The tense/lax distinction in vowels as a predictor of voicing cues for final, post‐vocalic stops

Abstract

Stop release [Wang, J. Speech Hearing Res. (1959)] and the presence/absence of a falling F1 transition [Wolf, J. Phonetics (1978)] have been shown to be effective cues to the voicing distinction in final, post‐vocalic stops in English. Neither, however, are necessary. Therefore, preliminary speech production data were gathered to determine the interrelationship of these cues. Spectrographic analysis indicates (1) that release is predicted by the tenseness of the preceding vowel: stops following tense vowels are released; those following lax vowels are not. (2) Besides serving as a cue to [± voice] in stops, release functions as a segment cue. Final stops following tense vowels are not predictable by the phonology of the language, whereas those following lax vowels are. That is, stressed, tense vowels can end a word in English, but stressed, lax vowels cannot. Thus, stop release is redundant after a lax vowel, but it has a high utility after a tense vowel in that it signals the presence of a stop. (3) stop release and a falling F1, both of which can signal [+ voice] in final stops, are in complementary distribution. Voiced stops following tense vowels are released and the vowel lacks a falling F1; voiced stops following lax vowels are unreleased and the vowel exhibits a noticeable falling F1. In sum, stop release (as a segment cue) and both release and a falling F1 (as voicing cues) are predictable on the basis of the tense/lax distinction in the vowel. In this case, then, the appearance of a particular acoustic cue is a function of the abstract phonological system of English.