Probabilistic Causality and Causal Generalizations

Abstract

Theorists of probabilistic causation have failed to distinguish between different tasks. One problem is to understand generalizations such as, “Smoking causes lung cancer,” “Seat belts save lives,” or “Just a spoon full of sugar helps the medicine go down.” Some causal generalizations, like the examples I have just given, are immediately practical. Other causal generalizations, such as those that are central in economics may be more theoretical. Whether immediately practical or not, causal generalizations are problematic, because the cause they purport to identify are not invariably accompanied by their effects. They are in this way irregular.

I owe a special debt to Ellery Eells and Elliott Sober, with whom I spent many hours discussing the issues raised in this essay, to Joonsung Kim, who devoted a chapter of his dissertation to criticism of an earlier version, and to Christopher Hitchcock for letting me appropriate so many of his ideas as well as for his criticisms of an earlier version of this essay. I am also grateful to Helen Beebee, Nancy Cartwright, Malcolm Forster, Huw Price, Peter Menzies, Charles Twardy, Jim Woodward, and John Worrall, who offered helpful comments on earlier drafts, to students in two seminars at the University of Wisconsin and to audiences at the University of Maryland, the London School of Economics, Duke University, Monash University, and Sydney University.