In this article, I provide a unified explanation for two puzzling phenomena concerning sentence-initial negation: the ban on True Negative Imperatives that is attested in many languages and the ban on sole negative markers in sentence-initial position in V-to-C languages. I argue that both phenomena can be explained once it is assumed after Han (2001) that operators encoding the illocutionary force of a speech act take scope from matrix C° and may not be outscoped by negation. Consequently, a morphosyntactically negative element can appear in a position in C° or SpecCP only if it is semantically non-negative or if it can reconstruct to a lower position.
展开▼
