One of the challenges faced by researchers working on conversational humour across languages is that the particular scientific metalanguage we use to talk about the phenomenon in question influences and shapes our understanding of it. The aim of this paper is to explore the import of such issues for research on conversational humour through an examination of the labels used by Taiwanese speakers of Mandarin Chinese when talking about what is broadly termed “teasing” in English. Our aim is to better understand the connotations of these various native terms, how they relate to each other, and how they are deployed by speakers when referring to “teasing” events. The study draws on interviews with native Taiwanese informants and an analysis of large web-based corpus of Mandarin Chinese, the zhTenTen17 Traditional Corpus, to show that “teasing” is conceptualised in complex ways by Taiwanese speakers of Chinese, and because of that the same “teasing” event is open to construal in different ways by those speakers. We conclude that metapragmatic studies of conversational humour in different languages are important if we are to avoid bleaching out important cultural properties of teasing and thereby distorting our objects of analysis.