Context effects on sensitivity, bias, and parsing phonetic information.

Project funded by NIH, NIDCD, R01-DC006241, 2004-2007. An application for renewal has been submitted.

This project tests the hpyothesis that incoming speech signals pass through a stage of linguistically naive, auditory processing before undergoing linguistically informed processing. It thus tests a model of speech perception which is autonomous rather than than interactive.

Three papers have been submitted for publication - all are being revised for resubmission - and a fourth is in preparation. Pre-prints will be made available here once the papers are accepted.

A paper laying out the guiding ideas and presenting preliminary supporting data is "From ears to categories: New arguments for autonomy" S. Frota, M. Vigario, & M. Freitas (eds.) Prosodies, Proceedings of the First Conference on Phonetics and Phonology in Iberia, (pp. 177-222), Mouton de Gruyter, (2005). The link is to the proofs of the published chapter.

"On the internal perceptual structure of distinctive features: The [voice] contrast" Kingston, J., Diehl, R. L., Kirk, C. J., & Castleman, W. A. Journal of Phonetics, 36, 28-54.

This paper reports the results of experiments using speech and non-speech analogues to test the hypothesis that acoustic properties that covary in the stops contrasting for [voice] integrate perceptually so as to enhance the contrast.

"The phonetics-phonology interface," Paul de Lacy (ed.), The Handbook of Phonology, (pp. 401-434), Cambridge: UK, Cambridge University Press.

This chapter describes my recent attempt to characterize this interface.

"Lenition," L. Colantoni & J. Steele (eds.) Proceedings of the Third Conference on Laboratory Approaches to Spanish Phonology, Cascadilla Press. The link is to the submitted final version of the ms.

This paper was developed from a plenary address given to the conference at the University of Toronto in September 2006. The paper argues that consonant lenition is not an attempt by speakers to minimize effort but instead a means they use to convey the position of the affected consonant inside a prosodic constituent rather than at its edge. That is, it is information about prosodic constituency.

"The phonetics of Athabaskan tonogenesis," S. Hargus & K. Rice (eds.), Athabaskan Prosody, John Benjamins Press. The link is to the submitted, camera-ready version of the ms.

This paper substantially revises an account of Athabaskan tonogenesis that I first advanced in my dissertation, The Phonetics and Phonology of the Timing of Oral and Glottal Events, University of California, Berkeley (1985). Tones developed in these languages from a contrast in the protolanguage between syllable-final glottalic versus non-glottal consonants. The paper accounts for two factors that complicate this sound change: (1) some of the present-day languages have developed low tones in syllables than originally ended in glottalic consonants, while others have developed high tones (with the opposite tone developing in all other syllables), and (2) when the syllable ends in a glottalic stop, the expected tone only develops when the vowel is short, and the tone that otherwise develops in syllables not ending in glottalic consonants develops when the vowel is long; however, when the syllable ends in a glottalic fricative or sonorant. the expected tone develops on long as well as short vowels. To explain how both low and high tones can come from the same source, I show that a glottalic consonant can be pronounced in two ways, one which causes the preceding vowel's voice quality to be creaky and its F0 low and the other which instead causes its voice quality to be tense and its F0 high. The second complication is a consequence of a difference in timing of the glottalic articulation relative to the oral articulation in stops versus fricatives and sonorants. A dramatic acoustic event occurs when an oral stop closure is released, the burst, which is absent when the oral articulation of a fricative or sonorant is released. The glottalic articulation "binds" temporally to the burst and is thus too far from the preceding vowel to affect its voice quality or F0 noticeably unless that vowel is short. In the absence of such an acoustic event to which the glottalic articulation can bind in a fricative or sonorant, that laryngeal articulation is freer to shift earlier in the consonant and even onto the preceding vowel. As a result, the voice quality and F0 of long as well as short vowels are noticeably altered.

"Segmental influences on F0: Controlled or automatic?" C. Gussenhoven & T. Riad, (eds.), Tones and Tunes, v. 2, (pp. 171-210), Mouton de Gruyter, (2007). The link is to the final version of the submitted ms.

This paper was developed from a presentation given at the Tone and Intonation in Europe conference on the island of Santorini in 2004. The experiments reported in it were designed to test the hypothesis that F0 differences in vowels differing in height or next to obstruents differing voicing are controlled rather than automatic consequences of these segments' other articulations. Previous studies had indicated that the cricothyroid muscle contracted more in higher than lower vowels, which suggests F0 is controlled independently of vowel height. The larynx is lowered in voiced obstruents compared to voiceless ones, which suggests that F0 is automatically lowered in adjacent vowels. Words containing the segments of interest were produced in prosodic contexts that differed in how they influenced to the intonationally determined F0 values. If F0 differed between vowels differing in height or next to obstruents differing in voicing across these prosodic contexts, that would suggest these differences are automatic. If the occurrence of the segmentally determined F0 differences instead depended on prosodic context, that would suggest that they are not automatic but controlled. Surprisingly, the results suggested that the vowel height F0 differences are automatic. The obstruent voicing differences were too small and variable for any firm conclusion to be drawn.