In Part III of Book I, entitled "Of Knowledge and Probability," Hume proceeds in a rather curious and idiosyncratic manner that will strike modern readers especially as very odd. Later on, we shall see that when he came to write the Enquiry Concerning the Human Understanding, he shifted to a mode of analysis that is much more in keeping with the discussions of analytic and synthetic judgments that were common in the early and middle twentieth century, and which derive in large measure from Kant. Indeed, this change is one of the two principal ways in which the doctrine of the Enquiry differs from that of the Treatise.
Let us keep in mind that Hume has only three elements to his analysis of "human nature": the diversity of perceptions that present themselves in consciousness; relations of one sort or another among those perceptions; and the habits and propensities of the mind to do something or other with those perceptions. Everything must be explained in terms of these three elements. Part I contains Hume's account of the perceptions [impressions and ideas, simple and complex, impressions of sensation and impressions of reflexion] and a preliminary description of the "gentle force" of association. Part III now examines the relations that perceptions bear to one another.
Hume identifies seven kinds of what he calls "philosophical relation," namely: resemblance, identity, relations of time and place, proportions in quantity or number, degrees in any quality, contrariety, and causation. He now divides these into those relations that "depend entirely on the ideas," and those that can be changed by a change in the order or manner in which they are presented to the mind. According to Hume, resemblance, contrariety, degrees in quality, and proportions in quantity or number are of the first sort. Regardless of the order or manner in which ideas present themselves to the mind, these four relations remain unaltered. Hence, Hume says, they and they alone can yield genuine knowledge. The other three, namely identity, relations of time and place, and most especially causation, can be altered, without any change in the ideas themselves, merely by the order or manner in which they appear to us.
This is [at least to me] an unexpected way of classifying relations, and it is worth taking a moment to be clear about what Hume is saying. Whether two ideas resemble one another [say, two shades of a color, or two human faces] is, he says, purely a function of the unalterable characteristics of the ideas, and is immediately apparent to us without any process of reasoning into which we could introduce doubt. Either two ideas strike us as resembling or they do not, regardless of whether A appears before or after B, or appears near to or far from B, or appears quite contemporaneously to B or separated by a lengthy period of time. The same is true for the relation of contrariety [which seems to be simply the negative of resemblance], and also relations of quantity and quality. Hume has no quarrel with anyone who wishes to claim that judgments of resemblance, contrariety, quality, and quantity can yield certainty and hence genuine knowledge, and almost immediately he leaves them behind, never to trouble with them again. It is the other three -- identity, time and place, and causation, and especially identity and causation, on which he will spend the next 200 pages.
As many modern readers of the Treatise have pointed out, Hume is in fact wrong about the group of four relations that are supposedly invariant under changes in the manner of presentation. All of us these days are familiar, as Hume was not, with so-called gestalt phenomena. Surround two identically matching dots of color with expanses of differing colors, and they no longer look identical. Draw a pair of manifestly parallel lines, append to them rows of little slanting appendages tilted in different directions, and they cease to look parallel. These facts do not really very much call Hume's arguments about identity and causation into question, but they are interesting, and suggest that the facts of perception are more complex than Hume supposed.
Hume now turns his attention to the idea of cause and effect, and immediately things get very interesting indeed. He begins [Section II] by noting something that is central to his epistemology. Of all seven relations, only causation carries me beyond the evidence of the senses, to objects and events not now present to me. To remain with his examples, I feel a great heat, and infer the existence of a fire, even though I do not now see a fire. I hear the sound of footsteps, and infer that someone is climbing the stairs to my apartment, even though no person is at this moment before my eyes. Hume is poignantly aware of the fact that our actual sense presentations are a thin and radically incomplete fragment of the world in which we imagine ourselves to be living. We are trapped in the subjectivity of consciousness, and only causal inference carries us beyond immediate presentations, fleshing them out with people and objects, with a world. The same is true with regard to time. What I perceive at any moment is a tiny slice of the temporal flow of perceptions that taken all together constitute my consciousness. Only causal inference enables us to infer past events from present perceptions. Hume's starting point is what came in later centuries to be called a "solipsism of the present moment."
For this reason, it is for Hume a matter of the greatest urgency to examine causal inference and determine two things: First, whether causal inference can be justified by any process of reasoning; and Second, if it cannot, how and why it is that despite the lack of the slightest rational justification, we firmly believe the causal inferences we make, and ground our entire lives upon them.
With the examination of the idea of causal connexion, Hume embarks on the core reasoning of Book I of the Treatise. For the first time, we see the enormous power of the elementary claim, advanced at the very start of Book I, that all of our ideas are copies of preceding impressions. If we are to understand the idea of causation, he says, we must "trac[e] it up to its origin, and examin[e] that primary impression from which it arises." This is the defining rhetorical and analytical move of the British Empiricist school, brought to its highest pitch by Hume. "Hunt the impression!" might be the motto inscribed on the banner under which Hume rides into battle.
When we cast about for examples of causes and effects -- or rather, for examples of what are commonly considered causes and effects -- we find that the idea cause does not arise from the properties of any particular object or event. Rather, it arises, Hume says, from pairs of resembling instances of objects or events that exhibit two characteristics: they are spatially contiguous and they are temporally continuous. Hume is here echoing the common scientific conviction of his day that there cannot be what was called "action at a distance." If one event causes a second, then the first must occur immediately before the second and in immediate proximity to it. If there is the least lapse of time between cause and effect, or the least spatial gap between the two, then we may be sure that there is some intervening event that is the effect of the first and the cause of the second.
But contiguity and continuity alone are not sufficient to yield the idea of causal connexion, for post hoc ergo propter hoc is a fallacy of reasoning, not a foundational principle. There is one more essential component of the idea of causation: necessity of connexion. [I realize that I have fallen into the habit of using Hume's spelling -- you will have to be patient with me, as I wander down memory lane.] Hume now poses two questions, quite different from one another, which must both be answered before we are done with the matter of causation:
"First, for what reason we pronounce it necessary, that every thing whose existence has a beginning, shou'd also have a cause? Secondly, Why we conclude that such particular causes must necessarily have such particular effects; and what is the nature of the inference we draw from the one to the other, and of the belief we repose in it?"
Tomorrow, we shall examine the most famous and consequential sceptical argument in the history of philosophy.