Testing Ordinary Meaning: An Experimental Assessment of What Dictionary Definitions and Linguistic Usage Data Tell Legal Interpreters
62 Pages Posted: 6 Nov 2018 Last revised: 10 Jul 2020
Date Written: August 1, 2019
Within legal scholarship and practice, among the most pervasive tasks is the interpretation of texts. And within legal interpretation, perhaps the most pervasive inquiry is the search for “ordinary meaning.” Legal interpretation—of contracts, statutes, wills, trusts, deeds, patents, regulations, treaties, and constitutions—regularly includes evaluation of how ordinary people would understand the text. Theorists and practitioners often treat the search for ordinary meaning as an empirical inquiry, aiming to discover facts about how ordinary people would understand language. To discover ordinary meaning, interpreters increasingly recommend as evidence a relevant term’s dictionary definition or its pattern of usage across various sources in an English-language corpus. However, the most central question about these sources of evidence remains open: Do these popular methods accurately reflect ordinary meaning?
To assess this question, this paper develops and employs a novel method of “experimental jurisprudence.” A series of experimental studies (N = 4,162) reveals systematic divergences among the verdicts delivered by modern concept use, dictionary use, and corpus linguistics use. For example, today people apply the concept of a vehicle differently from the way in which they apply modern dictionary definitions or modern corpus linguistics data concerning vehicles. The same results arise across levels of legal expertise—participants included 230 “elite-university” law students (e.g. at Harvard and Yale) and 98 United States judges—and across various terms and phrases, including “vehicle,” “labor,” “weapon,” “carrying a firearm,” and “tangible object.”
The paper elaborates several implications of these results. First, the results provide insight into what dictionaries and corpus linguistics suggest to legal interpreters. Drawing on insights from linguistics and psychology, I distinguish between “prototypical” and “broad” senses of the same term. For example, a car is a prototypical vehicle, while airplanes, bicycles, and canoes are less prototypical vehicles. An extensive criterion would include all of those entities as vehicles, while a prototypical criterion would include only cars. This distinction about language is well-known, but the experiments show that the distinction also illuminates ordinary meaning’s sources of evidence. That is, dictionaries and corpus linguistics often track only one of these criteria—dictionaries tend to track the broad criterion and corpus linguistics the prototypical one.
Second, I identify several fallacies of interpretation that are supported by the results. As one example, consider “The Non-Appearance Fallacy,” the mistaken assumption that the non-appearance of some use in a corpus indicates that this use is outside of ordinary meaning. Arguments committing this fallacy have great rhetorical strength: Across thousands of sources in our corpus, we could not find even one example of an airplane referred to as a “vehicle,” therefore the ordinary meaning of “vehicle” does not include airplanes. However, as the experimental results indicate, ordinary meaning sometimes diverges from ordinary use: People’s full understanding of language is not always reflected in recorded speech and writing, especially their understanding concerning non-prototypical category membership.
Third, I evaluate the findings’ significance for different theories of legal interpretation. First I consider certain formalist, textualist, and originalist views that are committed to the existence of a single ordinary meaning of terms like “vehicle” and phrases like “carrying a firearm,” one which is outcome-determinative without reference to further context, textual purpose, or even type of law (e.g. criminal vs. contract). The data suggest that popular methods of dictionary-use and corpus linguistics carry serious risks of diverging from ordinary understanding—conservatively estimated, 20-35%. And in some circumstances, even judges’ use of these methods carried extremely large divergence rates—between 80-100%. The results shift the argumentative burden to theorists and practitioners that rely on these tools to determine legal outcomes: In light of the data, these views must articulate and demonstrate a reliable method of interpretation.
Finally, I consider the results from the perspective of interpretive theories that are uncommitted to, or even skeptical of, the notion of a single “ordinary meaning” that determines legal outcomes across a range of cases and contexts. On these views, the findings illuminate two different criteria that are often relevant in interpretation: a more extensive criterion and a more narrow, prototypical criterion. Although dictionaries and corpus linguistics can help us assess these criteria, a hard legal-philosophical question remains: Which of these two criteria should guide the interpretation of terms and phrases in legal texts? Insofar as there is no compelling case to prefer one, the results suggest that dictionary definitions, corpus linguistics, or even other more scientific measures of meaning may not be equipped in principle to deliver simple and unequivocal answers to inquiries about the ordinary meaning of legal texts.
Keywords: legal interpretation, law and language, contracts, statutory interpretation, legal theory, legal philosophy, experimental jurisprudence, empirical legal studies, formalism, textualism, dictionaries, corpus linguistics
Suggested Citation: Suggested Citation