UvA

In current minimalist analyses dislocation (Move/Agree) is motivated by Full Interpretation: Move/Agree is said to check uninterpretable features. In this paper I address several of the background assumptions behind such analyses, and discuss several problems that go with them. I this paper I first demonstrate by discussing so-called doubling phenomena that the assumption that language exhibits semantically uninterpretable material is essentially correct. Second, I show that if language is thought to be an optimal solution to different interface conditions that are (to some extent) in conflict, language allows three equally economic marking strategies to express a particular semantic operator/property. These marking strategies are constituted by the three syntactic operations External Merge, Internal Merge (Move), and Agree. Finally, I conclude that the third marking strategy, Agree, calls uninterpretable formal features into existence (i.e. without such features Agree-relation creates uninterpretas cannot be intantiated in the first place).