Can knowledge underlying a simple perceptual-motor skill be unconscious? Three experiments (a) trained participants on a 4-choice reaction time (RT) task in which the stimulus on each trial was determined by a repeating 12-element sequence and (b) studied the extent to which participants' knowledge of this sequence was implicit, that is, unavailable for conscious access. Participants proved via an indirect test to have acquired knowledge of the sequence, because their RTs increased when the sequence was changed. To evaluate whether this knowledge was consciously accessible, participants were asked to perform an "objective" free-generation or recognition test of sequence knowledge. Results show that sequence knowledge is fully accessible on these objective tests. Moreover, it is demonstrated in this procedure that old-new recognition cannot be explained by unconscious attribution of perceptual-motor fluency. The question is raised whether distinct implicit (procedural) and explicit (declarative) forms of knowledge are acquired when participants learn a perceptualmotor skill.The possibility that knowledge can exist in an implicit form-whereby it is capable of affecting ongoing behavior without being accessible to consciousness-has been a major subject of investigation in the last few years. Laboratory experiments with normal participants and with braindamaged patients have yielded significant amounts of relevant evidence. In these so-called "implicit learning" experiments, participants acquire knowledge about some complex domain and are then tested for the degree to which this knowledge is consciously accessible.The sequential reaction time (RT) procedure, which is the focus of the present article, has proven to be a particularly useful tool for studying implicit learning. In the typical procedure, a stimulus appears in one of four locations (1-4) on each trial, and the participant presses a button corresponding to that location. The participant is given standard choice RT instructions, but the stimulus follows a predictable repeating sequence of about 10-12 trials in length. After many cycles of the sequence, participants can be shown chronometrically to have learned something about the repeating nature of the stimuli. This is established by unexpectedly transferring them to a random, nonrepeating

and M. Van der Linden (1997) presented evidence for 2 distinct mechanisms involved in artificial grammar learning. They suggested that after training on 32 letter strings (Experiment 2A), participants classify test strings using knowledge of the distributional statistics of letter chunks, whereas after training on 125 letter strings (Experiment 2B) they classify on the basis of knowledge of the rules of the grammar. This article offers an alternative unitary account of Meulemans and Van der Linden's findings. The authors show that information about grammatical rules and chunk locations was confounded in the test strings used in Experiment 2B and then present evidence that all of the data can be explained in terms of distributional knowledge, provided this includes knowledge of the positional constraints on chunks. Finally, the authors question the utility of traditional finite-state grammars for investigating abstraction processes, and suggest alternative methods. Reber (1967Reber ( , 1989 suggested that when participants memorize a representative sample of training strings generated from a grammar they abstract rules about the structure of the underlying grammar; moreover, in a later classification test they classify a string of letters as grammatical to the degree that it conforms to the abstract rules of the grammar. An opposing theory is that participants learn about the frequency of occurrence of chunks of letters in the training strings and that they classify a novel test string as grammatical to the extent that the string contains chunks that were present in the training strings (Dulany, Carlson, &

Four experiments explored the extent the extent to which abstract knowledge may underlie subjects' performance when asked to judge the grammaticality of letter strings generated from an artificial grammar. In Experiment 1 and 2 subjects studied grammatical strings instantiated with one set of letters and were then tested on grammatical and ungrammatical strings formed either from the same or a changed letter-set. Even with a change of letter-set, subjects were found to be sensitive to a variety of violation of the grammar. In Experiments 3 and 4, the critical manipulation involved the way in which the training strings were studied: an incidental learning procedure was used for some subjects, and others engaged in an explicit code-breaking task to try to learn the rules of the grammar. When strings were generated from a biconditional (Experiment 4) but not from a standard finite-state grammar (Experiment 3), grammaticality judgements for test strings were independent of their surface similarity to specific studied strings. Overall, the results suggest that transfer in this simple memory task is mediated at least to some extent by abstract knowledge.

No abstract

No abstract