ABSTRACT
Much remains to be discovered about the impact of specific types of experience and training on performance. In word and object recognition, for example, does experience facilitate performance because it consolidates abstract models or because it provides a larger repertoire of specific examples? And to what extent is expertise specific to the domain or modality in which it is acquired? Is practice with words specific to the modality in which they are shown, or does it generalize to other modalities? A mixture of review and new repetition priming data is brought to bear on these questions. The main findings are: first, repetition priming occurs cross-mod ally as well as intra-modally; second, intra-modal priming is sensitive to surface similarity; third, cross-modal priming is sensitive to word frequency; and, fourth, cross-modal priming is unaffected by the nature of the mapping relationship between the sub-lexical constituents of printed and spoken words. We conclude that repetition priming depends on information from two sources, tied to and independent of, respectively, presentation modality. We propose that representations associated with different input modalities are not integrated, even when they share formally equivalent sub-lexical constituents, while representations that are independent of input modality cannot be accounted for in terms of conceptual or semantic features. Instead, a hypothesis involving reference to production or response representations merits consideration. We present a framework that involves event-specific records in the perceptual and response or production domains, and we further propose that these components must be connected at a lexical, as distinct from sub-lexical or constituent level, for priming to occur.
