SPELLING-TO-SOUND EFFECTS IN SINGLE-WORD READING

Computational modelling of word-naming processes has suggested that the frequency of spelling-to-sound correspondence facilitates print-to-sound translation time and has given rise to predictions that are supported by empirical data (Brown, 1987a). This contrasts with earlier claims that regularity is the only spelling-to-sound correspondence attribute that determines word-naming time. Here we report further evidence for the effects of the frequency of spelling-to-sound correspondences on word-naming latency. Experiment 1 excludes alternative, orthographic neighbourhood interpretations of the putative spelling-to-sound effects. Experiment 2 shows that a word's number of rhymes does not affect naming latency for that word, thus excluding explanations in terms of output phonology. Experiment 3 replicates earlier findings that a word's spelling-to-sound enemies, as well as friends, determine naming latency. Thus both the regularity and the frequency of spelling-to-sound correspondences influence word-naming time. The implications for models of oral reading processes are explored.