Evidence for a 3-component model of speech production with phonology-extrinsic timing

This talk will present evidence supporting a model of speech production which includes phonology-extrinsic timing, and consists of 1) a Phonological Planning Component to plan the goals for an utterance, including the segmental and prosodic structure for an utterance and non-grammatical goals such as speaking quickly or in a particular style, 2) a Phonetic Planning Component to plan the quantitative details of the acoustic goals and how they will be achieved articulatorily, and 3) a Motor-Sensory Implementation Component to ensure that the goals are achieved on time. Supporting evidence from the literature includes findings of greater timing precision at movement endpoints compared to other parts of movements, suggesting the separate control of the timing of movement endpoints compared to other parts of movement. This evidence presents a challenge to models in which all parts of a movement trajectory are controlled by the same equation of motion, but is consistent with models in which 1) abstract, symbolic phonological representations map onto spatial and temporal characteristics of movement endpoints, 2) movements are planned to reach the endpoints on time, and 3) speakers give priority to the accurate implementation of the part(s) of movement most closely related to the phonological goals.