On the Structural and Typal Differences between Mass and Count Nouns: Evidence from Romance

There is an ongoing debate regarding the syntactic and semantic nature of the mass versus count-noun distinction. If we concentrate solely on the semantic side of the debate, we can distinguish two families of approaches. The first approach, represented, for instance, by Link (1983) and Chierchia (1998), among others, argues that the difference between mass and count can be reduced to homogeneity or vagueness. According to the other approach, the difference is typal (Krifka, 1989, Rothstein, 2010). This paper contributes to the debate by considering new data from Romance dialects which provide empirical support for the latter family of approaches. Our starting position is based on two assumptions. First, we assume – following Borer (2005) and Rothstein (2010), among others – that real world homogeneity (or cumulativity for that matter) does not necessarily correspond to a grammatical representation of homogeneity (or cumulativity, respectively). Furthermore, we assume that there is no one-to-one mapping between real world countability and semantic countability. Once we take the stand that homogeneity and countability are encoded in the grammar, the question that arises is whether the encoding happens at the level of the lexicon or whether it arises through derivation. Specifically, if the difference between mass and count is typal, does it mean that typally distinct nouns must have a distinct syntactic structure? The answer to such a question is not trivial. For example, Borer’s position is that a typal difference necessarily arises out of differences in the syntactic structure, whereas Rothstein defends the position that typal differences are in principle independent of structural differences, and that semantically distinct nominal structures can in principle be syntactically identical.