摘要：AbstractThe principle of counter-transitivity plays an important role in argumentation. It states that an argument x is stronger than another argument y, if the attackers of x are weaker than those of y. In this paper, we generalise this principle in order to compare and rank arguments. Two basic criteria, quantity-and quality-based, as well as their various combinations, are proposed to instantiate the principle. Based on these instantiations, a family of argument ranking semantics are then developed and studied. We show that these semantics enjoy different behaviours, and can provide different levels of ranking precision on arguments. Moreover, we reveal that Dung’s classical semantics also satisfy counter-transitivity, and can be seen as a three-level quality-based semantics.
关键词：KeywordsArtificial intelligenceArgumentationLogical designRanking semanticsDecision making
