Probabilistic Causation and Human Action

Abstract

Let us consider the following example, which is supposed to illustrate probabilistic causation. We conceive of a slot machine with 1000 different outputs. If a coin is inserted in the machine, it produces one of the outputs, 1,2,…, 1000, according to some (objective) probabilistic process. We may, for simplicity, assume that each output is realized with an equal probability of 1/1000, when a coin is inserted. In this example, we are then prepared to assert that inserting the coin (probabilistically) causes one or the other outputs to be realized. Furthermore, let us suppose that in fact output no. 617 (a chocolate bar) was realized on a particular trial. Then we say that inserting the coin on that occasion (probabilistically) caused the chocolate bar to come out. An explanation of the chocolate bar’s emerging (a “low-probability” event) can be given by saying that it was causally produced by my inserting a coin. (It did not come about because of the dysfunctioning of the machine, for example.)