1 Davidson, Donald, “Truth and Meaning,” in Readings in the Philosophy of Language, Rosenberg and Travis, eds. (Englewood Cliffs, N. J.: Prentice-Hall, 1971), p. 454.Google Scholar

2 For example, Chomsky, , Syntactic Structures (The Hague: Mouton, 1957), pp. 15·17.Google Scholar

3 It is, of course, open to doubt whether the natives will always Judge utterances as determinately significant and whether the grammarian will always be capable of detecting native acceptance. Still, doubts on this score cannot give way to total despair. Speakers (properly so called) of any language (properly so called) do distinguish its meaningful expressions from senseless noise and even if it's difficult to explain what this distinction amounts to, it is often easy to tell that they have, and what their opinion is. Obvious examples of acceptance will be found in any human community, and the grammarian's enterprise will draw its objectivity from its successes in accounting for the cases everyone can agree upon.

4 Adapted from the more standard ‘terminal string'. For comparison between this and other accounts of grammar, cf. Chomsky, “On the notion ‘Rule of Grammar',“ in The Structure of Language, J. Fodor & Katz, J., eds. (Englewood Cliffs, N. J.: Prentice-Hall, 1964);Google Scholar or “Topics in the Theory of Generative Grammar,” in The Philosophy of Language, Searle, J. R., ed. (London: Oxford University Press, 1971);Google Scholar or Quine, “Reply to Geach,” in Words and Objections, Davidson, D. & Hintikka, J., eds. (New York: Humanities Press, 1969). pp. 328-32.Google Scholar

5 Cf. Chomsky, , Aspects of the Theory of Syntax (Cambridge, Mass.: MIT Press, 1965),Google Scholar Chap. 1 passim.

6 Davidson, “Truth and Meaning,” p. 453.

7 Cf. Quine, , “Methodological Reflections on Current Linguistic Theory,” in Se· mantics of Natural Language, Davidson, D. & Harman, G., eds. (Dordrecht, Holland: D. Reidel, 1972), pp. 445-46.Google Scholar

8 Cf. Chomsky, Aspects, Chap. 1, Sec. 6.

9 As trivial as disjunction. See (D3) below.

10 I do not, of course, suppose that a recursive syntax of English would literally describe utterances in the way (D1) and (D2) do.

11 Chomsky, , Current Issues in Linguistic Theory, (The Hague: Mouton, 1964), p. 29Google Scholar and Chap. 2 passim. Cf. also Syntactic Structures, 6.1, and Aspects, 1.6.

12 Lest all this reifying talk about sentence-types raises any nominalist qualms, recall that Goodman and Quine made it one of the first of their steps toward a constructive nominalism to argue that talk of sentence-types is reducible to talk about actual particulars. Nominalist grammars fit neatly into the general form described on pp. 65-6. Goodman suggests that instead of speaking of tokens as instantiating the same sentence- (or word- or letter-) types the nominalist should speak of tokens as being “replicas”. “Being a replica” is an equivalence relation satisfied by any two tokens of the same expression-type. While Goodman allows that the replica relation is, despite its name, not a matter of mere physical similarity but is, rather, a conventionally determined feature of a language, he nowhere explains how we go about discovering if tokens are replicas in a language. Cf. “Steps towards a constructive nominalism,” Goodman, and Quine, , in Journal of Symbolic Logic, 12 (1947),Google Scholar and Languages of Art, Goodman (New York: Bobbs-Merrill, 1968). pp. 137-38.

13 Actually this simplifies enormously. Throughout the following I will adopt a very simple taxonomic model of phonology (cf. Chomsky, Current Issues, p. 78). I will ignore the complications of suprasegmental phonemes, distinctive features analysis, generative phonology, and even context sensitivity. None of these complications affects the point I want to make and all of the problems I will raise for the taxonomic model will apply equally well to more sophisticated approaches to the phonological problem.

14 Cf. Bloomfield, , Language, 57, 77;Google Scholar and Trubetzkoy, N. S., “Phonemes and How to Determine Them,” in Phonology, Fudge, E. C., ed. (Middlesex, England: Penguin, 1973), p. 53.Google Scholar

15 Cf. Seuren, Pieter A.M., Semantic Syntax (London: Oxford University Press, 1974), pp. 97–98.Google Scholar

16 Harris, Zellig, Structural Linguistics (Chicago: University of Chicago Press, 1951). pp. 32–33;Google ScholarHalle, Morris, “The Strategy of Phonemics,” in Word, 10 (1954), 200;CrossRefGoogle Scholar Putnam, “Some Issues,” pp. 92-93; and Chomsky, Syntactic Structures, p. 96.

17 Wedberg, Anders, “On the Principles of Phonemic Analysis,” Ajatus 26 (1964). 235-53.Google Scholar

18 Wedberg, “Phonemic Analysis,” p. 252.

19 Consider the language L *: let every acceptable utterance of L * consist of a sequence of two or more tokens of a set of phones {p, q, r, #} and suppose that any such sequence of phones in any order will be well-formed. Now interpret ‘p', 'q', ‘r’ as unstructured atomic sentences, and let the concatenation of their tokens in strings signify their conjunction. Thus ‘pqr’ is ‘p & q & r'. Let the concatenation of any phone with itself indicate its negation with the scope read as associating to the left. Thus ‘ppp’ is ‘-p & p'. Now let ‘#’ be a symbol which has no effect on the reading of any string save when it occurs flanked by other phones, in which case it has the effect of conjoining them. So ‘p#pp’ is p&-p'. L * contains propositional logic and notice that, while every two phones are every where interchangeable salva well-formedness, for every two phones there is a string containing one which, when the other phone is put in its place, yields a proposition logically inconsistent with the original (e.g., compare ‘ppp’ and 'pqp’).

20 Words and Objections, p. 94; cf. also Quine, , Philosophy of Logic (Englewood Cliffs, N. J.: Prentice-Hall, 1970), p. 16,Google Scholar and “Methodological Reflections,” p. 450.

21 If all else fails, Quine recommends (in “Methodological Reflections“) that we try individuating phonemes by failures of stimulus synonymy. This is a richer notion than assent but to discuss it here would take us far beyond our present purposes. Suffice it to note that stimulus synonymy is the strongest kind of equivalence in meaning allowed by Quinean semantics.

22 Syntactic Structures, p. 94. Cf. also Harris, Roy, Synonymy and Linguistic Analysis (Oxford: Blackwell, 1973),Google Scholar Chap. 2, passim.

23 Carnap, R., The Logical Structure of Language (London: Routledge & Kegan Paul, 1937), p. 1.Google Scholar

24 Philosophy of Logic, p. 26.

25 Ibid. p. 35.

26 I emphasize “source” since any contrast in meaning is the grammarian's business even if it turns out that the contrast doesn't pair with any phonetic difference but instead, e.g., with lexical ambiguity. One finds, by the way, some evidence that that individuation of expressions is tied to meaning in the practice of speaking of ambiguous words as “homonyms,” i.e. “different words with the same pronunciation.“ That we are thus willing to multiply expressions beyond their phonetic realizations testifies to our willingness (albeit a reckless willingness, from an ontological point of view) to treat difference in meaning as difference in ex· pression.

27 Cf. Harman, Gilbert, “Logical Form,” in The Logic of Grammar, Davidson & Harman, eds. (Encino: Dickenson, 1975).Google Scholar