In defense of the transitivity of causation

Some people think it's obvious that causation is transitive: The abrupt motion of the cue stick causes the cue ball to have motion M, the cue ball's having motion M causes the eight ball to have motion M' which causes to be sunk in the corner pocket. The abrupt motion of the cue stick causes the eight ball to be sunk in the corner pocket. QED

Others think there are counter-examples (my own flowery version): Heavier-than-normal April showers lead to brighter-than-normal May flowers. This draws larger-than-normal crowds of tourists which leads, sadly, to a forest fire. But heavier-than-normal showers don't lead to forest fires. QED

I suppose a physical causation person like Phil Dowe would try to explain away the appearance of transitivity by saying that the appearance of transitivity piggybacks off the fact that there's an uninterrupted line of a conserved quantity running from the cue to the eight ball. What I don't know--none of those folks I've read have discussed it in what I've read as far as I recall--is why they don't want, indeed, why they aren't stuck with transitivity.

Either a line of conserved quantity gets interrupted or it doesn't. If it does, then there's no causation across the gap and there's nothing to worry about (though there will be cases where there's *dependence* across a gap, so there will be disagreements with the counterfactual theorist here). But if there's not an interruption in the line from A to B to C then the case will be like the billiards case and so it'll be transitive.

There's probably a simple answer to this I'm just overlooking. (And I'm blogging here in bed this morning thinking about whether to swim at 8:30 or 11:30 and I just don't want to go downstairs and grab a book to find out (Hitchcock 2001 might say), so I'm content to record my present puzzlement (Dowe's SEP article on Causal Process doesn't mention transitivity).)

Lewis's defense of transitivity is that we are mislead in particular judgements by what is common for types. Typicallly, rain doesn't cause forest fires. But under the right circumstances it can (as in my flowery story above). I think this can be spun even better (later post) but I think this is basically correct.

What causes what is highly subject to background conditions. Lit cigarettes thrown on the ground cause forest fires, right? But under the right circumstances they can prevent them it seems. If a small fire started by a careless tourist causes underbrush to burn away a forest fire could be averted. In fact pre-burning is a *common* way to prevent forest fires. Most often setting small fires in the brush causes big fires in the woods.

But under the right circumstances things go otherwise. I think what's going on in the purported counter-examples is that they are--of course, they're attempted counter-examples--hand selected in such a way that there are highly unusual background conditions. This is what makes them plausible counter-examples, but it's also precisely what reveals why our intuitions are mislead.