Verb use in specific language impairment.

The aim of the present study was to use longitudinal data to provide a detailed profile of early word combinations by children with SLI. Three children with SLI were videotaped during mother-child interactions in the home over a 2-year period. The data obtained were compared to MLU-matched samples of normal language-learning children from Wells' (1981) longitudinal database, which provided a control for the linguistic measures used in this study. A range of analyses were carried out on controlled data samples in order to determine how the children with SLI's early utterances compared with those of MLU-matched normal language peers. The measures were specifically designed to assess the children's use of verbs and verb morphology because recent research has suggested that verbs may play a central role in the acquisition process, and children with SLI may have particular problems with verbs. We found that children with SLI used verbs less frequently, nouns more frequently, and were more input-dependent than their MLU-matched peers. The children with SLI used verb bare stems incorrectly more often than their MLU-matched counterparts. However, further analyses showed that this high frequency of incorrect bare stems may be at least partly due to the fact that children with SLI have particular difficulties using auxiliaries. Furthermore, the proportion of verb use that consisted of General All Purpose (GAP) verbs for children with SLI was similar to that of the MLU-matched children. The above findings were compared with those from other relevant studies of lexical diversity in children with SLI, and the potential implications of these data for theories of SLI language development were discussed, particularly with reference to Marchman and Bates' (1994) "critical mass" hypothesis.