To: Scott B.
From: Geoffrey Klempner
Subject: David Hume on the idea of causation
Date: 30th April 2008 11:53

Dear Scott,

Thank you for your email of 21 April, with your University of London Diploma essay in response to the question, 'What explanation does Hume give of the idea we have of a cause?'

Whenever you get a question like this, it is always a challenge to find something more to say than merely giving a precis of what the philosopher says. This is a point I may have made before. In fact, although you stick close to the text, you do attempt to provide your own take on what Hume is trying to do, in giving two definitions of 'cause'.

Generally, with a question like this, you will get more marks if you are able to offer a critique of the account which a philosopher gives. If you disagree or think that the account is inadequate in some way, then explain the reasons why. If you agree, then you can still offer objections that someone might have to the account and then, on the philosopher's behalf, give replies to those objections.

You say, 'It seems that the first definition is based on when we judge a cause to produce an effect. the second definition is based on why we judge a cause to produce an effect.' There is something right about this, but it needs to be made more precise.

The first part would be correct if you said, 'the first definition is based on when we *correctly* judge a cause to produce an effect.' The point is that Hume is stating what it is for it to be true to say that A caused B.

The second part has a certain ambiguity, because the phrase 'why we judge a cause to produce an effect' can refer either to the causal process in the mental machinery -- in Hume's case the story about impressions and ideas -- or to the justification which we give in support of our statement. To use a contemporary term, it is the difference between the 'logical space of causes' and the 'logical space of reasons'. Hume does not ever deny that there is a logical space of reasons. He can't do because all through the text he is giving reasons of one sort or another. His official theory, however, is that human rationality is ultimately explained by the theory of association of ideas.

The definition offered by Hume in terms of contiguity, priority and constant conjunction would, to use the contemporary jargon, be regarded as an attempt to give the 'truth conditions' of causal statements, or the 'necessary and sufficient conditions' for a causal statement to be true.

Why isn't that enough? If, and only if, it is true that A is contiguous with B, occurs prior to B and A and B are constantly conjoined (in the sense of a true universal generalization, not restricted in space or time) then A *is* the cause of B. There is nothing more to be said, on Hume's account. There is, of course the question of how we can ever know that A has caused B, and as you remark Hume gives a list of rules for 'judging causes and effects'. The upshot is that if ever A occurs without B then according to this definition of 'cause' we must assume there is something different in this case that distinguishes it from previous cases of A causing B that we have observed.

However, Hume still hasn't explained why it is that we *think* that there is more to causation than this, why we succumb to he philosophical prejudice or illusion that there exists some kind of intrinsic or necessary connection between A and B.

Hume's explanation, as you point out, follows his theory of ideas and impressions. However, it is not clear that he has succeeded in explaining how this extra factor arises.

If the first, 'philosophical' account is correct then applying causation to the mental realm, the second account of how an impression calls up a lively idea (i.e. the belief that B will follow, based on our perception of an event of type A) is exactly what you would expect. We are naturally disposed to form beliefs, based on our perception, and this natural disposition leads us to form true beliefs, at least enough of the time and provided that we are careful in applying the rules that Hume proposes.

But this still doesn't explain why we feel unsatisfied with the original definition. To parody Hume, looking into my mind I don't find any idea of 'necessary connection'. As a good Humean about causation, I believe that causal statements are, ultimately, true in virtue of universal laws and nothing more. Am I missing something?

So, your answer to the essay question will depend on how for you agree with this claim. In other words, in addition to giving Hume's explanation, the examiner would expect you to say something along the lines of (1) why you think *he* thinks that he needs to give the explanation that he gives, and (2) whether according to his own terms you think he has been successful.

All the best,

Geoffrey