False memory following rapidly presented lists: the element of surprise.

This article examines a false memory phenomenon, the Deese-Roediger-McDermott (DRM) effect, consisting of high false alarms for a prototype word (e.g., SLEEP) following a study list consisting of its associates (NIGHT, DREAM, etc.). This false recognition is thought to occur because prototypes, although not presented within a study list, are highly activated by their semantic association with words that are in the list. The authors present an alternative explanation of the effect, based on the discrepancy-attribution hypothesis. According to that account, false (and true) familiarity results when a comparison between expectations and outcomes within a processing episode causes surprise. Experiment 1 replicates the DRM effect. Experiment 2 shows that a similar effect can occur when participants are shown lists of unrelated words and are then surprised by a recognition target. Experiments 3 and 4 show that the DRM effect itself is abolished when participants are prevented from being surprised by prototypes presented as recognition targets. It is proposed that the DRM effect is best understood through the principles of construction, evaluation, and attribution.