Thursday, March 28, 2013

David Hume's account of causation

To: Alistair L.
From: Geoffrey Klempner
Subject: David Hume's account of causation
Date: 15th October 2009 10:10

Dear Alistair,

Thank you for your email of 7 October, with your essay for the University of London Modern Philosophy module, in response to the question, 'What was Hume trying to do in his account of causation? How successful was he in attaining his objectives?'

This is a gripping question for me, as I am currently involved in a dialogue with an ISFP Fellowship student over Hume's analysis of causation. I am Humean about causation (but prepared to be proved wrong) while my student is anti-Humean. The most difficult thing, however, is trying to determine just what the Humean view essentially is.

You make a stab at characterizing Hume's position dialectically, by identifying his view of causation as an 'empiricist' response to 'rationalist' views of causation. The problem with this is that you don't say what these views are. 'From a rationalist viewpoint we are asked to accept cause and effect as innate knowledge' tells me nothing.

You go on to talk about Hume's rejection of the view that the cause exercises a 'power' over the effect, or that cause and effect are linked by a 'necessary connection'. But what does this amount to?

If I put a saucepan of water on the stove and switch on the gas, is it false to say that the flame has the power to boil the water? Is it not necessary that the flame will heat the water? Hume denies neither of these things. On his objective account of causation (the first of your his two definitions which you quote on p.3) to say that heating water causes it to boil is to assert the truth of a universal generalization. From the proposition, 'The water in the saucepan was heated' and the universal generalization, it follows as a matter of logical necessity that the water boils. (This is essentially Carl Hempel's 'deductive-nomological' analysis of causation.)

Similarly, we can define the 'power' of the heat to cause water to boil in terms of this generalization. So what we are looking for, as the target of Hume's critique, is a false or illusory (from Hume's perspective) philosophical notion of 'power'. The question is, how does one give expression to this?

Pre-philosophically, we undoubtedly do have the notion that the process of cause and effect involves some invisible connection between the actual cause and the actual effect themselves. It may well be true that the instance in question may be derived from a generalization, but somehow this doesn't capture the 'actual-doingness' of the cause. All the generalization states is that one thing happens, then the other, and this sequence is lawlike (exceptionless, supporting counterfactual statements etc.).

Attempts have been made to construct thought experiments where an example of 'A followed by B' is lawlike and supports counterfactuals, but where, intuitively, A is not the cause of B. This might be one way to go. Unfortunately, putative examples tend to boil down to A being caused by C and B being caused by D, where C and D do fulfil this intuitive requirement.

Hume's second definition, his psychological account of the 'necessity' of the causal relation is, in effect, a diagnosis of our pre-philosophical intuitions about causation which seeks to account for them in a way which is consistent with the 'deductive-nomological' view.

What I would have liked to have seen in your essay is some appreciation that the problem of causation is difficult, 'gripping', in a way which makes this more than a historical exercise. I get the impression that you see Hume as successfully putting the case for a 'scientific' approach, clearing away the rationalist rubbish so that we can see the world more clearly. The problem is that this doesn't engage with the intuition that there is something more to causation than 'constant conjunction'.

Another question which is completely left hanging is the role of Hume's very strict empiricism in all of this. You start off alluding to Hume's theory of impressions and ideas, his analytical tool for understanding causation and other notions. Yet this same approach notoriously led Hume ('in the section 'Scepticism with regard to the senses') to doubt the meaning of the claim that ordinary spatio-temporal objects have 'distinct and continued' existence during periods where they are not perceived!

This leads to the suspicion that, maybe, the analysis of causation which Hume offers is similarly the product of a false methodology, an overly restrictive analytical tool. If not, why not? I don't think so, I think that the analysis of causation survives the rejection of Hume's 'radical' empiricism. But that's something that has to be shown.

Generally, I found this essay rather short on arguments and analysis. I get the impression that you have relied on your reading of the Treatise rather than exploring debates in the literature on Hume. These would have given you a useful perspective.

All the best,

Geoffrey