Beliefs Based on Evidence and Argumentation

In this paper, we study doxastic attitudes that emerge on the basis of argumentational reasoning. In order for an agent's beliefs to be called 'rational', they ought to be well-grounded in strong arguments that are constructed by combining her available evidence in a specific way. A study of how these rational and grounded beliefs emerge requires a new logical setting. The language of the logical system in this paper serves this purpose: it is expressive enough to reason about concepts such as factive combined evidence, correctly grounded belief, and infallible knowledge, which are the building blocks on which our notions of argument and grounded belief can be defined. Building further on previous work, we use a topological semantics to represent the structure of an agent's collection of evidence, and we use input from abstract argumentation theory to single out the relevant sets of evidence to construct the agent's beliefs. Our paper provides a sound and complete axiom system for the presented logical language, which can describe the given models in full detail, and we show how this setting can be used to explore more intricate epistemic notions.