順向侵入・逆向侵入の分析的研究

感応理論の研究 (65)

抄録

In serial rote learning, inter and intra serial intrusions occur very frequently, though their precise nature has been known little for the following reason: as to experiments traditionally made, the numbers of items correctly reproduced and failed to reproduce were the subject matters of discussion and less attention had been payed to the role of intrusion.
The present experiment was designed to analyse the phenomenon of intrusion in serial rote learning.
This experiment was conducted with the modified reproduction method originally known as the retained members method.
Each subject, after learning two (Exp. I, II) or three (Exp. III) series of alphabet letters, eighteen letters in one set, was asked to reproduce one particular series which had beforehand been designated by the experimenter.
The experiments were so designed that the retroactive intrusion could be observed from Exp. I (test A 4-6) and Exp. II (test B 2), in which S was required, after learning two series, to reproduce the preceeding one. In other experiments, Exp. I (test A 1-3) and Exp. II (test B 1), S was ordered to reproduce the succeeding series which follows the first one, to see how the proactive intrusion occurs. In Exp. III (test C 1), three series of the letters were used to find out the dual effects of retroactive and proactive intrusions by asking S to reproduce the intermediate series.
The following results were found through these experiments:
1) The amount of correct reproduction in the beginning part of series is large and decreases towards the end of series (Exp. I)
2) In overt interserial intrusions:
i) Proactive intrusions tended to appear more frequently than retroactive intrusions.
ii) Relative position of letters in stimulus series plays an important role, namely, the factor of transfer in interserial intrusion was revealed in this experiment.
3) In overt intraserial intrusions, proactive intrusions were frequently found at the first half part of a series, while the retroactive intrusions at the last half. Contrary to the case of interserial intrusions, retroactive intrusions seemed to be more frequent than the proactive in numbers.
4) What is hitherto called proactive and retroactive “inhibitions” can be explained through our present study as due to intrusions and in terms of competitions between intrusions and correct reproductive tendency.
5) The authors considered proactive and retroactive intrusion as proactive and retroactive reproductions respectively. Thus the possibility that the learning process can be explained by the mechanism similar to the perceptual induction was suggested.