Explain Hume's problem of induction and the new riddle of induction, and assess the main responses including pragmatic, probabilistic and Popperian replies

What this dot point is asking

SEAB wants you to explain David Hume's problem of induction, perhaps the deepest challenge in the philosophy of science, and to assess the responses to it. Since science relies on inferring from observed cases to unobserved ones, a successful argument that such inference cannot be justified threatens the rational standing of science itself. Your task is to state the problem exactly, add Goodman's refinement, and weigh the leading replies.

The answer

Hume's problem stated

Inductive inferences move from observed cases to unobserved ones: every observed emerald has been green, so the next will be green; bread has always nourished, so it will nourish tomorrow. Hume points out that every such inference relies on the assumption that the future will resemble the past, often called the uniformity of nature. The problem is how that assumption is justified.

The dilemma

The justification cannot be deductive, because there is no contradiction in supposing that nature changes its course tomorrow; a world in which the sun fails to rise is conceivable, so uniformity is not a logical truth. The justification cannot be inductive either, because to argue that nature will stay uniform on the grounds that it has been uniform so far is to use an inductive inference to justify induction, which is circular. With both options closed, induction has no non-circular justification.

Use versus justification

It is essential to state what Hume does and does not conclude. He does not say it is irrational or foolish to reason inductively; we cannot help doing so, and life depends on it. He says we cannot provide a rational foundation for it. Inductive expectation is, for Hume, a habit of mind produced by repeated experience, not a conclusion reached by reason. So the problem is about justification, not about whether to use induction.

Goodman's new riddle

Nelson Goodman sharpens the difficulty even for those who set Hume's problem aside. Define grue: an object is grue if it is examined before some future time and found green, or else is not so examined and is blue. Every observed green emerald is also grue, so the same evidence supports both "all emeralds are green" and "all emeralds are grue," which make conflicting predictions about unexamined emeralds. The riddle is why we are entitled to project "green" rather than "grue." Evidence alone underdetermines the choice, so even a licensed induction needs a principled account of which predicates are projectible.

The main responses

Pragmatic vindication, associated with Reichenbach, argues that if any method of prediction works, induction will work too, so using induction is a no-lose strategy; critics object that this shows induction is a reasonable bet, not that it is reliable. Probabilistic and Bayesian accounts recast induction as the rational updating of probabilities in the light of evidence; critics note they require prior probabilities and arguably smuggle in a uniformity assumption. Popper bites the bullet and denies that science uses induction at all: science proceeds by conjecture and attempted falsification, never by inductive confirmation; critics reply that corroboration still seems to guide which theories we rely on, which looks inductive in spirit. Goodman's own reply to the new riddle is entrenchment: we project the predicates that our language and past inductive practice have repeatedly projected.

Examples in context

Example 1. The turkey and the farmer. A turkey, fed every morning, inductively concludes that it will be fed every morning, until the day before a festival when its expectation is fatally disappointed. The fable dramatises Hume's point: a long run of confirming instances gives no logical guarantee about the next case, because the future need not resemble the past. It is a vivid reminder that inductive confidence is a habit, not a proof.

Example 2. Replication in science. Scientists trust that an experiment which yields a result here and now will yield the same result elsewhere and later, which is why replication matters. This trust is inductive and so inherits Hume's problem; yet science could not proceed without it. The example shows that the problem is not a reason to abandon science but a reason to understand its inductive commitments as a rational policy rather than a certainty.

Try this

Q1. State the dilemma at the heart of Hume's problem of induction. [6 marks]

• Cue. The uniformity-of-nature assumption cannot be justified deductively (its denial is not a contradiction) nor inductively (that is circular), so induction has no non-circular justification.

Q2. Explain the difference between Hume's problem and Goodman's new riddle of induction. [8 marks]

• Cue. Hume questions whether induction can be justified at all; Goodman, granting that, shows we must still explain why we project "green" rather than "grue," since the evidence fits both. The riddle concerns projectibility, not legitimacy.

Q3. Outline Popper's response to the problem of induction and one objection to it. [6 marks]

• Cue. Popper denies science uses induction: it proceeds by conjecture and falsification. Objection: corroboration still seems to guide which theories we rely on, which looks inductive, so the problem may not be dissolved.