Activation of classifiers in word production: insights from lexico-syntactic probability distributions

In speech production, speakers need to activate and select appropriate lexico-syntactic features of words to plan grammatically correct sentences. While previous studies have focused on situations with a single grammatically correct option, Mandarin Chinese classifiers present a case where multiple options are correct. In this study, we asked native Mandarin Chinese speakers to name target pictures using a picture-word interference paradigm while recording EEG. Distractor words with varying degrees of classifier distribution similarity were superimposed while controlling for dominant classifier congruency. Distractors with dissimilar classifier distributions resulted in a more positive P600-like effect, but no behavioural effect was observed, compared to distractor nouns having similar classifier distributions. Based on this result, we propose that, when producing a bare noun, multiple compatible classifiers are activated at the lexical level with the degree of activation being determined by their corresponding compatibility with the given noun.