Hume on Relations: Are They Real?

William James criticizes Hume for failing to adhere to the strictly empiricist method when he postulates discrete constituents of experience—which Hume calls perceptions—thereby making our experience a train of disconnected pieces. James argues that the discontinuity of experience in Hume results in part from his failure to recognize the immediate presence of relations in experience.¹ Emphasizing a continuity and unity of experience, James thus differentiates his empiricism from Hume's as being radical in the sense that it recognizes relations as 'real' parts of experience just as are things that are experienced to be so related.² This raises a question concerning the experiential status of relations in Hume: is James correct in accusing Hume of failing to notice the experiential reality of relations? Does Hume deny the experiential reality of relations entirely? Or does Hume downplay the experiential reality of relations and come to take them merely as products of the imagination, hence not 'real,' for certain theoretical reasons?

My main concern in this paper is with the status of relations in Hume, and in particular, the question of whether Hume does, and could, acknowledge any relations to be immediately present in our experience. I will argue that two different views on relations may be found in Book One of the Treatise of Human Nature. On the first view, some relations are immediately present along with related objects in the content of our experience, or in Hume's term, complex perceptions. On the second view, some relations are simply 'produced' or 'feigned' by the imagination in the mind. While there is textual evidence to show that Hume holds the first view, I will also argue that he ostensibly puts forth the second one when he carries out the theoretical analyses of perceptions taken as disconnected, independent atoms. Thus, James is correct to the extent that Hume denies the immediate presence of relations in the content of experience for theoretical reasons, but simultaneously wrong in saying that Hume fails to recognize it. It is Hume's metaphysical notion of perceptions as discrete and disconnected that forces him to deny, or at least downplay, the reality of relations in experience in the Treatise.

I Real Relations

For Hume, relations are complex ideas: he divides complex ideas into three kinds, 'Relations, Modes, and Substances' (T 1.1.4.7; SBN 13).³ In the next section, I. i.5, entitled 'Of relations,' Hume goes on to distinguish two types of relations: (1) natural relations in accordance with which ideas are associated with one another in the imagination; (2) philosophical relations which are judged to hold between ideas arbitrarily conjoined in the imagination. He writes:

The word RELATION is commonly used in two senses considerably different from each other. Either for that quality, by which two ideas are connected together in the imagination, and the one naturally introduces the other … or for that particular circumstance, in which, even upon the arbitrary union of two ideas in the fancy, we may think proper to compare them.

[T 1.1.5.1; SBN 13]

What does he mean by a relation being a 'quality' by which two ideas are associated and a 'particular circumstance' in which two ideas are arbitrarily united and compared? Is a relation an immediately given 'quality' in perception, like colors, tastes, smells, etc.? Is it immediately present in ideas united in the imagination so that it allows them to be 'compared'? Upon a close analysis, I argue that Hume presupposes—and sometimes asserts—that some relations are immediately present along with related objects in the content of perceptions, as well as between perceptions. But he also seems to restrict some relations to the domain of the imagination, seemingly suggesting that they are produced in the imagination and are thus not an empirically real part of perceptions. The first view, according to which relations are immediately present in perceptions, I will call the 'Real Relations' view, and the second view, I will call the 'Imagined Relations' view. While these two views are not consistent with each other if taken together for one and same...