Phonemes, segments and features

Abstract

As the token generative phonologist invited to comment on the article by Hickok (henceforth H.), I feel that it is incumbent upon me to both clarify some terms and to counter an assumption about the content of generative theories of phonology made in H.'s article. While I restrict myself most specifically to H.'s article, I hope that these comments are also of some use to others who, like H., aim to integrate the ideas of various traditions in models of speech processing.

Keywords:

• phonology
• phonological theory
• phonemes
• segments
• features

Notes

1. Thanks to Matt Goldrick for useful and encouraging discussion of these comments as I contemplated how to write them up, and to Matt and an anonymous reviewer for helpful feedback. Remaining errors are mine.

2. I am rather conveniently avoiding here the important difference between the structuralist and generativist conceptions of the phoneme, defined by a break in tradition and analysis that is commonly attributed to Halle (1959). When making this difference mattered a great deal, what I am calling a “phoneme” here would more properly have been called a “morphophoneme”. See Anderson (1985) for discussion of this history.

3. See Baković (2013) for in-depth discussion of particular versions of these and other analyses.

4. Similarly, while the features employed by Chomsky and Halle (1968) were described primarily in terms of their articulatory correlates, the authors are clear that “the acoustical and perceptual correlates” of features are given a back seat only “because such discussions would make this section, which is itself a digression from the main theme of our book, much too long” (Chomsky & Halle, 1968, p. 299).

5. Also to be noted here is the “substance-free” approach (Hale & Reiss, 2008), largely defined in opposition to the resurgence of interest in phonetic motivation noted in the text, denying that there is any phonetic content to phonological representations (auditory, articulatory or otherwise).