On Oct. 13, 2022, a handful of the country’s most conservative federal judges gathered inside a wine cellar at a luxury ski resort in Deer Valley, Utah. The surrounding mountains were ablaze with yellow aspens, and the speaker addressing the room joked that the judges were already planning their afternoon hikes.
Before they could roam the slopes, they would spend the morning learning about a tool that could supposedly revolutionize how judges interpret the law. It was called corpus linguistics, and it was simple on its face. A corpus essentially works like a search engine that returns every example of how a word or phrase was used in a select database of historical texts.
But the leading proponents of legal corpus linguistics see it as something more: a powerful new tool to shore up the legitimacy of the conservative legal movement. Now, judges claiming to be interpreting the Constitution as it was originally understood could wield the imprimatur of big data.
“In the beginning, we only had paper, hard copies. Remember those things called books?” the speaker, Josh Blackman, a prolific legal scholars on the right, said to the judges assembled in the wine cellar. “Computer technologies open an entire new world of research.”
The runaway success of conservatives’ decadeslong campaign to dominate the federal courts is not without its challenges. With a 6-3 stranglehold on the Supreme Court and Donald Trump-nominated judges dominating federal appeals courts, the right wing has increasingly pushed the legal view that the law must be interpreted based on “history and tradition.” Yet few have failed to notice how perfectly “history,” in these judges’ rendering, aligns with current Republican beliefs on issues like guns and abortion. Corpus linguistics offers one way to dodge these criticisms. A judge who could keyword-search millions of lines of historical text, the thinking goes, is a judge who could ward off accusations that his version of history was invented for a partisan outcome.
As James C. Phillips, a conservative legal scholar, has put it, “Corpus linguistics is the tool that originalists have been waiting for.”
Already, corpus linguistics has made cameos in some of Supreme Court Justice Clarence Thomas’ opinions and played a role in striking down the Biden administration’s public transportation mask mandate.
The huddle in at the ski resort was part of a well-funded effort to spread corpus linguistics even further. A new nonprofit called the Judicial Education Institute had picked up the tab. Since its incorporation in 2020, the Judicial Education Institute has raised roughly $1 million from Charles Koch’s network and Donors Trust, an anonymous right-wing funding network dubbed “the dark-money ATM of the conservative movement.”
The institute’s chair, James Heilpern, is a graduate of Brigham Young University’s J. Reuben Clark Law School, the intellectual birthplace of legal corpus linguistics and where he first became enthralled with its potential. “It’s not every day that something really new comes along that has the opportunity to transform the legal profession,” he has said.
Now a lawyer with the far-right legal advocacy firm Schaerr Jaffe, Heilpern has led a sporadic effort to promote corpus linguistics over the years — but never before with such deep-pocketed support.
The first retreat the Judicial Education Institute hosted ― in February 2022 ― unfolded over six days at The Greenbrier, an opulent, 11,000-acre West Virginia mountain resort owned by Republican Gov. Jim Justice. Under the sherbet glow of Dorothy Draper-designed wallpaper, at least 11 U.S. district and appeals court judges listened to speakers vaunt the potential of corpus linguistics in the mornings. In their free time, they could indulge in one of The Greenbrier’s four golf courses or miles of trails.
Deer Valley, the same ritzy resort where Gwyneth Paltrow skied into legal trouble, provided an even more exclusive backdrop. At least seven federal judges attended. One U.S. district court judge, Stephen McGlynn, reported his trip alone cost the Judicial Education Institute $5,558.
Multiple judges brought their family members. Judge L. Steven Grasz appears to have brought his daughter on his trip to The Greenbrier ― in a photo she posted on Instagram, she is posing in front of its White House-like exterior ― and McGlynn’s wife joined him in Deer Valley, he told HuffPost. In an email, Heilpern said the institute did not pay travel expenses for the judges’ guests.
All told, the Judicial Education Institute spent $211,596 on conferences, travel and speaking fees in 2022, the latest year for which the nonprofit reported its finances to the Internal Revenue Service. In his application to the IRS for tax-exempt status, Heilpern disclosed he plans to take an annual salary of $175,000 in the future.
Billionaire-bankrolled judicial trips are nothing new on the right. In 2021 and 2022, an investigation by The Lever found just two conservative organizations, George Mason University and the Federalist Society, paid to send more than 100 federal judges on a total of 251 educational retreats.
The retreats at The Greenbrier and in Deer Valley were a way to attract a powerful audience for the top proponents of legal corpus linguistics, something the speakers acknowledged openly.
“I’m always happy to talk to judges in particular, because you’re on the front line, right?” Blackman said in his presentation at Deer Valley. “Unless lawyers find our work useful, and eventually judges think it’s worth citing, our stuff exists in this sort of weird vacuum.”
Other speakers included Heilpern; Thomas R. Lee, a former Utah Supreme Court justice and the first judge to use corpus linguistics in a court ruling; and Phillips, the conservative legal scholar and a former Lee clerk.
HuffPost tracked which federal judges attended the institute’s retreats using their annual financial disclosure reports from 2022, the latest which are available.
Virtually every guest and speaker at these junkets was a member of the Federalist Society, the powerful conservative legal network best known for hand-selecting former President Donald Trump’s judicial nominations. And nearly every judge who attended was a Trump appointee — including some of his most controversial.
There was Patrick Bumatay, whose nomination elicited the sweeping opposition of civil rights groups for his work in Trump’s Justice Department, and Grasz, who received a rare unanimous rating of “not qualified” from the American Bar Association when Trump nominated him to the Eighth Circuit Court of Appeals in 2017.
Andrew Oldham, a Fifth Circuit Court of Appeals judge who upheld an extreme abortion pill ban, was a speaker. So was Jeffrey Sutton, a George W. Bush appointee, who late last year, teed up the Supreme Court to weigh in on bans of gender-affirming care for minors.
Judge Lawrence VanDyke of the Ninth Circuit, whom the ABA scorched as “arrogant, lazy, [and] an ideologue” biased against LGBTQ people, is a Judicial Education Institute advisory board member.
In an email, Heilpern said the trips included judges “across the political spectrum” and involved “modest travel expenses.”
“The Judicial Education Institute is a non-partisan, 501(c)(3) organization that is dedicated to providing accessible, high-quality continuing education for federal and state judges across the political spectrum,” he wrote. “Our focus is to introduce judges to effective research tools in order to aid them in their efforts to engage in neutral, effective judicial decision-making and to bolster the public trust in the legal system. In accordance with federal judicial ethics, we reimburse modest travel expenses (typically no more than $750), group meals, and hotel accommodations.”
The Judicial Conference, the government body that writes ethics rules for federal judges, allows judges to attend privately-funded educational seminars so long as they disclose their participation. Other than that, there is very little oversight. Judges are not required to disclose the cost — McGlynn did so by mistake — or their guests. And the Judicial Conference does not police the line between an educational opportunity and a luxury trip meant to influence or reward judges for towing the party line.
A spokesperson for the Judicial Conference, in response to questions for this story, pointed to its advisory stating judges should consider whether their attendance at an event “promotes public confidence in the integrity and impartiality of the judiciary.”
But advocates of judicial reform have called on the Judicial Conference to exercise stronger oversight.
“We do not begrudge federal judges for wanting to stay at nice hotels, and we understand that there needs to be a draw to get judges on a plane,” Fix The Court, a nonprofit watchdog, wrote to the Judicial Conference in August 2023. “But there’s nice, and then there’s nice.”
“My computer program tells me”
Corpus linguistics started out as an academic tool for linguists and is not inherently conservative. But it holds an obvious appeal for the conservative legal movement.
Under the guises of originalism and textualism, which both present as ways to interpret the law based on its “original public meaning,” right wing judges have handed down a series of increasingly unpopular rulings, eviscerating gun laws, abortion rights, and prohibitions on prayer in public schools.
Corpus linguistics offers a way to hand-wave accusations of partisanship by claiming they have an empirical way to divine a law’s “original public meaning.”
“The conservative legal movement has put a lot of energy into depicting legal interpretation as something that can be and often is empirical, objective and relatively simple, and not normative, political or policy-oriented,” said Anya Bernstein, a professor studying legal interpretation at the University of Connecticut School of Law who has written critically of legal corpus linguistics.
“The idea is that a judge who is doing it ‘right’ doesn’t really have to make a difficult decision. … Legal corpus linguistics, because of its scientific imprimatur, and because it looks quantitative and objective, gives people a way of saying, ‘It’s not that I think this is how the law ought to be read — my computer program tells me this is how these words are originally used.’”
A few months after The Greenbrier retreat, U.S. District Judge Kathryn Kimball Mizelle, a Trump appointee in Florida, deployed corpus linguistics to strike down the Biden administration’s public transportation mask mandate.
Part of Mizelle’s justification was that “sanitation” — a key word in the 1944 law that gives the CDC its authority to make public health rules — historically described the act of making something clean. “Wearing a mask cleans nothing,” Mizelle wrote.
For unclear reasons, Mizelle’s financial disclosure from 2022 is not yet public, and she did not answer when asked if she attended a Judicial Education Institute retreat.
“Corpus linguistics is the tool that originalists have been waiting for.”
- James C. Phillips, a conservative legal scholar
Proponents of corpus linguistics in the courts claim the tool can result in rulings that are less partisan, not more.
“One of the great virtues of corpus linguistics is that it is falsifiable,” Lee and Stephen C. Mouritsen, a BYU professor who helped Lee first pilot corpus linguistics, wrote in a recent law review article. “If the judge has engaged in cherry-picking or motivated reasoning, a replicated search will reveal this and expose the judge to public criticism.”
Yet the overwhelming majority of judges who have favorably cited corpus linguistics were Republican appointees, according to Kevin Tobia, an assistant professor at the Georgetown University Law Center who tracked the spread of corpus linguistics in a 2021 study.
Lee, who doesn’t have a formal position with the Judicial Education Institute but is a fixture at their retreats, is the person most responsible for corpus linguistics taking off in the courts. Starting in 2011, as a judge on the Utah Supreme Court, Lee used corpus linguistics in a steady drip of rulings, and the practice slowly trickled out to other judges and lawyers at the state level.
Then, in 2018, Supreme Court Justice Clarence Thomas mentioned corpus linguistics in two separate dissents.
In the more notable case, on whether police can obtain a suspect’s location data from his cell phone provider without a warrant, Thomas argued that criminal suspects have no “expectation of privacy” in part because he could not find the exact phrase in any linguistic databases from the time of the founders. “At the founding,” Thomas concluded, “‘search’ did not mean a violation of someone’s reasonable expectation of privacy.”
Thomas’ dissents coincided with a surge of interest. In 2019 and 2020, corpus linguistics popped up in multiple opinions from federal circuits and state courts, with Heilpern sometimes traveling to those states to offer one-day corpus linguistics trainings.
“I have had conversations with dozens of other judges around the country who are actively trying to learn more and looking for the right case to use it on,” Heilpern wrote in the letter he uses to introduce himself to potential trainees.
The Supreme Court hasn’t relied on corpus linguistics in a majority opinion. But it has come up in oral arguments, with Judge Amy Coney Barrett demonstrating that she is familiar with the concept.
In 2022, in Kennedy v. Bremerton, a case on whether a high school football coach had a right to hold public prayers on the field after football games, the court ruled 6-3 in favor of the coach, citing “history and traditions.” The majority opinion, by Justice Neil Gorsuch, pivoted off a lower court dissent by Judge Ryan D. Nelson, who said a search of the founding-era corpora “found no evidence that prayers or religious practices in schools were considered an establishment of religion at the time of the Establishment Clause’s ratification.”
Many of the judges citing corpus linguistics in recent years have some connection to the Judicial Education Institute or have attended one of its retreats.
Although it is not clear where Clarence Thomas first heard of corpus linguistics, Lee clerked for Thomas on the Supreme Court in the mid-’90s, and in 2010, Thomas swore Lee in to the Utah Supreme Court.
Nelson is a former research assistant of Lee’s and attended the Deer Valley retreat. A. Marvin Quattlebaum Jr., a Trump appointee to the Fourth Circuit Court of Appeals, attended The Greenbrier retreat. Amul Thapar, a Trump appointee on the Sixth Circuit, has attended BYU’s semi-regular corpus linguistics conferences. U.S. District Judge Lee Rudofsky is a member of the Judicial Education Institute advisory board.
Other board members include Judge Diarmuid O’Scannlain, a semi-retired Reagan appointee who became a leading conservative voice on the liberal Ninth Circuit Court of Appeals, and Robert George, a Princeton University scholar and a powerful connector on the religious right, who helped convene the 2010s backlash against same-sex marriage and abortion rights. “If there really is a vast right-wing conspiracy,” the conservative Catholic journal Crisis once joked, “its leaders probably meet in George’s kitchen.” George and O’Scannlain did not return requests for an interview.
Rudofsky said he has no involvement with the organization’s fundraising.
“I have worked hard at each stage of the organization’s development to ensure that I have followed all relevant judicial ethics canons,” he wrote in an email. “The training JEI provides is among the best training I have ever received as a judge in terms of rigor, information, fairness and ideological neutrality. And I believe corpus linguistics will, in the near future, be an important (or at the very least useful) tool in the toolkit we judges use to interpret legal texts.”
Pick and choose history
Conservative justices aren’t the only ones to have dabbled in using corpus linguistics or a version of it. As far back as 1998, Justice Stephen Breyer performed a quick-and-dirty analysis of how the word “carry” was used throughout history to determine if a person who had a gun in their vehicle “carries a firearm.”
“Language is not inherently conservative or liberal,” Tobia said. “If our goal is to understand a text, this tool could be useful.”
Tobia is a proponent of more liberal justices adopting its use because textualism and originalism have become the predominant modes of analyzing the law.
But putting corpus linguistics in the hands of judges after a few days’ training, he added, comes with clear downsides.
Linguists train for years to perform unbiased searches and constantly reevaluate the correct way to interpret the results. Mark Davies, a retired Brigham Young University linguist, told The Verge in 2022 he was troubled by the idea of “naive judges … spending their lunch hour doing quick-and-dirty searches of corpora.”
The search for a law’s “original public meaning” invokes the idea of one, singular public, when different groups of people have diverging interpretations of the law, Bernstein said. If those people happened to be enslaved, women or poor, they are not likely to have left an extensive written record that made it into a digitized corpus.
“A lot of what legal corpus work tends to do is allow you to pick and choose your history,” she said. “You pick and choose your corpus, you pick and choose what you’re putting into the software, you pick and choose your results.”
Knowing what people in the founding era believed, she added, doesn’t justify governing based on their standards.
Speaking in the wine cellar, Blackman raised these problems before waving them away.
“Look, we acknowledge the process has not been perfect,” he said. “Lots of people were excluded ― women, slaves and others did not have the right to vote.” But the only alternative, he said, was to be subjected to the whims of the Supreme Court. “The alternative is the rule of nine.”
Legal corpus linguistics, he suggested, would have its day just like originalism and textualism. In 1985, the liberal Supreme Court Justice William J. Brennan Jr. dismissed these concepts as “arrogance cloaked as humility.” Three decades later, Justice Elena Kagan gave a speech in which she declared, “We’re all textualists now.”
“Who won this debate?” Blackman asked. “Scalia. Originalism. They won. It took 30 years. … But we got to the goal line.”