disagreed. What if, asked Fuller, the city council installed as a memorial a non-functioning World War II truck in the park? Fuller suggested that this truck would be permitted by the rule. More importantly, Fuller thought that what explains this judgment is the rule’s purpose. This debate spawned decades of legal philosophical debate about the role of text and purpose in interpretation.

In recent years, empirical legal scholars have turned attention to this hypothetical. Experimental jurisprudence studies have examined ordinary people’s judgments about the “no vehicles” rule.¹⁸⁷ For example, one study confirms that there are easy and hard cases: Most laypeople, law students, and judges agree that a car is a vehicle; but all populations are divided about other entities, like drones and canoes.¹⁸⁸ Interestingly, the rates of disagreement are similar among populations; for example, judges are just as divided as laypeople are over whether a canoe is a vehicle.¹⁸⁹

One might conclude from the similarity among the populations (e.g., cars are vehicles; crutches are not) that there is a stable core of ordinary meaning. This is the result, presumably, taken by scholars who use these survey results as a baseline against which to assess the accuracy of new tools, like word embeddings¹⁹⁰ and ChatGPT,¹⁹¹ in measuring ordinary meaning. Alternatively, one might emphasize that the disagreement about some cases (e.g., canoe) reinforces that it is hopeless to achieve interpretive determinacy—at least in hard cases.

A different interpretation of these data is that they underscore the importance of context and illustrate the limits of relying exclusively on individual words in interpretation (and tools and paradigms that emphasize word meaning, like dictionaries and early approaches to legal corpus linguistics). Interpreting “no vehicles may enter the park” requires consideration of more than just the word vehicle. Empirical studies demonstrate this effect. Laypeople evaluate what is a vehicle under the rule “no vehicles in the park” differently from their evaluation of what is simply a vehicle.¹⁹² Moreover, changing the background context affects people’s understanding of the category boundaries of artifact nouns like vehicle in a rule “no vehicles in the town square.”¹⁹³

This Article’s case study in artifact nouns enriches this second, contextual interpretation. The meaning of artifact nouns is context dependent, influenced by the object’s functional potential and design. For rules that regulate conduct, these facets—functional potential and design—are often closely connected with the purpose of the rule. For example, imagine “no vehicles may enter the park” is passed for the purpose of keeping the park safe. Can a bicycle enter the park? A bicycle may not be considered a vehicle in all contexts; however, its design and functional potential characteristics are such that it can move so quickly as to injure someone. In our hypothetical example, context—in particular, policymaker objectives—identifies these dimensions as the relevant facets of vehicle, which in turn may suggest a particular interpretive result, namely that the “no vehicles” rule prohibits bicycles. Of course, if noise or emissions are the policy concerns, this cuts the other way.

Other theoretical frameworks suggest similar interpretive outcomes. Hart suggested that purposes play a role outside of the core of settled meaning, in the penumbra.¹⁹⁴ And Eskridge proposes that “[t]ext and purpose are like the two blades of a scissors; neither does the job without the operation of the other.”¹⁹⁵ Recent work in cognitive science concludes that laypeople use both a rule’s text and purpose when evaluating rule violation.¹⁹⁶

This Article’s theory is subtly, but importantly, different from these other views that emphasize context and purpose. Our point is not that a rule’s purpose informs people’s application of a rule,¹⁹⁷ or that purpose can override text in cases of conflict. Our point is that there is a more fundamental problem with the “text vs. purpose” dichotomy.

This “text vs. purpose” dichotomy is at the heart of many discussions of the Hart-Fuller debate. Consider, for example, legal philosopher Fred Schauer on the World War II Truck example, asking does a non-functioning truck count as a vehicle for the purposes of the rule?:

The war memorial made out of a functioning military truck really was a vehicle…It was Fuller’s point that language could not…ever be sufficient to produce a core or clear case, because in at least some instances the clear application of clear language would nonetheless produce an absurd result. Only by always considering the purpose behind the rule, Fuller believed, could we make sense of legal rules and indeed of law itself.¹⁹⁸

This is one of many examples illustrating the traditional understanding of the debate about vehicles in the park: First, we begin with some clear cases of vehicles (e.g., trucks). Both Hart and Fuller endorse this conclusion.¹⁹⁹ Hart thinks that for these clear or easy cases, text alone resolves interpretation.²⁰⁰ Fuller disagrees, claiming that we always need to consider purpose, even for some “easy” cases, such as the example of a non-functioning truck, which is a truck.²⁰¹

We seek to question this central text vs. purpose dichotomy. For example, we reject the assumption that a truck is simply within the core of the meaning of vehicle in all contexts. On Schauer’s account above, this is a starting point that both Hart and Fuller share. Everyone agrees that a truck is a vehicle. The disagreement simply concerns whether text (linguistic meaning) trumps purpose.

To elaborate our account, recall our hypothetical law with the text “no vehicles may enter the park,” passed for the purpose of keeping the park safe. Now imagine that a commemorative, non-functional truck is installed in the park. One way to analyze this case is to posit a conflict between text and purpose:

Text: The rule is violated because any truck is a vehicle.
Purpose: The rule is not violated because this truck does not threaten the safety of people in the park.

Did the truck violate the rule by entering the park? The traditional understanding of the Hart-Fuller debate would see the answer “yes” as evidence of purposivism. A truck is simply a vehicle, as the quotation from Schauer above illustrates. What leads us to say that this truck is permitted is a desire to override the clear linguistic meaning with consideration of purpose or consequences.

This Article’s theory of artifact nouns and their facets of meaning supports a different interpretation. If the World War II truck is allowed by the rule, it is not necessarily because people elevate the rule’s purpose over its clear text, which includes all trucks. Instead, we propose the meaning of vehicle (i.e., the “text”) is specified by context. And the stated purpose of the rule is relevant context in understanding what vehicle means. In other words, people construe vehicle to exclude this non-functional truck because they understand the in-context meaning of vehicle to make relevant a vehicle’s functionality—not because the rule’s purpose overrides the common truck-inclusive meaning of vehicle.

Similarly, if bicycles are prohibited by the “no vehicles” rule in our hypothetical, it is not necessarily because the rule’s purpose fills in gaps left open in its “core” textual meaning, nor because considerations of purpose are weighed against the rule’s plain text. These explanations presuppose that the words of the rule are sufficient to determine a linguistic interpretation, one which may be augmented or overridden by purposive considerations. But out of context, there is no complete interpretation of vehicle in “rules about vehicles” and hence no interpretation of such rules.²⁰² In our “no vehicles” hypothetical, text and purpose must jointly conspire to produce an interpretation: That is, purposive considerations resolve an indeterminacy in vehicle which is not otherwise resolvable in the provided context. This suggests that, at least insofar as artifact nouns are concerned, the “text-purpose dichotomy” assumed by previous accounts offers an incomplete, and perhaps even misleading, starting point for theorizing about legal interpretation of artifact nouns.

Finally, we note a second important way in which “text” and “purpose” are complementary, rather than oppositional, notions when it comes to artifact noun interpretation. With artifact nouns, context can indicate the relevance of design (i.e., an object’s intended purpose) and functional potential (i.e., its possible purpose). These facets of meaning are relevant to the textual or linguistic analysis of artifact nouns. In this sense, purposive considerations are at the heart of an artifact noun’s meaning.

In concluding this discussion, let us return briefly to Hart and Fuller.²⁰³ Hart claimed that for some cases (“easy” cases, in a core of settled meaning), no consideration of morals or purpose is necessary. Fuller claimed that purposive reasoning is needed in all cases.²⁰⁴ Although Hart and Fuller debated “legal interpretation” generally, their extended discussion concerned one rule about vehicles in the park; we limit our conclusions to that case and the class of interpretive problems centered on artifact nouns.

At first our theory might seem to provide new support for Fuller. We have argued that artifact nouns like vehicle are necessarily context-dependent. In other words, there is no “core of settled meaning” for the term vehicle, in the sense that even some trucks or cars could fall outside of the contextual meaning of vehicle. We always need to consider context to specify the term’s meaning.

However, insofar as Schauer’s reading of Fuller is right, and Fuller agreed that the World War II truck is always within the meaning of vehicle,²⁰⁵ our results chart a third path forward. In some cases, in which it appears that purpose trumps text, what really occurs is that the linguistic context indicates the relevance of an artifact noun’s design or potential function. Assume that someone agrees that a commemorative, non-functioning truck does not violate the “no vehicles in the park” rule,²⁰⁶ passed for the purpose of reducing park emissions. This is not necessarily a demonstration of intuitive purposivism, in which the person elevates a rule’s purpose over its text. Instead, our theory posits that this is a demonstration of the person’s sophisticated attention to context. The non-functioning truck is simply not within the contextual meaning of vehicle.

2. Intensional vs. Extensional Definition

The analysis here also illuminates the relevance of the distinction between intensional and extensional definition, which divides textualists on the Supreme Court.²⁰⁷ Terms like vehicle can be defined extensionally through the set of entities falling under the term. For example, extensionally defined, vehicle denotes the set of car, truck, bus, etc. An intensional definition characterizes vehicle through its features: For example, a vehicle is an entity that can self-propel and carry an object. In this section, we briefly explain how this case study in artifact nouns further illustrates the usefulness of intensional definition.

Consider the “original meaning” of the disputed provisions in VanDerStok (i.e., the 1968 meaning). The Gun Control Act was passed in 1968, and the provisions disputed in VanDerStok were not affected by later amendments. “Gun parts kits” were not manufactured and sold until decades later. An extensional approach to the original meaning of firearms might suggest that gun parts kits are not within that meaning. This approach might be buttressed by research demonstrating that it was rare or uncommon to refer to “gun parts kits” or parts as firearms in 1968.

In contrast, an intensional approach might begin by identifying the features that characterized firearm, such as those that have a high cue validity for firearm in 1968.²⁰⁸ For example, the feature “having a beak” has a high cue validity for the category of birds: Many (all) birds have beaks, and most non-birds lack beaks. For firearm, the GCA’s definition itself identifies plausible original features that have a high cue validity: a weapon which will expel a projectile through the action of an explosive; a weapon which is designed to expel a projectile through the action of an explosive; and a weapon which may be readily converted to expel a projectile through the action of an explosive. These features identify weapons that are designed to or may readily be converted to fire (e.g., an unloaded gun or a gun separated from its magazine), overlapping with the original 1968 extensional definition of firearm. But these features also identify weapons that fit the intensional criteria but which no one would have pointed to in 1968, such as firearms that were only invented later.

The existence of intensional definition counts against an “originalist” argument that seeks to reject the government’s view in VanDerStok on the basis that the 1968 meaning of firearm excludes these kits simply because no one associated such kits with “firearms.” One of us has proposed that, for textualists, intensional definition, rather than extensional definition, is generally a more promising approach to interpretation.²⁰⁹ The details of this debate take us too far from the analysis offered in this Article, but the more modest point here is that originalism about statutory meaning does not require rejecting category members that no one “expected.” One can adopt intensional originalism.

Readers familiar with debates about constitutional originalism will recognize a correspondence between “original expectations originalism” and extensional definition and “original public meaning originalism” and intensional definition.²¹⁰ Consider questions like: Could the original meaning of arms in the Second Amendment include arms that no one would have expected at the Founding; could the original meaning of “unreasonable searches” include searches of cell phones; could the original meaning of a statute from the 1850s concerning vehicles apply to airplanes invented decades later? It is easy to answer “yes” to these questions with intensional definition. An original 1850 intensional definition of vehicle could include airplanes, even if the original extensional definition excludes them. In constitutional law, modern public meaning originalists have gravitated towards the intensional approach, although recent emphasis of “history and tradition” pulls some analyses in the direction of a more extensional analysis.

Again, here we do not defend intensional over extensional definition for interpretation—although it is difficult to square many originalist applications with a purely extensional theory of original meaning. However, recognizing the availability of intensional definition provides an avenue to accommodate the arguments of this Article within an originalist approach to statutory meaning.

Considering intensional definition also underscores a second point about context sensitivity. Our analysis of artifact nouns is highly context sensitive, and some might object that language is not. Some things cannot be firearms in any context. And firearm calls to mind a certain thing, a gun, suggesting that its meaning does not entirely depend on context.

Our account is consistent with these intuitions. The first intuition is that there are some things that cannot be category members in any context. A hunk of metal is never a firearm, and an egg is never a motor vehicle. The second intuition is that these nouns seem to have a stable conceptual core that licenses generalizations. For example, vehicle calls to mind a certain set of features, and we are inclined to say that a truck is a better example of a vehicle than a bicycle even if a bicycle is also a vehicle in some contexts.

Our theory of artifact nouns is highly context sensitive, but it also accommodates these intuitions. First, consider the intuition that some entities are not category members in any context such as a hunk of metal as a firearm. Insofar as none of the facets of the artifact noun’s meaning are applicable to this entity, this conclusion is consistent with our theory. Category members can also be ruled out when context indicates the relevance of certain facets of meaning. Recall that in the context of the GCA, a toy gun or cake that looks like a gun frame is not a firearm. In other contexts that emphasize certain formal and constitutive properties, those entities would be firearms. Moreover, in other contexts, the facets of meaning will be specified in a way that rules out certain entities. For example, imagine that in the middle of a war, a soldier yells to another, “I’m all out of ammo, pass me a firearm.” The functional facet in this context is not the same as it is in the GCA. Here, the soldier refers to a firearm that will immediately shoot. Finally, the modal “readily” in “readily be converted to” is necessarily context sensitive. It would be strange to understand this term to imply that only objects readily convertible in 1968 are included. One way to interpret this phrase is as a hearer-centric term,²¹¹ communicating that originally the meaning of “readily” should be completed by its current hearer. Another possibility would be to give the term an intensional meaning.

Second, consider the intuition that artifact nouns have a stable conceptual core. This idea is premised on intensional definition—there are essential features that define these nouns. This proposal is intuitive and runs through Schauer’s analysis of the “no vehicles” hypothetical: There is something that vehicle calls to mind like a car, so the term must have a stable core of context-invariant meaning. We agree with the first part of this claim. The term vehicle calls to mind certain properties, and it is sensible to use these properties to endorse statements about artifact nouns that cut across contexts: A car is generally a better example of a vehicle than a wheelchair. An explanation for this phenomenon is prototype theory,²¹² which posits that certain category members are seen as more “central” than others. One way to elaborate this idea is through the possession of category features, in a way that is similar to the proposal to use category features as a basis for intensional definition. On this theory of artifact nouns, there is still substantial context sensitivity. In some contexts, a non-functioning truck is a vehicle, but in others it is not. But there is also a compatible explanation for our sense that, in general, a truck is a good example of a vehicle.

VII. Conclusion

Extending a scholarly tradition at the intersection of law and linguistics,²¹³ this Article employed linguistic theory and empirical methods to enrich current textualist theory and analysis.

As a specific case study, we answer a question about the GCA’s meaning—the question at the center of Garland v. VanDerStok. The Supreme Court’s decision will impact the regulation and accessibility of firearms across the country.

More broadly, this analysis illustrates how textualists should reason about artifact nouns. When it comes to the artifact nouns in VanDerStok, like weapon, firearm, frame, receiver, and those elsewhere like vehicle or vessel, context informs interpretation by identifying the relevant essential facets of those nouns’ ordinary meanings. We showed that this analysis has implications not only for the interpretational disputes in VanDerStok but also for theories of statutory interpretation more generally.

We supported this theoretical analysis with new empirical data which demonstrates that ordinary Americans use and comprehend artifact nouns in a way that is both flexible and systematic. This Article documents over one hundred examples of people describing weapon parts kits as a “firearm,” “weapon,” “rifle,” or a similar term. It also presents an original survey experiment, supporting the same interpretation. These findings support the broader analytical framework explored in this article.

The Article has concrete practical implications for the Supreme Court’s analysis of VanDerStok, a case which has profound implications for the regulation of “ghost guns” or “gun parts kits,” which have been involved in several mass shootings. This Article’s extended case study also has implications for statutory interpretation practice, clarifying the interpretation of artifact nouns in many other cases, and theory. For example, the results illustrate how textualists should supplement the common “word sense disambiguation” paradigm. Finally, the results have broader implications for legal philosophy, providing a deeper explanation for a resolution between textualists and purposivists.

[*] Fritz Postdoctoral Fellow, Georgetown University

[**] Professor of Linguistics, Stanford University

[***] Professor of Computer Science, Brandeis University

[****] Professor of Linguistics and Computer Science, Georgetown University

[†] Professor of Law, Georgetown University Law Center. For excellent editorial assistance, thanks to the Harvard Journal on Legislation, especially Marissa Medici and Mia Berman; for outstanding research assistance thanks to Alexa Bieber, Connor Berry, Brendan Gallomore, Jake Irving, and Jack O’Hara; and for helpful comments, great thanks to Joseph Blocher, Beth Levin, and Richard Schmalbeck.

View PDF Version