Duke Center for Firearms Law
Duke Law logo

Legal Corpus Linguistics and the Meaning of “Bear Arms”

By on July 16, 2021 Categories: , , ,

Over the past decade, research into the ordinary meaning of constitutional terms has been supplemented by corpus linguistics. There is obvious value in examining large databases of historical texts to determine how a particular group of people used a particular word or phrase at a particular time.

The text of the Second Amendment protects the right to “bear Arms.” The majority and dissenting justices in District of Columbia v. Heller disagreed over how the phrase “bear Arms” was understood in 1791. Justice Scalia, writing for the majority, read the phrase broadly to include protection for the carrying of firearms apart from military service (what Justice Scalia called its “natural” meaning). Justice Stevens, writing for the dissenting justices, read the phrase narrowly to protect only the carrying of firearms in connection with military service (what the majority and dissent called its “idiomatic” meaning). Both the Scalia and Stevens opinions relied on multiple original sources to support their conclusions, but, at the time, those sources were limited in number.

Since Heller, the creation of two databases—the Corpus of Founding Era American Usage (COFEA) and the Corpus of Early Modern English (COEME)—has enabled researchers such as Dennis Baron, Neal Goldfarb, Josh Blackman, James Phillips, and Josh Jones to analyze how the phrase “bear arms” was understood during the founding era (1760-99).

These researchers generally agree that “bear arms” was used mostly in its idiomatic or military sense during this period, but not solely or exclusively so. Baron concluded that the natural use of the phrase was “almost always” in a military sense, while Goldfarb found that nearly 95 percent of all uses of “bear arms” conveyed the idiomatic sense relating serving in the military. Blackman and Philips, examining a smaller sample, found that the “overwhelming majority” of uses of were in the military context. Josh Jones, using somewhat different search and coding parameters, found that the figurative or specialized military sense of “bear arms” was used in 66 percent of relevant uses, the literal carrying sense in 21 percent, including both military and civilian contexts, and the remaining 13 percent were too ambiguous to place in either category.

Heller rejected the view that “bear Arms” in the Second Amendment exclusively means to carry arms while serving in a state-organized militia. Has Heller’s conclusion been undermined by legal corpus linguistics (LCL)? There are several points to consider in answering this question.

First, there is no question that “bear arms” was used during the founding period to describe carrying weapons in individual and civilian contexts, outside of service in an organized militia or other military unit. Here are several examples:

  • William Robertson’s 1770 history of the reign of Charles the Fifth, emperor of Germany, which was published in America, refers to “women, orphans, and ecclesiastics, who could not bear arms in their own defence.”
  • Timothy Cunningham’s 1771 popular English legal dictionary of the period, which was found in Jefferson’s library, gives this example of the usage of “arms”: “Servants and labourers shall use bows and arrows on Sundays, & c. and not bear other arms.”
  • James Madison proposed an anti-poaching Bill for Preservation of Deer to the Virginia legislature in 1785, which had been written by Thomas Jefferson in 1779. Anyone convicted of killing deer out of season faced further punishment if, in the following year, he “shall bear a gun out of his inclosed ground, unless whilst performing military duty. The illegal gun carrier would have to return to court for “every such bearing of a gun” to post additional good-behavior bond.
  • The 1795 epic poem M’Fingal by lawyer John Trumbull reads: “A soldier, according to his directions, sold an old rusty musket to a countryman for three dollars, who brought vegetables to market. This could be no crime in the market-man, who had an undoubted right to purchase, and bear arms.”
  • Charles Brockden Brown’s 1799 novel, Edgar Huntly: or, Memoirs of a Sleepwalker, states, “I fervently hoped that no new exigence would occur, compelling me to use the arms that I bore in my own defence.”
  • John Leacock, well-known Philadelphia businessman, patriot, and playwright, wrote the following line for the character Paramount in the patriotic drama, The Fall of British Tyranny: or, American Liberty Triumphant, which was printed in Philadelphia, Boston, and Providence: “I shall grant the Roman Catholics, who are by far the most numerous, the free exercise of their religion, with the liberty of bearing arms, so long unjustly deprived of, and disarm in due time all of the Protestants in their turn.”

Given these and other examples of “bear arms” being used outside the military or militia context, the LCL debate over the meaning of “bear arms” turns largely upon the clashing arguments that set frequency against context found in Kevin Tobia’s excellent chart in the opening blog post in this series.

Second, some LCL researchers have overclaimed that because “bear arms” was used most often in a military or militia sense, this is its meaning in the Second Amendment, and therefore Heller was wrongly decided. It is not surprising that “bear arms” was used mostly in a military context during the founding period, given the Revolutionary War and subsequent debates about a standing army and state militias. But it does not necessarily follow that the Second Amendment uses this narrow idiomatic or specialized sense, rather than a more general sense of carrying arms in either civilian or military contexts.

The use of frequency data in LCL is more complex than those who suggest it is determinative in Heller. It can show us 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, but frequency data by itself is not conclusive. Frequency data must be supplemented by relevant legal context.

Third, the most relevant legal context for the ordinary meaning of “bear arms” in the Second Amendment is found in several contemporary state constitutions, as well as a proposed amendment to the federal Constitution from Pennsylvania. These founding-era sources used the phrase “bear arms” to describe the right to carry weapons for non-military purposes. They are especially significant because they occur in the context of defining the scope of constitutional rights to arms.

The Anti-Federalist minority 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.” Additionally, there were multiple complaints from citizens in Massachusetts that their state’s constitutional right to arms was too narrow because “the people have a right to keep and bear arms as well for their own as the common defense.” 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.” Both the proposed amendment proffered by the Pennsylvania Anti-Federalists as well as multiple state constitutional protections extended the right to “bear arms” to both civilian self-defense as well as to service in the militia.

Congress obviously was aware that the Pennsylvania Anti-Federalists and early state constitutions used “bear arms” in a sense that did not refer exclusively to service in a state militia. But did their omission of a modifying phrase signal to ordinary readers that the term “bear arms” had a narrower military meaning? That’s what Justice Stevens and the Heller dissenters believed. In his Heller dissent, Justice Stevens points out that

[h]ad the Framers wished to expand the meaning of the phrase “bear arms” to encompass civilian possession and use, they could have done so by the addition of phrases such as “for the defense of themselves,” as was done in the Pennsylvania and Vermont Declarations of Rights. The unmodified use of “bear arms,” by contrast, refers most naturally to a military purpose, as evidenced by its use in literally dozens of contemporary texts.

But this fails to account for how ordinary readers would understand the meaning of “bear arms” given Congress’ deliberate decision not to modify the phrase. The fact that you don’t modify something also can indicate what it means. Congress rejected a proposal to add the narrowing modifier “for the common defence” after “the right of the people to keep and bear arms.” Since two early state constitutions protected the right to bear arms only “for the common defence” (Massachusetts) or “for the defence of the state” (North Carolina), the decision not to limit the scope of the right to an exclusive military or militia context supports the Heller majority’s broader reading of “bear arms.”

Fourth, while LCL provides one among many useful tools for determining the meaning of words or phrases as used by original speakers, it has various methodological limitations. One limitation is that legal researchers must formulate search parameters and make coding decisions that inevitably are influenced by the researcher’s own judgment, intuition, and biases.

For example, Neal Goldfarb concluded 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. When another researcher, Josh Jones, included and coded these and other uses omitted by Goldfarb, he found that “while the specialized sense of bear arms (i.e., serving in the military or engaging in collective armed conflict) appears to have been used significantly more often than the carrying sense of bear arms, the latter still appears to have been used more often than past research may have suggested.” Jones’s own research reveals that “over one-fifth of recorded uses of bear arms in COFEA employed the phrase in the carrying sense. While the literal carrying sense should still be recognized as the minority sense, it is not an all-but-nonexistent sense like Goldfarb and others suggest.”

This and other limitations with LCL methodology in the Second Amendment context suggest that other tools should not be abandoned. As Jones points out,

[b]ecause the carrying sense of bear arms was still common at the time of the Founding (and . . . it may have been even more common than this Note’s principal findings suggest), the Court may need to continue to rely on other textual and historical tools (such as the grammar cannon, states’ Second Amendment equivalents, early American firearm regulations and practices, the Second Amendment’s connection to the English Bill of Rights, etc.) to determine the original understanding and current legal scope of the Second Amendment. Corpus linguistics data are most reliable when they are harmonized with other tools of interpretation.

Corpus linguistics is a wonderful tool for constitutional interpreters, but it is not a complete tool. At this point, the verdict is still out on the probative value of LCL in understanding the original meaning of the Second Amendment.