Abstract
A research strategy is presented that uses the siblings of autistic children to investigate the possibility that autism involves a social deficit that can be detected prior to language development. Underlying the strategy is the assumption that a social deficit is reflected in the failure of the infant sibling and his or her mother to establish coordinated interpersonal timing, i.e., to mutually influence the temporal patterns of each other's vocal behaviors. The measurement of such mutual influence is accomplished by an automated microanalysis of the interactive behavior, the results of which are subjects to a time-series regression analysis. Various implications and limitations of the strategy are discussed.