Hume

We will be reading Hume's Enquiry, which he wrote after the failure of the Treatise of Human Nature (Hume had to review it himself and remarked that it fell dead-born from the press). The Treatise is a deeper, richer book, but it is not nearly as clear as the Enquiry. Many key passages of the Enquiry, moreover, are taken directly from the Treatise.

The Enquiry is an odd book, in that sections are introduced by rather literary, rather than straightforwardly philosophical, passages. Section I is a good example. In this section, Hume states his task as that of reconciling two distinct ways of doing philosophy. On the one hand, there is an easy and obvious (and therefore popular) philosophy full of lessons in virtue, as in Cicero. It treats of particular cases, primarily, and is concerned with human activity.

The second sort of philosophy, practiced by Aristotle and Locke, among others, is abstruse and accurate. It studies the human being as a reasoner rather than doer. Its issue is metaphysics, which yields the most general first principles of things. Unfortunately, it is not popular for several reasons. One is that the deep thinking it requires plunges people into melancholy. Hume himself was afflicted with this depression, which he found he could only relieve by socializing with his friends. Another is misunderstanding. Abstruse philosophy is seen as a cover for ignorance or superstition. The solution is to keep the spirit of accuracy but make it easy, not abstruse. Such was Hume's goal.

Like Locke and Berkeley before him, Hume undertook an "mental geography," mapping the mind and its operations. The inhabitants of the mind are ideas and impressions. The fundamental difference between them is said to be the degree of liveliness or vivacity, impressions being the more lively. Hume made the brash claim that this difference correlated with the kind of mental operation involved, so that thoughts are always duller than sensations. However, this is tricky business, since the imagination (say in a dream) yields ideas which are apparently more lively than, say, the impressions of sense had when one is drowsy.

Ideas are one and all copies of impressions. This can be seen by paying attention to the ideas one has, and recognizing the impressions which are their source. It can also be shown from the fact that a person born without the right sense organs never has the corresponding impressions, and always lacks the associated ideas. So there is a general maxim that all ideas are copies of impressions.

To this is opposed the notorious "missing shade of blue" example. A person might have had impressions of all but a single, very specific, shade of blue. Hume grants that he could have the idea of this shade without the corresponding impression, but that this singular case should not affect the general maxim. It is hard to see what Hume meant here, e.g., did he mean that the maxim is merely general or not universal?

There is a footnote about the innate ideas controversy (see the first lecture on Locke). Insofar as impressions are original, they can without harm be called "innate," but the question of their being inborn is "frivolous." Apparently, Hume thought that the idea of inborn ideas is too ludicrous to require refutation.

Having given the origin of ideas, Hume turned to ways in which they make their appearance in the mind. It is obvious that they occur systematically and not randomly, so Hume looked for patterns in their occurrence and attendant "principles of association" of ideas. Two relatively minor principles are those of resemblance and contiguity. In both cases, ideas we have remind us of others. The key principle is cause and effect.

In Section IV, the cause/effect principle is subject to a destructive skeptical critique. It begins with a division of the objects of inquiry into relations of ideas and matters of fact. In many cases, the mind is able to gain knowledge merely by examining ideas which it has before its view, as in mathematics. This knowledge, be it intuitive or demonstrative (cf. Locke), is certain because it is about the ideas alone. Whether or not a triangle exists, what I demonstrate of it holds good for the ideas which are the material and end of the demonstration. On this point, compare Locke, p. 94 (IV,IV,6).

Matters of fact, on the other hand, go beyond our ideas to the world. They are not intuitively or demonstrably certain. There is always the possibility of their falsehood, because we can always conceive one without the other. It is just as distinctly conceivable that the sun will not rise tomorrow than that it will, for example. Again on this point, compare Locke, p. 109 (IV,XI,9).

Now if we are to get knowledge of matters of fact beyond what we are presently perceiving or what we remember, we must look elsewhere, specifically, to the relation of cause and effect. But how do we gain knowledge that the relation holds?

Upon what are our beliefs about causal connections based? Two possibilities have generally been acknowledged: reason and experience. Some philosophers have claimed that we have some kind of rational insight into necessary connections between events, but Hume countered this claim by stating that the opposite of any outcome is possible. Suppose I believe that the moving cue ball will cause the eight-ball to move upon impact. Still, I can conceive that the motion of the cue ball stops and the eight-ball does not move in the slightest. My conception does not depend on their being anything unusual about the situation. Both balls are made of wood, the table surface is felt overlaying slate, the gravitational situation is that of the planet earth. Everything might be equal, but the outcome is different. Consider Adam, just after his creation. He would have no idea what the billiard balls would do.

Adam's problem is that he lacks experience. Even if the outcome could be different, experience shows that it is always the same. Thus when I observe a constant conjunction between exemplars of two types of events, I believe, on the observation of the first, that the second will follow. Nothing could be more reasonable than that, you will say.

But at this point Hume invoked the argument that made him famous. The belief is not based simply on the experience of the constant conjunction between the two. It is also based on the expectation that the patterns previously to repeat themselves will continue to repeat themselves. That is, everything is based on a belief that nature is uniform with respect to repetition. This conclusion seems innocuous until it is asked why we believe it is true.

Once again, the choices are reason or experience. And once again, reason is of no help, since we can conceive of the end of the uniformity of nature just as we can of its continuance. So we could only appeal to experience: nature has always proved to be uniform, so it shall continue to be so. But here the trap is sprung. Why do we believe that the pattern of repetition of patterns will repeat itself? Because it always has? But then we need to justify a still-broader belief, in a never-ending process. Our experience is only of use when we add to it a belief in uniformity, but experience cannot justify this belief.

Hume did not stop at this skeptical conclusion, however. For we do have the belief, whether it is justified or not. So he tried to explain where the belief comes from: it is formed as the result of custom and habit. The repeated observation of uniformity engenders a belief in uniformity. Of course, we can always conceive the opposite, as the above arguments claimed. Thus we need to find a difference between conception and belief. Hume located it in the superior force of the idea of the ensuing event when the idea of its constant companion is given.

Hume's positive statement of the nature of beliefs based on cause and effect was that it is an idea enlivened by the presence of another idea which it has been constantly conjoined in the past. This account can be generalized to cases in which the conjunction is not so constant. Suppose A has been conjoined to B 60% of the time and to C 40% of the time (B and C being incompatible). Then there will be a corresponding degree of liveliness of the ideas of B and C upon the occurrence of the idea of A. This is the basis of probability, which can be understood as degree of belief based on observed frequencies (a combination of two modern notions of probability).

So probability is just as much based on regularity as is cause and effect. In neither case is chance, a real randomness in the world, involved. What we call chance is simply lack of a discernible pattern, which results in a lack of belief. This ignorance can (at least in principle) be dispelled through a more thorough investigation of causes (in Hume's own sense of 'cause').

In a kind of reprise of his arguments against the rationality of causal inference, Hume next turned his attention to the idea of a necessary connection, which philosophers traditionally had placed at the basis of causality. Locke had held, for example, that the powers of bodies are necessary connections, and his skeptical claims about the limits of our knowledge of bodies stemmed from his pessimism about our ever knowing what these necessary connections are (IV,XI,9, p. 109).

The problem is that there is no impression of a necessary connection. Observe any two physical events, and only a succession will be found. The same holds for events in the mind. For example, Berkeley held (Sec. 28) that he had a notion of a power of his will to call up or discard ideas. But there is no impression of this power (especially given Berkeley's doctrine that ideas cannot resemble anything active), only a succession in the mind.

Hume advanced a number of specific considerations to undermine any claims to our knowledge of necessary connections, either in the mind's relation to the body or to its relation to its own ideas. The gist of these objections is that if we had knowledge of the power, we would know much more than we actually do. We have no knowledge of the mechanisms involved, why the mind is limited in the was it is, etc.

In fact, the production of anything by anything else takes on the character of a miracle. Malebranche had given up on Descartes' claims about the causal efficacy of bodies, both on one another and on the mind, giving it all over to God. But he claimed (echoing Berkeley) that this diminishes God, in the sense that God would be an inept creator if He had to intervene so that even the slightest thing could happen in the world. Indeed, recourse to God is rank and inexcusable speculation.

Another argument against Malebranche touches Berkeley's metaphysics. Even if we say that the necessary connection is between God and things, we still have no idea in what it consists. Berkeley had claimed that we can understand God's powers insofar as we understand our own. But Hume gives a way to turn the argument on its head: we do not understand how our minds have the power to affect ideas, and so we have none of how God's does.

The best course to take is that reluctantly adopted by Newton in his discussion of attraction (see above). Newton had held that attraction is a phenomenon in that bodies are observed to move toward one another in proportion to their masses and the square of their distance. This constant conjunction would qualify as a causal relation, on Hume's grounds. Newton's lament that he could not discover the means by which attraction takes place is misguided, given Hume's conclusion. Search as we will, we shall never find a power or active force.

Still, Hume had to explain why we so naturally conclude that there is a power operating when there are uniform correlations between events. He chalked it up to a habit of the mind. So we feel ourselves making a connection between ideas and this is "the sentiment or impression from which we form the idea of power or necessary connection."

The neatness of Hume's position may be merely apparent, however. Consider this passage from the Treatise (I,III,XIV) "Either we have no idea of necessity, or necessity is nothing but that determination of the thought to pass from causes to effects and from effects to causes, according to their experienced union." Now what exactly is a "determination" of a thought? In Section III, Hume had written of a "principle of connection," and in the Treatise he had explained them this way. "Here is a kind of attraction, which in the mental world will be found to have as extraordinary effects as in the natural, and to shew itself in as many and varied forms. Its effects are every where conspicuous; but as to its causes, they are mostly unknown, and must be resolved into original qualities of human nature, which I pretend not to explain" (I,I,V). It seems a bit too convenient of Hume to be able to appeal to this unknown "kind of attraction" when his targets, such as Locke, are not permitted this kind of appeal. Note that the passage from the Treatise did not reappear in the Enquiry.

The remainder of the Enquiry consists in applications of Hume's revisionary account of beliefs based on the relation of cause and effect. One obvious application is to the issue of human liberty. One must wonder about Hume's bold assertion that everyone agrees about this, and that the dispute is only about words, especially given his discussion of determinism in the last part of the section.

In its essential, Hume's doctrine was the same as Locke's, i.e., that liberty is lack of external constraint. One is at liberty to do what one desires so long as nothing hinders action intended to satisfy the desire. Acting freely may involve practical reasoning: one calculates the optimal course of action based on what one desires and the beliefs one has about courses of action that would most likely satisfy them. But the desires and beliefs themselves are not arrived at by reason.

Locke had held that liberty is the power to act or forbear to act in accordance with the directives of the will. Hume could not accept this statement of the matter at face value, since he reinterpreted the meaning of 'power.' It is not that we find within us some spring or cause of action, but only that there is a constant conjunction between the existence of certain desires (motives) and the carrying out of certain actions.

Understood this way, human action is no different from any other kind of action in nature. In both cases we find constant conjunctions which induce us to believe that future patterns of activity will resemble those past. In fact, if some action apparently deviates, we tend to be suspicious that there is some deception involved. A number of years ago, anthropologists claimed to have discovered a primitive tribe of people whose peaceful ways set them apart from all other known human societies. Many people immediately suspected fraud. As Hume was to argue in the next section ("Of Miracles"), the very uniformity of human behavior is itself an argument against any deviation from it.

Hume contended (one wonders how ingenuously) that everyone agrees about the uniformity of human behavior. So if they would just get clear about what a "cause" is, they would recognize that human causality is just a variety of natural causality. It is clear enough that people reject the claim as stated, but Hume just marks this down to a misunderstanding about the nature of a cause. If they think that human agency is the product of a will which is a power to initiate activity, then the criticism of Locke's view applies.

Descartes held that he could not find anything forcing his will in any direction, and concluded from this that his will is free. Hume turns this view on its head: one can find nothing because there is nothing to find. The will is not a power in the sense Descartes thought it was.

Another objection is that we actually feel a certain looseness in our willing. Compare the feeling we have when we observe a coin turning over in the air, it seems open as to which side it will come down on. But this kind of feeling is to no purpose if it concerns what will be; it only indicates ignorance. The so-called feeling of looseness is post hoc, or after the fact. I feel that I could have done things differently because I can conceive of a different outcome and associate it with an idea of myself. Still, I can only test whether this could be by doing something else. If I raised my arm once, I can keep it at my side now. Hume notes rightly that the trial would prove nothing. What motivates me to keep my arm down now is the desire to show that I am free, which is different from why I raised my arm in the first place.

Hume's view, like Locke's, is a "soft" form of determinism, in which liberty is a function of no external restraint on action, while the action itself follows inevitably from the earlier state of the agent. A natural objection is that the deterministic side presents an impediment to moral evaluation, to reward and punishment. How can someone be held accountable for something which is as predictable as the fall of a stone.

As usual, Hume turned the question on its head. How could someone be held accountable for actions that are not predictable? If they are not, they are probably not the outcome of any principle. Is the builder of Olson Hall to be held accountable for the collapse of the building when it is hit by a meteorite? Am I to be held responsible if I perform an act entirely out of character, e.g. if I had unknowingly ingested PCP? Attributions of praise and blame, administration of reward and punishment, both demand a constancy of character, and Hume's account of human agency gives constancy central location.

Still, there is a nagging question. If my character is determined entirely by prior conditions, e.g., my bad environment, am I to be held responsible for what I do "in character?" This question bedevils us in contemporary society, but it also posed a theological challenge. It seems that human actions are the outcome of the creation of the world itself, and thus, that the ultimate source of blame for bad actions would have to be God. We will discuss this issue in the next lecture.

Hume recognized that his forward-looking sense of freedom is threatened when one looks backward. That is, if my present actions are the repetitions of patterns that together make up my character, it may be that my character itself is not the outcome of any of my choices. Nowadays, we look to genetics or the environment as molding our characters without our co-operation, but in Hume's day the most burning issue was the role of God in all this.

In particular, if things are as they are due to divine decree, then it seems that God is ultimately responsible for everything that takes place. And if there is evil in the world, then it is brought about by God, which is an impiety. On the other hand, it seems absurd to deny that there is evil in the world.

Hume vigorously defended the absurdity of denying that there is evil in the world. He noted that some philosophers, most notably the ancient Stoics, have held that everything that happens is for the best. (Leibniz obviously falls into this category as well.) But Hume noted that this lofty philosophical thought is of no consolation to persons who are really suffering. We can apply Hume's theory of belief here, and recognize that the vivacity of suffering cannot be overcome by a pale, abstract thought, so once cannot really believe that there is not something terribly wrong when one is suffering.

Then what about the impiety, that God is the author of evil? Here Hume disappointingly retreats into his skeptical shell. Philosophers have disputed endlessly about the issue, to no avail. Some, for example, have held that evil is the work of creatures who were made with free will. God's role is that of creating the world in spite of the knowledge that it would be done, but this does not constitute God's actually being the author of evil. Hume recognized the futility of this kind of argumentation, pronouncing the whole issue a mystery. This conveniently gave him a way out of this version of the problem of determinism: God need not be said to be the author of evil. But one is left unsatisfied, especially wanting to know about the mundane form of determinism. If a pattern of constant conjunction between a combination of external environment and genetic coding could be discovered, so that all the actions of a person are perfectly predictable as a result, where is there room for liberty?

The section on the reason of animals does not require a great deal of comment. The general principle is that Hume's theory of our beliefs based on cause and effect is corroborated by the belief-forming mechanisms of animals. They do not reason abstractly, yet they form beliefs based on cause and effect in essentially the same manner we humans do, through habit formed in the face of constant conjunction. It should be noted that Hume afforded far higher status to non-human animals than did most philosophers of his time (Leibniz was a notable exception). Descartes' declaration that animals are robots represents the other side.

The section on miracles was an addition to Hume's corpus; it does not have a prototype in the Treatise. In class, it was asked whether Hume was in any danger given his attack on the justification of belief in miracles. The answer is that he was not. I think in part this is due to his close ties with powerful protectors, notably the ambassador to France, whom he served as secretary (reluctantly, he wrote, because he would have rather have been doing philosophy). He was also very guarded in the exact nature of his remarks and in what he published. The inflammatory Dialogues Concerning Natural Religion was not published in his lifetime. What nasty things Hume had to say about religion was largely confined to insulting Roman Catholicism.

In some ways, Locke, though asserting that there is a rational proof of God's existence, was more radical than Hume. For he had written that faith must not contradict reason. Hume's skepticism served him well (as in the argument about whether God is the author of evil just discussed). He did not have to hold faith up to a standard of rationality, since he had shown that reason confers no justification on any matter of fact. Further, by relegating religious matters to faith alone, Hume exploited an anti-rationalist undercurrent that was quite strong in medieval and reformation Christianity.

Returning to the subject at hand, belief in miracles, we see Hume employing his usual tactics. If we understand how probability really works, then we will see that there is always overwhelmingly strong evidence against the occurrence of any miracle. Recall that belief is a very vivid idea. There are degrees of vivacity short of belief as well. The disparity is accounted for by the relative diminution in vivacity brought about by contrary cases. There is "proof" only when a pattern of occurrences is utterly uncontradicted by contrary cases, and not otherwise.

The key to the argument about miracles is that there is a built-in array of cases contrary to any testimony about the occurrence of a miraculous event. Even a very credible witness to an event cannot overcome this onslaught. The problem is that a miracle is an event contrary to the laws of nature, and a law of nature is a pattern of occurrences for which there is a "proof," to which there are no contrary instances. So the very nature of a miraculous event unleashes a flood of evidence against its ever occurring. Washed away in the flood is the credibility of the person who testifies to it. As Hume stated, for such a one to be believed, it would have to be an even greater miracle that he has uttered a falsehood.

In Part II of the Section on Miracles, Hume sought to show in specific ways why testimony about the occurrence of a miracle has never credible enough to overcome its built-in disadvantage. The disadvantage, it will be recalled, is that the whole weight of experience, and hence probability, is against any exception to a law of nature, which a miracle would be. Although interesting, the section does not contain a great deal of philosophical content. There are many observations about human nature, which is prone to the propagation of incredible stories.

It may be worth noting how Hume tried to deal with the hard cases. In that of Cardinal de Retz, the evidence in favor was overwhelming, but the Cardinal concluded that the story carried its falsehood on its face. That is, he did not have to consider the evidence at all. The same holds for the miracles at the tomb of Abbé Paris, where the "absolute impossibility or miraculous nature of the events . . . in the eyes of all reasonable people will alone be regarded as a sufficient refutation."

A final tidbit in this section concerns Hume's artful concession to Christian faith, which he held is not dependent on the credibility of miracles. Despite the fact that it clashes with reason (in the form of reckoning probabilities), people believe in miracles nonetheless. This, Hume notes with supreme irony, is itself a continued miracle!

Section XI, "Of a Particular Providence and of a Future State," continues in the application of Hume's results to commonly-held beliefs. Here it is the nature of God that is at issue. Since we cannot experience anything not given to the senses, we can only infer from the intelligence and design of the world to an intelligent designer. But there are several problems with the inference from the outset.

First, we are entitled to infer no more in the cause than in found in the effect (the flip side of Descartes' principle that there must be as much in the cause as in the effect). In the Dialogues Concerning Natural Religion, Hume noted that the ill condition of the world is consistent with various hypotheses about its designer or designers: that it is a novice or someone with diminished skills, or that it is a committee that came up with an inferior compromise.

Second, we cannot draw any inferences from the nature of this cause other than what we had to put into it originally. So we cannot make any claims about any future acts of the deity, for example. "We cannot mount up to the cause to infer any new effects, beyond what is observed."

There is a discussion of evil in the world, but I will not examine it, in the interests of time. Nor will I comment on the final point, which states that causal inferences can only be to species, not individuals.

In Section XII, Hume discussed the various forms of skepticism, culminating in a description of the ways in which he himself was a skeptic. In Part I, he distinguished between antecedent and consequent skepticism. In the former case, we bring a skeptical attitude to our subject-matter before the investigation of knowledge begins. In the latter, we discover the infirmities of our faculties of knowledge ex post facto.

Descartes is celebrated for his antecedent skepticism, which was designed to keep him from affirming anything which might be wrong, hence purging his system most decisively of any false beliefs. Even if one could bring one's self to such a state of doubt (and Descartes admitted that he had his problems doing this), the method could not work, on Hume's view. For a thorough application of the method would dig us into a pit from which escape is impossible. If we question all our faculties, we have nothing left on which to rely to get us out of the state of hyperbolic doubt. Another form of antecedent skepticism is moderate, enjoining us to free ourselves from prejudice before beginning an investigation.

The consequent skepticism finds our mental faculties to be unsound in practice. The senses are the usual target here. Although we often misinterpret the deliverances of the senses, Hume noted, we can correct these by reason. This corrected use of the senses is the criterion of truth or falsehood for judgments concerning sensible objects. And we naturally place our faith in the senses.

Unfortunately, we also naturally place our faith in the view that the senses reveal things just as they are. That is, what is presented to the senses, images, is taken to be identical to the objects of the senses. This naive realist view is countered by the philosophical knowledge that there is a distinction between the fleeting perceptions of the mind and the uniform and independent things they are supposed to represent. We cannot tell whether the perceptions resemble the existences. And philosophy has no resources to get us out of this predicament. We cannot realistically rely on the veracity of a supreme being, as did Descartes, because we cannot prove that God exists.

Berkeley, of course, had a way out, but Hume dismissed him as a skeptic. The reason, I think, is that Hume thought that no one could actually believe what Berkeley proposed in the name of common sense. At most, we are made irresolute by what is essentially a skeptical argument. Instead of trying to find some rational way out of the mess, we should recognize that reason is helpless. We should rely on instinct instead. As we have seen, this is the heart of Hume's positive program.

Next Hume set his sights on abstract reasoning. Here, at least, should be an area immune from skeptical doubt. But there is a problem, i.e., the well-known paradoxes of the infinite. Geometry tells us that extension is infinitely divisible. But it also tells us that there are finite line segments. The paradox is in the relation between the two: how can a finite quantity be made up of an infinite number of parts? Leibniz had concluded from this problem that extension itself is only an idealization, and not something real. Hume's way out is to follow Berkeley and deny that we have abstract general ideas. The impressions and ideas of sense are only finitely divisible.

Hume's solution to the paradoxes of the infinite was to adopt Berkeley's view that there are no abstract general ideas. The objects of mathematics are only concrete impressions and ideas in the mind, and these are only finitely divisible. This is a kind of minimizing approach, denying to the science some of its objects in favor of others which are safe, which will not lead to paradox.

By comparison, consider Russell's paradox, which he discovered in the first decade of this century. The German logician Frege had given axioms for set theory, for which he had high hopes. Each of the axioms seemed intuitively to be true. One of them was that corresponding to every property is a set (axiom of comprehension). Russell noticed that the following property must therefore correspond to a set: the property a set has of not belonging to itself. So the set of people in the classroom does not contain itself; its only members are persons in the room.

Now consider the set of all sets that do not belong to themselves (the Russell set). This set has the following paradoxical characteristic: it belongs to itself if and only if it does not belong to itself. So apparently evident principles have plunged us into paradox. I think this situation parallels the problems with the infinite. One popular solution to Russell's paradox is the conservative use of axioms yielding the existence of sets. Weaker versions of comprehension will not allow the existence of the Russell set. This minimizing approach is similar to the strategy Hume used. (Note, however, that there is more compelling force to Russell's set theory than to the paradoxes of the infinite, since it is closer to pure logic than is geometry.)

Having shown the trouble reason can get into even regarding its proper subject-matter, relations of ideas, Hume reiterated the skeptical conclusion he had reached concerning the role of reason in beliefs about causes and effects. This result would have disastrous consequences if pressed to the extreme (as Hume's woeful "Conclusion of the Book" in the Treatise indicates. All would be destroyed, and we would have no motive to act at all. Thus Hume, like earlier skeptics, recognized that what is admitted in theory cannot be accepted in practice. Once we escape from the philosopher's attic, we look back with amusement at the infirmities we had found.

In the last part of the book, Hume finally recommended a version of skepticism, namely, academical or mitigated skepticism. It has as its enemy dogmatism, the unjustified assertion of judgments that go beyond matters of fact and relations of ideas. Mitigated skepticism pricks the pretensions of the arrogant intellectuals, revealing the emptiness of their metaphysical inquiries. Philosophy should stick to methodical description of common life, fallen from its lofty pretensions. Reason and demonstration can reveal only relations of ideas; matters of fact cannot be demonstrated. After canvassing some of the current areas of study, Hume concluded by suggesting that we commit to the flames any work which does not contain abstract reasoning concerning quantity or number or experiential reasoning concerning cause and effect.

Back to Lecture Menu

Eighteenth Century Philosophy Home Page