On the So-Called Idea of Causation

The argument of this paper may be summarized as follows. Causal propositions (propositions of the type “x causes y”) are ambiguous. Such a proposition may have any one of three meanings (possibly more; but three is enough for this paper). The ambiguity, however, is of a rather odd kind. Sense I, which is historically the original sense, is presupposed by the others, and remains strictly speaking the one and only “proper” sense. When we assert propositions containing the word cause in senses II and III, we are “saying” one thing and “meaning” another; we are describing certain things as if they were things of a kind which we do not actually believe them to be. This always has an element of danger in it: the danger of inadvertently beginning to “mean” what one had only intended to “say”, i.e., of thinking that things are what we describe them as if they were. This danger is much worse when our “metaphors” get “mixed”. This is what has happened with the so-called “idea of causation” from the time of Kant onwards. It is a confusion of certain characteristics belonging to sense II with certain others belonging to sense III. Nothing can be done, therefore, towards clearing up our minds about causation, by merely analysing the idea as it stands and detailing the various elements it contains; for these elements are mutually contradictory. We must carry the process further, by segregating the elements under different heads, and distinguishing these as different “senses” of the word. But even this is not enough. A further step in the process is needed: namely a critical discussion of each “sense” taken singly. When this is done it will be found that the best way of avoiding confusion will be to restrict our use of the word cause to occasions on which it is used in its “proper” sense, No. I; that on the occasions on which we use it in sense II we should be wise to use instead the terminology of means and ends; and that when we use it in sense III we should do better to speak of “laws” and their “instances”.