Variation in anticipatory coarticulation with changes in clarity and rate.

This study tests the hypothesis that the relative timing, or coarticulation, of articulatory movements at VC and CV boundaries is influenced by both the listener's requirement for clarity and the speaker's strategy to economize effort. Movement and acoustic data were collected from 7 subjects who spoke in three conditions: normal, clear, and fast. It was predicted that fast speech would show more coarticulation and clear speech would show less coarticulation than normal speech. The speech materials were designed to investigate coarticulation in the movements of the upper lip and tongue. They consisted of repetitions of [iC(n)u] utterances embedded in carrier phrases, where the number of consonants, n, ranged from 1 to 3. Analyses focused on kinematic measures and the amount of coarticulation (overlap) of the /i-u/ transition movement with the acoustic interval of the /i/. The consonant-string duration was longest in the clear speaking condition and shortest in the fast condition. Compared to the normal condition, peak velocities were higher in the fast and clear speaking conditions, indicating increased effort. The influences of speaking condition on coarticulation and on the formants of the /i/ were small. Thus, even though there was evidence of increased effort in the clear and fast conditions, the hypothesized effects of a trade-off between clarity and economy of effort were minimally evident in formant values for /i/ and measures of coarticulation.