ABSTRACT
An extension to the activation-verification model (Paap, Newsome, McDonald, & Schvaneveldt, 1982) is described that assumes both lexical and non-lexical pathways to pronunciation. The extended model provides an account of pronounceability effects in the lexical decision and naming tasks. The non-lexical pathway is also used to explain why lexical factors, like word frequency, generate smaller effects in naming than in lexical decision. This view is contrasted with Balota and Chumbley’s (1984) argument that frequency effects in the lexical decision task overestimate the role of familiarity in word recognition and that naming provides the better baseline. The experiments reported in this paper show that equivalent frequency effects are obtained with both lexical decisions and naming responses when the naming task requires lexical access before the onset of pronunciation.
