A formal framework for causal modeling and argumentation

Abstract

We develop a framework for causal modeling that features a language based on causal defeasible rules, a semantics based on order-ofmagnitude probabilities, and a proof-theory based on the interaction of arguments. The framework extends and integrates logical, probabilistic and procedural modeling languages such as logic programs with negation as failure, Qualitative Bayesian Networks and Axelrod’s cognitive maps.