Posner and Segall's Evasion of the Sense-Reference Distinction
Judge Posner and Eric Segall have each recently replied (in a 4-page co-authored piece at the end of the collection here, and in a longer solo piece by Segall here) to Will Baude and Stephen Sachs's recent work on originalism (a 6-page co-authored piece in the same collection, Baude alone here, Sachs alone here and here, and a forthcoming co-authored piece here). Baude and Sachs hold that our constitutional law must "lawfully derive from the law of the founding." I have a bit of sympathy for Posner and Segall's terminological quibble; Baude and Sachs's view might more usefully be called meta-originalism, rather than originalism as such, because it can encompass those who think that we have always had a common-law, intergenerationally-authored Constitution--David Strauss, Jed Rubenfeld, or Ronald Dworkin, say. Unlike Posner and Segall, though, I think Baude and Sachs's original-law meta-originalism is both true and important.
One important flaw in Posner and Segall's recent contributions is their failure to engage with Baude and Sachs's use of the sense-reference distinction, which philosophers of language have used at least since the time of Frege to explain how linguistic categories work. I wrote at length about the distinction over ten years ago, though, as Baude and Sachs kindly note, "too few scholars have read" my work. Some philosophers have, to be sure, criticized the distinction (and others have defended it), but Posner and Segall simply ignore it. That seems "obviously wrong, yet eminently curable," in Posner's words.
The basic idea is that building cars does not, in itself, change the meaning of the word "car"--the "sense" of the word. But building cars does change something about the word "car": its "reference," or collection of things picked out by the term. So, if someone asks me how many cars there are, I need to know (a) what he means by "car" and (b) the automotive facts. Both linguistic and automotive facts change all the time, of course, but they do so in different ways and for different reasons. If we want to find out how many cars there are in America, we have to do so (paraphrasing Brown) "in the light of [cars'] full development and [their] present place in American life." Current automotive facts are essential, even if we are merely applying a word that someone spoke long ago (e.g., in verifying or falsifying a prediction long ago that someday there will be a certain number of cars in America).
My view is that original sense, but not original reference, is interpretively binding. This is how the meaning of words can be both fixed but abstract. The original sense determines a function from possible worlds to outcomes; current facts tell us what possible world we are in, which we then plug into the sense-determined function, producing an outcome. As Baude put it at p. 2356, footnoting my paper, "a word can have a fixed abstract meaning even if the specific facts that meaning points to change over time."
After quoting this statement, Segall asks incredulously at p. 40, "How can the meaning of words be simultaneously 'fixed' and 'abstract?' " But Segall does not then engage the sense-reference distinction on which Baude and I rely, but instead jumps to disputes over the interpretation-construction distinction, on which originalists have differed regarding how to deal with vagueness, ambiguity, and kindred phenomena. But vagueness is not the same thing as abstraction. A term is "abstract," as Baude uses the term, if its reference can change without the term's meaning changing as well. On the other hand, a term is vague if it has a fuzzy, unclear boundary; these are not the same thing. I offer my own take on how to deal with vagueness and lack of clarity here and here, but it is important to see that Baude plainly does not use the interpretation-construction distinction to explain how meaning can be both fixed and abstract. The sense-reference distinction does that, whatever we think about construction and interpretation.
An appreciation for the sense-reference distinction would have prevented Posner and Segall's misreading of Brown v. Board of Education, which I paraphrased above. Posner and Segall think (at p. 110) that in stressing the need to consider the "full development" and "present place" of education, "The Court in Brown expressly rejected the idea that history mattered." But the relevance of present facts does not entail the irrelevance of historical facts. Indeed, Brown itself says that historical sources "cast some light," but "not enough."
History tells us that (as I see it) the Fourteenth Amendment bans second-class citizenship, but current facts tell us what rights are in fact given generally to citizens (what counts as "the very foundation of good citizenship," as Brown puts it) today. Building new cars does not change what it means to be a "car"; neither does building new privileges of citizens change what it means to be a privilege of citizens of the United States. To get outcomes, we need both (a) the historically-fixed function from possible worlds to outcomes and (b) a specification of which world we're in. Our need for reference-yielding current facts in (b) does not undermine our simultaneous need for historically-fixed sense in (a).