To: Mark S.
From: Geoffrey Klempner
Subject: Hume's two definitions of causation
Date: 30th March 2010 11:56

Dear Mark,

Thank you for your email of 21 March, with your essay for the University of London Modern Philosophy: Descartes et. al. BA module, in response to the question, 'Why did Hume provide more than one definition of 'cause'? How do his definitions differ, and what is the significance of those differences?

I find myself disagreeing both with you and with Garrett on this question. I agree that it seems excessively strained to regard the two definitions of a cause -- an object followed by another 'where all the objects similar to the first are followed by objects similar to the second', and 'whose appearance always conveys the thought to that other' -- as 'extensionally equivalent'. But I am also not convinced by your explanation, according to which the two definitions are merely intended to give two 'incomplete' views of the same object, that of the 'ordinary person' and that of the 'philosopher'.

Hume's statement that 'it is impossible to give any just definition of cause except what is drawn from something extraneous and foreign to it', looks, superficially, like an admission that there is something down there, a secret connection which philosophy cannot describe other than indirectly. But this is clearly wrong. There's nothing 'down there'. Lawlike connection is all there is, there is nothing else to 'causation'.

Except for the fact, of course, that human beings are gifted with the ability to discern at least some of the lawlike connections that obtain in nature. That is because of our natural constitution, and in particular the lawlike generalizations which describe the workings of the human mind.

I find this an incredibly deep idea. To appreciate its significance it is necessary to go back to Descartes, and the evil demon scenario. Descartes is not only interested in the extreme case where the evil demon gives me 'experiences' of a world which does not really exist. The question of the possibility of scientific knowledge crucially turns on the human capacity to form causal beliefs -- to infer explanations -- and the seemingly miraculous correspondence between the way the world works and the way our minds work in figuring out how the world works.

This is a miracle, for Descartes, because the mind is not part of the natural world. There are no limits on how a soul can generate thoughts within itself because souls don't have 'workings', they are simple and indivisible, not made of 'parts' in the way that the objects of our perception and judgement are. The solution has to be that 'God is no deceiver'. Souls are designed to form true representations of the world, provided that human beings exercise their powers of judgement responsibly.

With Hume all that is swept away. Human beings are part of the natural world, and the human mind obeys natural laws -- laws of cause and effect -- just as the objects of its perception do.

If you look at the problem of defining cause from this point of view, it is clear that there is no alternative for Hume but to deny the 'secret connection'. God is out of the picture. All we have are the principles of the association of ideas to go on.

It would be legitimate to say that in giving his two definitions, Hume has one eye on the philosophical 'illusion' concerning the nature of causation which he is rejecting. The second definition forms part of a naturalistic explanation of the aetiology of that illusion. However, if this is all Hume wanted to say then it would be curious that he gives the two definitions in the way that he gives them.

I am not sure about your 'huge pile of objects' analogy. What is real, are the lawlike connections that render science possible. We can leave aside for now the question which I've raised earlier concerning how 'realist' we should be about the 'truth' of these lawlike statements. 'Cause' is difficult to define because of the inherent vagueness in the idea of 'similarity' -- or, what amounts to the same thing, the impossibility of fully defining 'ceteris paribus' conditions. The first definition covers all the 'causes' there are, nothing is left out, but even if per impossibile we knew all the laws of nature, there would still be a gap between that knowledge, and the knowledge of all the causes in the universe, because of the inherent fuzziness in the definition of a cause.

The second definition explains why we need that inherently fuzzy notion. Human beings are in the world, interacting with it on a piecemeal basis, so what comes first in terms of knowledge is particular causal statements rather than general laws.

Perhaps this is what you were trying to get at in relating the two definitions of cause to what Hume says about 'natural' and 'philosophical' relations. There is something important here to explain, namely, how it is that through the medium of language human beings have been able to extend the capacity for causal reasoning with which they are naturally endowed. Hume's rules for judging causes and effects is the first step in that explanation. Nature provides us with the capacity to develop language (or 'form ideas'), which in turn expands our powers of explanation. That was the main thing that interested Locke. But Hume saw far deeper into the problem of how it is that human beings are able to acquire knowledge of the external world.

All the best,

Geoffrey