Justified argument revision in agent dialogue

In certain dialogue protocols, when an agent concedes information that conflicts with an existing commitment, its opponent can force it into retracting that existing commitment, as well as other commitments from which the retracted commitment can be inferred. Belief revision studies how a minimal amount of information can be given up in the face of new, conflicting information, with prominent theories in this field being applied belief sets, expressed in propositional logic. While many dialogue games also find their roots in propositional logic, relatively recent advances have seen these games adapted for use with systems of argumentation, which offer more sophisticated models of reasoning and conflict beyond propositional inference and classical negation. In this paper, we define and describe the properties of two operators for the modification of a knowledge base in a system of structured argumentation. We then go on to present a method by which an agent can determine the minimal change when applying these operators in order to justify the addition or removal of propositions to bring about consistency in its commitment store.