People generally display an important amount of gestural and motoric activity when speaking. Since recent data have shown the limits of an explanation of this activity in terms of nonverbal or bodily communication, the present study attempted to explore what would happen if subjects were impeded from making the principal movements they normally perform during a conversation. Subjects were led to hold a 50-minute conversation while sitting in an armchair devised to restrain their movements of the head, arms, hands, legs, and feet during part of the experiment. The main dependent variables consisted of nonverbal activity in body zones that remained free to vary: eyebrows, eyes, mouth, and fingers. During the phase of movement restriction, highly significant increases in activity were recorded in these zones, with subsequent return to base levels when the subject recovered free movements. Also, significant interactions of conditions of movements and subject's conversational role (speaker vs. listener) were observed for most of the variables. Samples of dialogues submitted to a computerized technique of content analysis revealed a significant decrease in the vividness of imagery during movement restriction.