Can our reliance on induction be rationally justified, or does Hume's argument show that it cannot?
explain Hume's problem of induction and evaluate proposed responses, including the appeal to the uniformity of nature and pragmatic justifications
A focused QCE Unit 3 answer on the problem of induction. Covers David Hume's argument that inductive inference cannot be justified without circularity, the role of the uniformity of nature, and responses including pragmatic justification, Popper's falsificationism and the limits of each.
Reviewed by: AI editorial process; not yet individually human-reviewed
Have a quick question? Jump to the Q&A page
Jump to a section
What this dot point is asking
QCAA wants you to grasp the deepest challenge to inductive reasoning. David Hume argued that although we constantly reason from observed cases to unobserved ones, this inference cannot be rationally justified without arguing in a circle. You need to reconstruct Hume's argument precisely, state the role of the uniformity of nature, and evaluate the main responses. This is a high-value link between the reasoning strand and the theory of knowledge.
The answer
What induction assumes
Every inductive inference, from "the sun has always risen" to "this drug has always worked," assumes that unobserved cases resemble observed ones: roughly, that nature is uniform across time and space. Without this assumption, past observations would tell us nothing about the future. Hume's question is simple and devastating: what justifies the assumption that the future will resemble the past?
Hume's argument
Hume, in An Enquiry concerning Human Understanding (1748), argues by elimination. A justification of induction must be either demonstrative (deductive, from relations of ideas) or probable (empirical, from matters of fact).
- It cannot be demonstrative, because there is no contradiction in supposing the future will differ from the past. A world where bread suddenly stops nourishing is conceivable, so no deductive proof rules it out.
- It cannot be probable (empirical) without circularity. Any argument that induction has worked before, so it will work again, already uses induction. It assumes the very uniformity it is trying to establish.
Since both routes fail, Hume concludes that our inductive habits rest not on reason but on custom or habit: having seen events conjoined repeatedly, the mind is conditioned to expect the one on seeing the other. Induction is psychologically irresistible but rationally unfounded.
The circularity at the heart of it
The core point is the circle. To justify "the future resembles the past," you cannot say "it always has," because that is itself an inductive inference relying on the principle in question. This is why the problem is so stubborn: any empirical defence of induction presupposes induction.
Responses
- Inductive justification ("it has worked, so it will work"): rejected by Hume as circular.
- The uniformity of nature as a premise: you can make induction valid by adding "nature is uniform" as a premise, but then you must justify that premise, which faces the same dilemma.
- Pragmatic or vindicatory justification (Hans Reichenbach): we cannot prove induction will work, but if any method works, induction will; so it is rational to use it as our best bet. Critics object that this does not show induction is reliable, only that nothing else is guaranteed better.
- Falsificationism (Karl Popper): science does not rely on induction at all; it proceeds by bold conjectures tested by deductive falsification. Critics reply that we still need induction to trust that a theory which has survived testing will keep working (the problem of corroboration), so Popper may dissolve the problem only verbally.
- Dissolving the problem (P. F. Strawson): asking whether induction "as a whole" is justified is like asking whether the law is legal; being reasonable just means, in part, proportioning belief to inductive evidence. Critics say this redescribes the practice rather than vindicating it.
Why it matters
If induction cannot be justified, the entire edifice of empirical science and everyday prediction rests on something other than reason, which is a striking sceptical result. Yet no one can actually abandon induction. The problem therefore sharpens what we mean by rational justification and feeds directly into debates about scientific method and the theory of knowledge.
Try this
Q1. Reconstruct Hume's dilemma about justifying induction. [4 marks]
- Cue. Justification is demonstrative or probable; not demonstrative (non-uniform future conceivable), not probable (circular), so neither works.
Q2. Explain why adding "nature is uniform" as a premise does not solve the problem. [3 marks]
- Cue. That premise itself needs justification and faces the same demonstrative-or-circular dilemma.
Q3. State one objection to Popper's claim that science avoids induction. [2 marks]
- Cue. We still seem to rely on induction to trust that a corroborated theory will keep working.
Exam-style practice questions
Practice questions written in the style of QCAA exam questions on this dot point, with worked answer explainers. The year tag is the paper they imitate, not the source.
QCAA 20227 marksExplain Hume's problem of induction and evaluate the appeal to the uniformity of nature as a response to it.Show worked answer →
A 7 mark response reconstructs Hume's dilemma, then assesses the uniformity response.
Hume's argument. Induction assumes unobserved cases resemble observed ones (the uniformity of nature). A justification must be demonstrative or probable. It is not demonstrative, because a non-uniform future is conceivable without contradiction. It is not probable without circularity, because arguing "induction has worked, so it will work" already uses induction. So induction rests on custom, not reason.
The uniformity response. One reply adds "nature is uniform" as an extra premise, making the inductive inference deductively valid.
Evaluation. This only relocates the problem: the new premise itself needs justification and faces Hume's same dilemma (not demonstrable, and provable only by a circular appeal to past uniformity). So the response does not escape the circle; it postpones it.
Markers reward an accurate reconstruction of the dilemma and a clear evaluation showing the uniformity premise inherits the original problem.
QCAA 20235 marksCompare Reichenbach's pragmatic justification of induction with Popper's falsificationism as responses to Hume, and assess whether either solves the problem.Show worked answer →
A 5 mark response states each response and judges it.
Reichenbach (pragmatic vindication). We cannot prove induction will work, but if any method succeeds at predicting, induction will, so it is the rational bet. Limit: this shows induction is no worse than alternatives, not that it is reliable, so it concedes Hume's sceptical point.
Popper (falsificationism). Science avoids induction by making bold conjectures tested through deductive falsification; theories are never confirmed, only refuted or corroborated. Limit: critics argue we still rely on induction to trust that a well-corroborated theory will keep working (the problem of corroboration), so the reliance is hidden, not removed.
Assessment. Neither fully solves the problem: Reichenbach accepts the lack of proof, and Popper arguably smuggles induction back in. Each manages rather than refutes Hume.
Markers reward an accurate account of both responses and a critical judgement of each.