長さの継時比較における時間誤差の因子分析 AL理論の観点より

Abstract

There have been long controversies among the theories of “time-error”, for example, the theory of memory image (Fechner-Lehman), the trace theory (Köhler), the theory of assimilation (Lauenstein), the theory of Einstellung (Woodraw), and so on. However, a unitary conclusion will be given to the controversies in a near future. It seems that all these theories have already carried out their missions, and a new idea must be presented to give a definite unitary explanation to the “time-error”. Thus, Freeman and Sharp (1941) wrote that this new idea must be such one as to give a definite unitary explanation to the problems of “time-error” which will change according to the (1) magnitude of time-separation between two stimuli, (2) sort of the stimuli in its back-ground, (3) levels of intensity of the stimuli, (4) repetition of experiment.What will be the new idea which will be able to explain the “time-error” unitarily? However, the author believes that it is necessary to reexamine the results of researches concerning these problems by applying factor analysis, before constructing a new idea, because, so many data and theories have been accumulated that interpretation of foctors may be easy with the help of these informations. The factor analysis may be one of the best ways of giving a definite unitary explanation to the problems of “time-error”. The present experiment is designed from this viewpoint.A straight line which was illuminated by weak light was presented to a subject in a dark room. The length of the standard stimulus was 150mm, and variable stimuli were 140, 142, 144……158, 160mm. Ss was always presented prior to Sv. The time-separation (P) was as follows: P…0.5, 1, 3, 5, 8, 12sec.The subjects were fifteen students who were requested to make nine category judgments for six times in each time-separation. The adaptation level (A) of each subject was computed by Helson's short-cut method. The “time-error” (TE) of each subject was computed by the formula TE=A-Ss. The centroid method was applied to analyse the inter-correlation coefficient matrix.Tentative conclusions are as follows:(1) when the time-separation is 12sec., all the factor loadings are not high and fall within ±. 40.(2) The first factor has high positive loadings on P=0.5 and 3sec.(3) The second factor has high positive loadings on P=5 and 8sec.(4) The third factor has high negative loadings on P=1 and 3sec.(5) The fourth factor has not any high loadings on P=1 and 3sec.(6) The effects of time-separation may be classified into two groups, 1) short time-separation which is equal to or less than 3sec., 2) long time-separation which is equal to or longer than 5sec. The first and third factors may be called “the short time-separation factors” which have high (positive and negative) loadings on the short time-separation and low loadings on the long time-separation. The second factor may be called “the long time-separation. factor” which has high loadings on the long time-separation and has low loadings on the short time-separation.