ABSTRACT
Let us call the neural system for controlling motor action in the world system A, and the neural system for controlling speech articulation system B; then the sensory-motor level of representation arises from system A and speech from system B. Sensory-motor ideas become important for speech production in the following way. For one to have syntagmata, it is necessary for systems A and B to converge so that A provides the plan, or scheme, ofthe action and B provides the channel of execution of this scheme. In this way the dual character of sensory-motor representations, as simultaneously part of meaning and part of action, is exploited. Such AB fusions appear at an early point in ontogenesis and remain, we suppose, the fundamental mechanism of speech coordination from the semantic level with adult speakers as well (McNeill, 1975). [This role for sensory-motor action schemes can be related to the description of action control given by Bernstein (1967), Greene (1972), or Turvey (1975); according to them the control of action is accomplished through a hierarchy of processes each level of which manages only a few degrees of freedom, and in these hierarchies the highest level can be identified with the scheme of the action.]
Thus, as speech production proceeds, speakers are presumed to be constantly activating sensory-motor content (A), which they do on the basis of meaning relationships, in order to access and control the articulatory system (B). We should, therefore, expect that measures of the coordination of speech output (phonemic clauses, the place and kind of dysfluency, etc.) will coincide with units of sensory-motor meaning.
