Chief Justice Roberts not sure ‘what to make of’ linguistics tool

U.S. Chief Justice Roberts arrives for President Biden's address to a joint session of the U.S. Congress at the U.S. Capitol in Washington
Supreme Court Chief Justice John Roberts arrives to attend President Joe Biden's first address to a joint session of the U.S. Congress inside the House Chamber of the U.S. Capitol in Washington, U.S., April 28, 2021. REUTERS/Jonathan Ernst

(Reuters) - Based on oral arguments on Wednesday, it appears that the U.S. Supreme Court is not yet quite ready for the corpus linguistics revolution.

I told you earlier this month about a brand-new study that applied a “revolutionary” linguistics tool to the key statutory phrase at issue in ZF Automotive US Inc v. Luxshare Ltd. That's the Supreme Court case posing the question of whether U.S. courts can order discovery in commercial litigation being conducted overseas.

The key phrase, in case you’ve forgotten, is “foreign tribunal.” Luxshare, which is seeking U.S. discovery from ZF to support a prospective fraud claim before German arbitrators, argues that the phrase, which appears in a 1964 law, encompasses private arbitration proceedings. ZF contends that “foreign tribunal” means only a governmental or quasi-governmental proceeding, not a private, commercial arbitration.

In their March 12 study, law professor James Phillips of Chapman University and linguist Jesse Egbert of Northern Arizona University used an interpretive tool called corpus linguistics to examine how the phrase has been used contextually. Their paper, which is slated for publication in the Virginia Law Review Online, searched for the phrase in several vast databases, or corpora, of text, including legal databases. Coders then analyzed search results to determine whether the phrase referred to private proceedings.

I’m drastically oversimplifying the professors’ methodology, but their conclusion was easy enough to understand: The study did not find a single instance, preceding the enactment of the 1964 law, in which the phrase was clearly used to describe a private commercial arbitration. The finding was dramatic enough that ZF’s lawyers at Latham & Watkins highlighted the study in the company’s reply brief, filed just days after the study was published.

Phillips told me earlier this month that he was sure the study would crop up during oral arguments at the Supreme Court. The real breakthrough, he said, would be if the justices eventually cited the analysis in their opinion, showing that the court understood the value of corpus linguistics as a tool of statutory interpretation.

Phillips was right in his first prediction. The Supreme Court explicitly discussed corpus linguistics during Wednesday’s argument – apparently the first time that’s happened, according to Phillips. But it looks like Chief Justice John Roberts, at least, still needs to be convinced that the court should rely on corpus linguistics analysis.

“Yeah, I don't quite know what to make of that,” Roberts said when ZF counsel Roman Martinez said the study disproved any assertion that common usage of the phrase extended to private arbitration. “That's -- that's something new,” Roberts said. “Have we relied on that source before?”

Martinez explained that corpus linguistics is a more scholarly and systematic approach to methodology the Supreme Court has previously used to examine the common usage meaning of statutory text. Even Roberts, Martinez said, relied on database search and analysis in his majority opinion in 2011’s FCC v. AT&T Inc, which turned on the definition of “personal privacy” in the context of a corporation’s privacy rights.

“I don't think it's methodologically new,” Martinez said. “I think it's just a little bit more scholarly, a little bit more reliable. They use Latin words, which maybe makes it a little scarier in some way, but I think it's the same basic idea.”

Justice Amy Coney Barrett showed familiarity with corpus linguistics, pointing out that the 6th U.S. Circuit Court of Appeals and the Utah Supreme Court have both relied on searches and analysis of a database called the Corpus of Historical American English, a collection of more than 100,000 newspaper and magazine articles, play and movie scripts, books and other texts. Barrett also agreed with Martinez that the U.S. Supreme Court has relied on similar, albeit informal, database surveys to inform its interpretation of statutory text.

Justices Elena Kagan and Brett Kavanaugh, though, seemed to push back. “I'm all for, you know, being serious about language when there's something to be serious about, but I don't know,” Kagan said. Kavanaugh challenged Martinez to explain why the court should focus on the phrase as a whole, rather than looking at the words “foreign” and “tribunal” distinctly.

Luxshare’s counsel, Andrew Rhys Davies of Allen & Overy, responding to a similar question from the chief justice, said the corpus linguistics analysis of the entire phrase proved only that “foreign tribunal,” as a phrase, “didn't really have a meaning as of 1964.” The court, Davies said, should therefore look at its own precedent defining “tribunals” to encompass private arbitration.

Phillips told me via email that it’s a milestone for corpus linguistics even to have been discussed explicitly during oral argument. “That's progress for a methodology that the court has been doing on a more basic level for a long time without calling it corpus linguistics,” he said.

As for the chief justice’s question about the novelty of relying on the Corpus of Historical American English, Phillips pointed out that the study relied on four other text databases as well, including sources familiar to the justices, such as Westlaw and a repository of Supreme Court opinions.

“If the court wants to ignore the small bit of our analysis from COHA because a majority opinion has never cited to COHA before, that’s one thing,” he said in a statement. “But we don’t see why it would ignore the analysis from sources it regularly relies on.”

There are, to be sure, a lot of complexities in the use of corpus linguistics as a tool of statutory interpretation. In a paper last year for the Georgetown Law Journal, for instance, law professors from Brigham Young University discussed the potential pitfalls of relying on general corpora for legal analysis. They also cited the concern that “bias and intuition lurk behind corpus linguistics’ cloak of empiricism.” Lawyer and linguist Neal Goldfarb, who filed a Supreme Court amicus brief rooted in corpus linguistics way back in 2010, told me he fears that the black-and-white conclusion of the Phillips and Egbert study masks underlying analytical decisions that, in Goldfarb’s view, cast doubt on that conclusion.

We’ll see later this year if the Supreme Court has figured out how to make sense of it.

Read more:

‘Revolutionary’ linguistics tool awaits star turn in Supreme Court arbitration case

Our Standards: The Thomson Reuters Trust Principles.

Opinions expressed are those of the author. They do not reflect the views of Reuters News, which, under the Trust Principles, is committed to integrity, independence, and freedom from bias.

Thomson Reuters

Alison Frankel has covered high-stakes commercial litigation as a columnist for Reuters since 2011. A Dartmouth college graduate, she has worked as a journalist in New York covering the legal industry and the law for more than three decades. Before joining Reuters, she was a writer and editor at The American Lawyer. Frankel is the author of Double Eagle: The Epic Story of the World’s Most Valuable Coin.