ABSTRACT
Roediger and McDermott’s comprehensive review of the implicit memory literature (1993) included a puzzling pair of findings: If a prime item is studied in massed fashion (i.e., longer study or successive presentations), the prime benefit on a later perceptual-implicit memory test (e.g., lexical decision) does not increase beyond that provided by a brief or single presentation. However, if the prime item is repeated in spaced fashion, the prime benefit increases with the number of presentations (see Jacoby & Dallas, 1981; Roediger & Challis, 1992). These findings are puzzling because explicit memory tests show that performance increases with the number of both massed and spaced presentations (albeit more so for spaced presentations). Starting with Shiffrin and Steyvers (1997) we had been developing a theory that included a key role for context to account for implicit and explicit memory and the relation between these. Long-term priming, for example, was explained in large part by the assumption that event-study produced not only an explicit trace (incomplete and noisy), but also additional context storage in that event’s knowledge trace (if one existed; see Schooler, Shiffrin, & Raaijmakers, 2001). However, the findings highlighted by Roediger and McDermott did not fit that developing model, and led us to look deeper into the role of context and the mechanisms by which it affected memory. Now a decade and more after Roediger and McDermott put implicit memory in context, we believe it fitting to report our subsequent attempts to place context in context.
