Văn bản này: Corpus-based and computational approaches to discourse anaphora