Thursday, March 02, 2006

Hume on the Components of Causal Inference

What do we do when we engage in causal inferences? One of Hume’s important claims is that chains of causal inferences must be ‘fixed’ in some impression or memory. Hume establishes this by a regress argument. If Hume is right, any causal inference will involve the following:

(A) An originating impression of the thing
(B) A transition to the idea of the connected cause or effect
(C) The idea of this connected cause or effect

What is more, causal inferences can induce belief, and this is an important feature of them, given that they are only way (besides impressions) that we can gain the knowledge that something exists. Hume uses these insights to divide his discussion in the Treatise into the following parts:

1) The analysis
1a) The nature of the impressions (1.3.5)
1b) The nature of the inference (1.3.6)
1c) The nature of the idea or belief (1.3.7)
1d) How belief is induced (1.3.8)

5) Relevant objections
5a) How can Hume’s analysis distinguish causal inference from other inferences? (1.3.9)
5b) How can Hume’s analysis account for the immense and important effects of belief? (1.3.10)

6) Confirmation of Hume’s analysis by application to different types of causal reasoning
6a) ‘Probability of chances’ (1.3.11)
6b) ‘Probability of causes’ (1.3.12)
6c) ‘Unphilosophical probability’ (1.3.13)

At this point, Hume uses his analysis to return to his original question, namely, the nature of the necessary connection between cause and effect. One advantage of laying out the argument in this way is that it enables us to see just how beautifully organized the argument really is. It sprawls considerably, so on a first reading, or, for that matter, on multiple readings, it is very easy to regard it as one long ramble. In fact, it is a brilliantly organized argument, despite its apparent digressions. Hume has a particular analysis of causal inference, which is important for answering the question he wants to answer. He recognizes, however, that the analysis is likely to be controversial, so he does his best to show just how strong it can be by answering the chief objections and showing how fruitful it is. It is, if I may interject a personal note here, a breathtaking instance of one of the great rewards of doing history of philosophy; at times one comes across an argument, perhaps unclear or rambling or apparently confused, that, laid out correctly, shows itself to be truly beautiful. If this is rambling, would that more philosophers rambled!

Of course, I think for various reasons that Hume's analysis fails, particularly with regard to the induction of belief; but it's a genuinely noble attempt.