Two Types of there-sentences and Feature Specification

This paper aimed to account for unexpected accusative case on a pronominal associate in there-sentences (there was him). It is unexpected that under the long-standing assumption in generative grammar, agreement coincides with case assignment. Since there-associates appear to agree with T in number (e.g. There was/*were a dog), they are expected to be valued as nominative case. Furthermore, such a pronominal associate with accusative case is not available in there-V type sentences (*There arrived him). In this paper, I propose a multiple-there hypothesis to account for different behaviors of there-V and there-BE. In the proposal, I argue that there-s in there-V and there-BE are base-generated in different positions due to their different “grammatical roles” (semantically null expletive vs. subject argument). Based on the distinction, I further argue that there-s have different feature specifications and show that the proposed system captures different behaviors between there-sentences with respect to sub-extraction and control.