Filling the perceptuo-motor gap

Introduction The session under focus deals with phonetic articulatory-acoustic variation and with the link between phonetic detail and phonological modelling. The underlying question was to estimate the impact of phonetic detail on the phonological status of a given unit, and to attempt to better understand how details are produced by the speaker or exploited by the listener to access (encode or decode) the phonological level. But what is over and over at work in the four papers of this session is the possibility that there could exist a gap between the speaker's intention and the listener's perception, and that the phonetic variation is in some sense contained, produced or at least made possible by this gap. The present discussion will be focused on the perceptuo-motor gap. In an initial section, I will briefly recall how speech communication theories deal with the perceptuo-motor link. Then, I will discuss each of the four papers of the session, around a single question – What does a listener know about a speaker's gesture? – that is, what does the corresponding study tell us about the perceptuo-motor gap. Finally, I shall conclude around the theory that we have developed at ICP, called PACT (for " Perception-for-Action-Control Theory ") in an attempt to show how it could indeed contribute to " fill the perceptuo-motor gap ". The debate between auditory and motor theories of speech perception is ancient but still quite vivid. signal processing / pattern recognition problem, which should be considered in reference to the characteristics of the acoustic input, and the properties of the auditory system. The way signals are produced by the speech motor system is not considered as relevant for solving the 1991) assume that the listener recovers the speaker's gesture. The focus is not put here on auditory processing, but rather on this recovery mechanism, conceived in some sense as " integral " , that is the phonetic percept is the speech gesture. Simplifying somehow, auditory theories, considering speech perception without action, posit a " no-link " between perception and action; while motor theories, considering speech perception without audition (this is most clearly expressed in the " speech is special " view, according to which audition does not intervene per se in the processing of speech gestures: see, e.g., Whalen and Liberman, 1987; Whalen et al., 2006) posit a " full-link " between perception and action. Hence in both kinds of …