Memory for associative history of a conditioned stimulus

The common assumption that the current, but not the past, associative status of a conditioned stimulus (CS) is represented in memory is examined with respect to various behavioral phenomena within the existing literature and is found to be inadequate. This assumption is frequently expressed as associative path independence, which posits that the memorial effects of stimulus events occurring during a conditioning trial depend only on the associative strength of the CS at the initiation of that trial. Twelve laboratory phenomena are reviewed in which behavior is readily interpretated in terms of subjects remembering past as well as present associations. The possibility that behavior is influenced by past associative states as well as current associative strength is contrasted with several alternative explanations, including the suggestion that the current associative state of a CS is defined by variables in addition to associative strength. However, these alternatives are only partially successful without additional assumptions that increase their complexity. Contrary to many prevailing models, we conclude that subjects appear to retain the associative history of a CS.