ABSTRACT
Clinical observations by I. L. Janis and W. R. Russell on retrograde amnesias induced by electroconvulsive shock (ECS) or other trauma suggest that recently formed associations are especially subject to disruption. If true, this is of basic importance to learning theory. It suggests that there is a process of fixating memory traces which operates for a period after each learning episode. In addition to repression, two related sampling artifacts may contribute to clinical impressions of a temporal gradient of retrograde amnesia. Memory for events immediately before the trauma is likely to be sampled in class intervals of time which are shorter than those used for memory for earlier events. However, in the experimental designs of Duncan and the others any fear or conflict produced by the avoidance response’s association at the shorter intervals with some aversiveness of the ECS itself would produce the same effects as any amnesia.
