On the status of temporal cues to phonetic categories: Preceding vowel duration as a cue to voicing in final stop consonants

Abstract

The duration of the preceding vowel has been called a primary and even necessary cue to voicing in final stop consonants. The results of this investigation suggest that in natural speech, vowel duration differences are probably neither necessary nor adequate cues to this distinction and that voicing during closure may be required to disambiguate final voiced stops. The stimuli were 52 one-syllable words recorded by two speakers and subjected to an analog-to-digital process and to linear predictive coding. Deletions, compressions, and expansions of segments of these 104 syllables produced 521 stimulus items which were randomized and presented to 12 adult listeners who judged the syllables to end in a voiced or voiceless stop. Though syllable duration was a more significant cue to the voicing feature than was vowel duration, syllable duration was not a necessary cue in that, even at the extremes of syllable length, syllables with final transition and/or final segment information intact did not effect a crossover in voicing judgment. Syllable duration was not an adequate cue to the voicing feature in that syllables without final transition and final segment information were not heard as better than 60 percent voiced at any syllable duration. By contrast, voicing during closure determined the voicing decision across the full range of syllable durations.