Sunday, November 22, 2026

Hume and Popper on the Problem of Induction

A bridge has carried the morning train every day this year. We expect it to carry the next one. What makes that expectation reasonable, and what would it take to turn it into a general claim? David Hume’s problem of induction begins with the distance between what we have observed and what we expect to happen.

Hume on reasoning from experience

In the fourth section of his Enquiry Concerning Human Understanding, Hume asks how experience can justify conclusions about events we have not observed. A contradiction does not follow from supposing the next train will fail to cross. An appeal to the past success of similar inferences already assumes that past experience is a guide to what comes next. Hume goes on to explain the force of those expectations through custom. We acquire them by repeatedly encountering events together.

What Popper does with the problem

In Conjectures and Refutations, Popper accepts that repeated observations cannot establish a universal law by deduction. He proposes a different account of scientific method. A scientist puts forward an explanation and works out what should happen if it holds under specified conditions. Observations can then be used to challenge the explanation. The conjecture may have been suggested by earlier experience, but its origin and the severity of its test are different questions.

What a counterexample can show

Consider the claim that every sample of a certain material expands when heated under specified conditions. Each sample that expands is compatible with the claim. A reliable case of contraction would be incompatible with it. In practice the result depends on the sample, the measurements and the conditions of the test. Those details help determine what the observation counts against.

The practical question for Popper is which available explanation has stood up to the most informative criticism so far. Scientists can use it provisionally and still leave open the possibility that a later test will expose a problem. This is why falsifiability matters. The explanation has to tell us what would count against it.

Further reading · Chalmers on Deriving Theories from Facts · Hume’s Enquiry, section IV · Read Popper’s opening chapter