ABSTRACT

This book concerns the automatie resolution of anaphors, which is a crucial task in the understanding of naturallanguage by computers. Before introducing concepts central to the book in Chapter I, I shall discuss why 'understanding' natural language is so difficult and hint at how computer systems attempt this task by analysing the input at different levels. The sketch provided here should help the reader to see where anaphora resolution fits in the bigger picture of Natural Language Understanding. I shall briefly review the levels of linguistic and extralinguistic analysis, and the various relevant forms of knowledge.