| Recently, researchers have begun investigating the extent to which readers have strategic control over phonological processing in word recognition. Davelaar, Coltheart, Besner and Jonasson (1978) reported that a homophone effect was present in a lexical decision task (LDT) when standard nonwords (e.g., SLINT) were used as distracters, but not when pseudohomophones (e.g., GRONE) were used. The homophone effect involves longer response times for homophones (e.g., MAID/MADE) than for control words and is considered evidence of phonological processing. Davelaar et al. concluded that when pseudohomophones were used, readers de-emphasized phonological processing in order to make their LDs more easily. Further, Pugh, Rexer, and Katz (1994) found that in a LDT inhibition for FREAK-BREAK word pairs turned to facilitation when pseudohomophones were included among the distracters. Pugh et al. also concluded that when pseudohomophones were included participants adopted a strategy of de-emphasizing phonological processing. Both Pugh et al.'s and Davelaar et al.'s experiments, however, contained methodological problems. In Experiment 1 these problems were corrected for Pugh et al.'s paradigm and the issue of strategic de-emphasis of phonological processing was re-examined. Results showed phonological inhibition for FREAK-BREAK pairs even when pseudohomophones were used. Experiment 2 was an exact replication of Pugh et al.'s experiment but, again, inhibition was observed with pseudohomophones, and thus Pugh et al.'s results were not replicated. In Experiment 3 the methodological problems in Davelaar et al.'s experiments were corrected. Results showed larger homophone effects with pseudohomophones; the opposite effect to that reported by Davelaar et al. The purpose of Experiment 4 was to investigate whether these effects are moderated by the extent to which the distracters were word-like. The results showed that processing was affected by the extent to which the distracters were word-like but, again, larger homophone effects emerged when pseudohomophones were presented. The purpose of Experiment 5 was to examine whether other phonological effects (regularity, homography) show the same pattern. Although homophone effects were again larger with pseudohomophones, regularity and homography effects were not. Several new conclusions are drawn about the nature of phonological processing and its strategic control in LDT. |
