Feelings of familiarity are not direct products of memory. Although prior experience of a stimulus can produce a feeling of familiarity, that feeling can also be aroused in the absence of prior experience if perceptual processing of the stimulus is fluent (e.g., Whittlesea, Jacoby, & Girard, 1990). This suggests that feelings of familiarity arise through an unconscious inference about the source of processing fluency. The present experiments extend that conclusion. First, they show that a wide variety of feelings about the past are controlled by a fluency heuristic, including feelings about the meaning, pleasantness, duration, and recency of past events. Second, they demonstrate that the attribution process does not rely only on perceptual fluency, but can be influenced even more by the fluency of conceptual processing. Third, they show that although the fluency heuristic itself is simple, people's use of it is highly sophisticated and makes them robustly sensitive to the actual historical status of current events.

No abstract

No abstract

B. W. A. Whittlesea and L. D. Williams (1998, 2000) proposed the discrepancy-attribution hypothesis to explain the source of feelings of familiarity. By that hypothesis, people chronically evaluate the coherence of their processing. When the quality of processing is perceived as being discrepant from that which could be expected, people engage in an attributional process; the feeling of familiarity occurs when perceived discrepancy is attributed to prior experience. In the present article, the authors provide convergent evidence for that hypothesis and show that it can also explain feelings of familiarity for nonlinguistic stimuli. They demonstrate that the perception of discrepancy is not automatic but instead depends critically on the attitude that people adopt toward their processing, given the task and context. The connection between the discrepancy-attribution hypothesis and the "revelation effect" is also explored (e.g., D. L. Westerman & R. L. Greene, 1996).

Many investigators have observed that the feeling of familiarity is associated with fluency of processing. The authors demonstrated a case in which the feeling of familiarity did not result from fluency per se; they argued that it resulted instead from perceiving a discrepancy between the actual and expected fluency of processing (B. W. A. Whittlesea & L. D. Williams, 1998). In this article, the authors extend that argument. They observed that stimuli that are experienced as strongly familiar when presented in isolation are instead experienced as being novel when presented in a rhyme or semantic context. They interpreted that result to mean that in those other contexts, the subjects brought a different standard to bear in evaluating the fluency of their processing. This different standard caused the subjects to perceive their performance not as discrepant, but as coherent in one case and incongruous in the other. The authors suggest that the perception of discrepancy is a major factor in producing the feeling of familiarity. They further suggest that the occurrence of that perception depends on the task in which the person is engaged when encountering the stimulus, because that task affects the standard that the person will apply in evaluating their processing.