Exploring metaphorical polysemy with Multiple Correspondence analysis: A corpus-based study on the predicative hei ‘black’ in Chinese

This paper provides a corpus-based, statistical analysis to explore the semantic (dis)similarities of four metaphorical meanings of the Chinese color term hēi ‘black’ regarding its predicative usages. With the Behavioral Profiles approach, 379 instances were manually annotated with 35 contextual features proposed from three categories, including lexical-collocational, morphosyntactic and semantic, and discourse information, to capture the contextual variations of hēi. Based on the annotated data, the Multiple Correspondence analysis (MCA) technique is then used to visualize the semantic distribution of the four meanings of hēi and their strength of associations with the distinctive features. It is found that the semantic (dis)similarities of the four meanings of hēi are well profiled by the MCA results, which demonstrates the effectiveness of the MCA method in studying metaphorical polysemy.