Passage A:

For most of the twentieth century, a judge seeking the ordinary meaning of a statutory term relied on two instruments: the dictionary and her own linguistic intuition. Both are now under challenge from legal corpus linguistics, a method that mines vast electronic databases of naturally occurring text to determine how a word was actually used by ordinary speakers at the moment a statute was enacted. Where a dictionary catalogues the senses a word can bear, a corpus reveals which sense predominated in ordinary usage, and how often. A judge asking whether "carry" in a firearms statute ordinarily denoted conveyance on the person or transport in a vehicle need no longer guess; she can observe the distribution of usages across millions of contemporaneous sentences. Proponents advance two claims. The first is epistemic: corpus analysis replaces the judge's idiosyncratic intuition, an unreliable sample of one, with systematic evidence drawn from the very linguistic community whose conventions the statute presupposes. The second is democratic: because a statute binds ordinary citizens, its words should be given the meaning those citizens would most naturally attach, and only corpus data can measure that meaning without circularity. On this view the method does not merely assist interpretation; it disciplines it, exposing readings that a judge's untutored ear might have found plausible but that the linguistic record shows to be marginal. The dictionary tells us what a word could mean; the corpus tells us what, in ordinary hands, it did.

Passage B:

That a word was used in a particular way ten thousand times tells us that the usage was common. It does not tell us that the usage is what a competent reader would assign to the word as it appears in a statute. This gap - between frequency and meaning - is the persistent difficulty confronting the enthusiasts of legal corpus linguistics. A general corpus records how speakers deploy language across recipes, sports reporting, and private correspondence; a statute is none of these. Its readers approach it already knowing that its terms may carry technical, term-of-art, or context-narrowed senses that ordinary frequency counts will drown out rather than reveal. The method also rests on choices that are anything but mechanical. Which corpus, which date range, which search string, and which of the returned examples count as relevant are all decisions an advocate can make strategically, and small variations can reverse the apparent result. More fundamentally, the appeal to ordinary usage presupposes what it purports to establish: that a statute should be read as an ordinary reader would read it, rather than as the enacting legislature meant it. That premise is contestable, and no quantity of usage data can settle it, for the data become relevant only once the premise is granted. Corpus linguistics is therefore a genuine instrument, capable of dislodging confident but mistaken intuitions. What it cannot do is relieve the interpreter of the prior, frankly normative judgment about whose meaning governs. The tool defers that question; it does not answer it.