Characteristics of errors in the recognition and reproduction of geometrical figures

Abstract

The Wulf-Koffka hypothesis of dynamic activity within the memory trace assumes that the trace is a more or less unstable organization of field forces which tends to shift toward equilibrium. After a critical review of literature as well as through his own experimentation, Hebb, however, concluded that there is complete lack of factual support for this hypothesis It is purpose to examine in this study whether or not Hebb's conclusion is justifiable.
Procedure : Three kinds of figures-angle, src and parallel-lines-were used as stimuli. The methods used were the recognition method afte Hebb (Exp. I), the discrimination learning method after Gellerman (Exp. 2), the psychophysical method modified for recognition experiment (Exp. 4), and the reproduction method (Exp. 3). In the recognition experiment the S was first told that he would be shown a figure which he was to remember exactly, and that the accuracy of his memory would be tested. After the removal of the stimulus figure he was shown a list of recognition figures, and told to pick out the one which had been shown. In the reproduction experiment the S was asked to reproduce the stimulus figure by drawing.
Subjects : Subjectg employed were kindergarten children and children of first grade in primary school, ranging in age from 3 years and a half to 7 years, totalling four hundreds and fifty children in number, and one hundred and twenty college students.
Results : The results obtained were as follows :
1) In the recognition experiment, larger angles than those of the standard stimuli were picked out when the original ones were comparatively small (underestimation of the recognition figures) ; whereas smaller angles were selected when the original ones were large (overestimation of the recognition figures). With regard to the arc this tendency was nqt so clear (Exp. 1).
2) The same tendency was found in the recognition experiment using the discrimina-tion learning method (Exp. 2). This was clearly observed in immediate recognition and the tendency was not significant after 24-hours.
3) In reproduction of the angle the small ones were reproduced larger and the large ones smaller. The phenomenon corresponds with the under-and overestimation in recognition process. In case of the arc the small arcs were reproduced larger and the large ones smaller.
4) When the size of angle was varied from small to large after the manner of psychophysical method, a gradual change from underestimation to overestimation of the recognition figures was observed (Exp. 4).
5) The same tendency as the one in Exp. 4 was found with the parallel-line-figure. The distance between the parallel lines became narrower when the length of lines was shorter, while it became broader when its length became longer. The tendency was more distinct in the parallel-line-figure with the distance of 20mm than the one with the distance of 40 mm (Exp.5).
Summarizing the above mentioned results, we can conclude that, both in recognition and reproduction, underestimation of the recognition figures is dominant when the angle, the arc and the length of lines are comparatively small, while overestimation is dominant when they are large. This was commonly observed both with children and adults.
We presume that this phenomenon can be explained in terms of “absolute impression”. As absolute impression of “big” or “small” are predominant in our comparative process, we make judgment on a figure regardless of its real size, resulting in the erroneous estimation.
In our experiments, the validity of the Wulf-Koffka hypothesis of dynamic activity in the trace was not proved as was the case with Hebb. Thus, we are inclined to reject it in the explanation of memory process.