大きさの円対比錯視の呈示条件に関する実験的研究

The purpose of this study is to verify the effects of the stimulus presentation in the size illusions of contrastive circles (Fig. 1.) These effects are examined by using three kinds of stimulus-condition on the above size illusions and by using two kinds of trial-repetition. Magnitude of illusion (MI) of a center circle (CC) was measured under the following three stimulus-conditions ; they are, (1) the distance (D) between surrounding circles (SCs) and the CC (D=0.75, 5, and 10mm), (2) the diameter ratio (R; R=1/2, 1/1, and 2/1) of the SCs to the CC, and (3) the number (N; N=1, 2, and 4) of the SC. On the other hand, the variations of the MI were investigated by repeating the experimental session 200 times over an extended period of 265 days in Exp. I and 16 times a day in Exp. II using the same stimulus-conditions with one subject. Two personal computers (NEC : PC-9801XA) were used for presenting many kinds of standard stimulus (SS) and comparison stimulus (CS) with the observation distance of 115cm, and also for controlling efficiently the responses of the subject (Fig. 2). The subject was instructed to compare the apparent size of the SS presented randomly at the center of one color-display-monitor (NEC : N-5923) and that of the CS (a single circle) presented at the other display (NEC : N-5923). The spatial distribution of the SS to the CS was alternated in the order of L (SS : left, CS: right) – R (SS : right, CS : left) – R – L. The size of the CS was changed through the method of limits. The results verified the effects of three stimulus-conditions on the MI and also the ascending trends of the MI gained by repeating the experimental sessions with distributed (Exp. I) and concentrated (Exp. II) observations. (1) The CC of the size illusion of contrastive circles was clearly overestimated under the R=1/2 (SC<CC) but clearly underestimated under the R=1/1 (SC=CC) and the R=2/1 (SC>CC) during the early stages of the repeated observation. This tendency was observed consistently using the combinations with other two stimulus-conditions (Fig.3, 4, and 6). The above tendency was enhanced as the number of the SC increased. However, surrounding the CC by the increased SCs was not necessary in the size illusion of contrastive circles, because the contrastive change in the apparent size of the CC was verified in the case of the N=1 (“pure” size illusion of contrastive circle) and the N=2 (Baldwin illusion), in which the SCs do not surround the CC (Figs. 1C and D). In spite of the comparatively limited range of figure presentation inside the display, the consistent variation of the MI was gained as a function of the distance between the SCs and the CC (Figs. 3, 4, and 6). With four SCs (Ebbinghaus illusion), the magnitude of overestimation of the CC decreased clearly as the distance between the SCs and the CC enlarged with the R=1/2 (SC<CC) and the R=1/1 (SC=CC). In this case, the magnitude of overestimation decreased and that of the underestimation increased moderately in the above condition with the R=2/1 (SC>CC). The MI variations caused by the differences of the above distance between the SCs and the CC were compared with those of the Delboeuf illusion (Fig.8) studied by Ogasawara (1952) in regard to the distance between the circumference of the CC and the outer circumference of the SCs. Since the patterns of the MI variations in the Delboeuf illusion were similar to those estimated by Ogasawara’s results (Fig. 9), it may be concluded that the Ebbinghaus illusion was due to the “contrastive judgment” under the effect of the cohesive interaction (assimilation) between the SCs and the CC. (2) In the repeated observations in Exp. I and II, the MIs under all three kinds of stimulus-condition tend to show “an ascending trait (shift to overestimation)” in the process of the repetition. As shown in Figs.5, 7, and 10, the magnitudes of overestimation under the R=1/2 increased as a function of the repetition (trial). Contrarily,