Abstract
The purpose of the present investigation was to study the problem of what is learned in discrimination reversal learning. In experiment I, 6 Japanese monkeys were trained on object discrimination problems. Each problem consisted of three stages; original learning, prereversal experience and reversal learning. The original learning continued to a criterion of 18 out of 20 correct responses, followed by 10 additional trials. There were three conditions of prereversal experience; N-P, P-N and C (control). For condition N-P, Ss received 10 rewarded prereversal trials to the object that had initially been negative but which was to be positive during the subsequent reversal trials : and, conversely, for the condition P-N, Ss received 10 non-rewarded trials to the object that had initially been positive but which was to be negative during the reversal. For the control condition, Ss received no prereversal experience between the original and the reversal learning. Each S was given 3 problems, involving different prereversal experience. The order of the presentation of the 3 problems was different by Ss and was determined according to a latin square. The experiment consisted of the duplication of the above procedure, using the same Ss, with the interval of about 7 months.Performance of each S on each problem is shown in table 1. Analysis of variance of number of errors during reversal learning revealed superior performance in conditions N-P and P-N to the control condition. The results support the duoprocess theory which emphasizes the effects of both reward and nonreward. In experiment II, 30 kindergarden children, 46 years old, were tested on pattern discrimination problems. After the preliminary problem, Ss received the main problem, which consisted of three stages; original learning, prereversal experience and reversal learning. After each S reached the criterion of 10 consecutive correct responses in the original learning, he received 4 trials of the prereversal experience, which were different by groups. Ss of group N-P, group P-N, and the control group received the corresponding prereversal experience, followed by the reversal learning. The results in Table 2 show the superior reversal performance of Ss of group N-P, to that of Ss of the other 2 groups. They support the uniprocess theory which emphasizes the effect of reward. The results of the above two experiments on Japanese monkeys and children, along with the previous data on rats and squirrel monkeys, reveal the systematic species differences in the nature of reversal learning.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have