ABSTRACT
In this chapter, the authors describe a model of speech motor control which uses a dynamical system to infer movements from a sequence of motor commands. They explore the model output to real movement observed during natural speech. The tongue also provides many opportunities for coarticulation effects and so provides a challenging problem for understanding the gestural intention structure of speech. In addition to the authors modeling efforts, they have also collected a small amount of data from two other areas of speech science that provides further support for their approach to the control of tongue movement. The authors modeling results do not offer any clear answers, but suggest that modifications in the intention score would be a likely place to adjust the speaking rate. This speech error is consistent with the fundamental basis of the authors model, because it can reasonably be attributed to the simple deletion of one element of the intention score—the glottal approximation gesture.
