Diagnostics of phonological lexical processing: pseudohomophone naming advantages, disadvantages, and base-word frequency effects.

Phonological lexical access has been investigated by examining both a pseudohomophone (e.g., brane) base-word frequency effect and a pseudohomophone advantage over pronounceable nonwords (e.g., frane) in a single mixed block of naming trials. With a new set of pseudohomophones and non-words presented in a mixed block, we replicated the standard finding in the naming literature: no reliable base-word frequency effect, and apseudohomophone advantage. However, for this and two of three other sets of stimuli--those of McCann and Besner (1987), Seidenberg, Petersen, MacDonald, and Plaut (1996), and Herdman, LeFevre, and Greenham (1996), respectively--reliable effects of base-word frequency on pseudohomophone naming latency were found when pseudohomophones were presented in pure blocks prior to nonwords. Three of the four stimulus sets tested produced a pseudohomophone naming disadvantage when pseudohomophones were presented prior to nonwords. When nonwords were presented first, these effects were diminished. A strategy-based scaling account of the data is argued to provide a better explanation of the data than is the criterion-homogenization theory (Lupker, Brown, & Colombo, 1997).