A Retrospective View of Masked Priming: Toward a Unified Account of Masked and Long-Term Repetition Priming

We propose a general, unified account of masked and long-term priming, based on the assumption that prime events create a memory resource that can be recruited to assist with subsequent target processing. This account is contrasted with the classical account of priming, in which it is assumed that a prime proactively creates a temporary change in the cognitive system (e.g., through spreading activation or opening a lexical entry) that enables more efficient or accurate encoding of a target that is related in some way to the prime. In support of the classical, abstractionist account of priming, masked priming has been dissociated from long-term priming on three different bases: (1) lack of masked priming of nonwords, (2) additivity of masked priming and word frequency, (3) masked priming is short-lived. We closely examine these dissociations and describe recent evidence that challenges each of them. These challenges serve to support the proposition that masked and long-term priming have a common basis. Moreover, we review evidence showing that the recruitment of a prime may or may not occur, depending on factors such as need for the prime resource (as determined by target processing difficulty) and validity of the prime (as determined by prime-target relatedness proportion). These novel results are readily interpreted within a retrospective account of masked priming.