No abstract

We investigated the programming of generalization and maintenance of correspondence between verbal and nonverbal behavior in a preschool setting. Four children participated in a series of multiple-baseline designs. In Experiment 1, delayed reinforcement of verbal behavior effectively controlled maintenance of correspondence with previously trained responses and also resulted in generalization of correspondence to one untrained response. As the latter effect was limited, Experiment 2 was a further assessment of the effects of delayed reinforcement of generalization of correspondence to untrained responses, and consistent generalization was shown. Experiment 2 also showed that generalization, if lost, could be recovered through use of "booster training," in which the original contingencies were reinstated for a brief period. Experiment 3 provided replications, with two additional children, of the effects of delayed reinforcement on maintenance of correspondence. Results are discussed in terms of using delayed reinforcement as an indiscriminable contingency.

A correspondence training procedure was used to develop consistency between children's verbalizations and their subsequent behavior across increasingly remote settings and time. The interval of time between the verbalizations and the opportunity to engage in several target behaviors was systematically increased across four preschool settings. Probes of generalized verbal control of home behaviors were conducted throughout training and showed that generalization was obtained in the absence of any salient externally imposed contingencies after the children had reliably come under the control of verbalizations about preschool behaviors.