earticle

영어

Within Chomsky’s Minimalist Program it is assumed that morpho-syntactic features fall into two types: interpretable and non-interpretable. The latter must be checked within the Computational System, that is, within core syntax, whereas the former need not be checked. This seems to follow from the Minimalist assumption that uninterpretable features may not survive at an interface and that therefore some sort of erasure by checking is needed to prevent a derivation from crashing. Since interpretability roughly means “having a semantic interpretation”, this view seems to provoke a number of hitherto unanswered questions: One problem is that standard model theoretic semantics does not provide us with any interpretation of pronominal features like singular or third person, although these features need not be checked whenattached to an NP
(or DP). The reason is that singular atomic reference to 3rd person is given “cognitive preference” in standard semantics: within this theory there is no way to interpret the respective features. Another problem is posed by numerous examples showing that features that need not be checked do not have the semantic interpretation one might expect. In order to overcome these
problems, I propose a combination of five different modules of grammar to get the desired results, resorting to (a) assumptions about unmarkedness in morphology; (b) assumptions about the semantics of pluralization; (c) pragmatic assumptions about the choice between singular and plural; (d) grammatical restrictions on feature agreement; and (e) a certain way of
interpreting pronominal features.