« Ashlee Paxton-Turner: Continental Originalism
Michael Ramsey
| Main | Matthew Franck on "McCulloch v. Maryland at 200"
Michael Ramsey »


A Response to David Weisberg on Corpus Linguistics
James Cleith Phillips

[Ed.: We're pleased to have this guest response from James Cleith Phillips, Assistant Professor of Law at Chapman University Dale E. Fowler School of Law and one of the nation's leading experts on corpus linguistics in legal interpretation.]

David Weisberg has commented on a recent article I posted to SSRN wherein my co-author and I applied a corpus linguistic methodology to determine the ordinary meaning of key terms of the Antiquities Act of 1906. Our paper looked at the usage of the following phrases or words as they were used in the time leading up to and including the first decade of the 20th century:

  • situated upon (the) land(s) (searching for various forms of the verb situate)
  • upon the lands
  • lands
  • on land
  • land
  • parcel(s) of land

We searched for these phrases and words in the Corpus of Historical American English. And after reading hundreds of results, we could find no clear evidence that Americans from 1810-1909 commonly (if at all) used these phrases or words to refer to the bottom of the ocean. We concluded that the ordinary meaning of these terms would not include so-called “land” underneath the sea.

Mr. Weisberg criticizes not so much our research but the enterprise of using a corpus linguistic methodology itself for questions of ordinary meaning (his title supports this broader criticism). He finds it “false” that the dominant sense of a word or phrase in a context relevant to a statute is the ordinary meaning. In his view, “[f]requency does not determine or even alter meaning.”

As we noted in our paper, how exactly one should operationalize ordinary meaning is a theoretical question scholars and jurists have yet to definitely answer. If ordinary meaning is the meaning a word or phrase would have had to ordinary folks in a particular context, then understanding how people use language is an important insight into how they understand language. After all, just because a word or phrase can take on a particular meaning doesn’t mean it is usually understood to have that meaning in that context. Otherwise, language would constantly be ambiguous.

Mr. Weisberg rejects this and puts forth his own methodology using an example of the term misdemeanors from the Constitution’s text. He contends that while a search of the term in U.S. materials would reveal that in the context of “high crimes and misdemeanors” it would mean “serious, substantial wrongdoing,” it also had the common usage of “a relatively minor, insubstantial infraction.” Thus, in his view, the “frequency of one usage would not have obliterated or changed the meaning of the less-frequent usage.” So, the President could be impeached for “crossing against [a traffic] light.” (How could he know how common the criminal law usage was without resorting to a corpus? And when he refers to frequency, is he referring to overall or in a particular context?)

This example demonstrates at least three errors, which, unfortunately, are all too common. First, it ignores the linguistic principle of “non-compositionality.” Sometimes “the combination of words has a meaning of its own that is not a reliable amalgamation of the components at all,” such as for good or at all. Idioms are a specific type of this, as are cliches and frozen metaphors. (This is why the Supreme Court adopted the Anti-Dissection Rule in analyzing trademarks a century ago.) Phrases need to be understood as units—perhaps they will mean nothing more than the sum of their parts, but perhaps not.

Second, his misdemeanors example ignores context. The Constitution allows impeachment for high crimes and misdemeanors. Thus, the most relevant sense of the word misdemeanor is the sense with which it was used in the context of impeachment. To use the sense not used in the impeachment context would lead to a misreading of the Constitution just as if one argued that the phrase corruption of blood could be understood at the founding to mean one had a blood ailment.

Third, the misdemeanors example relies on a methodology that appears to never eliminate senses, which would lead to communication chaos. Just because a word or phrase has multiple possible senses doesn’t mean each one is activated in a given context. And with statutes, it would make the laws dangerously broad, particular those laws that use words with dozens of senses.

Next in his criticism, Mr. Weisberg appears to reject textualism in favor of some form of intentionalism (this is bolstered by a later sentence where he claims, “There is absolutely no reason to believe that Congress intended to limit the Antiquities Act to objects on dry lands and to exclude objects on submerged lands….”) While conceding that land or lands would be infrequently used to refer to the seafloor, he argues that the legislators probably never gave the distinction much thought and that there is no evidence they deliberately thought to exclude the seafloor sense from the statute. Without perhaps realizing it, he has stumbled upon another interesting theoretical question that textualism has yet to answer (as we noted in our paper): what to do in situations when a broader sense can include a narrower sense, but we think the narrower sense is the ordinary one in that context? Still, his criticisms are of the ordinary meaning inquiry and textualism itself. To the extent textualism’s ordinary meaning inquiry is concerned about notice, then requiring statutes to incorporate rare understandings that few if any would have understood seems deeply problematic. There is no real evidence that Americans around 1906 would have understood the Act to include seafloors, whatever the intent of Congress (and Mr. Weisberg seems to concede it wasn’t really thinking about seafloors either).

Finally, Mr. Weisberg asks, “If that’s what the legislators meant, why isn’t that what they enacted” That is, why didn’t the 1906 Act provide that the president could declare monuments on “dry lands” of the federal government?” Is anyone going to argue that, in 1906, the phrase ‘dry lands’ was not in common use?” Obviously the phrase was in use, though it appears just 3 times from 1900-1909 in the corpus, as compared to lands, which appears 1241 times. Furthermore, when used in those three instances in the decade the Act was adopted, it was in reference to land that did not get much rain and either had to be sowed with a type of wheat suited for such, or irrigated. Of course, dry land could have the meaning that Mr. Weisberg points out the Oxford English Dictionary gives it—land above water. But what he ignores is that nearly every time (that we examined) that Americans around the turn of the 20th century used the word lands or lands (when it wasn’t an irrelevant sense such, as the verb to land or as a synonym for a country), they used it to mean dry land. That was its common or ordinary meaning.

And Mr. Weisberg is silent on the other term from the act we analyzed, parcels of land, and how anyone could read that a statute that authorized Congress to reserve “parcels of land” to protect the objects made national monuments “situated upon the lands” the federal government owned or controlled. No one in the early 20th century used parcels of land to refer to the ocean or seafloor.

In sum, whatever Congress intended, we are stuck with the ordinary meaning of the words it used in the 1906 Act. And people didn’t use these words and phrases to mean anything under the ocean. They therefore would not have understood the Act to cover such. How do we know how they would have understood these words phrases? By how they used them. And if the way a word is most commonly used doesn't determine its meaning, what does?