The covert repair hypothesis: prearticulatory repair processes in normal and stuttered disfluencies.

Self-repairing of speech errors demonstrates that speakers possess a monitoring device with which they verify the correctness of the speech flow. There is substantial evidence that this speech monitor not only comprises an auditory component (i.e., hearing one's own speech), but also an internal part: inspection of the speech program prior to its motoric execution. Errors thus may be detected before they are actually articulated. In the covert repair hypothesis of disfluency, this internal error detection possibility has been extended with an internal correction counterpart. Basically, the covert repair hypothesis contends that disfluencies reflect the interfering side-effects of covert, prearticulatory repairing of speech programming errors on the ongoing speech. Internally detecting and correcting an error obstructs the concurrent articulation in such manner that a disfluent speech event will result. Further, it is shown how, by combining a small number of typical overt self-repair features such as interrupting after error detection, retracing in an utterance, and marking the correction with editing terms, one can parsimoniously account for the specific forms disfluencies are known to take. This reasoning is argued to apply to both normal and stuttered disfluency. With respect to the crucial question concerning what makes stuttering speakers so greatly disfluent, it is hypothesized that their abilities to generate error-free speech programs are disordered. Hence, abundant stuttering derives from the need to repeatedly repair one's speech programs before their speech motor execution.