言語弁別学習における言語強化の組合せ

概要

論文の詳細を見る
Previous studies which examined the effects of verbal reinforcement combinations upon the performances of two-alternative verbal-discrimination learning showed that the Right-Wrong (RW) and the Nothing-Wrorlg (NW) combinations were superior to the Right-Nothing (RN) combination (e. g., Spence, 1964). Three experiments were done to examine the nature of Nothing (N) which was combined with R or W by comparing with the nature of N which was not combined with R or W. The subjects in the three experiments were undergraduate students. In the Experiment 1 RN, NW, and NN conditions were compared in a within-subjects design. Fifteen cards of nonsense-syllable pairs were presented one by one. On 6 out of 15 cards the subject's responses were reinforced under RN condition, on 6 cards they were reinforced under NW condition, and on the remaining 3 cards they were not given any reinforcement (NN condition). In the first trial, half of the responses on RN or NW cards were reinforced by experimenter's saying "Right" on RN cards or "Wrong" on NW cards irrespective of the subject's responses. After the first trial 9 acquisition trials were given in 9 different orders of the cards, and the responses to the three items which had been assigned as correct on the first trial were reinforced according to the conditions. The results obtained were as follows : (1) the mean number of correct responses in RN and NW conditions increased with trials, but did not differ from each other, and (2) the mean numbers of response repetitions to N-item were the greatest in NN condition and the least in RN condition. In the Experiment 2 RN and NW conditions were compared in a between-subjects design. Twelve cards of nonsense-syllable pairs were devided into 8 cards for the RN or NW condition and 4 cards for NN condition. The subject's responses were reinforced by "Right" or "Wrong" in the same manner as the Exp. 1. The results obtained were as follows : (1) the mean number of the correct responses in RN and NW groups increased through trials, and that in NW group was greater than that in RN group at the last three trials, and (2) the mean number of response repetitions to N-item tended to increase in NW condition and decrease in RN condition, but that in NN condition in both groups was maintained at a high level through trials. In the Experiment 3 the independent RN, NW, and NN groups were compared in a between-subjects design. Twelve cards of nonsense-syllable pairs were used. In the first trial the responses were reinforced by saying "Right" on RN cards or "Wrong" on NW cards. The other procedures were the same as the Exp. 1 and 2. The results obtained were as follows : (1) the mean number of correct responses in RN and NW groups increased through trials, but did not differ from each other, and (2) as were found in the Exp. 1 and 2, the response repetitions to N-item increased in NW condition and decreased in RN condition, and they were maintained at a high level in NN condition. These results suggest that the subject has a strong tendency to repeat the responses which he selected on the first trial whether the responses were reinforced or not. These findings throw some question on the Spence's hypothesis (1966, 1970) that the subject has a tendency to regard N as R. As the verbal discrimination learning is regarded as a type of rote learning, the first decision of the responses seems to have an influence to the response tendency in the latter trials. The common findings that the NW group shows better performance than the RN group may be explained by assuming that the subject in NW group must repeat the N responses for avoiding W but the subject in the RN group need not avoid the N responses.
奈良教育大学の論文

著者

玉瀬耕治
奈良教育大学

言語弁別学習における言語強化の組合せ

スポンサーリンク

概要

著者

関連論文

スポンサーリンク