Philosophy › Reason, experience & Enlightenment › David Hume
Hume's problem of induction
1739–1748· Idea or argument· Knowledge, logic & science
Well documentedSolid records, texts or physical evidence.
Hume's argument that no amount of past experience logically proves that the future will resemble the past.
See it on the Philosophy map →The idea
Idea: we cannot justify reasoning from observed cases to unobserved ones without going in a circle. Reasoning: the sun has risen every day so far, but nothing in logic guarantees it rises tomorrow. Saying 'nature is uniform' just restates the claim, and 'induction has worked before' uses induction to prove induction. Hume's answer: we rely on habit, not reason. Main objection: Karl Popper said science does not need induction; it tests and tries to falsify bold guesses.
Why it matters
It challenges the footing of science, forecasting and every model built on past data, and drove work by Kant, Popper and modern probability theorists.
Think of it like
Like underwriting a tenant on ten years of on-time rent: reasonable, but the record alone cannot guarantee next month's check.
Where it comes from
Set out in Hume's Treatise (1739, Book 1) and Enquiry Concerning Human Understanding (1748, Section 4). Russell's chicken, fed daily until the day its neck is wrung, is a famous 1912 restatement.
Still debated
No solution commands agreement; Bayesian, pragmatic and 'dissolve the problem' replies all have defenders, and Nelson Goodman's 1955 'grue' puzzle made it harder.
Connections (3)
▸Francis Bacon and the scientific method championed the induction questioned in Hume's problem of inductionMostly agreed
The connection
Bacon made induction, reasoning from many observed cases to general laws, the method of the new science. Hume later asked what justifies that step at all, showing that past observations cannot logically guarantee future ones. Bacon's confident method thus became the target of Hume's famous puzzle.
When and how
Bacon's New Organon, London, 1620; Hume's Treatise of Human Nature, 1739, and Enquiry, 1748.
How sure are we?
Hume subtitled his Treatise an attempt to bring the 'experimental method' into moral subjects, and its introduction dates experimental natural philosophy from 'my Lord Bacon'.
Debates and open questions
Bacon's induction, built on eliminating rival explanations, is more careful than simple generalizing; whether Hume's problem hits it equally hard is discussed.
▸David Hume posed Hume's problem of inductionWell documented
The connection
Applying his rule that ideas must trace back to experience, Hume asked what justifies expecting the future to resemble the past. Not logic, since the opposite is conceivable; not experience, which would be circular. He concluded we rely on custom or habit, a fact about human nature, not a proof.
When and how
Treatise of Human Nature, Book 1 (1739), restated more clearly in the Enquiry Concerning Human Understanding (1748), Section 4.
How we know
Hume's texts state the argument plainly; the name 'problem of induction' came later, but philosophers from Kant to Popper credit him with it.
Debates and open questions
Whether Hume was a skeptic about induction or simply explaining how it works naturally is one of the main debates in Hume scholarship.
▸Hume's problem of induction was answered by Karl Popper and falsifiabilityWell documented
The connection
Hume showed that no number of past observations proves a general law: the sun rising every day so far does not guarantee tomorrow. Popper accepted this and said science never needs induction. It makes bold guesses and tries to refute them; one failed prediction disproves a theory by plain logic, so science runs on deduction.
When and how
Popper set out his answer in The Logic of Scientific Discovery (Vienna, 1934) and later in Objective Knowledge (1972), which titles a chapter 'my solution of the problem of induction'.
How we know
Popper explicitly frames his work as an answer to Hume in these books.
Debates and open questions
Many philosophers think Popper did not escape induction: trusting a well-tested theory to hold a bridge up tomorrow still assumes the future will resemble the past.