Reading Pseudohomophones: Implications for Models of Pronunciation Assembly and the Locus of Word-Frequency Effects in Naming

Abstract

Pronunciation performance under speeded conditions was examined for various kinds of letter strings, including pseudohomophones (e.g., TRAX), their real word counterparts (e.g., TRACKS), and a set of nonword controls (e.g., PRAX). Experiment 1 yielded a pronunciation advantage for the pseudohomophones relative to the controls, which was largest among items having few or no orthographic neighbors. No effect of base word frequency was found on pseudohomophone pronunciation. Experiment 2 ruled out an account of the pseudohomophone advantage based on differences between pseudohomophones and controls in initial phonemes. Experiment 3 established the existence of a large frequency effect on pronunciation of the base words themselves. These results are taken to suggest that whole word representations in the phonological output lexicon are consulted in the course of assembling a pronunciation. Implications of this conclusion for theories of lexical involvement in the assembly of phonology are discussed. In contrast to all current accounts of word-frequency effects, the present results also suggest that representations in a phonological output lexicon are insensitive to word frequency. A distinction between lexical representations and the connections between lexical representations is made, and it is suggested that these connections are a source of word-frequency effects in naming. © 1987 American Psychological Association.