課題解決事態における硬さについて

抄録

In experiments on rigidity, Ss learn one set of responses and then are required to shift to another set of responses. Rigidity is defined in terms of failure to make such a shift. There are several different types of shifts. Buss A. H. found that, compared with reversal shift, which involved a reversal of positive and negative cues, nonreversal shift, which involved a change in the stimulus dimension, retarded the subsequent learning. He explained that in the learning of subsequent discrimination, nonreveral shift had fortuitous partial reinforcement of the set that had been learned previously.
To confirm this Buss's explanation, we investigated the effects of one reversal shift and three non-reversal shifts upon learning of subsequent discrimination. Four groups of Ss learned two series of discrimination problems. Ss were presented with two stimuli simultaneously, and were required to choose one of theme.
One group learned a size discrimination in Series 1 and the reversed size discrimination in Series 2 (reversal shift). Another three groups learned a size discrimination in Series 1 and a form discrimination in Series 2 (non-reversal shift). In Series 2. these three groups differed from each other in regard to the possibility of fortuitous partial reinforcement of size discrimination learned previously. Values of such possibility for the three groups were 25%, 50% and 75%, respectively. If Buss's account is right, we should be able to predict that the learning of Series 2 will be more difficult in proportion to the degree of such possibility in Series 2.
Our results confirmed the prediction. The actual percentages of partial reinforcement in Series 2 were 0% (reversal shift), 15.3%, 33.2%, and 65.0%, respectively. Learning of Series 2 was the easiest in reversal shift, and was more difficult in proportion to the percentage of partial reinforcement.