Does the artificial grammar learning paradigm involve the acquisition of complex information?

Abstract

The purpose of this article is to present a series of experiments originally aimed to study the respective role of fragment versus complex (rule-based) knowledge in artificial grammar learning, and to discuss some methodological constraints in artificial grammar learning tasks. What we are concerned with here is (a) the balance of different chunk-strength measures between grammatical and nongrammatical test items, (b) the importance of the number of items presented during the study phase, and (c) the importance of using a no-learning control group, with some insights on how these control subjects actually perform in the classification task. We also report an experiment conducted with the purpose to replicate the complex learning effect we observed in previous experiments, using a task in which grammatical and nongrammatical test items differed only according to their adherence to the rules. Finally, we present data obtained in young and elderly subjects, suggesting that implicit learning abilities are preserved in the elderly.