Vowel length and closure duration in word-medial VC sequences

Abstract

Previous research has shown that English vowel length varies depending on the voicing characteristic of the following consonant. For stop consonants, closure durations also vary as a function of consonantal voicing. Generally, vowel-stop sequences containing voiced consonants show longer vowel durations and shorter closure durations than similar sequences containing voiceless consonants. These previous studies have focused on stressed vowels in monosyllabic or bisyllabic words. Very little research has examined the effects of postvocalic voicing on unstressed vowels. In the present study, we examined the influence of postvocalic voicing on vowel and closure durations in VCV and VCV sequences. Subjects produced sentence pairs containing target words contrasting in intervocalic consonantal voicing (e.g., adopt/atop, tabbing/tapping). Both stressed and unstressed vowels tended to lengthen before voiced consonants. However, the vowel-lengthening effect was not as consistent for unstressed vowels as for stressed vowels. Closure durations following stressed vowels were longer for voiceless stops than for voiced stops. However, voicing effects on closure duration were inconsistent after unstressed vowels. The results have implications concerning perceptual cues for intervocalic voicing, the syllabification of intervocalic consonants, and the nature of syllable-internal structure.