ABSTRACT
Theories of word identification and priming or context effects have been powerfully influenced by two principal assumptions. One is that the representation of word concepts can be modelled in terms of a single representational unit for a single concept. The other assumption is that activation of one concept spreads automatically to other concepts that are directly connected to it. My claim is that although these assumptions have been supported by a number of empirical results, a number of recent findings constitute a serious challenge to the validity of these assumptions.
