Shaping Mands in Mute Autistic Children : A Program for Training Functional Vocal Topographies

Abstract

In this study, three autistic children who emitted few verbal responses were trained to emit vocal mands for food. The training program consisted of three main steps. First, the children were trained by peer modeling and differential reinforcement to make some non-vocal gestures (such as finger pointing) that functioned as mands. Second, these non-vocal gestures were shifted to vocal ones through a modified time-delay procedure that differed from the original time-delay procedures in terms of the schedule of reinforcement. That is, with the modified procedure, the children could not get food in the training situation until they emitted the target response (a vocal response) spontaneously in the modified time-delay situation. Third, after the training was completed, generalization and maintenance tests were conducted to assess the efficacy of the program. In generalization test, the extent of generalization from the training setting, where the children could not get food by themselves, to a probe setting, where they could easily do that, was assessed. In addition, maintenance of the newly shaped mands was checked after 2 months. The results indicated that after the non-vocal mands were shaped, the modified time-delay procedure was effective for producing vocal mands in all the children in the study. But the extent of generalization and maintenance was poor for one of the three subjects. The results were discussed in terms of differences in the history of reinforcement for each subject's mand repertoire.