USING ARGUMENTATION FOR IDENTIFYING NON-AMBIGUOUS INTERPRETATIONS OF NATURAL LANGUAGE

We introduce Argumentation Theory as a setting for identifying non-conflicting interpretations of natural language discourses. Following existing work on textual entailment, our solution relies on extracting entailments and contradictions from texts. We cast the former and latter into arguments and attacks from Argumentation Theory, respectively. Finally, we examine different types of semantics and select the preferred semantics as the appropriate characterisation which yields un-ambiguous interpretations of a discourse.