The repertoire and timing of gestures accompanying speech were compared in children with specific language impairment (SLI), aged 5-10 years, in typically developing peers (CA), individually matched on age and nonverbal IQ, and in younger language-matched (LM) children. They were videotaped in two tasks, recounting a cartoon and describing their classroom. Three types of gestures were coded -iconics, deictics and beats -and the synchrony of these gestures with speech was examined in terms of number of words encompassed, grammatical speech category at gesture onset, and relationship of iconic gestures to the concept expressed in speech. All groups used more deictic gestures in the classroom description task. SLI children differed from the comparison children only in their use of iconic gestures. They produced somewhat more of these, used them more often to replace words, and began them more often on a noun phrase object. Otherwise, language proficiency, at least as measured by standardized tests, did not appear to impact the gestural system. The fact that, for all groups, most iconic and deictic gestures began on the noun phrase subject indicates a close synchrony between gesture and speech onset.

Fifteen children with specific language impairment (SLI) were matched in gender, age, and nonverbal mental age to one comparison group and in gender and standardized expressive language age to another comparison group. Their spontaneous speech was recorded while they narrated a cartoon and described their classroom. No group differences were found in mean length of utterance in morphemes or in syntactic complexity. SLI children had a lower percentage of correct irregular past tense forms but were equivalent to controls on regular past tense forms and noun plurals. Lexical diversity in terms of the number of different verbs expressed in the past tense and the number of different nouns pluralized did not differ. Some problems in generating complex sentences were found in the speech of SLI children. The importance of task constraints is stressed in explaining the discrepant performance of SLI children on spontaneous versus standardized measures of language, as well as the contribution of auditory memory limitations.