Effect of syllable-initial voicing on temporal characteristics of words

Abstract

As part of an ongoing study of the voicing distinction in syllable-initial stop consonants, we measured segmental durations of consonant–vowel–consonant (CVC) words beginning with voiced stop consonants (e.g., ‘‘big,’’ ‘‘duck,’’ ‘‘gap’’) and matched CVC words beginning with voiceless stop consonants (e.g., ‘‘pig,’’ ‘‘tuck,’’ ‘‘cap’’). Target words were produced in randomized lists together with numerous filler words. Results from the four speakers measured to date revealed a consistent pattern. As expected, initial voiceless stops were associated with a longer voice-onset-time (VOT) than initial voiced stops. In addition, initial voiceless stops were associated with a shorter vowel duration (measured from the onset of voicing) than initial voiced stops. Thus, the change from voiced to voiceless stop entailed both an increase in VOT and a decrease in vowel duration. However, the magnitude of the decrease in vowel duration was insufficient to compensate fully for the increase in VOT, resulting in a net increase in the duration of the CV portion of words beginning with voiceless as opposed to voided stops. The results are discussed in terms of implications for models of speech production and syllable planning, as well as implications for studies of speech perception. [Work supported by NIH/NIDCD.]