Helen Beebee

Hume is traditionally credited with inventing the ‘regularity theory’ of causation, according to which the causal relation between two events consists merely in the fact that events of the first kind are always followed by events of the second kind. Hume is also traditionally credited with two other, hugely influential positions: the view that the world appears to us as a world of unconnected events, and inductive scepticism: the view that the ‘problem of induction’, the problem of providing a justification for inference from observed to unobserved regularities, is insoluble. Hume on Causation is the first major work dedicated to Hume’s views on causation in over fifteen years, and it argues that Hume does not subscribe to any of these three views. It places Hume’s interest in causation within the context of his theory of the mind and his theory of causal reasoning, arguing that Hume’s conception of causation derives from his conception of the nature of the inference from causes to effects.

According to ‘regularity theories’ of causation, the obtaining of causal relations depends on no more than the obtaining of certain kinds of regularity. Regularity theorists are thus anti-realists about necessary connections in nature. Regularity theories of one form or another have constituted the dominant view in analytic Philosophy for a long time, but have recently come in for some robust criticism, notably from Galen Strawson. Strawson’s criticisms are natural criticisms to make, but have not so far provoked much response from regularity theorists. The paper considers and rebuts Strawson’s objections. For example, Strawson claims that if there were no necessary connections in nature, we ought continually to find the regularity of the Universe surprising. I argue that the fact that the Universe is regular is something we take ourselves (fallibly) to know, and hence, in the light of this knowledge, its continued orderliness is not at all surprising.

Singularists about causation often claim that we can have experiences as of causation. This paper argues that regularity theorists need not deny that claim; hence the possibility of causal experience is no objection to regularity theories of causation. The fact that, according to a regularity theorist, causal experience requires background theory does not provide grounds for denying that it is genuine experience. The regularity theorist need not even deny that non-inferential perceptual knowledge of causation is possible, despite the fact that such knowledge would sometimes allow us to make inferences about what happens in far-off places and times

Scepticism concerning the idea of causation being linked to contingent chance-raising is articulated in Beebee’s challenging chapter. She suggests that none of these approaches will avoid the consequence that spraying defoliant on a weed is a cause of the weed’s subsequent health. We will always be able to abstract away enough of the healthy plant processes so all that’s left is the causal chain involving defoliation and health. In those circumstances, there will be contingent chance-raising. Beebee’s conclusion is that we should reject the idea of contingent chance-raising and just accept that all causation involves chance-raising. This involves the reclassification of some intuitive cases of causation as causal processes without causation but rather hindering (a distinctive kind of process). It seems clear from this discussion and from the brief earlier remarks about the status of prevention that the classification of types of causal processes and the characterization of their link to causation are matters of some importance.

Hume's two definitions of causation have caused an extraordinary amount of controversy. The starting point for the controversy is the fact, well known to most philosophy undergraduates, that the two definitions aren't even extensionally equivalent, let alone semantically equivalent. So how can they both be definitions? One response to this problem has been to argue that Hume intends only the first as a genuine definition—an interpretation that delivers a straightforward regularity interpretation of Hume on causation. By many commentators' lights, however, this is a bug rather than a feature: such an account of the two definitions leaves necessary connection out of Hume's story about the meaning of "cause" ..