Semantic diversity in Chinese words: Exploring its role in word recognition, lexical ambiguity, character complexity,...

Semantic diversity (SemD) quantifies how (dis)similar a word's various contextual usages are across linguistic experiences. While extensively validated in English, its application to Chinese-a morpho-syllabic language where most words consist of multiple characters-remains unexplored. Here, we present the first comprehensive SemD database for 25,460 simplified Chinese words spanning one-to-ten characters using context vectors from latent semantic analysis. We evaluate its cognitive validity through five lenses. First, Chinese SemD exhibits systematic relationships with dictionary-based ambiguity, human ratings, and key psycholinguistic variables. Crucially, unscaled SemD shows stronger associations with all factors than scaled variants. Second, higher SemD facilitates recognition speed and accuracy beyond baseline predictors in mega-study lexical decision, with a stronger benefit for low-frequency words-replicating English findings in an ideographic system. Third, both lexical ambiguity types-polysemy and homonymy-positively predict SemD, indicating that words with multiple related senses or unrelated meanings are associated with more diverse contextual usage. However, behavioral and SemD-based analyses convergingly show that only polysemes demonstrate a processing advantage and higher SemD relative to unambiguous controls, whereas homonyms do not differ from controls on either measure. Fourth, words with more characters exhibit lower SemD, extending the Mutual Annotation and Constraint account whereby additional morphemic cues narrow Chinese words' interpretation. Fifth, cross-linguistic analyses demonstrate that English-Chinese translation equivalents with higher semantic alignment exhibit lower mean SemD and smaller between-language SemD disparities. These findings establish SemD as a psychologically meaningful measure of Chinese lexical ambiguity that bridges distributional, morphological, and cross-linguistic perspectives on the mental lexicon.