ABSTRACT

In this chapter I attempt to elucidate the phenomenon of implicit learning by discussing the nature of implicit knowledge. In so doing I make a distinction between two senses in which so-called implicit knowledge could be said to be implicit. Research concerned with implicit learning typically has construed implicit knowledge as unreportable knowledge. However, within this framework there are two possibilities. First, implicit knowledge might be implicit in the sense that it is knowledge that is unreportable for some reason, despite being explicitly represented. Second, implicit knowledge might be implicit in the sense that it is knowledge that is not merely unreportable, but knowledge that also is represented implicitly. The relevant literature has not clearly distinguished between these two possibilities, either conceptually or empirically. In the experimental study of implicit learning, attention has been focused largely on whether the knowledge in question is available for verbal report, with less concern for the underlying representational format of that knowledge. In contrast, the enterprise of modeling implicit learning has been concerned to construct models that learn without explicitly representing what they learn. As a result, it is unclear how the phenomenon of implicit knowledge should be conceived. Is it simply an unreportable form of the reportable knowledge that subjects display, not qualitatively different in format from reportable knowledge, or is it qualitatively different in format to reportable knowledge? The latter possibility appears theoretically more tractable, because if so-called implicit knowledge is qualitatively different to reportable knowledge, in the sense of not being explicitly represented, then this may explain the lack of reportability of the knowledge that is represented. However, if implicit knowledge is explicitly represented, despite being unreportable, then it is difficult to explain why such knowledge should be unreportable. Furthermore, knowing that unreportable knowledge is implicitly represented severely constrains the kind of processes and models that can be postulated to account for the phenomenon of implicit knowledge and implicit learning.