Abstract: The inclusion of an uncertainty component in argumentation formalisms can be easily motivated by the need of capturing multi-agent, adversarial scenarios, in which agents might not be sure of other agents’ argumentative information. Up until now, uncertainty has been mainly studied from the perspective of abstract argumentation, i.e. ignoring the internal structure of arguments and the conflict relation among them. In this talk, I will discuss some ideas and results about structured argumentation frameworks with qualitative (non-probabilistic) uncertainty about some of its elements (e.g., inference rules, premises or preference among arguments). The main aim of the research is to discover whether these structured models are a correct instantiation of their “corresponding”, already-studied, abstract ones. Most of the current results are negative, showing that a simple kind of uncertainty at the level of components (rules, premises of preferences) can generate more complex forms of uncertainty at the abstract level (arguments and conflicts). I will discuss whether these results can be used as a reason against the uncontrolled flourishment of abstract formalisms, as already claimed from many sides.
The talk is based on ongoing work with Carlo Proietti.
edev 