David Hume

In brief

David Hume (1711 to 1776) took the empiricist principle that all knowledge comes from experience and followed it, without flinching, to conclusions that unsettled everything built on it. If every idea must trace back to some impression of the senses, he asked, then where is the impression of causal power? We see the first billiard ball strike the second and the second move, but we never see a force passing between them, only one event followed by another, over and over. Causation, Hume argued, is not something we observe in the world; it is a habit of expectation the mind forms after repeated pairings. From the same principle he drew three more results that have never stopped biting: that no argument can justify our confidence that the future will resemble the past (the problem of induction); that when we look inward for the self we find only a bundle of fleeting perceptions and never the "I" that supposedly has them; and that no conclusion about what one ought to do can be validly deduced from premises about what merely is. He is often called the greatest philosopher to write in English, and the most destructive. His skepticism did not lead him to despair but to a mild, sociable acceptance: reason cannot ground our everyday beliefs, yet nature makes us believe anyway, and we get on.

The life

Hume was born in Edinburgh in 1711 into a family of modest Scottish gentry and grew up at Ninewells, the family estate in the Borders. His father died when he was a child. He entered the University of Edinburgh around the age of ten or eleven, as was then usual, and left without a degree. His family intended him for the law; he found it repellent and instead read philosophy and literature obsessively. In his late teens he suffered a mental and physical breakdown, brought on, by his own later account, by the intensity of his studies and the strain of trying to think his way through the whole of human understanding at once.

He spent the years 1734 to 1737 in France, mostly at La Fleche, the town where Descartes had been schooled, and there, still in his twenties, wrote A Treatise of Human Nature. He published it anonymously in 1739 and 1740 with enormous hopes and famous disappointment: it "fell dead-born from the press," he wrote, attracting little notice and less understanding. He blamed the presentation rather than the ideas, and spent much of the rest of his career recasting the same arguments in more inviting form, first as the Enquiry Concerning Human Understanding (1748) and then the Enquiry Concerning the Principles of Morals (1751), the latter of which he judged his best work.

He never held a university chair. He was twice passed over, at Edinburgh in 1745 and Glasgow later, blocked by clergy who regarded him, correctly, as a religious skeptic and possibly an atheist. He supported himself as a tutor, a diplomat's secretary, a librarian to the Faculty of Advocates in Edinburgh (which gave him the materials for his hugely successful multivolume History of England), and later as an embassy official in Paris, where he was lionized by the salons and the philosophes. He was for a time a friend of Jean-Jacques Rousseau and brought him to England, before the friendship collapsed into public quarrel. His closest and steadiest friend was the economist Adam Smith. Hume was famous in his own lifetime chiefly as a historian and essayist; his philosophy was recognized as epochal only later. He died in Edinburgh in 1776, cheerful and composed to the end, an equanimity in the face of death that scandalized the pious and moved James Boswell, who came to watch for signs of a deathbed conversion and found none. His most direct assault on religion, the Dialogues Concerning Natural Religion, he prudently left for publication until after he was gone.

The full treatment

The empiricist starting point: impressions and ideas

Hume inherited from John Locke and George Berkeley the view that the mind begins empty and is furnished by experience. He sharpened it into a tool. All the contents of the mind, he held, are "perceptions," which divide into impressions (the vivid, immediate deliverances of sensation and feeling) and ideas (the fainter copies of impressions that memory and imagination work with). His governing rule: every simple idea is a copy of some prior impression. This gives him a test. Confronted with a supposed idea, ask which impression it derives from. If none can be produced, the idea is empty, a word with no genuine content behind it. Hume turns this test on the grandest concepts in philosophy, and several of them fail.

Causation as constant conjunction

The most consequential target is causation. We speak constantly of one thing causing another, and we treat this as knowledge of a real connection, a power in the cause that produces the effect. Hume asks for the impression. Watch as closely as you like: you see the cue strike the ball and the ball roll; you do not see any causal power, any necessary tie, any "must." You see only three things, contiguity in space, succession in time, and constant conjunction (this kind of event has always, in your experience, been followed by that kind). The idea of necessary connection, the "glue" that seems to hold cause to effect, corresponds to no impression in the objects. Where does it come from? From the mind's own habit. After the eye has seen many pairings, the mind forms a tendency to expect the second when it meets the first, and it projects this internal feeling of expectation outward, mistaking a fact about itself for a feature of the world. Causation, so far as we can know it, is regular succession plus a habit of inference. This is the argument that separates the mere observation that two things go together from any claim that one produces the other, the distinction a later entry on correlation and causation turns into a working rule.

The problem of induction

The causation argument opens straight into a deeper one. All our reasoning about matters of fact beyond present observation, that bread will nourish tomorrow as it did today, that the sun will rise, that fire will burn, rests on the assumption that the future will resemble the past, that nature is uniform. Hume asks what justifies that assumption. Not logic: there is no contradiction in supposing the course of nature might change, so the uniformity of nature cannot be proved as a relation of ideas. Not experience: any appeal to the fact that nature has been uniform so far, and therefore will continue to be, already assumes the very principle in question, that the past is a guide to the future. The argument is circular. So the belief that grounds every prediction, every inductive inference, every expectation we act on, has no rational foundation. We hold it not because we have reasons but because custom makes it irresistible. This is the problem of induction, and it remains, by wide agreement, unsolved.

The bundle theory of the self

Hume turns the impression-test inward. We assume there is a self, a continuous, unified "I" that persists through change and owns all our experiences. Hume looks for the impression of it. When he enters most intimately into what he calls himself, he reports, he always stumbles on some particular perception (a sensation, a feeling, a thought) and never catches himself without a perception, never observes the bare self that supposedly underlies them all. What we call the mind is a bundle or collection of perceptions in perpetual flux, succeeding one another with inconceivable rapidity, like a theatre across which images pass. There is no stage, only the passing scenes. Personal identity, then, is not something we perceive but something the imagination confects out of the resemblance and causal linkage among perceptions. Hume was uneasy with this conclusion, and admitted in an appendix to the Treatise that he could not fully make it work; but the bundle theory has been a live option in the philosophy of mind ever since, and it strikingly anticipates the Buddhist doctrine of no-self, arrived at independently.

The is-ought gap and morality from sentiment

In Book III of the Treatise, Hume records that writers on morality reason along in the ordinary way, establishing facts about God or human affairs connected by "is," and then, imperceptibly, the connective becomes "ought." Since "ought" expresses a new relation, he insists a reason should be given for a deduction that seems altogether inconceivable, how this new relation can follow from others entirely different from it. This single paragraph became the is-ought problem, treated at length in its own entry. Its positive counterpart is Hume's own moral theory: since reason only tells us what is, and morality moves us to act, morality cannot be founded on reason alone. We approve of benevolence and condemn cruelty not because we have deduced their wrongness but because of a feeling, a sentiment of approval or disgust rooted in sympathy. Hence his most quoted and most misread line: "Reason is, and ought only to be the slave of the passions." He does not mean we should be irrational. He means reason is instrumental: it can find the means to our ends and correct our beliefs about facts, but it cannot by itself set an end or move the will. The passions supply the ends; reason serves them.

The critique of miracles

In Section X of the first Enquiry, Hume gives a rule for weighing testimony to the miraculous. A miracle is by definition a violation of a law of nature, and a law of nature is established by uniform, exceptionless experience. So the evidence against any miracle is the maximum possible: everything we have ever observed. The evidence for it is the testimony of witnesses, which is always fallible, and which we know from experience is corrupted by wonder, by religious zeal, by the human love of surprise and the tendency of miracle reports to flourish among the credulous. Hume's principle: no testimony is sufficient to establish a miracle unless the falsehood of the testimony would be more miraculous than the event it reports. In practice, he argues, that bar is essentially never cleared, because it is always more probable that witnesses are mistaken or lying than that a law of nature was suspended. He does not claim to prove miracles impossible; he argues that no reasonable person could ever be justified, on testimony, in believing one occurred.

Lineage

Hume stands at the end of the classical British empiricist line: Locke (1632 to 1704), who grounded all ideas in experience; Berkeley (1685 to 1753), who used empiricism to dissolve material substance; and Hume, who used it to dissolve causation, the self, and the rational credentials of induction. Behind him lay the Newtonian ambition to find, for the mind, laws as exact as Newton had found for bodies; Hume subtitled the Treatise "an attempt to introduce the experimental method of reasoning into moral subjects." In ethics he descends from the moral-sense school of Francis Hutcheson (1694 to 1746), against the rationalists Samuel Clarke and William Wollaston. He belongs to the Scottish Enlightenment, alongside his friend Adam Smith, whose theory of moral sympathy is deeply Humean.

His largest downstream effect was on Immanuel Kant (1724 to 1804). Kant recorded in the Prolegomena (1783) that it was Hume's treatment of causation that "interrupted my dogmatic slumber and gave my investigations in the field of speculative philosophy a quite new direction." Kant accepted that Hume was right that causal necessity cannot be found in experience, and refused to accept the skeptical conclusion; his response, that causation and other categories are contributed by the mind as the very conditions of possible experience, is the hinge of the Critique of Pure Reason. In the twentieth century Hume was claimed as an ancestor by the logical positivists and by Karl Popper, whose falsificationism is an attempt to live with the problem of induction by giving up on justifying it.

The strongest case for it

Hume's power is that he does not smuggle in premises; he takes the empiricist principle everyone in his tradition already accepted and refuses to stop applying it when it becomes uncomfortable. Grant that all knowledge of matters of fact comes from experience, and the results follow: you cannot experience necessary connection, you cannot experience the uniformity of the future, you cannot experience the bare self. The arguments are not tricks. Three centuries of effort have not produced an agreed refutation of the induction problem, and the demand that anyone moving from is to ought show their work has become basic intellectual hygiene. His analysis of causation cleared the ground for the modern regularity and probabilistic theories of causation and for the statistical maxim that correlation is not causation. And his temperament is a model: having reasoned his way to conclusions that make ordinary life look groundless, he does not pretend to disbelieve what he cannot help believing. Nature, he says, is too strong for skepticism; we return from the study to backgammon and dinner, believe in bodies and causes and selves, and the philosophy earns its keep by teaching modesty about how much of this we can defend.

The strongest case against it

The most influential answer is Kant's: Hume is right that causal necessity cannot be read off experience, and wrong to conclude that it is therefore merely a habit. On Kant's account, causation is a category the understanding brings to experience, without which no coherent experience of an objective world would be possible at all; so the principle is not derived from experience but is a condition of it, and is secure in a way Hume never allowed. This concedes Hume's negative point and reverses its force.

On the self, Hume's own appendix confesses failure, and critics from Thomas Reid (1710 to 1796) onward pressed the obvious rejoinder: the very act of noticing that perceptions succeed one another presupposes a single subject present through the succession to notice it; a bundle cannot be aware of itself as a bundle. Reid, founder of the Scottish school of common sense, argued more broadly that Hume's whole "way of ideas," the assumption that we perceive only our own perceptions rather than the world, is a philosopher's mistake that generates its own skepticism, and that the beliefs Hume declares unjustified (in an external world, in causes, in the self) are original, rational endowments of the mind that no argument can or should overturn.

On induction, the problem is conceded rather than solved, but many hold that Hume set an impossible standard. Reliabilists argue that a belief-forming method can be justified by being reliable, whether or not we can produce a non-circular argument for its reliability, so induction may be justified even if Hume's demand for a proof cannot be met. Peter Strawson (1919 to 2006) argued that to ask whether induction as such is rational is a confusion, because being reasonable about matters of fact just means proportioning belief to evidence in the inductive way; there is no external standpoint from which to grade the standard itself.

On morality, defenders of moral realism reject the claim that value rests on sentiment. If moral judgments merely express feelings, they ask, why do we reason about them, correct them, and hold that some are mistaken? And on miracles, critics from Richard Price (1723 to 1791) to modern probability theorists have argued that Hume's argument, treated as a formal claim about weighing evidence, is too strong: consistently applied, it would forbid belief in any sufficiently rare event on testimony, including well-attested ones, and it appears to assume the falsity of the miraculous in setting up the balance of probabilities.

Where it stands now

Hume is more central to living philosophy than almost any figure of his century. The problem of induction is still taught as unsolved and still shapes the philosophy of science, from Popper's attempt to dispense with induction to the ongoing work on confirmation and the "new riddle" posed by Nelson Goodman in 1955. The is-ought gap is standard equipment in metaethics and in the criticism of arguments that read duties off nature, science, or custom. The bundle theory is a serious option in debates over personal identity, and Derek Parfit's influential reductionist view of the self is recognizably in its line. Regularity and probabilistic theories of causation trace to Hume, and his instrumental picture of reason underwrites much of decision theory and the "reason is the slave of the passions" strand in moral psychology. His empiricism is no longer accepted whole, and Kant's response set the terms for two centuries of reply, but the pattern holds: philosophers define their positions by how they answer Hume. That is the surest sign of a thinker who changed the questions.

Test yourself

Pick a belief you act on without hesitation: that dropping this cup will break it, that you are the same person who woke up this morning, that tomorrow will broadly resemble today. Now try to give the reason, one that does not quietly assume what it sets out to prove. If you find you cannot, and yet cannot stop believing it either, you have arrived exactly where Hume did, and felt the gap between what we can justify and what we cannot help.