Pure Gold, Buried Long Ago in the Book-Review Columns

Exactly 70 years ago this month, an obscure little note on the philosophy of the cognitive and linguistic sciences appeared in a highly unlikely place. Looking back at it today, it seems a marvel to me.

It was a comment on some remarks in an obscure book review in the International Journal of American Linguistics (IJAL), a journal for descriptive work on the indigenous languages of the Americas. The author of the critical comment was the notable American linguist Charles F. Hockett (1916–2000), and his title was “A Note on ‘Structure.’"^*

Hockett was commenting on an assertion that the structure of a language exists only in the statements made by the linguist doing the description: “The structure of a given language or language corpus does not exist until it is stated.”

This is what philosophers call antirealism about grammatical structure. It says in effect that there aren’t really any prepositions, verb phrases, relative clauses, or subjects. They figure only as linguists’ inventions. In the real world, there are people talking, but there is no such thing as grammatical structure.

Hockett’s first thought was that, at least on a first reading, the statement was ridiculous. On deeper reflection, he decided, it was perhaps only misleading. The fact is that two views were in conflict at the time, and Hockett thought they should not be. Later, in 1952 (also in an IJAL book review, as it happened), the linguist Fred Householder dubbed them the “hocus pocus” view (the antirealist one, holding that linguists classify in ways that create patterns, but the patterns aren’t a real part of nature) and the “God’s truth” stance (which says there is actual linguistic structure in sentences that the linguist is supposed to uncover).

But then Hockett asserts that a scientific analysis cannot just catalog regularities in a corpus of collected utterances; it must “account also for utterances which are not in [the] corpus": The linguist “must be able to predict what other utterances the speakers of the language might produce.”

And simply exhibiting patterns is not enough. The patterns must be real, because the linguist’s analysis “parallels what goes on in the nervous system of a language learner, particularly, perhaps, that of a child learning his first language.” That was the point at which Hockett’s note is truly striking. Back when he was writing, linguists did not claim to be studying brain structure. That is a very modern view.

Hockett acknowledges that utterances the child produces may at first not have the structure that other people’s have; but “the child eventually reaches the point of no longer making ‘mistakes’.”

Of course, he concedes, there will still be “lapses” (even adult speakers sometimes have to apologize for a blunder or repeat part of their utterance to correct themselves:

But by the time the child has achieved linguistic adulthood, his speech no longer contains errors; for he has become an authority on the language, a person whose ways of speaking determine what is and what is not an error.

What is so striking about all this is that it appeared in October 1948. It sounds like something from 20 years later. It is a view that anyone acquainted with modern linguistics would associate with Noam Chomsky.

Chomsky certainly decided early in his career that linguistics and psychology were destined to have close interrelations; he touched on it in some lectures he gave at the University of Michigan College of Engineering under the heading “Linguistics, Logic, Psychology, and Computers” in the summer of 1958. But he only really began developing his ideas about how psychology and linguistics were linked in the 1960s, particularly in the first chapter of Aspects of the Theory of Syntax (1965) and in Language and Mind (first edition 1968). What’s remarkable is that Hockett, a linguist whose work was almost completely overlooked once Chomsky’s views had taken hold, was expounding views about linguistic patterns being neurophysiologically real 20 years before Language and Mind, and 10 years before the Michigan lectures.

Notice that Hockett’s reference to the distinction between “lapses” in speech production and the absence of “errors” in a person’s command of their native language is really just Chomsky’s celebrated distinction between performance and competence, put forward in Aspects of the Theory of Syntax (1965). Hockett was sketching the distinction 17 years early. Given Chomsky’s extraordinary prestige in linguistics today, this is a remarkable finding.

Students in the course on the philosophy of linguistics that I’m co-teaching with my philosopher colleague Brian Rabern this semester were instructed to read Hockett’s little note. They’re lucky they didn’t have to discover it for themselves, buried in the 1948 book-review columns of an Amerindianist journal. The only reason I know about it is that it was rescued from obscurity by an American electrical engineer and cryptanalyst who turned to linguistics after World War II: Martin Joos. He spotted Hockett’s note and reproduced it in pages 279–280 of his indispensable anthology Readings in Linguistics I (ACLS, 1957; University of Chicago Press, 1966), to the great benefit of historians of linguistics.

^* Charles F. Hockett, “A note on ‘structure’ [review of de Goeje by W.D. Preston],” International Journal of American Linguistics 14.4 (October 1948), 269–271.