For several reasons, I’ve been thinking a lot about causation recently. And one thing that has come up in a few places is the variety of causal talk that we have. Consider the following two claims.
bq. (1) Andy turned the TV on.
(3) Andy caused the TV to be turned on.
The two claims are fairly clearly distinct. Imagine that I have the habit of turning on the TV any time someone drives a red car down my street. Andy drives a red car down my street, so I turn the TV on. Then (3) is true but (1) is false. I’m pretty sure (1) entails (3), but it is clearly the case that (1) can be false and (3) true.
The reason I’m bringing this up is because of a discussion I had with Zoltan Szab{o’} about where claims like (2) fit into the equation.
bq. (2) Andy made the TV turn on.
I think (after being prodded by Zoltan in this direction) that (2) is strictly weaker than (1) and strictly stronger than (3).
The latter claim should be easy enough to prove. It seems in the case I gave above (3) is true, but (2) seems false. But anyone who causes a TV to turn on makes it turn on, so (2) is strictly stronger than (3).
It’s the former claim that I worry about. Anyone who turns on a TV makes it turn on, so the entailment of (2) by (1) isn’t a problem. The interesting question I think is whether (1) can be false and (2) true. Here are three cases that suggest it might be true.
bq. (4) Andy discovers my habit of turning on the TV when a red car drives down the street. Desiring that the TV be turned on, he drives down my street in a red car. I observe him, and this causes me to turn the TV on.
bq. (5) Andy asks me to turn on the TV. I do so.
bq. (6) Andy, who is President, orders me, a lowly soldier, to turn on the TV. Since I don’t feel like being court-martialled for insubordination, I do so.
In (4) it is clear that Andy doesn’t turn on the TV, but it isn’t clear (I think) that he makes the TV turn on. Perhaps the case is similar enough to our original case that he doesn’t do so.
In (6) it is clear that Andy does make the TV turn on, but perhaps he does turn on the TV. Note that if we change the verb, we get cases where we will clearly use a causitive construction. If Andy commanded me not to turn on the TV, but to kill various prisoners, then we’d say that he killed them. At least, that’s what we do say about dictators who kill thousands, or millions, of their subjects.
I suspect (5) is the best case for one where Andy makes the TV turn on without turning it on. It would be odd, not impossible but odd, to deny either of these claims. So I’m inclined to think that (2) is strictly intermediate between (1) and (3).
The next question is how much this three-way distinction matters to familiar debates about causation. I’m inclined to think it matters a lot, in two respects. (The rest of this is going to be very speculative.)
First, whether a particular event is a causing of something phi, a making something phi, or a phi-ing matters, I suspect, to our intuitions about the causal relevance of other events that are causally connected to it.
Second, causation, understood as the relation being picked out in (3), might be much less philosophically and practically interesting than the making relation, or the relation that is relevant to causitives, like _killed_ or _turned on_. Call that latter relation R. I think both the making relation and R occur much more commonly in regular thought and talk than the causal relation, and that both concepts are acquired at an earlier age than the general causal concept. And I suspect, though I have no idea how one would prove this, that the making relation and R figure in more analyses than the causal relation itself does.