Cognitive compositional semantics using continuation dependencies

Abstract

This paper describes a graphical semantic representation based on bottom-up 'continuation' dependencies which has the important property that its vertices define a usable set of discourse referents in working memory even in contexts involving conjunction in the scope of quantifiers. An evaluation on an existing quantifier scope disambiguation task shows that non-local continuation dependencies can be as reliably learned from annotated data as representations used in a state-of-the-art quantifier scope resolver, suggesting that continuation dependencies may provide a natural representation for scope information.