Where is the locus of difficulty in recognizing foreign-accented words? Neighborhood density and phonotactic probability effects on the recognition of foreign-accented words by native English listeners

This series of experiments (1) examined whether native listeners experience recognition difficulty in all kinds of foreign-accented words or only in a subset of words with certain lexical and sub-lexical characteristics---neighborhood density and phonotactic probability; (2) identified the locus of foreign-accented word recognition difficulty, and (3) investigated how accent-induced mismatches impact the lexical retrieval process. Experiments 1 and 4 examined the recognition of native-produced and foreign-accented words varying in neighborhood density with auditory lexical decision and perceptual identification tasks respectively, which emphasize the lexical level of processing. Findings from Experiment 1 revealed increased accent-induced processing cost in reaction times, especially for words with many similar sounding words, implying that native listeners increase their reliance on top-down lexical knowledge during foreign-accented word recognition. Analysis of perception errors from Experiment 4 found the misperceptions in the foreign-accented condition to be more similar to the target words than those in the native-produced condition. This suggests that accent-induced mismatches tend to activate similar sounding words as alternative word candidates, which possibly pose increased lexical competition for the target word and result in greater processing costs for foreign-accented word recognition at the lexical level. Experiments 2 and 3 examined the sub-lexical processing of the foreign-accented words varying in neighborhood density and phonotactic probability respectively with a same-different matching task, which emphasizes the sub-lexical level of processing. Findings from both experiments revealed no extra processing costs, in either reaction times or accuracy rates, for the foreign-accented stimuli, implying that the sub-lexical processing of the foreign-accented words is as good as that of the native-produced words. Taken together, the overall recognition difficulty of foreign-accented stimuli, as well as the differentially increased processing difficulty for accented dense words (observed in Experiment 1), mainly stems from the lexical level, due to the increased lexical competition posed by the similar sounding word candidates.