Using Immediate Memory Span to Measure Implicit Learning1

Implicit learning is the process of acquiring knowledge about the structure of a complex stimulus environment largely independent of conscious awareness of either the process or the products of acquisition. In the present study, a new method of measuring implicit learning using immediate memory span capacity was developed. This method avoids many of the conceptual and methodological pitfalls found in standard implicit learning tasks. Subjects were presented with sequences that had been generated by an artificial grammar and were asked to reproduce the sequences by pressing buttons on a customized response box. After being exposed to these sequences, subjects showed a selective increase in memory span for novel sequences governed by the same grammatical constraints, despite being largely unable to display explicit knowledge about grammatical sequences in a recognition memory test. Individual differences in the degree of implicit learning co-varied with general measures of memory span, indicating that individuals with larger immediate memory capacity are better able to learn and subsequently exploit the higher-order sequential dependencies that compose the structure of an artificial grammar. Implications for current models of implicit learning, the relationships between immediate memory span and sequential learning, and individual differences in implicit learning and language processing are discussed.