My paper is now out in @HarvLRev! Originalist & textualist theories seek the “original public meaning” of legal texts. This paper takes those theories on their own terms, testing whether their tools actually reflect how Americans understand language. https://harvardlawreview.org/2020/12/testing-ordinary-meaning/ 1/

On a common originalist & textualist view, the Constitution and statutes should be interpreted as the ordinary public first understood them. But how should courts uncover this “ordinary meaning”? 2/

Just this week Chief Justice Roberts raised one possible solution: “take a poll of 100 ordinary speakers of English and ask them what it means” (Facebook v Duguid oral argument) 3/

There’s a lot to be said for polling ordinary people. Originalism & textualism direct us to ordinary people’s understanding of language because it is democratic, promotes reliance values, and makes law more public. Here’s J. Gorsuch on those values (in his Bostock opinion) 4/

But originalism doesn't rely on polls. (Of course, we can’t poll past generations.) Instead, it searches for original meaning with tools like dictionary definitions and “legal corpus linguistics” (“LCL”) 5/

This article tests whether use of those tools reflects what ordinary people report language to mean *today*. It finds that those two tools can recommend wildly different answers from each other, and they also diverge from ordinary poll responses. 6/

As an example, take the “no vehicles in the park” hypo. Justice Kavanaugh recently recalled this familiar hypo (in his Bostock dissent). He notes that the “ordinary meaning” of the term “vehicle” does not encompass baby strollers. 7/

Let’s assume (for the sake of argument) that it’s sensible to talk this way: Terms like “vehicle” have an “ordinary meaning.” I polled ordinary people: Is a baby stroller a vehicle? Most agreed with J. Kavanaugh: ~75% say no. 8/

But now take a different example. Is a canoe a vehicle? People were divided about this question (~50-50). But judges answering with a dictionary gave a clear answer: 95% say yes. And judges answering with LCL data strongly endorsed the opposite result: Only 10% agreed. 9/

Originalist judges sometimes rely on dictionaries, sometimes on LCL-style analyses & anecdotes about what's the most “common” way to use a term. Those methods can deliver conflicting results. Flexibility about which tool to use allows flexibility in judicial decision making. 10/

Moreover, neither tool systematically reflected what ordinary people report about meaning. E.g. Today, 70% of people say an airplane is a vehicle, but only 20% of LCL users came to that same conclusion. 11/

These experiments reveal divergences among dictionaries, LCL, and ordinary people’s understanding of language *today*. The results also speak to the use of these tools in originalist projects. 12/

E.g. We can’t poll 18th century Americans, but these results also raise concerns that dictionary-use and LCL analyses are not straightforward reflections of constitutional “original public meaning.” 13/

There are several very thoughtful responses to the paper:
from @ASKrishnakumar (soon appearing in @HarvLRev)
from Neal Goldfarb https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3553016
& from Justice Lee & Stephen Mouritsen https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3727504 14/

Finally, this article is part of larger trend in “experimental interpretation,” which intersects with the new field of “experimental jurisprudence” 15/

On the former, check out:
@omribenshahar & @liorjs: https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2905873

@Jammacleod1: https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3729005

@ShlomoKlapper: https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3593917 16/

And for some “XJur” beyond legal interpretation:
@rosesomm: https://www.yalelawjournal.org/article/commonsense-consent
@scottjshapiro & Knobe: https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3544982
@RaffDonelson & Ivar Hannikainen: https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3283952 17/