A study was designed to test 2 alternative hypotheses--a symbolic hypothesis and an executive function hypotheses--for the imitation and pantomime deficits found in previous studies of autism. The subjects were 17 adolescent high-functioning subjects with autism spectrum disorders and 15 clinical comparison subjects who were matched on chronological age and verbal IQ. Meaning and sequence were manipulated in facial and manual imitation tasks. Sequence was manipulated in the pantomime and control tasks. Recognition memory and motor control tasks were matched to the experimental tasks. The results provided no support for the symbolic deficit hypothesis; meaning aided rather than hindered the performance of the group with autism. Partial support for the executive deficit hypothesis was found. There were no group differences on motor control tasks, and few on the memory control tasks, arguing against deficits in motor initiation, basic motor coordination, or visual recognition memory.