Letters to Language

Language accepts letters from readers that briefly and succinctly respond to or comment upon either material published previously in the journal or issues deemed of importance to the field. The editor reserves the right to edit letters as needed. Brief replies from relevant parties are included as warranted.

A historical corrective

January 22, 2005

To the Editor:

These remarks concern intellectual credit for certain ideas in Ch. 6 of my book, Skeptical linguistic essays (New York: Oxford University Press), published in January 2004, credit which should have been given but due largely to faulty memory on my part was not. I am greatly indebted for this opportunity to set the record straight in the pages of Language.¹

In going through some old boxes of papers in mid-2004, I discovered a handwritten letter to me from a former lifetime member of the Linguistic Society of America (LSA) named Lester A. Rice. The secretariat of the LSA has since informed me that, unfortunately, Mr. Rice died in 1994. He was apparently one of the students in a large syntax class I gave at an LSA summer institute at Indiana University in the summer of 1964. I reproduce here most of his letter, including everything relevant to the issues at hand. I note that he indicates in his letter that he had dropped out of linguistics and was then living in Spain.

Letter from Lester A. Rice to Paul M. Postal (May 15, 1986)

Dear Paul:

With regard to Terry’s and your The vastness of natural languages,² which I haven’t seen and know only through Abbott’s review in Lg.,³ I would like to add a couple of observations which I believe bear on vastness and even greater vastness.

Notice that in indirect quotes in English the quoted material appears in well-formed English, generally. However, in direct quotes there seems to be little or no constraint that the quoted material be well-formed, or in English, or limited in any way as to length, or even be ‘linguistic’ (this letter if we include written material). In particular, I can cite a sentence recently published in an article in Lg.,⁴ slightly paraphrased:

The man asked the woman what she had for lunch and she answered [sakatsho:ri nii] and he replied [A:ke ki’ na:’a tsi nikanutaraku] and a further discussion of soup followed.

Since there is no restriction on what follows the words ‘answered’ and ‘replied’, we are free to insert at each of these points the entire set or collection (for instance) of the sentences of ns; etc. And in general a simple sentence of any one NL [natural language] can contain, in this way, at once, all other NLs, in addition perhaps to much else.

Even on a level of sentence production and interpretation, folklore tales permit a narrator to create personages and their dialog; in this way a linguistically sophisticated narrator has the power and knowledge in principle to represent or imply representing in some fashion all NLs. (In finite terms, this would mean by token and by ‘Selbstverstandlichkeit’.) Thus, ‘He said “X”’, where X represents all earthly NLs past present and future, plus all NLs which by their nature will never cross the human mind, is merely a single S of English. It’s an impeccable and simple S of English. Each NL must at once contain every other NL.

Paragraph two of the letter makes abundantly clear that Mr. Rice had observed one of the key factual bases for the conclusions of Ch. 6 of my book. He had noticed in particular that there are well-formed sentences in natural language L some of whose parts are not composed of elements drawn from any conceivable lexicon or morpheme listing of L. He gave such an English sentence in which the two parts not drawn from the English lexicon each involve elements of the Iroquoian language Mohawk. In Ch. 6 I cited several other classes of facts which support parallel inferences about natural language and concluded inter alia that on these bases alone, natural languages cannot have generative (proof-theoretic) grammars...