Extending semantic relation classification to presupposition relations between verbs
In contrast to typical semantic relations between verbs, such as antonymy, synonymy or hyponymy, presupposition is a lexical relation that is not very well covered in existing lexical resources. It is also understudied in the ﬁeld of corpus-based methods of learning semantic relations. But presupposition is very important for the quality of automatic semantic and discourse analysis tasks. In this paper we present a corpus-based method for learning presupposition relations between verbs, embedded in a discriminative classiﬁcation approach for ﬁne-grained semantic relations. The focus of the present paper is to discuss methodological aspects of our approach including the choice of resources and data sets, the selection of features for classiﬁcation, and design decisions regarding the annotation of ﬁne-grained semantic relations between verbs.