ABSTRACT
INTRODUCTION The identification of written words involves analysis of orthographic structures aimed at extracting phonologic, semantic, morphologic, and syntactic linguistic information. This information is used to integrate words into meaningful phrases and sentences that, in normal verbal communication, convey informative messages. Although there are models of visual word recognition which stipulate that letter strings may gain access to their meaning directly (e.g., Jared & Seidenberg, 1991), our premise is that access to meaning is mediated by the
phonological lexicon. According to this premise, letters activate orthographic and phonological units which are associated with nodes in a semantic network.
