The Volokh Conspiracy

Mostly law professors | Sometimes contrarian | Often libertarian | Always independent

corpus linguistics

Corpus Linguistics and the Second Amendment

Support for the right to bear arms for all purposes


Corpus linguistics is the scholarly technique of searching historic databases to gather information on the use of important words or phrases. In the pending U.S. Supreme Court on the Second Amendment to right to bear arms, New York State Rifle & Pistol Association v. Bruen, a pair of amicus briefs purport to apply corpus linguistics to the Second Amendment. The briefs say that they prove that individuals have no right to bear arms, and that even if such a right exists, it is tiny. This post examines the claims in the briefs.

This post is co-authored by Campbell University law professor Gregory Wallace. Professor Wallace and I are two of the five co-authors of the just-published third edition of the textbook Firearms Law and the Second Amendment: Regulation, Rights, and Policy (Aspen, Wolters Kluwer). Professor Wallace wrote the textbook's section on corpus linguistics and spoke at a colloquium on the subject at the Duke Center for Firearms Law.

One of the amicus briefs is on behalf of three professors of linguistics—Dennis Baron (U. Illinois), Stefan Th. Gries (U. Cal. Santa Barbara), and Jason Merchant (U. Chicago)—and one law professor, Alison LaCroix (U. Chicago), who has written about corpus linguistics and founding era documents. It was filed by attorneys for Morrison & Foerster. The other brief is by and for Washington, D.C., attorney Neal Goldfarb. Goldfarb describes himself as "an attorney with an interest and expertise in linguistics, and in applying the insights and methodologies of linguistics to legal interpretation." His brief asks the Supreme Court to call for supplemental briefing on the corpus linguistics issues and to hold the Bruen case over to the next Term for argument on those issues. The arguments in both briefs are similar.

To be clear, we do not criticize corpus linguistics as a methodology. Volokh Conspiracy Professor Randy Barnett is, according to both the Goldfarb and Professors briefs, a pioneer in the field of corpus linguistics. But at the time, he did not identify as a corpus linguist; he just thought he was doing legal research.

Professor Barnett believes that the original meaning of the Second Amendment includes a right to bear arms for personal defense. In the New York State Rifle & Pistol Association case, he joined an amicus brief to that effect, filed by Kopel and other lawyers.

The persuasiveness of corpus linguistics claims depends on understanding words in context, considering all relevant sources, and classifying usages accurately. Some corpus linguists do so better than others.

James Madison's meanings of "bear arms"

Goldfarb accurately cites an essay by an eminent etymologist showing that by the end of the 1600s, the French import carry largely had supplanted bear as word for personally transporting an object. Philip Durkin, Borrowed Words: A History of Loanwords in English 408 (2014).

The Goldfarb and Professors briefs report the results of their searches for the phrase "bear arms." They say that most often, the phrase was used in some sort of military context.

According to the Professors, no evidence from the Founding Era shows an "individualized connotation" such as "bear a rifle," "bear a musket, "bear rifles," or "bear muskets." They overlooked an obvious example—an example from the man who wrote the Second Amendment. In 1785, James Madison, serving as a member of the Virginia House of Delegates, introduced an anti-poaching "Bill for Preservation of Deer." The bill had been drafted by a Committee of Revisors, whose members included Thomas Jefferson, George Wythe (signer of the Declaration of Independence and the Constitution; Chancellor of Virginia; distinguished law professor who taught Jefferson, John Marshall, and St. George Tucker), and Edmund Pendleton (previously Speaker of the House of Delegates and later first President of the Virginia Supreme Court of Appeals). The bill can be found in volume 2 of The Papers of Thomas Jefferson 443-44 (Julian P. Boyd ed., 1950).

While allowing unlimited hunting on one's own land, the bill established seasons for deer hunting elsewhere. Upon conviction, the poacher would have to pay a fine and to post bond for good behavior. And for the next year, he would be forbidden to carry a long gun outside his property. If he did carry a long gun off his property, the bond would be forfeit, and he would have to post a new bond:

Whosoever shall offend against this act, shall forfeit and pay, for every deer by him unlawfully killed, twenty shillings, one half thereof to the use of the commonwealth, and the other half to the informer; and moreover, shall be bound to their good behavior; and, if within twelve months after the date of the recognizance he shall bear a gun out of his inclosed ground, unless whilst performing military duty, shall be deemed a breach of the recognizance, and be good cause to bind him anew, and every such bearing of a gun shall be a breach of the new recognizance and cause to bind him again. (emphases added).

At the time, "gun" meant a long gun, as distinct from a "pistol." Noah Webster, An American Dictionary of the English Language (1828) (defining gun, and explaining, "But one species of firearms, the pistol, is never called a gun"). So the bill would not have forbidden convicted poachers to carry handguns in public.

In the poaching bill, we see how the Father of the Constitution and author of the Bill of Rights (Madison), the author of the document that created the United States (Jefferson), America's leading law professor (Wythe), and a future chief justice of Virginia (Pendleton) understood the word "bear." The word could mean to carry a gun for personal reasons, such as hunting. It could also mean to carry a gun while serving in the militia ("whilst performing military duty"). And it could encompass both meanings in the very same sentence.

No wonder Madison chose the word "bear" when he wrote the Second Amendment. It was the perfect word to express the right to arms for all personal use and for community service in the militia.

The Goldfarb and Professors briefs are based on the faulty premise that "bear" can only mean one thing at a time. Either it means personal use or it means community military use; it can't mean both at once. James Madison didn't accept the false dichotomy, and neither did the Virginia Committee of Revisors.

Counting wrong

Having created the dichotomy, the two briefs announce their quantitative results for uses of "bear arms." The military uses are the more common. The professors classify their dichotomy as "individual" versus "collective." In their system, plurals go into the "collective" category. An example from their brief is "Slaves were not permitted to bear arms." This is the sort of evidence from which they conclude that "the right to keep and bear arms" is solely a "collective" right, and not an "individual" right. In fact, collectives are not exclusively militaristic, and plurals often are used to describe two or more individuals. The quoted sentence tells us that slaves, as individuals, were forbidden to bear arms, and that slaves, as a class, were forbidden to bear arms. That sentence, like the Second Amendment, is not only about collectives or only about individuals. It is about both.

The Alleged Ambiguities

We examined the four quotes from the Professors brief that they showcased as examples of ambiguity. The Professors summarize the meaning of their four quotes:

Even examples that were "at best ambiguous" still "suggest[ ] a military or quasi-military sense of bearing arms."  Id. [Baron's Hasting Const. L.Q. article] at 512.  More importantly, these examples show  that  the  bearing  of  arms  was  subject  to  regulation, including limitations on where those arms might be carried. (second brackets added).

From the renowned philosopher Samuel Clarke's A New Description of the World (1689), the Professors brief quotes the line: "'A Peasant in this Country (unless in time of great Danger or Invasion) is not suffered to bear Arms.'[1689]." Here's the longer version of what Clarke wrote, describing conditions in Poland:

Towns are very scarce, the People dwell in Hutts of Straw and Loam, with Holes at the top of them to let in the Light, and give vent to the Smoak, living in miserable Poverty, as do (for the major part) the Peasants or inferiour People, being in a manner Slaves to the great ones, whose Tenants they are, though the Richer Sort are very Profuse and Expensive, rather Prodigal than Liberal, Impatient of Injuries, Delicious in Diet, and costly in Attire, often shaving their Heads, except one Lock, which they preserve with great care, being generally good Soldiers, and much Glory if they can kill a Turk in Battle, and bring off his Head, in Token of which, they wear Feathers in their Caps.

Theft in these Parts is very rare, especially to be committed by a Native Polander, and all Crimes are severely punished: As for the Religion they Profess, in Relation to the Government, it is that of the Romish Church, though the Reformed way of Worship is allowed and tolerated; and the King is of late Elective; the Women are tolerably Fair and well Proportioned, very Witty and Ingenious, great admirers, and observers, of their Husbands, and very neat in their Houses. A Peasant in this Country (unless in time of great Danger or Invasion) is not suffered to bear Arms; and when the Gospel is Read in the Churches, the Gentry and Nobility draw their Swords, in token that they are ready to defend it with their Lives.

Clarke, at 77.

According to Clarke, Polish peasants cannot bear arms in ordinary life. They only bear arms during wartime. The gentry and nobility, however, do carry swords, and show them off in church. Accustomed to routine arms bearing, the upper classes are good fighters. The tenant farmers live in squalor while their armed landlords are hot-tempered and spend their wealth ostentatiously. The peasants are "in a manner Slaves to the great ones."

Unambiguously, Clarke uses "bear arms" for personal carry in ordinary situations, and carry during wartime. The people who cannot ordinarily bear arms are the slaves of those who do.

In the Professors brief, the Clarke quote is said to show that "bear arms" connotes lots of government control. And so it does, in an unfree nation like Poland, where most of the people are forbidden to bear arms. The Second Amendment adopts the opposite policy for the intended opposite result: When "the people" can "bear arms," there can be "a free State."

Here is a decree from about 1330 by King Edward III, which the Professors brief labels "ambiguous":

"That no person shall use or bear any Arms within London, and the Suburbs, or in any place between the said City and Pallace of Westminster, nor in no other part of the Pallace by Land or by Water, except such of the Kings people, as he shall appoint to keep the Kings peace." [1657].

The brief's "1657" cite is misleading. The quote is from King Edward III in the third year of his reign, which began in 1327. It was reprinted in a 1657 book that collected royal historical documents. King Edward's decree clearly tells people, as individuals, not to walk around London carrying swords or other weapons.

Again, "bear arms" includes personal carrying, and carrying while in government service. That is why the king's broad language against "bear any Arms" needed to state an exception for people in government service.

In the purported dichotomy between personal self-defense and collective self-defense, the phrase "defence of his Family" would seem unambiguously about the former. But the Professors brief says that the next quote is ambiguous:

"That every Person who will go for Ireland on these Conditions, shall out of his first share of Money, buy for himself and every Relation and Servant that he carries with him (who are able to bear Arms,) a good Musket, or Case of Pistols for the defence of his Family." [1690].

This comes from Sir Richard Buckley, The Proposal For Sending Back The Nobility And Gentry Of Ireland. Buckley proposed that Irish nobility living in England should move back to Ireland. They would receive enormous financial grants from the English government, if they brought certain quantities of Irish land under cultivation. Pursuant to the rest of the sentence (not quoted in the brief), the Irish nobles moving back to Ireland had to acquire their guns in England, not in Ireland:

without the producing of which Arms in Ireland, and the making Oath that he himself bought the same in England, he shall have no Right to receive any of the said Forty Five Thousand Pounds in Ireland.

Buckley also proposed inducements for ordinary English farmers to move to Ireland. Besides a financial bonus, they would receive "a good Musket or Case [matching pair] of Pistols at Chester and Bristol."

The Buckley proposal plainly promotes the bearing of arms for personal defense. In 1689-90, the people of Ireland revolted against English control, in the Jacobite Rebellion. In Ireland, English immigrant farmers and English-allied nobles were not exactly popular, so of course they would need arms for personal defense. All their guns should come from England; the ruling country could have gun commerce, but not the colonized country.

The final example of ambiguity in the Professors brief is the 1776 Pennsylvania Constitution:

That the People have a Right to bear Arms for the Defence of themselves and the State, and as standing Armies in the Time of Peace are dangerous to Liberty, they ought not to be kept up: And that the Military should be kept under strict Subordination to, and governed by, the Civil Power." [1776].

Justice Stevens did not think the Pennsylvania Constitution was ambiguous. In his Heller dissent, he wrote that if the Second Amendment had copied the Pennsylvania text, the Second Amendment clearly would have included personal defense, as well as a militia right. 554 U.S. 570, 647 (2008) (Stevens, J., dissenting).

The four "ambiguous" quotes discussed above were the four that the Professors brief selected to showcase their methodology. None of the quotes is ambiguous. Every one recognizes individual bearing of arms, and three of the four also recognize collective bearing of arms. If all the Professors star examples are inaccurate, can one trust the accuracy of their other classifications?

The literal Second Amendment

The Neal Goldfarb brief acknowledges that Goldfarb's claims were criticized in Josh Jones, Comment, The "Weaponization" of Corpus Linguistics: Testing Heller's Linguistic Claims, 34 BYU J. Pub. L. 135, 171 (2019). However, writes Goldfarb, "when one puts aside the uses categorized by this researcher as ambiguous, one sees that he categorized only 21% of the relevant uses as unambiguously literal."

By "unambiguously literal," Goldfarb means individuals carrying weapons for personal reasons. This is contrasted with what Goldfarb calls the "idiomatic" use of "bear arms" to connote a military situation. Because military 79% is more than individual 21%, Goldfarb argues that the former is the sole meaning of "bear arms" in the Second Amendment.

On the other hand, it is usually reasonable in constitutional interpretation to read the text as including at least its literal meaning. The First Amendment guarantees "the freedom…of the press." Literally, it is the freedom to own a printing press, and a rejection of earlier English law that individuals could only own a printing press if the government granted them a license. Edward  Lee,  Guns  and  Speech  Technologies:  How  the  Right  to  Bear Arms  Affects  Copyright  Regulations  of  Speech  Technologies,  17  Wm.  &  Mary  Bill of  Rts. J. 1037, 1059–64 (2009). In addition to the literal meaning, freedom of the press includes protection of many forms of communication, not only those that involve printing. For example, the First Amendment severely limits prior restraints on the distribution of all publications, including those written by hand, even though prior restraints do not interfere with people buying and using presses. Kopel, The First Amendment Guide to the Second Amendment, 81 Tenn. L. Rev. 417 (2014)

The Fourth Amendment guarantees "The right of the people to be secure in their…houses…" Literally, the guarantee applies to houses. Additionally, the guarantee applies to places that are "houses" in a more figurative sense—such as camping tents, apartment buildings, and hotel rooms.

The unambiguously literal meaning of "the right of the people peaceably to assemble" is individuals gathering in a single location. But "assembly" also has other meanings, such as association in many locations. The millions of members of the worldwide Christian denomination known as "The Assembly of God" never all literally assemble in the same place. As members of a common faith, they are an "assembly" in a figurative sense. The U.S. Supreme Court has accurately recognized that the right of association is a necessary implication of the literal right of assembly. See NAACP v. Patterson, 357 U.S. 449, 460 (1958).

Constitutional rights language can encompass both the literal and the figurative. Scholars could search the historic databases and count how many times "press" is used literally, to refer to printing presses. Then the scholars could count how many times "press" is used figuratively, to refer to diverse forms of oral or written communication. The scholars could then announce that whichever usage comes in first is the only meaning of "press" in the First Amendment. These First Amendment scholars would be following the example set for the Second Amendment by the Goldfarb and Professors briefs.

The correct use of frequency data in corpus linguistics is different from the simplistic Professors and Goldfarb briefs. It is not just a matter of counting up uses and announcing that whatever use comes in first is the exclusive meaning. Frequency data can show the range of possible candidates for the ordinary meaning of the term, and perhaps even suggest a presumptive meaning if one sense is overwhelmingly predominant, and that sense necessarily excludes other senses. Frequency data by itself is not conclusive. Frequency data must be supplemented by relevant legal context.

Legal context

The most relevant context for the meaning of "bear arms" in the Second Amendment is contemporary state constitutions and proposals for constitutional revision. Founding Era constitutional sources used the phrase "bear arms" to describe the right to carry weapons for non-military purposes.

The Anti-Federalist dissenters report from the Pennsylvania constitutional ratifying convention in 1788 proposed an amendment to the Constitution protecting the people's "right to bear arms for the defence of themselves and their own state or the United States, or for the purpose of killing game." Three state constitutions—Pennsylvania, Vermont, Kentucky—written during the Founding period protected the people's right to "bear arms in defense of themselves and the state." Six additional state constitutions written prior to 1820 (Ohio, Indiana, Mississippi, Connecticut, Alabama, Missouri) contained protections for the right of the people to "bear arms in defense of themselves and the state" or the right of every citizen to "bear arms in defense of himself and the state." All of these plainly understood the right to bear arms to include civilian self-defense as well as to service in the militia.

Note the difference between the state constitutions and the 1788 Pennsylvania proposal. All of them list reasons for the exercise of the right. All of them name the personal reason of personal defense, and the community reason of collective military defense. Only the 1788 Pennsylvania proposal enumerates an additional personal reason, namely hunting. So it as least plausible that in the nine states, the state constitutional right to bear arms did not include hunting. Vermont and Pennsylvania had separate constitutional provisions for the right to hunt.

Suppose you wanted to write a constitution that protected the right to bear arms for all purposes: self-defense, militia service, hunting, other shooting sports, and so on? Then you would just write "bear arms," without any subsequent words of limitation. That is exactly how James Madison and the Committee used "bear" in their anti-poaching bill. Convicted poachers lost their right to bear long guns anywhere off their lands, for all purposes, including hunting, target shooting, and personal defense. The ability of the convicts to bear arms in the militia was saved only because the bill included an express exemption for military service.

When the Second Amendment was being debated in the U.S. Senate, Congress rejected a proposal to add a narrowing modifier: "for the common defence." If Congress had added the modifier, then interpretation of the Second Amendment would be more complex. Two state constitutions had similar qualifiers: Massachusetts ("for the common defence") and North Carolina ("for the defence of the state"). Later, some new states included similar language. The main line of 19th century state interpretation of such clauses recognized an individual right of all citizens to own and carry firearms. Several cases, especially in the latter 19th century, said that the right was only for militia-suitable arms—such as large handguns, rifles, swords, or cannons—and not for arms that (in the courts' understanding) would not be used in a militia—such as small handguns or Bowie knives. A minority of courts, such as Tennessee in Aymette v. State, said that the right to keep and bear arms for the common defense included everyone keeping arms, but bearing arms was only for the militia. See Kopel, What State Constitutions Teach about the Second Amendment, 29 N. Ky. L. Rev. 827 (2002); Kopel, The Second Amendment in the Nineteenth Century, 1998 BYU Law Review 1359.

The various state constitutions show different ways to limit, or not limit, the right to bear arms. The Second Amendment, with no qualifiers, takes the broadest approach.

Neal Goldfarb states that at least 95 percent of all uses of "bear arms" between 1760 and 1799 conveyed the idiomatic sense relating serving in the military. But Goldfarb excluded from his analysis all uses of "bear arms" when the phrase appeared in the text of a proposed or ratified constitutional provision at the federal or state level protecting that right, because he believed these uses irrelevant to his analysis.

He concedes that "[w]ith respect to the Second Amendment, the most contextually similar uses of bear arms are the state constitutional provisions." But he goes on to say that "those uses can't offer us much guidance" because the term "right of the people" used in those provisions may refer to a collective right, thus "in defense of themselves" could be understood as "the right of collective self-defense through service in the militia," rather than an individual right to keep and bear arms for self-defense. However, not a single court in any state with constitutional language like "bear arms in defense of themselves and the state" interpreted the right as encompassing only militiamen.

Goldfarb's retreat into a thinly-conceived ambiguity to dismiss the Second Amendment's most precise context—use in constitutions—suggests that his research may be skewed toward a particular result.

Private war

There is a final reason we doubt the accuracy of the briefs' claims about the numbers of usages in the supposedly mutually exclusive categories of personal self-defense and collective self-defense—with the latter based on a context of war: In the usage of the time, "war" included personal self-defense.

John Locke wrote that a criminal who attempts to murder, rob, or put an individual under the criminal's "Absolute Power, does thereby put himself into a State of War with him." In response, the defender "may destroy a Man who makes War upon him…for the same Reason, that he may kill a Wolf or a Lion." John Locke, Second Treatise of Government §§16-18 (1690).

In the most influential international law treatise of all time, Hugo Grotius explained:

if a Man is assaulted in such a Manner, that his Life shall appear in inevitable Danger, he may not only make War upon, but very justly destroy the Aggressor; and from this Instance which every one must allow us, it appears that such a private War may be just and lawful.

Hugo Grotius, 2 The Rights of War and Peace 397 (Richard Tuck ed., 2005, reprint of 1737 English translation) (1625). Likewise, "What we have hitherto said, concerning the Right of defending our Persons and Estates, principally regards private Wars; but we may likewise apply it to publick Wars, with some Difference." Id. at 416 (defensive private and public war are both morally permissible; public war may be undertaken for "revenging and punishing Injuries," but private war may not).

Swiss professor Jean-Jacques Burlamaqui was the first to declare a natural right to pursue happiness. His writings much influenced the American Founders. Burlamaqui argued that personal defense is essential to preservation of peaceful society, "otherwise the human species would become the victims of robbery and licentiousness: for the right of making war is, properly speaking, the most powerful means of maintaining peace." Jean-Jacques Burlamaqui, 2 The Principles of Natural and Politic Law 224 (Thomas Nugent trans., 2d ed. 1763) (1747 & 1751) (Pt. IV, ch. 1, ¶11).

Thus, to keep and bear arms for personal defense meant being prepared for a type of "war." The Goldfarb and Professors briefs give no indication that the authors have any awareness of the fact.

Corpus linguistics can be a valuable tool for legal scholars. Future scholars intending to employ corpus linguistics can usefully study the New York State Rifle & Pistol Association amicus briefs as models of errors to avoid: ignoring usages that don't support an author's theory, failure to understand that a words can have multiple meanings at once, separating phrases from context that clearly shows their meaning, not considering the most precisely relevant context (here, the use of words in constitutions), and imposing twenty-first century usage (e.g., "war" is national defense but not personal defense) on earlier generations who used words differently from how modern Americans do.