ArgAF: The Argumentation Annotation Framework

Abstract

Many text documents contain natural argumentation. In order to automatically process such documents, in particular to find and analyze the arguments they contain, there is need for annotated corpora. Two existing corpora are analyzed in this thesis, revealing some limitations. A new annotation framework called the ArgAF is designed, relying on a combination of existing data models. The ArgAF might serve as a basis for creating a new annotated argumentation corpus, which is intended to overcome the limitations of the existing corpora.