Semantic contributions to word naming with artificial lexicons

Date of Completion

January 2006


Language, Linguistics|Psychology, Experimental|Psychology, Cognitive




Inconsistencies in spelling-sound correspondence in English offer unique opportunities to examine the cognitive workings of the reading system. Prior studies have exploited these inconsistencies to demonstrate that meaning plays a significant role in resolving pronunciations from the written form of words. Strain et al. (1995, 2002) have shown an interaction between word regularity and imageability whereby highly imageable exception words are read aloud as quickly as words with regular pronunciations. However, these studies have suffered several criticisms, including failure to control for potentially confounding correlations between word regularity and age-of-acquisition. The current study introduces a technique, the trained pseudoword lexicon, which provides a means of controlling age of acquisition and other factors by experimental design, rather than statistical means. Experiment 1 presents evidence that this paradigm does indeed produce results consistent with classic regularity effect findings. Experiment 2 replicates findings of semantic contributions to single-word naming when a variety of factors are controlled by using the trained pseudoword lexicon. Implications of these findings are discussed for the interactive triangle and dual-route models of reading, the results being consistent with the predictions of a triangle model. Potential future applications of the trained pseudoword lexicon are also addressed. ^