Accessing the mental lexicon in spoken word production: Masked priming effects in picture naming.

Abstract

This dissertation investigated the process of lexical access in spoken word production by using a picture naming task which involves very similar processes. Experiment 1 showed that significant repetition priming effect was obtained in this task when the prime was heavily masked and was unavailable to conscious report. In addition, the repetition effect was independent of word frequency. However, a pattern of frequency attenuated priming effects was obtained in Experiment 2 when the prime was unmasked and was named about 10 minutes previously. These results suggest that the masked repetition effect is lexical in nature, whereas the unmasked effect is contaminated by non-lexical sources, such as auditory episodic memory. Experiment 3 showed that the masked repetition effect was independent of the neighborhood density of target names, but the masked form-priming effect was highly constrained by the density. Further, Experiment 4 showed that once the form-related prime became phonologically identical to the picture name the form-priming effect was no longer constrained by the density. In order to distinguish which processing component (lexicalization or production) was responsible for the elimination of the constraint, a picture-fragment matching task was used. Experiment 5A showed that in the matching task, repetition effects were significant and independent of neighborhood density. In contrast with Experiment 4, Experiment 5B showed that the form-priming effect was highly constrained by the density in the matching task. These experiments suggest that (1) the process of phonological encoding is automatic and extremely fast; and (2) the phonological encoding is a necessary process for production but not for lexicalization. In addition, no masked associative priming was obtained in either a picture naming task (Experiment 6) or a picture categorization task (Experiment 7), suggesting that masked priming effects obtained in the present picture processing tasks were not due to facilitation occurring at the semantic or conceptual level. Finally, the robust frequency effects established in picture naming tasks were severely weakened when a picture-fragment matching task was used (Experiment 8). This pattern of results suggests that frequency influences mainly name production rather than name retrieval in picture naming.