Generative Generative Phonology

  • D. Terence Langendoen
Part of the Cognition and Language: A Series in Psycholinguistics book series (CALS)


The syntactic component σ of a generative grammar G can be thought of as a 4-tuple (N, T, A, R), in which N is the nonterminal vocabulary, T the terminal vocabulary, A the axioms, and R the syntactic rules. It is often assumed that the elements of T, the words of a grammar, can simply be listed in a word-dictionary. However, they cannot, as elementary considerations of the productivity of the rules of word formation immediately show us. To take a very simple example, consider the process whereby the prefix re can be joined to English verbs, to form new, morphologically complex, verbs. That process is not only productive, in the sense that given practically any verb in English, a new verb can be formed from it by prefixing re; it is also recursive, since re can be added freely to verbs already containing the prefix. Thus we have, in English, such infinite sets of words as {analyze, reanalyze, re-reanalyze, re-re-reanalyze, .... Consequently, the elements of T cannot simply be listed. They must be generated.1


Generative Grammar Consonant Cluster Unstressed Syllable Phonological Component Full Vowel 
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. 1.
    Not only the morphological structures, but also the semantic structures, of the elements of T must be generated. How this is done is beyond the scope of this essay; for discussion, see Langendoen (1982).Google Scholar
  2. 2.
    I assume that each morpheme has an underlying phonemic shape, which may be altered by application of morphophonemic rules, the discussion of which is also beyond the scope of this essay.Google Scholar
  3. 3.
    I take the claim that there is no phonotactic limitation on the length of English morphemes as not requiring detailed justification. There is, to be sure, a longest meaningful morpheme in English, but its length (whatever it is) is not a consequence of a phonotactic limitation.Google Scholar
  4. 4.
    Once again, we omit consideration of the semantic properties of morphemes. We may assume that one of the tasks of the semantic component of a generative grammar is to generate a set of semantic structures, a finite subset of which is associated with morphemes of the language by an arbitrary pairing.Google Scholar
  5. 5.
    The phonological component ø of G is thus very much like the ‘phonological grammar’ of Householder (1959).Google Scholar
  6. 6.
    The notation is that of the X̄-theory of phrase-structure grammar, proposed originally by Chomsky (1970) and developed in detail by Jackendoff (1977).Google Scholar
  7. 7.
    We give the morphemes first as they are spelled in English and second in their underlying phonemic forms. Phonemic representations, when not given in tree-diagrammatic forms, are enclosed within solidi.Google Scholar
  8. 8.
    Note that this is a restriction of the underlying phonological forms of morphemes, not on their surface forms. The contracted form m (from am,/ám/) is not coronal, and the contracted form r (from are, /ár/) is not anterior.Google Scholar
  9. 9.
    In those dialects in which forms like scour are disyllabic, the analysis is as in (i): (i) scour, /skáwr/.Google Scholar
  10. 10.
    The problem then arises of accounting for the fact that the morphemes represented by /yūl/ are associated with the same semantic structure. I assume that this is not a problem that requires a solution within English grammar but that it is a problem for linguistic theory. Essentially what is needed is a principle from which it follows that identical sequences of phonemes in a language receive the same pairings with semantic structures. We may presume that this principle (whatever it is) establishes the general range within which so-called ‘free-variants’ can vary both phonemically and phonotactically.Google Scholar
  11. 11.
    In syntax, nouns (like consonants) are inserted before verbs (like sonants), and modifiers are inserted before heads; cf. Chomsky (1965, chapter 2).Google Scholar
  12. 12.
    I am indebted to Janet Fodor for this insight. I assume that full vowels are more sonorant than reduced vowels, which are more sonorant than glides, which are more sonorant than liquids, which are more sonorant than nasals, which are more sonorant than aspirates, which are more sonorant than fricatives, which are more sonorant than affricates, which are more sonorant than stops.Google Scholar
  13. 13.
    One important consequence of the relative sonority principle is that initial consonant clusters like str in (3j), /stríkt/, and final consonant clusters like mps in glimpse /glímps/, must be analyzed as ternary-branching. Thus we predict that the insertion of certain head consonants may depend on the prior choice of both the modifier that precedes it and the modifier that follows it. As a case in point, consider the rule schema that inserts the head t in initial consonant clusters. If no modifier precedes, then either r orw can follow, as in (3b) /tránz/, and twist, /twíst/. However, if s precedes, then only r can follow (/stwíkt/ is not a possible English morpheme; at best it will be heard as a misrendering of /strikt/).Google Scholar
  14. 14.
    Except that anS 2 that is headed by a sonant that is not a full vowel is necessarily metrically weak, as in (3a), /nәs/.Google Scholar
  15. 15.
    Liberman and Prince’s trees actually analyze words, not morphemes. It is beyond the scope of this essay to consider the analysis of English word stress and metrical structure; see Langendoen (1982) for discussion.Google Scholar
  16. 16.
    We also assume the converse principle of Chomsky and Halle (1968), namely, that syllables headed by full vowels that fall in an unstressed position in a word are reduced. Thus, for example, the vowel i that heads the second syllable in (4b), /āábil/ is reduced in the word able. On the other hand, in the word ability, it is the vowel ā of the first syllable that reduces thanks to this principle.Google Scholar
  17. 17.
    Except to point out that certain investigators prefer to think that bisyllabic morphemes like (4e), /mónē/, have a single phonotactic analysis, in which the intervocalic consonant simultaneously belongs to both syllables (see Kahn, 1976, for a detailed justification of this position). Obviously, such a treatment cannot be accommodated within the framework presented here, since we assume that the phonotactic structures of morphemes are generated by constituent-structure grammars. In defense of our position, it may be noted that it treats the analysis of morphemes like /mónē/ on a par with the analysis of (3f), /yūúl/, and (4h), /rṑbΛ́st/ or /rṓbΛ́st/, namely as falling within the scope of a principle of free variation (see note 10). This is a generalization that is not expressible in frameworks like Kahn’s.Google Scholar


  1. Chomsky, N. Aspects of the theory of syntax. Cambridge, Mass.: M.I.T. Press, 1965.Google Scholar
  2. Chomsky, N. Remarks on nominalization. In R. A. Jacobs and P. S. Rosenbaum (Eds.), Readings in English transformational grammar. Waltham, Mass.: Ginn, 1970.Google Scholar
  3. Chomsky, N. & Halle, M. The sound pattern of English. New York: Harper and Row, 1968.Google Scholar
  4. Halle, M. Phonology in generative grammar. Word, 1962, 18, 54–72.Google Scholar
  5. Householder, F. W. On linguistic primes. Word, 1959, 15, 231–239.Google Scholar
  6. Jackendoff, R. S. X-syntax: A study of phrase structure. (inguistic inquiry monographs, 2). Cambridge, Mass: M.I.T. Press, 1977.Google Scholar
  7. Kahn, D. Syllable-based generalizations in English phonology. Bloomington, Ind.: Indiana University Linguistics Club, 1976.Google Scholar
  8. Langendoen, D. T. The grammatical analysis of texts, In S. Allén (Ed.), Text processing: Proceedings of Nobel Symposium 51. Stockholm, Sweden: Almquist-Wiksell, 1982.Google Scholar
  9. Liberman, M.,& Prince, A. On stress and linguistic rhythm. Linguistic Inquiry, 1977, 8, 249–336.Google Scholar

Copyright information

© Springer Science+Business Media New York 1984

Authors and Affiliations

  • D. Terence Langendoen
    • 1
  1. 1.Linguistics Program, Brooklyn CollegePh.D. Program in Linguistics, Graduate School of the City University of New YorkUSA

Personalised recommendations