Speeded tests of implicit knowledge

Three experiments tested the hypothesis that implicit and explicit tasks involve distinct modes of processing. Ss observed rule-ordered letter strings and were asked either to memorize the strings or to try to discover the underlying rules. In Experiment 1, they then made well-formedness judgments of novel strings under long-deadline and short-deadline conditions. Rule-discovery Ss, but not memory Ss, were impaired by the short deadline. In Experiment 2, all Ss made 'similarity' judgments of the novel strings instead of the traditional 'rule-based' judgments; there were now no differences between the rule-discovery and memory groups. In Experiment 3, Ss explicitly instructed in the rules were significantly more impaired under short deadlines than were memory Ss. An analysis of decision times to individual strings for the rule-trained versus memory groups also showed qualitative differences consistent with the implicit-explicit distinction.