AN INVESTIGATION OF THE USE OF SPEECH ACTS IN AUTISTIC, MENTALLY RETARDED, AND NORMAL CHILDREN (PRAGMATICS, LANGUAGE DEVELOPMENT)

Nineteen autistic children were compared with eighteen normal and nineteen mentally retarded controls, on the use of speech acts thought to develop in the second through fourth years of life. The subjects were matched on mean length of utterance and pre-tested for non-linguistic mental age. Two half-hour spontaneous speech sessions were videotaped for each subject. In the first session, the child interacted with a familiar person, while in the second spontaneous speech session an experimenter who was naive to the specific experimental hypotheses interacted with all subjects. Data from these sessions were analysed in terms of Dore's classification of speech act types and in terms of the overall level of syntactic development as measured by Lee's Developmental Sentence Test. In another session, each subject was administered Snyder's structured pragmatic tasks which were designed to elicit imperatives and declaratives.;Results indicated that both autistic and retarded subjects showed a restricted range of speech act types in the spontaneous situation. The autistic subjects differed from both control groups in their distribution of responses across speech act categories: Autistic subjects produced fewer requests for information, internal reports and explanations, but produced more egocentric speech than the control subjects. In addition, the performance of the autistic subjects seemed to be affected by the familiarity of the interviewer to a greater extent than was that of the control subjects. When interacting with an unfamiliar person, the responses of the autistic groups were more similar to the normals. On Snyder's structured pragmatic tasks autistic subjects showed less communicative intent than controls. There were no group differences in any other aspects of the structured tasks, indicating that in the structured situation the autistic subjects were able to produce declaratives and imperatives at the same level as matched controls.