UC San Diego

The goal of the current work was to investigate the influence of neighborhood density (ND) on speech production. ND is an index of phonological similarity and refers to the number of meaningful words (neighbors) present in a language that differ only by adding, deleting, or substituting a phoneme in any word position (Vitevitch & Luce, 1998, 1999). Prior studies of ND have yielded conflicting findings for production, reporting both inhibitory and facilitory effects of ND (Heisler, 2004; Vitevitch, 2002). In order to discern whether words may act more as competitors or facilitators (or potentially neither) during speech production, three experiments were conducted with 39 preschoolers and 46 adults with typical development: Experiment 1 explored influences of ND during children's picture-naming, Experiment 2 considered effects of ND on adult word repetition, and Experiment 3 discerned the influence of ND during word learning in both age groups. Analyses of production included segmental and whole-word level errors, as well as the lexical nature of production errors. Results revealed varying effects of ND depending upon the task and group. Specifically, children's naming was facilitated by ND at the semantic level, yet unaffected by ND at the phonological level. Adult word repetition was facilitated by ND in terms of repetition accuracy and the nature of erred repetitions. Finally, an inhibitory effect of ND was found during word learning by children, while adults were seemingly unaffected by the ND manipulation during word learning. Taken together, the results suggest that words do not simply act as competitors or facilitators during speech production; rather, their interactive nature likely depends on the elements of a task and the developing status of the lexicon. Neighbors of a word appeared to generally act as facilitators until a threat was posed, such as when acquiring novel words that were similar in phonological composition to existing words in the lexicon. This research indicates that incorporating ND into experimental or treatment paradigms should be used with caution and should be based on the demands of a task as well as the participants involved