Derivational and representational views of minimalist transformational grammar

Abstract

In this paper we cast the long-standing debate in transformational grammar over movement derivations vs. well-formedness conditions on chains in a new light. We claim that the argument itself is misguided: a transformational grammar should have both a derivational and representational interpretation, connected by soundness and completeness results. Second, we argue that the proper form of the representational interpretation is as an axiomatization of T-markers rather than P-markers (e.g., LF structures). Little of significance in representationalist approaches is lost by this move, however. Antecedent government conditions on chains, for example, can still be easily stated in terms of the kinds of T-markers that we propose. Many well-known problems for representational approaches are avoided, however. In addition, from a derivational perspective, global constraints on derivations become simple structural well-formedness conditions on T-markers. We offer a rigorous formalization of a simple system of T-markers and show its equivalence to a simple minimalist derivation system.