Towards Computational Models of Natural Argument using Labelled Deductive Systems

During the last decade computational models of argument have emerged as a successful approach to the formalization of commonsense reasoning, encompassing many other alternative formalisms. Common elements can be identified in such frameworks along with a number of particular features which make it difficult to compare them with each other from a logical viewpoint. This paper presents a unifying approach to computational models of argument using Labelled Deductive Systems (LDS), a rigorous but flexible methodology which has been developed to formalize complex systems as logical frameworks with labelled deduction capabilities. In the context of defeasible argumentation, we show how labels can be used to represent arguments as well as argument trees. In particular, we will describe the wide range of possibilities allowed by an LDS-based formalization, such as capturing fallacies, formalizing arguments in social contexts, weighing arguments, and formalizing argument-based consequence operators.