The Problem of Induction

The problem, posed most influentially by David Hume in the Treatise of Human Nature and again in the Enquiry Concerning Human Understanding, of what could possibly justify inferring a general or future claim from past observations alone. Hume noted that reasoning of the form the sun has risen every day so far, therefore it will rise tomorrow, is not a matter of logical necessity, since its denial involves no contradiction, and cannot be justified by appeal to past experience either, since that would use induction to justify induction and argue in a circle. He concluded that our confidence in such inferences rests not on reason but on custom or habit, a psychological disposition formed by repeated observation rather than a rational guarantee. The problem has never received a solution every philosopher accepts: responses range from Karl Popper's proposal to replace induction with a logic of falsification, to attempts by later philosophers to justify induction pragmatically or probabilistically, and the difficulty remains a standard starting point for any course in epistemology or philosophy of science.