District of Columbia v. Heller, meet legal corpus linguistics. Opponents of the decision are excited to have a new lever that might dislodge it. Proponents of the methodology are excited to have a great victory to prove their methodology’s worth. But in my view most of this excitement is premature.
Legal corpus linguistics might be able to dislodge even Supreme Court precedent. Legal reasoning is only as good as its premises. Though judicial decisions frequently contain claims about facts – historical facts, linguistic facts, facts about the world – they do not make those facts true. If those premises are wrong, and if they are material to the decision’s conclusions, then the decision should be reconsidered.
At the same time, reconsidered does not always mean “overruled.” First of all, under longstanding principles of precedent, an established doctrine is presumptively correct. That presumption can be rebutted by evidence, but the evidence must rise to the level of clear error.
Second, of course a decision might turn out to be right for not-quite-the-right reasons. So even if we are convinced that a decision materially relied on a clearly erroneous premise, that does not mean that the decision was wrong. Rather it means we must return to square one to ask what the right decision would be, reasoning from the correct premise.
So we must ask whether corpus linguistics shows any reasoning in Heller to be clearly, materially, erroneous. And if so, we must ask what the correct analysis would be and whether it would change the holdings of Heller.
So what about Heller? In that case, the Court reached three major legal conclusions. It concluded that the Second Amendment right was not contingent on active service in an organized militia. It concluded that handguns are protected by the Second Amendment because they are commonly used for civilian self-defense. And it concluded that certain regulations of firearms would be forbidden (such as DC’s trigger-lock requirement) and others, not before the court, would be permitted (such as the regulation of particular weapons and persons).
In the course of supporting those conclusions, the Court invoked at least five historical premises. These include: (1) the conclusion that the prefatory clause does not limit the scope of the operative clause; (2) the conclusion that a “right” belonging to the “people” is an individual right; (3) the conclusion that “keep arms” included having weapons in one’s home; (4) the conclusion that “bear arms” does not connote participation in a structured military organization; (5) the conclusion that individual constitutional rights can be regulated largely based on historical analogs rather than public interest balancing tests.
There are many ways of criticizing this analysis. Perhaps Heller applied the wrong methodology. Or perhaps its erred on various historical grounds. For instance Saul Cornell has criticized (1) on historical, but non-corpus-based, grounds; Jud Campbell’s work suggests that (5) may be wrong. Some of these premises are probably closer to questions of historical law than historical language. So it appears to be that corpus linguistics arguments are especially aimed at (2) or (3) or especially (4).
On the legal corpus linguistics arguments I am only an interested consumer, not an expert. But as I read the arguments so far, most of Heller’s premises have not yet been dislodged, and the results are less dramatic than Heller’s critics hoped.
The evidence on premise (2) (“the right of the people”) appears weak. While there are examples in the corpus of a “right of the people” being used collectively, Heller’s chief reason for rejecting such a reading was constitutional context: other provisions of the Constitution (viz, the First and Fourth Amendments) use the “right of the people” to refer to an individual right, so it is likely that the Second Amendment did so as well. This is a good example of the limits to legal corpus linguistic analysis. The use of a phrase in other contexts cannot do much to rebut a claim made from the context of a particular document.
The evidence on premise (3) (“keep arms”) appears even weaker, with no showing that it clearly meant something else.
That leaves premise (4) (“bear arms”). Here there seems to be abundant evidence – much more evidence than one would expect from reading the Court’s opinion in Heller – that most uses of the phrase in the corpus appear in a military context.
This finding might be enough to dislodge Heller’s premise (4), but that is not certain, for at least two reasons:
First, the fact that a phrase usually appears in a particular context is only indirect evidence about its meaning in the context under examination. (Consider this parable.) That is especially true when the Court relies on context-specific evidence for its reading, such as Justice Scalia’s argument that mixing non-idiomatic and idiomatic uses is analogous to the “grotesque” phrase “he filled and kicked the bucket.”
Second, many instances in the corpus demonstrate a military context, but not all of them demonstrate a military (or military-only) meaning of the phrase. Perhaps “bear arms” meant “fight a war” or “serve in the militia.” But perhaps it merely meant “carry and/or use a gun (in the way prototypically associated with soldiers).” The second definition would undermine far less of Heller than the first.
This brings us to the final upshot. If Heller was wrong, even clearly wrong, about premise (4) it does not follow that the dissent’s view, or a “collective rights” view, was correct. Even a somewhat generous reading of the legal corpus linguistics evidence against Heller could still support an individual right to keep weapons, as well as an individual right to use weapons that have some military connection. This might suggest less protection for handguns and more for military rifles (or even for body armor). In other words, perhaps corpus linguistics indicates that Heller erred in denying the military flavor of the Second Amendment right without erring in concluding that it was an individual one.