History of Western Philosophy - Bertrand Russell [381]
Hume is commonly accused of having too atomic a view of perception, but he allows that certain relations can be perceived. 'We ought not,' he says, 'to receive as reasoning any of the observations we make concerning identity, and the relations of time and place; since in none of them the mind can go beyond what is immediately present to the senses.' Causation, he says, is different in that it takes us beyond the impressions of our senses, and informs us of unperceived existences. As an argument, this seems invalid. We believe in many relations of time and place which we cannot perceive: we think that time extends backwards and forwards, and space beyond the walls of our room. Hume's real argument is that, while we sometimes perceive relations of time and place, we never perceive causal relations, which must therefore, if admitted, be inferred from relations that can be perceived. The controversy is thus reduced to one of empirical fact: Do we, or do we not, sometimes perceive a relation which can be called causal? Hume says no, his adversaries say yes, and it is not easy to see how evidence can be produced by either side.
I think perhaps the strongest argument on Hume's side is to be derived from the character of causal laws in physics. It appears that simple rules of the form 'A causes B' are never to be admitted in science, except as crude suggestions in early stages. The causal laws by which such simple rules are replaced in well-developed sciences are so complex that no one can suppose them given in perception; they are all, obviously, elaborate inferences from the observed course of nature. I am leaving out of account modern quantum theory, which reinforces the above conclusion. So far as the physical sciences are concerned, Hume is wholly in the right; such propositions as 'A causes B' are never to be accepted, and our inclination to accept them is to be explained by the laws of habit and association. These laws themselves, in their accurate form, will be elaborate statements as to nervous tissue—primarily its physiology, then its chemistry, and ultimately its physics.
The opponent of Hume, however, even if he admits the whole of what has just been said about the physical sciences, may not yet admit himself decisively defeated. He may say that in psychology we have cases where a causal relation can be perceived. The whole conception of cause is probably derived from volition, and it may be said that we can perceive a relation, between a volition and the consequent act, which is something more than invariable sequence. The same might be said of the relation between a sudden pain and a cry. Such views, however, are rendered very difficult by physiology. Between the will to move my arm and the consequent movement there is a long chain of causal intermediaries consisting of processes in the nerves and muscles. We perceive only the end terms of this process, the volition and the movement, and if we think we see a direct causal connection between these we are mistaken. This argument is not conclusive on the general question, but it shows that it is rash to suppose that we perceive causal relations when we think we do. The balance, therefore, is in favour of Hume's view that there is nothing in cause except invariable succession. The evidence, however, is not so conclusive as Hume supposed.
Hume is not content with reducing the evidence of a causal connection to experience of frequent conjunction; he proceeds to argue that such experience does not justify the expectation of similar conjunctions in the future. For example: when (to repeat a former illustration) I see an apple, past experience makes me expect that it will taste like an apple, and not like roast beef; but there is no rational justification for this expectation. If there were such a justification, it would have to proceed from the principle 'that those instances, of which we have had no experience, resemble those of which