Abstract

In this talk I will argue that some prominent approaches to scope underspecification that are aimed at providing efficient procedures to represent scopal ambiguites lack the necessary expressive power to fully account for ambiguities of natural language. I will then show that a more fundamental problem underlies this observation, namely that expressively complete formalisms must necessarily run into some version of the combinatorial explosion problem, which they were initially designed to solve.

For more information, see http://www.illc.uva.nl/dip/.