Building a wide coverage dynamic grammar

Incrementality is relevant for language modeling, speech recognition, and language generation, but requires a shift of the mainstream perspective to be addressed in most current models of syntactic processing, which underlies all NLP tasks. In this paper we devise a dynamic version of Tree Adjoining Grammar (DVTAG) that encodes a strong notion of incrementality directly into the operations of the formal system. After discussing the basic features of DVTAG, we address the issue of building of a wide coverage grammar and present novel data for English and Italian.