Abstract This study investigates the complete functional range of numeral classifier systems in Southern Sinitic languages (Cantonese, Min and Hakka) from a new perspective, paying special attention to the functional range of specific classifiers (whose use is limited to a specific set of nouns) and general classifiers (which can often co-occur with any count noun) and the differences between them, based on spoken corpus data and existing literature. The results show that numeral classifiers in Southern Sinitic languages are characterized by extreme multifunctionality and considerable cross-linguistic variation. All classifiers in all the Southern Sinitic languages share the primary function of individuation, while classifiers in each language also display different additional functions such as marking (in)definiteness and possession. Moreover, the general classifiers can further take on various functions, including but not limited to relativization and nominalization. With their additional functions, general classifiers end up being ‘forerunners’ in the grammaticalization of numeral classifiers in Southern Sinitic languages. Their functional expansion is driven by syntactic need, cultural practice, high frequency and semantic generality. Furthermore, the case study on the functional development ‘numeral classifier > adnominal marker > nominalizer > stance marker’ based on spoken corpus data confirms the robustness of this grammaticalization chain.
Qi Huang (Thu,) studied this question.