The relative time course of neighborhood and lexical effects

Abstract

We have previously demonstrated that neighborhood density can have effects similar to those of lexical status in phoneme identification tasks. In a typical task demonstrating a lexical effect, subjects might hear two series, one ranging from ‘‘bag’’ to ‘‘pag,’’ while the other varies from ‘‘pal’’ to ‘‘bal.’’ Subjects are more likely to classify ambiguous stimuli from each series as members of the category that makes a word (they would classify ambiguous items as ‘‘b’’ in the bag–pag series but as ‘‘p’’ in the pal–bal series). In a similar task where none of the four endpoints were words, subjects classified ambiguous items as members of the category that had a greater neighborhood density. The current study examines the relative time course of these effects. We created both lexical and neighborhood series using the identical initial contrast, and presented them to the same set of listeners. Neighborhood effects occurred primarily in the fastest responses, while lexical effects occurred in the medium and slower responses. That is, neighborhood effects appear earlier in processing, and dissipate quickly, whereas lexical effects take longer to build up. These results will be discussed in terms of their implications for models of word recognition.