Implicit acquisition of abstract knowledge about artificial grammar: Some methodological and conceptual issues

In this reply to comments by Reber (1990) and Mathews (1990) on Perruchet and Pacteau (1990), the authors argue further against the claim that subjects, given a set of gramatical strings, implicitly abstract their constituent rules. On the one hand, the fragmentary knowledge of bigrams and trigrams, which accounts for performance on standard tests of grammaticality, may be formally described as abstract but can hardly be conceived of as the end product of active abstraction. On the other hand, transfer tasks may reveal some genuine abstraction process. Evidence for the implicitness of this process, however, is highly debatable. Incidental instructions are not a reliable guarantee of automaficity of abstraction in the study phase. Still more damaging to the views of Reber and Mathews is that abstraction may occur during the transfer phase while subjects are engaged in explicit cognitive activities.