ASA 127th Meeting M.I.T. 1994 June 6-10

4pSP34. Influence of silence duration distribution in perception of stop consonant clusters.

Krishna K. Govindarajan

Michael A. Cohen

Cognitive and Neutral Systems Dept., Boston Univ., 111 Cummington St., Boston, MA 02215

The effects of the silence duration between two different stop consonants in /VC[sub 1]--C[sub 2]V/ tokens were examined. Earlier results using synthetic stimuli had shown that the decision boundary between hearing one and two stop consonants varied as a function of the range and frequency number of tokens) of the silent interval [B. H. Repp, Haskins Lab. Stat. Rep. Speech Res. SR-61, 151--165 (1980)]. Using real speech stimuli, The Repp (1980) experiment was replicated, showing that the results hold for individual subjects. Extending this result, the overall variance of the silence duration distribution was manipulated while maintaining the same mean, for two different distributions. The result shows that the overall variance does not significantly affect the boundary, and that listeners must be adapting to the mean silence duration. The final experiment examines this adaptation by presenting trials of sequential tokens to determine the relative weighting of the former token(s) on the judgment of the final token in the trial. The results of this experiment will be presented in relation to the adaptation process. Implications for adaptation in other durational cues, and influence of cue trading will be discussed. [Work supported by AFOSR.]