Ellipsis presents an ideal testing ground for understanding the intricate relationship between form, meaning, and context. There are two common competing conceptions of how verb phrase ellipsis (VPE) can be resolved: in the first, ellipsis is resolved by constructing unpronounced syntactic representations at the ellipsis site; in the second, ellipsis can be resolved by consulting the semantic/discourse information present in the antecedent, without the mediation of any syntax at all. In this talk, I present some experimental studies to examine how English VPE is resolved. In the first set of studies, a series of syntactic priming experiments showed that VPE triggers the structural priming effect. Although these findings are consistent with the hypothesis that resolving VPE involves accessing syntactic representations, it is not conclusive due to uncertainties associated with the syntactic priming paradigm. In the second set of experiments, we more directly probe the interaction between discourse context and ellipsis-antecedent identity relation by systematically manipulating these two information sources. Our results suggest that the best account of VPE interpretation is a hybrid probabilistic account that incorporates both discourse pragmatic and syntactic information. In the final set of studies, we turn to deaccenting, a phenomenon that raises parallel questions as VPE resolution. We show that deaccenting actually imposes stronger requirements on the identity relation between the deaccented material and its antecedent, and any pragmatic effect is better modeled as pragmatic accommodation.