Abstract:
It is hypothesized that segmental speech movements are programmed to achieve perceptual objectives, by realizing articulatory and acoustic goals. As demands for clarity and rate change, movement kinematics and coarticulation will be adjusted in a trade-off between clarity and economy of effort. In a preliminary test of these ideas, articulatory and acoustic signals have been recorded from two speakers under different conditions of clarity and rate. Effort-related measures are calculated from kinematic data on the tongue body, tongue blade and lips; degrees of labial and lingual coarticulation are measured from timing data, and correlates of clarity are calculated from acoustic data. Initial results indicate a complicated picture, with differences among articulators and between the two subjects. In general, one mediating factor seems to be the influence of basic principles of motor control, reflected in such relationships as duration versus movement distance and velocity versus distance. The results will be discussed in relation to the idea that speakers use an internal representation of relations between articulatory and perceptual domains to help manage the trade-off between listener-oriented perceptual requirements and a speaker-oriented objective to conserve effort. [Work supported by NIDCD.]