TRANSFER IN ARTIFICIAL GRAMMAR LEARNING - A REEVALUATION

This article covers methodological and theoretical issues in artificia l grammar learning. Arguments that such tasks are mediated by abstract knowledge (e.g., A. S. Reber, 1969, 1990) are based primarily on evid ence from transfer experiments, where the surface vocabulary is change d between learning and test items. Because of a number of methodologic al concerns, the small magnitudes of artificial grammar learning effec ts generally are difficult to interpret. Possible solutions are offere d here. Furthermore, even reliable transfer effects imply neither that subjects have acquired abstract knowledge of the underlying grammar n or that they are performing a process of abstract analogy from memoriz ed whole exemplars. Models that learn only surface fragments of the tr aining stimuli and perform abstraction at test rather than during lear ning are wholly consistent with transfer phenomena.