DISCOVERING CAUSAL RELATIONSHIPSH
Jon Williamson · Oxford University Press eBooks · 2004
There are two main proposals for discovering causal relationships: a hypothetico-deductive and an inductive approach. Neither account is fully satisfactory. Various specific approaches are discussed, including constraint-based induction, Bayesian induction, information-theoretic induction, and Shafer's causal conjecturing. These approaches all take the causal Markov condition for granted and so may fail when the condition fails.