Development of response generalization gradients.

Publisher Website
Google Scholar
Abstract
Seventy-eight subjects (Ss) practiced a perceptual-motor task in which the responses were movements of a lever into 6 horizontal radial slots. Analysis of the 1st error made upon the appearance of each new stimulus showed that during the 1st 10 trials the errors were distributed about equally in the 5 incorrect slots which formed a bidirectional gradient of spatial similarity around the correct slot. During the last 10 trials the frequency of errors, although considerably diminished, was directly proportional to the first 2 degrees of similarity, but not to the 3d, in both the clockwise and counterclockwise directions from the correct response. It was held that these error gradients were response generalization gradients developed by a process of differential rate of extinction of errors. A rate of extinction of errors which is inversely proportional to similarity was offered as a possible substitute for the hypothesis of parasitic reinforcement in accounting for response generalization gradients. However, it was pointed out that the experiment is not crucial with regard to parasitic reinforcement.
Keywords

This publication has 0 references indexed in Scilit: