Lasting Representations and Temporary Processes
University of Oxford
Of the many possible ways of classifying memory, this chapter uses only one; the effects on memory of interference from other events that happen to the person. We can find types of memory that are impaired by event A but not by B, whereas a different class of memories may show impairment by B but not by A. Consequently, there must be something different about the changes in the nervous system that hold the representations of each type. The two types need not, of course, be physically distinct, in location in the brain or in the type of physiological coding used. They are, however, computationally distinct; they do different things. As we shall see, a classification based on this approach may well parallel classifications of other types, and does not exclude them.
In traditional experiments on interference, A and B might correspond to different types of material. A task in which an event A occurs may affect memory for earlier material that also includes A, but not for memory that only includes B. Nowadays, I expect, many of us, influenced by Endel Tulving, would say that A and B are cues that have each formed a separate and specific code with the items that are to be remembered. The intervening task damages one kind of memory, and not the other, because it affects only retrieval of the code that is elicited through cue A. The fact that recall and recognition show different effects of interference becomes very reasonable. It does not now disturb us, although it used to be a matter of great concern.
By using interference in this way, we could, if we wished to do so, classify memories as those that depend on A and those that depend on B. There would be little interest in doing so, when A and B are cues idiosyncratic to one experiment.