What's the point of talking? Auditory targets and communicative goals

Abstract

Human speech production is a complex action requiring minute control over the articulators and sensitivity to the surrounding environment. Computational and empirical work has attempted to identify the specific neural mechanisms and cognitive processes that allow us to reliably produce speech sounds. This work has established that humans can use their perception of the auditory and somatosensory consequences of their actions to guide subsequent speech movements. However, speech predominantly takes place in a communicative context, and this context is also known to modulate the way that people speak: human voices are highly flexible. In this paper, we try to unite the traditional motor control conception of internally defined acoustic and somatosensory goals with linguistic research showing that talkers respond and entrain to their conversational partners. We provide an overview of the theoretical and empirical work surrounding the use of sensory feedback monitoring in speech production and discuss practical constraints that have limited more naturalistic investigations into dyadic interaction. To conclude, we argue that the variability of results seen in the speech motor control literature reflects a more complex underlying neural architecture, and an overarching communicative goal that supersedes specific phonetic targets.