THE EFFECTS OF NUMBER OF REINFORCEMENTS ON THE STRENGTH OF A GENERALIZED INSTRUMENTAL RESPONSE

Abstract

The importance of the principle of stimulus generalization in behavior theory, and especially in stimulus-response theories of behavior, is generally recognized. The principle has been used to explain such diverse phenomena as the occurrence of sudden or insightful solutions of problems, and the maladaptive persistence of behavior that is not rewarded or is even punished. Despite the crucial role that stimulus generalization has played in stimulusresponse accounts of behavioral phenomena, important empirical data are still missing from the literature (12). The principle has recently taken on added significance through its application by experimenters to the explanation of results obtained from studies of the discrimination learning performance of children (I, 2, 3, 4, 7, 8, I3). In some of these experiments, the related principles of mediated generalization and acquired distinctiveness of cues (ii) have also been applied. Thus, it is desirable to devise experimental conditions and apparatus that are appropriate for children and which may also be used to verify and extend some of the important findings in the area, now based primarily on the study of infrahuman and human adult subjects. One of the important variables assumed to affect the generalization of a response is the number of reinforcements given on the training stimulus. Razran (12) reports that the evidence from classical conditioning with infrahuman and human adult subjects demonstrates fairly conclusively that an increase in the number of reinforcements on the training (conditioned) stimulus results in an increased amount of generalization to the test stimuli. A recent study with rats by Margolius (io) reports an increase in the strength of a generalized instrumental response with increased numbers of reinforcements on the conditioned stimulus.