| Publication | In Burnham, D., Luksaneeyanawin, S., Davis, C., & Lafourcade, M. (eds.) Interdisciplinary approaches to language processing. Bangkok, NECTEC. |
| Abstract | The term voicing is used as a label for both a phonetic property and a phonological feature, sometimes leading to confusion. The most salient phonetic aspect of voicing is audible glottal pulsing. Voicing in a part of an utterance interests us here only if its presence is in opposition phonologically to its absence. Although glottal pulsing is the dominant excitation source in speech, there are also such noise sources as turbulence and transients. In different contexts and across languages, phonological voicing distinctions entail various combinations of these sources as well as such concomitant traits as differences in fundamental frequency upon consonant-release, preconsonantal vowel duration, and intervocalic closure duration. Much perceptual research has been done with synthetic speech and manipulated natural speech. In my earliest research with Leigh Lisker on the acoustics of voicing distinctions, it was convenient in working with a variety of languages to focus on initial position, so the concept of voice onset time (VOT) came to the fore. This is the time of the onset of voicing relative to the release of the initial consonant. Actually, the broader concept of voice timing is relevant to initial, word-medial, and utterance-final positions. Although voice timing by itself is a powerful mechanism for perceptual differentiation of voicing states, research has shown that the concomitant traits mentioned above can also play a role in perception. Voice timing is broadly applicable in languages of the world; yet there are some languages in which non-temporal characteristics intersect with that dimension and thus must be handled and processed separately. |