Hume's Problem of Induction

Every morning you expect the sun to rise. You expect bread to nourish you rather than poison you, and a dropped cup to fall rather than hover. These expectations feel so secure that calling them into question seems perverse. David Hume, writing in the 1730s, asked the perverse question anyway, and the answer he reached has unsettled philosophers ever since.

Hume's question is not whether the sun will rise tomorrow. It is whether we have any rational ground for believing it will. We have observed sunrises many times. From this we conclude that sunrises will continue. But notice the shape of that inference. From a finite collection of past cases, we leap to a conclusion about cases we have not yet observed, including cases in the future. What licenses the leap?

Hume considers the obvious answer: nature is uniform. The future will resemble the past because that is simply how the world works. But how do we know this principle of the uniformity of nature? Not by pure reason — there is no contradiction in imagining a world where, starting tomorrow, bread suffocates and stones float. The denial of uniformity is strange, but not incoherent. So the principle is not a truth of logic.

Then perhaps we know it from experience. We have observed, again and again, that nature has been uniform. But this is precisely where Hume springs the trap. To argue that nature will continue to be uniform because it has been uniform so far is to use an inductive inference to justify induction. We are assuming the very principle we were trying to establish. The argument is circular, and no amount of additional observation can break the circle.

This is the problem of induction in its sharpened form. Inductive reasoning cannot be justified by pure reason, because its conclusions are not logical necessities. And it cannot be justified by experience without presupposing itself. There is no third source of justification on offer. Hume concludes that inductive inference, the engine of nearly all empirical knowledge, has no rational foundation.

This sounds like a counsel of despair, but Hume does not despair. He turns from the question of justification to the question of explanation. Why, given that we have no rational warrant for induction, do we go on inducing? His answer is that we are creatures of custom. Repeated exposure to one event followed by another produces in the mind a habit of expectation. When we see the flame, we expect the heat; when we hear the thunder, we brace for the lightning's aftermath. This expectation is not a piece of reasoning we perform but a disposition the mind has acquired. Nature, Hume says, has not left it to our uncertain reasoning to discover the connections that keep us alive.

It is crucial to see what Hume has and has not done here. He has offered a psychological account of why we make inductive inferences. He has not offered a justification for them. The habit of expectation explains our behavior; it does not show that the behavior is rational. A reader who finishes the passage thinking, "so induction works because it is built into us," has missed the move. That induction is built into us is exactly compatible with its having no rational ground.

Later philosophers have responded in many ways. Some have tried to vindicate induction pragmatically: even if it is unjustified in Hume's sense, it is the best strategy available, and any alternative would do worse. Others have argued that Hume's standard of justification is too strict — that demanding non-circular grounds for our most basic inferential practices sets a bar nothing could clear, and that we should reconsider the standard rather than condemn the practice. Still others have tried to dissolve the problem by reframing what induction is supposed to do.

What none of these responses can do is meet Hume on his own terms and win. The circle he identified is real. The question he leaves us with is whether circularity at the foundations is a catastrophe, a curiosity, or simply the human condition.