

Actual causes are particular — “Sam developed cancer because he smoked heavily”, “The pilot’s error was the cause of the plane crash, not the mechanical failure”, “Robert missed the meeting due to his alarm not going off”. They are not especially useful for making predictions, because they lack generality. But, looking backwards, they are the basis for assigning responsibility for outcomes and explaining why events unfolded as they did. This tech report defines actual causality in terms of contingencies, within the causal model formalism. causaLens’ own Hana Chockler’s work on responsibility and explanation is unpacked within this formalism.