Disfluency clusters of children who stutter: relation of stutterings to self-repairs.

The purpose of this study was to account for the frequency, type, and possible origins of speech disfluency clusters in the spontaneous speech of 3- to 6-year-old children, 30 who stutter and 30 who do not stutter. On the basis of the Covert Repair Hypothesis (Postma & Kolk, 1993), which suggests that stutterings are the by-products of self-repairs or self-corrections of speech errors, three hypotheses were tested in attempts to account for the frequency and location of stutterings within speech disfluency clusters. Sequences of various types of speech disfluencies in utterances containing disfluency clusters were collected from audio/videotaped conversations between each of these 60 children and their mothers. Three types of speech disfluencies--overt self-repairs, covert self-repairs, and within-word disfluencies ("stutterings")--and the disfluency clusters they comprised, were identified and analyzed frame-by-frame. Results indicated that children who stutter produced significantly more stuttering-stuttering clusters (e.g., "I-I-I w-w-want ..." or "w-w-waaaant") and that, although children who do not stutter occasionally produced stutterings, they never produced stuttering-stuttering clusters. Furthermore, children who stutter produced significantly more stuttering-repair clusters, whereas children who do not stutter produced significantly more repair-repair clusters. Within the disfluency clusters of children who do not stutter, stutterings were more likely to follow an overt self-repair produced at a relatively fast speaking rate (6.6 sylls/sec). Findings are taken to suggest that stuttering-stuttering clusters may help differentiate between children who do and do not stutter, and that speech errors, self-repairs, and speech disfluencies influence one another within and between adjacent sounds, syllables, and words in what appears to be a nonhappenstance and theoretically important fashion.