ABSTRACT
The covert repair hypothesis (CRH) of stuttering (Postma & Kolk, 1993) considers disfluencies to be the result of covert self-monitoring and selfrepair of speech errors. In this chapter, we consider how well this hypothesis accounts for an interaction between lexical type and position in a phonological unit on stuttering frequency (Au-Yeung, Howell, & Pilgrim, 1998). We show that the CRH predicts this interaction when it is supplemented with a formal model of the time-course of self-monitoring, which relates observed symptoms to the moments in time when errors in speech plans are intercepted and repaired.
