Posted Apr 2025
No. 127 Provided as a service to the firm’s clients and friends. April 2024 |
by Andrew J. Petersen
Word View as World View
By Andrew J. Petersen
Years ago, I studied literary criticism. Literary criticism developed under various methodologies, including textualism, structuralism, deconstruction, postmodernism, colonialism, psychoanalytic criticism, feminism, formalism, and other “isms.” I attended the Modern Language Association’s annual conference and encountered a dizzying array of interpretative theories. (This was the time when literary criticism was drunk on Foucault, Chomsky, Bakhtin, Lakoff, Lacan, and Stanley Fish, among others.)
This was before computer research. Our research tools required a greater investment of time. Writing a paper on death and madness in Hamlet required consulting a Shakespeare concordance and concordances of contemporary playwrights, such as Christopher Marlowe and Thomas Kyd. Concordances and dictionaries, like the Oxford English Dictionary, were helpful for cross-referencing words, motifs, and tropes.
Technology has significantly transformed research. I first encountered what we now refer to as “corpus linguistics” at a writing conference. At that time, the presenters had developed a database (corpus) of early American English texts. By searching for words or phrases, the computer would generate a lineage and display relationships, somewhat like Westlaw or Lexis, but extending far beyond legal texts.
“Corpus linguistics” in legal interpretation examines the context of word usage at the time a law was enacted. The “corpus” refers to a specific collection of texts, while “linguistics” is the study of this collection. This represents a “big data” approach to textualism. The best-known “corpus” is the Corpus of Contemporary American English (COCA). Another “corpus” focuses on Historical American English (COHA). Both are available online for free. (If you’re curious about when these were created, where they originated, and why this work was pioneered at a particular religious university, feel free to call me.)
In statutory interpretation today, the stated goal of textualism is to “determine the ordinary public meaning.” Bostock v. Clayton Cty, 140 S. Ct. 1731, 1738 (2020). This prioritizes “original public meaning” over every other interpretive concern, including context, intent, purpose, or policy.
Three years ago, the Arizona Supreme Court referred to “corpus linguistics” in a workers’ compensation case: Matthews v. Indus. Comm’n, 254 Ariz. 157 (2022). The court held that “personal injury” in Arizona’s constitution did not encompass a purely emotional distress claim. Justice Bolick discussed the application of corpus linguistics in the opinion and cited Thomas R. Lee & Stephen C. Mouritsen, Judging Ordinary Meaning, 127 Yale L.J. 788 (2018). Justice Timmer dissented, referencing her own historical sources for defining “personal injury,” including dictionaries, legal treatises, and children’s fiction from 1915. (There is confusion on what should be included in the corpus. Original public meaning should not equate to ordinary lay meaning. After all, it is legal texts we are interpreting.)
Two weeks ago, in one of my client’s cases, the Arizona Supreme Court again used “corpus linguistics.” Garibay v. Johnson, (March 13, 2025). The court interpreted the words in the levy statute “guilty of any misconduct” (language enacted in 1901) as not encompassing plaintiffs’ claims for negligence or gross negligence. My client, the constable, was entitled to judicial immunity for any negligence or gross negligence that may have occurred while serving a writ. The court considered an 1897 lexicon and two legal dictionaries, one from 1904 and another from 1910. The court then considered corpus linguistics and looked at COHA concluding: the prevailing common usage of “misconduct” during the relevant period connotes intentional rather than negligent conduct. I agree with that statement, but do not believe the levy statute applied to the tort claims asserted by plaintiffs.
In both the Matthews and Garibay cases, Justice Timmer expressed reservations. She dissented in Matthews regarding the court’s interpretation of “personal injury,” emphasizing that other texts include emotional distress in this definition. While she concurred in Garibay, she suggested to “pump the brakes” on corpus linguistics. She referenced several law review articles that criticize corpus linguistics and concluded that this type of analysis should first be addressed by the parties and their linguistic experts.
Justice Timmer’s dissent and concurrence highlight the inherent subjectivity in defining the “corpus,” the parameters used to conduct the search, and whether the results of the search are meaningful. See Evan C. Zoldan, Corpus Linguistics and the Dream of Objectivity, 50 Seton Hall L. Rev. 401 (2019); cf. Thomas R. Lee and Stephen C. Mouritsen, The Corpus and the Critics, 88 Univ. Chicago L. Rev. 275 (2021) (responding to and recognizing criticism).
These are significant concerns. I believe we undermine our jurisprudence when we use corpus linguistics by examining a collection of unrelated texts, as if the answer resides within the crowd, especially when looking beyond legal texts. Early legislatures included attorneys or those with a legal background, as well as legislative staff who utilized legal cyclopedias and borrowed statutes from other jurisdictions.
Why abandon years of accumulated wisdom in statutory interpretation by seeking truth from a multitude of texts? Corpus linguistics serves merely as a tool when courts evaluate and then disregard it. This is evident in Justice Scalia’s Second Amendment opinion in Heller. He dismissed the interpretation by linguistic scholars who submitted an amicus brief on “keep and bear arms.” Scalia initiated a revolution in the interpretation of legal texts by examining the original public meaning. One of the strongest advocates of textualism and “public meaning” was not captivated by linguistics and rejected its use, as Justice Stevens pointed out in his dissent.