Revisiting form typicality of nouns and verbs A usage-based approach

Research has shown that, in English, the mapping between a word's form and its syntactic category is not entirely arbitrary. Tough formal differences between lexical categories are subtle, adults are sensitive to them and access this knowledge when retrieving or manipulating grammatical category information. Studies of form typicality have so far exclusively investigated unambiguous (or disambiguated) wordforms. We test the prediction that form typicality also affects visual processing of ambiguous wordforms, with formal features correlating, not with a form's designation as a particular category, but with a form's probability of being used as a particular category. Our results indicate that "form discrepancy", a measure of how well a form's category usage matches up with its form (i.e. typically nouny forms associated with high probability of usage as a noun), is a signifcant predictor of lexical decision response time. Tese data are in line with models in which category is not specifed for roots in the lexicon but rather assigned within syntactic or semantic context, and show that distributional information about grammatical category usage is automatically accessed in visual word processing.