ABSTRACT
In recent years, acquisition-deficit accounts of cue competition (e.g., Rescorla & Wagner, 1972) have been challenged by expression-deficit accounts of cue competition (e.g., the comparator hypothesis; Miller & Matzel, 1988). The comparator hypothesis was originally proposed as a response rule for the expression of Pavlovian associations that posits that associative learning is determined by spatiotemporal contiguity, whereas expression of that learning is determined by relative associative strengths. The comparator hypothesis has received support from studies of posttraining deflation and (with qualification) inflation of the competing cues that yield retrospective revaluation of responding to target stimuli. However, recent models of associative learning that also predict retrospective revaluation have revitalized acquisition-deficit accounts of cue competition (e.g., Dickinson & Burke, 1996; Van Hamme & Wasserman, 1994), such that the comparator hypothesis is no longer unique in predicting recovery from cue competition. This chapter reviews studies that are supportive of both the comparator hypothesis and retrospective revaluation models of cue competition, as well as recent studies that differentiate between acquisition-focused models that predict retrospective revaluation and the comparator hypothesis. An extended multilayer version of the comparator hypothesis is presented, and the importance of encoded temporal information for the comparator process is discussed.
