Trial News

In a case involving statutory interpretation of the Employee Retirement Income Security Act (ERISA), two Sixth Circuit judges disagreed about the benefits of “corpus linguistics” in separate lengthy concurrences to a decision affirming summary judgment for a defendant employer on the basis that the plaintiff’s state law claims were preempted. One judge argued that corpus linguistics—which uses a searchable database to uncover specific examples of how a particular word was commonly used at a particular time—could enhance statutory interpretation of the ordinary meaning of words or phrases; the other judge said that it would introduce too much subjectivity and potentially lead to more confusion. (Wilson v. Safelite Grp., Inc., 2019 WL 3000995 (6th Cir. July 10, 2019).)

Plaintiff Dan Wilson sued his former employer, Safelite Group, Inc., over a deferred compensation plan the company’s board adopted before the company was bought to avoid certain payments that would increase taxes for employees under the plan. When Wilson left the company several years later, a tax audit determined that he owed significant taxes and penalties because a portion of his deferred compensation did not comply with tax requirements for deferred compensation plans. Wilson’s suit alleged state law breach of contract and negligent misrepresentation claims, and the defendant countered that ERISA preempted the plaintiff’s claims because the plan was an employee pension benefit plan under §3(2)(A)(ii) and not an exempted bonus plan under U.S. Department of Labor regulations.

The lower court agreed with the defendant that the plan was an employee pension benefit plan. Wilson appealed to the Sixth Circuit, which devoted most of its opinion to examining the statute’s language and interpreting its meaning. Section 3(2)(A)(ii) of ERISA provides that a plan that “results in a deferral of income by employees for periods extending to the termination of covered employment or beyond” is an employee pension benefit plan. The court noted that ERISA’s purpose is “to establish a ‘uniform regulatory regime’ for plan administration that protects monies belonging to plan beneficiaries while such funds are held and managed by others.” Any state law claim that “duplicates, supplements, or supplants” ERISA’s enforcement mechanism is preempted because it conflicts with Congress’s intent to make ERISA the only remedy in such cases.

The court focused on the meaning of “results” in the statute and on Wilson’s argument that this term encompasses a “requirement” for an employee’s deferral of income. Referring to U.S. Supreme Court opinions and dictionary definitions as well as the use of “require” in other sections of ERISA, the court concluded that “require” is too stringent an interpretation and that “results” merely means that it is an “effect, issue, or outcome” of the plan, which would encompass the deferred compensation plan at issue.

The court also disagreed with the plaintiff’s assertion that the statute does not cover plans that allow withdrawal of deferred income before employees leave a company based on the “termination” language used in the statute. The plaintiff interpreted it to mean “until termination,” but the court found instead that the use of “periods extending to the termination” means “up to or beyond termination,” thus encompassing the deferred compensation plan as a pension benefit plan. The court concluded that because of this statutory interpretation, the plan fell under ERISA. Thus, it affirmed summary judgment for the defendant.

The in-depth parsing of the statute’s language led two of the judges on the three-judge panel to write concurrences addressing the use of corpus linguistics in statutory interpretation. Judge Amul Thapar argued for expanding the statutory interpretation toolbox, since “words often have multiple permissible meanings,” which can complicate how judges determine “ordinary meaning” when reading the plain language of a statute. Advocating for the use of corpus linguistics—which relies on searchable databases to show how words are used at specific points in time—Thapar explained that “these databases, available mostly online, contain millions of examples of everyday word usage (taken from spoken words, works of fiction, magazines, newspapers, and academic works).” By collecting usage examples for these databases, lawyers and judges could develop a more robust picture of how a word was used and generally understood at the time a statute was enacted. He noted that “its foremost value may come in those difficult cases where statutes split and dictionaries diverge,” and corpus linguistics can act “as a cross-check on established methods of interpretation (and vice versa).”

Thapar then tested the method on the case at issue, searching the Corpus of Historical American English database for uses of the term “results in” from the 1960s and 1970s around the time ERISA was enacted and listing several examples that support the court’s reasoning that “results in” is not synonymous with “requires.” He repeated the test with “extending to” and found that most of the results also supported the court’s conclusion about the termination portion of the statute.

In a separate concurrence, Judge Jane Stranch argued that corpus linguistics “invites judges to perform the same kind of subjective decision making that the concurrence otherwise cautions us to avoid.” She noted that the databases contain many examples of a word’s usage, leading to judges deciding which examples are the most accurate or relevant to inform their analysis of the word’s ordinary meaning at the time. It also leaves up to judges which entries are irrelevant and involves an inherent bias toward the types of usages that are most likely to be in the database in the first place, such as those used by newspapers and not necessarily by the average American at the time. Instead, Stranch argued that trained lexicographers are the best source for a word’s meaning, and that judges already have these experts’ tried-and-true tool: a dictionary.

Stephen Mouritsen, a University of Chicago Law School professor who has written on the use of corpus linguistics in statutory interpretation, said, “Judges generally begin their interpretive task by looking for the ordinary meaning of the language of the law. And they often end there out of respect for the notice function of the law or deference to the presumed intent of the lawmaker. Courts often fail to take into account a variety of questions that have bearing on the meaning of an utterance, including its context, historic usage, and the speech community in which it was uttered. Using the tools of a linguistic corpus, we can measure the comparative frequency of a given sense of a given word in a given context. By incorporating corpus methods into the search for ordinary meaning, we can turn a largely intuitive and opaque inquiry into an empirical and transparent one.”