VOT variability in stop-consonant productions by bilingual speakers of English and Mandarin Chinese

Abstract

It has been observed [e.g., L. Williams, Percept. Psychophys. 21, 289–297 (1977)] that some bilingual speakers of Spanish and English will use language-appropriate VOT values in elicited sets of English and Spanish stops. As part of an ongoing cross-language project in repeated-measures acoustic phonetics, six native speakers of Mandarin Chinese (three females and three males) were asked to produce repetitions of English and Chinese stops in a nonsense /ha-'Ca/context. For the Chinese stops, five separate series were collected, with the stressed syllable pronounced on each of five Chinese tones. Productions were digitized via pulse code modulator, stored on video tape, and VOT segments identified and measured using waveform analysis software developed in our laboratory. Measurements of mean VOTs for each talker indicate that English and Mandarin Chinese are similar in range of absolute VOT values, clear VOT distinctions between voiced/voiceless stops, and inconsistent use of VOT to distinguish place of articulation. Further, VOT variability patterns suggest that for some of these talkers, control of VOT timing may be sensitive to tonal context, and this sensitivity may be reflected in productions of English as well as Chinese stops. [Work supported by AFOSR.]