Hume on causation

Class notes for 9 February

Main points

We use causal inferences in order to form almost all of our beliefs about things we cannot presently observe. This includes our beliefs about the future, as when we make predictions, and the past, as when we explain why things happen.

We also draw a sharp distinction between causal relations and coincidences. A cause makes its effect happen; it makes it necessary that the effect comes about.

Hume’s broad contention is that we cannot draw this distinction: we don’t understand how one thing makes another necessary and so every inference we make is just a matter of probability.

In today’s class, we talked about why he claimed that our causal inferences are not based on reason.

There are two kinds of reasoning: that involving relations among ideas and that involving matters of fact.

The first cannot be the source of our causal inferences because it is impossible to imagine its results being otherwise: you can’t imagine 3 x 5 = 16, e.g.

The second fails because all we know about matters of fact is what we have observed. In order to reason that past observations are relevant to the future, we would have to have reason to believe that the future will resemble the past (in the relevant respects).

But there is no way of establishing that the future will resemble the past. It is not certain or necessary that this be so. And we cannot infer that it will be so on the basis of our past experiences because establishing that the past is relevant to the future is precisely what is at issue.

That’s Hume’s argument, in any event.

Nathana’s challenge

Nathana, and Jon, expressed dissatisfaction with Hume’s argument.

They said that Hume’s argument rests on the assertion that we can conceive of causal regularities changing: the cue ball rolls backwards after hitting the eight ball, gravity flings us off the face of the Earth, the coyote keeps running off the cliff and never falls down (even when he looks down), etc.

But, they said, he hadn’t shown this. They granted that it seems that we can imagine these things happening. But, they said, there are two explanations for that.

  1. It’s possible that they might change.
  2. We just don’t understand why, in fact, they cannot change. If we did, we wouldn’t think we could conceive of it.

They said that Hume simply assumes that (1) is the correct explanation without dismissing (2).

If they can bolster the case for believing (2), they will have a tough objection on their hands.