Update versus activation of semantic memory

Two processes were studied which can alter the speed with which a concept is retrieved from semantic memory: activation and update. Activation has short-lived effects which spread to semantically related concepts. Update has relatively permanent effects, and it would seem that update cannot spread nearly as much as activation. Data from previous studies suggest that a clear separation of the durations of activation and update is possible. The present study was designed to investigate whether activation and update are two separate and distinct processes by determining if their magnitudes can be independently manipulated. In the present study, world-knowledge questions were presented visually. Response time was measured from the onset of the visual question to the subject's spoken answer. The answer-production task required the subject to recall an answer, e.g., "Where do wild kangaroos live?" The verification task required a "yes" or "no" answer, e.g., "Do wild kangaroos live in Australia?" By hypothesis, a verification task and semantically related primes would result in strong activation and weak update, whereas an answer-production task and identical primes would result in strong update. Experiment 1 partially supported these hypotheses by indicating that the answer-production task and identical primes produce a strong update effect. However, a strong update effect was also found for the verification task and for semantically related primes. In Experiment 2, all available questions were used to investigate the critical combination of the verification task with semantically related primes. With the power thus increased, semantically primed verification questions showed strong activation and weak update. This outcome supports the conclusion that activation and update are separate processes, in that their magnitudes can be manipulated independently.