Building on the first session, this session goes on to explore the role of laws of nature to back up causal relations which underpin the idea of illness aetiology. If Hume’s arguments are broadly right and then causation is an essentially general notion. But Hume also highlights a different element: the counter-factual claim that if the cause event had not happened then the effect event would not have happened either. (There is an immediate objection to this which is that the effect event might have occurred to which we will return.) Thus we have two immediate questions to answer.
What is the connection between Hume’s first (and third) and his second definition of causation? What is the relation between the generality of causation and the counter-factual claim?
What kind of generality underpins causal connections?
The essential reading from the Textbook addresses both these questions.
JL Mackie set out an intuitive model of what we mean by saying that one event caused another event: the INUS condition. The event is an Insufficient but Necessary part of an Unnecessary but Sufficient condition. If that does capture what we mean by talk of something being the cause of an effect it addresses the worry about Hume’s second definition. In the context of the broad (unnecessary but sufficient) condition, had the cause-event not occurred then neither would the effect. the cause is necessary for the effect in that context. Thus it may also seem that Mackie’s account is based on Hume’s second definition. It is not, however, because Mackie thinks that such claims about necessity and sufficiency are underpinned by lawlike generalities. This prompts the second question, above: What kind of generality underpins causal connections?
The general problem is that some generalities seem to be accidentally true associations between variables where, intuitively, the first factor is not the cause of the second. But the problem is that post-Hume, we cannot simply say that if the appeal to generalities is supposed to explain what we mean by ‘cause’. We need to earn the right to distinguish between generalities which are genuine lawlike connections and mere accidentally true generalities (‘accidents’) without appealing to the notion that one thing makes another thing happen.
There is another clue to this second problem suggested by Hume’s second definition of cause. Laws of nature – such as f=ma – also support counter-factuals. Had a net force f been exerted on this mass m (although in fact it was not) then it would have accelerated at rate a. That serves as a test for what we take to be a law but merely postpones the question: why do laws but not accidents support counter-factuals?
Essential reading
Fulford, K.W.M., Thornton, T. and Graham, G. The Oxford Textbook of Philosophy and Psychiatry Oxford: Oxford University Press chapter 15 from p412 to the end of session 1 on p417 (including readings).
Further reading
Mackie, J.L. (1993) ‘Causes and conditions’ in Sosa, E. and Tooley, M. (eds.) Causation, Oxford: Oxford University Press: 33-50
Papineau, D. (1987) ‘Laws and accidents’ in MacDonald, G. and Wright, C. (eds.) Fact Science and Morality, Oxford: Oxford University Press:189-218
Yet further reading
There are other approaches to causation than the nomological (or lawlike) approach discussed here. Nancy Cartright discusses the connection between causation and raising the probability of an effect. David Lewis argues that Hume's second definition can be taken to be fundamental as long as one can think of the relative nearness of different possible worlds: different ways things might have been. These are discussed in the later section of the textbook but which argues, nevertheless, that a background system of laws is still necessary to pick out causes. You may not agree, of course.
Fulford, K.W.M., Thornton, T. and Graham, G. The Oxford Textbook of Philosophy and Psychiatry Oxford: Oxford University Press chapter 15 session 2 (including readings).
This session’s lecture slides are here.
Reflections on this session are here