The Kinesthetic Aftereffect: Ritual versus Requisites

Abstract

Measured under a variety of conditions, the critical feature of the induction period for the kinesthetic aftereffect was found to be a distance between thumb and forefinger different from the width of the test block, and attended to by the subject. A constant-stimulus technique (the group version of the up-and-down method) yielded aftereffects comparable to those obtained by the traditional adjustment method. Across testing days, there was a declining aftereffect with high reliability of pre- and postinduction scores but low reliability for the differences between scores (the aftereffect). The original studies of Kohler and Dinnerstein (1947) showed that the judged width of a test block could be altered by rubbing an inducing block, a phenomenon called the kinesthetic aftereffect. A more or less standard procedure has evolved in which subjects make several baseline judgments of the test block by grasping it with thumb and forefinger while using the thumb and forefinger of the other hand to find a position that feels equally wide along an elongated comparison wedge. Next, the one hand is shifted from the test block to the inducing block, which differs in width. After a period of rubbing the latter (induction period), the felt size of the test block is again assessed via the wedge. Data averaged across subjects generally show a size change that is termed contrast: the felt width of the test block after the induction period is altered in a direction opposite to the width of the inducing block. There are methodological problems associated with this wedge/adjustment method. An 'error of anticipation' is found; the subject stops too soon as he moves along the wedge (Costello, 1961; Blitz, Dinnerstein, and Lowenthal, 1966). It is usual for four judgments to be taken-two ascending the wedge, two descending the wedge-so that the average 'gets rid of' the error. The error may be statistical in that it is tied to the biased sampling procedure dictated by the adjustment method. Averaging is then justified. (See Herrick, 1969, who discusses the problem as it applies to the method of limits.) On the other hand, the sampling bias may be confounded with subject factors, which are disguised by the averaging process. In addition to the error, the time required to gather