Causation

Causation is a concept that is at the heart of the subject of social power, of social relations more generally, and of questions of legal and moral responsibility and prediction in the social and natural sciences. In very general terms, to ask for the cause of some particular event is to ask what brought it about. That is, a cause is a relation between events, processes, or entities in the same time series, in which one event, process, or entity, C, has the efficacy to produce or be part of the production of another, the effect (or outcome) x. C is causal for x if C somehow makes a difference to the occurrence of x. In social contexts, C is an action performed by a person.

Arguments among philosophers about what actually constitutes this efficacy—or “making a difference”—are often concerned with whether the efficacy is a necessity requirement (if x occurred, then C preceded it) or sufficiency requirement (if C occurs, then x will follow) or both. There are three senses for each of these. In descending order of stringency, for necessity we have

• Strict necessity C is necessary for the occurrence of x whenever x occurs.
• Strong necessity C was necessary for x on the particular occasion.
• Weak necessity C was a necessary element of some set of existing conditions that was sufficient for the occurrence of x.

• For sufficiency:
• Strict sufficiency C is sufficient by itself for the occurrence of x.
• Strong sufficiency C was a necessary element of some set of existing conditions that was sufficient for the occurrence of x.
• Weak sufficiency C was an element of some set of existing conditions that was sufficient for the occurrence of x.

There are three remarks to be made here. First, we can instantly dispense with the concept of weak sufficiency for the simple reason that it is trivially satisfied by any condition C by simply adding C to an already existing set of conditions that is sufficient for x, although C is irrelevant as regards x. That is, weak sufficiency can fail to satisfy the simple counterfactual test of efficacy known as the “but for” test in a very unacceptable way. According to this test, C was a cause of some result x only if but for the occurrence of C, then x would not have occurred.

Second, if C is strictly necessary for x, then C should be ascribed causal status for x, and if C is strongly necessary for x, then C should also be ascribed causal status for x. In both cases, C satisfies the “but for” test. Yet strict necessity is inadequate to cover causal relations in general because it will often not be fulfilled. It would mean that evidence of an electrical short-circuit that combined with other conditions and resulted in a house burning down would not qualify for causal status given that it is not true that for all instances of a house burning down there must have been an electrical short-circuit. Strict sufficiency is also inadequate because it too may fail to exist: it is rare that a single condition C is sufficient for x. Strict sufficiency also rules out ascribing the electrical short-circuit causal status for the house [Page 97]burning down given that it alone cannot lead to the fire; the presence of flammable material, and so forth, is also called for. Moreover, that C may be strictly sufficient for x and was present on the occasion of x's occurrence does not imply that C actually caused x; x may have brought about by another strictly sufficient condition that was also present on the occasion.

...