Formant frequencies of some fixed-mandible vowels and a model of speech motor programming by predictive simulation

Formant frequency data are reported for Swedish vowels produced both with fixed and unconstrained mandible. Measurements were made at the first glottal pulse to confine interpretation of the results to nonauditory feedback mechanisms. Results indicated that in spite of physiologically unnatural jaw openings, subjects were able to produce F patterns within the ranges of variation of the normal vowels. Results are explained by hypothesizing that the “instantaneous” learning of highly unfamiliar tasks, such as compensatory articulation of fixed-mandible vowels, is possible because normal speech motor programming is indeed “compensatory” rather than due to either speakers drawing upon past similar experience or invoking special motor mechanisms distinct from those of natural speech. That is, it operates in a context-sensitive mode to achieve listener-oriented goals. Since “contexts” constitutes an essentially infinite class of events the programming must be “creative,” or capable of handling conditions never before experienced. A model of speech production is proposed that achieves increased plasticity of motor control by incorporating an “upstairs” predictive component used to simulate “downstairs” processes of execution.