Why We Believe in Induction: Standards of Taste and Hume's Two Definitions of Causation

Why We Believe in Induction: Standards of Taste and Hume's Two Definitions of Causation Bennett W. Helm A major puzzle in the literature on Hume's Treatise ofHuman Nature is that of why he gives two definitions of causation and what the relation between the two is. Given the amount ofsecondary Uterature on the subject, it is somewhat striking that two interrelated elements of Hume's account have received so little attention. The first is the distinction of causation into the natural and the philosophical relations: although many have tried to give accounts of why Hume presents two definitions of causation, it is often not clear in these accounts that the one definition is of causation as a natural relation and the other is of causation as a philosophical relation.1 Where the distinction is taken into consideration, it is usually misunderstood or taken as "an obfuscatory complication" that is better left out.2 The second element is that in making causal inferences "we must follow our taste and sentiment," 3 where the appeal to taste here, as in morality and aesthetics, does not imply that each person's taste is as good as everyone else's. Rather, Hume makes it clear both in book 3 of the Treatise and in his essay, "Of the Standard of Taste," that there are standards to which one's judgements of taste (in moraUty, aesthetics, or causation) must conform. My purpose in this paper is to attempt to provide an interpretation of Hume's account of causation that brings these two elements explicitlyinto the foreground. This interpretation is, to a greater extent than usual, a reconstruction of Hume's account of causal inference, drawing, as I have indicated, from a range of texts not normally associated with his discussion of causation. As such, it should be considered as exploratory in nature, perhaps focusing too single-mindedly on these two elements in an attempt to make out as strong a case as possible for their relevance in understanding Hume's account of causal inference. To begin, recall Hume's two definitions ofthe relation of causation: 1. We may define a cause to be 'An object precedent and contiguous to another, and where all the objects resembling the former are plac'd in Uke relations of Volume XLX Number 1 117 BENNETT W. HELM precedency and contiguity to those objects, that resemble the latter*. 2. ? CAUSE is an object precedent and contiguous to another, and so united with it, that the idea of the one déterminée the mind to form the idea of the other, and the impression ofthe one to form a more Uvely idea ofthe other'. (T 170) How do these two definitions fit together? Are they two parts of one definition, or are they two distinct definitions that are to be used for distinct purposes? From the immediate context, it really is not clear how we are to take them, and different things Hume says seem to provide evidence for different interpretations. Thus, on the one hand, Hume says that the two definitions "are only different, by their presenting a different view ofthe same object" (T 170), which seems to indicate that the two are interconnected aspects of a single definition and so cannotbe understood independently. That the idea ofnecessity, which Hume clearly takes to be crucial to an understanding of causation (T 77), seems to be entirely missing from the first definition might be taken as support for this reading. But, on the other hand, Hume also seems to offer each definition as one that can stand alone, though with defects that may cause it to be "rejected" and the other "substituted" for it. Thus, Hume says immediately following the first definition, Tfthis definition be esteem'd defective, because drawn from objects foreign to the cause, we may substitute this other definition in its place,"andproceedstogive the second definition(whichis,however, similarly defective). This suggests that, although neither definition is perfect, they are the closest we can come to an adequate formulation of the relation of causation, and each definition has its own domain of application within which it is most appropriate. Although these twointerpretations seem tobe in...