Viability of VOT as a parameter for speaker identification: Evidence from Hong Kong

This research explores VOT as a speaker-specific property within the context of English- Cantonese bilingualism in Hong Kong. Utterances collected from five individuals for /p, t, b, d/ vary over two languages and four emotional states. Results show that VOT means by themselves appear to be generally useless as a speaker-specific property because there is as much inter-speaker variation as there is intra-speaker. However, in this paper, we have been able to show that the profile of VOT shifts of each phoneme across the two languages is speaker-specific. Extrapolating from this, profiles of VOT shifts across other parameters like emotional states and even vowel adjacency would be likewise speaker-specific. Herein lies the viability of VOT for speaker-identification.