Are there “weak” definites in bare classifier languages?

Abstract

This paper motivates a new view on the typology of definiteness that integrates (quasi-)names. The primary data is drawn from Cantonese and Bangla, where both bare classifier constructions and bare nominals are recruited for definite expressions. We argue that these bare nominals, while often analyzed as the so-called “weak”/unique definites in other languages, are indeed name-like expressions akin to the quasi-name Mom in English, in contrast with the definite descriptions denoted by bare classifier constructions. We propose that quasi-names, as well as proper names, are derived by a definite determiner that encodes a functional relation between the discourse participants and the referent. We further discuss cases where quasi-names compete with definite descriptions and proper names. The findings not only suggest that names should be brought into the picture, but also shed light on how pragmatic principles interact and determine the choice of referring expressions.