ABSTRACT
The similarity between study and test contexts has long been known to be an important determinant of performance in recall and other explicit memory tasks. Recall is more likely to be successful when the study and test contexts match than when they are different. Current theoretical views of implicit memory, on the other hand, would not expect context to be an important determinant of performance: their explanations of implicit memory phenomena typically involve low level perceptual, or simple cognitive, processes—such as the activation of a pre-existing concept—which are conceptualized as being immune to contextual manipulations. Indeed, except for certain special circumstances, context effects have rarely been found in implicit memory. In this chapter we present data from a number of experiments, using both lexical decision and word completion paradigms, that explore further the role of context in implicit memory. The data show context to be an important determinant of performance only if it changes the perceived sense, or meaning, of a word between study and test. Context manipulations remain ineffective if they do not change the perceived sense of a stimulus, for example if words with few meanings are used. Our preferred account of these results is a sense-specific activation view which postulates that each word has multiple senses, and that the context determines which sense is activated upon stimulus presentation. Implicit memory is observed when the context at test re-instates the sense that was encoded at study.
