In two previous posts we looked at David Hume’s argument that causation should be analyzed in terms of regularities between kinds of events. We found that his argument has difficulty distinguishing true causal relations from accidental regularities and from non-accidental regularities which involve a common cause. Therefore, something more is needed for a reductive analysis of causation to work.
A standard alternative to Hume’s view is to analyze causation in terms of counterfactual dependence. To say that A caused B just is to say that (in similar circumstances) if A had not occurred, then B would not have occurred either. B must depend on A in this sense; it is not enough for B merely to be correlated with A.
For example, saying that an ignition of gunpowder caused an explosion just is to say that (in similar circumstances) if the ignition had not occurred, the explosion would not have occurred either. The explosion must depend on the ignition in this sense; it is not enough for the explosion merely to be correlated with the ignition.
The counterfactual account makes up for what is apparently lacking in Hume’s account because it can distinguish true causes from mere regularities which do not involve causation:
First, consider an accidental regularity: when a silent clock pointing to the hour is regularly followed by another clock striking the hour, we know the former does not cause the latter because there is no counterfactual dependence. In similar circumstances, when the silent clock fails to point, the other clock will still strike.
Second, consider a non-accidental regularity: a scientist observes two chemicals form in a test tube. These two events always occur in rapid succession whenever he adds a catalyst to the test tube. How can the scientist know what is causing what? Well, if (in similar circumstances) both chemicals fail to form when he withholds the catalyst; and if (after adding the catalyst) either chemical still forms when he intervenes to block the other from forming, then the scientist can conclude that the catalyst is probably the common cause. Why? Because the formation of the two chemicals is counterfactually dependent on the catalyst and their formation is not likewise dependent on each other, despite their being regularly connected to each other.
To sum up, the counterfactual analysis of causation distinguishes true causes from mere regularities on the basis of what would happen in similar circumstances, when the supposed cause is absent. So far so good. But this leaves us with some questions: what exactly are the “similar circumstances” spoken of here? And how should they be understood? In my next post, I will discuss how the philosophical work of David Lewis addresses these questions.