A wing of the atlas · how knowledge is made

Philosophy of science

Science is the most reliable method we have for understanding the world — and explaining exactly why has proved surprisingly difficult. What separates it from other confident systems of belief? What justifies generalising from the experiments we ran to the ones we didn't? Philosophy of science asks what science actually is, and the answers have been less tidy than science's success suggests.

Philosophy of science examines the nature, methods, and limits of scientific knowledge. Central questions include what distinguishes science from non-science, how observation justifies general theories, whether science progresses cumulatively or through revolutions, and whether scientific theories describe reality or merely predict observations.

The positions · at a glance
The central questions
Position 01

What counts as science?

The demarcation problem

What separates science from pseudoscience? Popper proposed falsifiability — a theory is scientific only if it forbids something observable.

Position 02

Why trust generalisation?

The problem of induction

No number of observations logically guarantees the next case. Hume showed this has no non-circular solution, and it remains open.

Position 03

How does science change?

Cumulative or revolutionary

Kuhn argued science proceeds through periods of normal work punctuated by paradigm shifts that reframe the questions themselves.

Position 04

Is science objective?

Values and evidence

Theory choice involves judgement, and observation is shaped by expectation. Whether that undermines objectivity or merely describes it is disputed.

Science works. That is not seriously in dispute, and any philosophy of science that implies otherwise has gone wrong somewhere. The interesting problem is that when philosophers try to specify precisely why it works — what rule scientists follow that others do not — the tidy answers keep failing on close inspection.

Take the obvious proposal: science is what is confirmed by observation. Popper's objection is that confirmation is cheap. Astrology, conspiracy theories and pseudo-medicine all accumulate confirming instances happily, because a sufficiently flexible theory can absorb any observation. What distinguishes a scientific claim, he argued, is that it forbids something — it sticks its neck out and could be shown wrong.

RevolutionHow science actually changes.

Thomas Kuhn complicated the picture by looking at history rather than logic. Science, he argued, does not accumulate steadily. Most of the time researchers work within a paradigm — a shared framework of assumptions, methods, and exemplary problems — doing what he called normal science, which is essentially puzzle-solving within accepted rules. Anomalies accumulate, are set aside, and eventually reach a point where the framework itself gives way. The resulting shift is not merely a change of belief but a change in what counts as a question worth asking.

Kuhn's controversial claim was that competing paradigms can be incommensurable — lacking a fully neutral standard by which to compare them, since the standards themselves are part of what changes. Critics argued this makes theory choice irrational or merely sociological. Kuhn resisted that reading, holding that good reasons operate in such choices even where no algorithm settles them, which is arguably how most difficult judgement works.

RealismDoes science describe reality?

A final divide. Scientific realists hold that successful theories are approximately true and that their unobservable entities — electrons, fields, genes — really exist; the alternative, they argue, would make science's predictive success an inexplicable miracle. Anti-realists reply that the aim of science is empirical adequacy rather than truth, and point to the historical record: theories that were highly successful, such as the luminiferous ether, turned out to posit entities that do not exist. If past successful theories were wrong about what exists, present success is weaker evidence than it appears.

None of this undermines science, and it is worth saying so plainly, because these debates are sometimes misused for exactly that. That we cannot fully articulate why a method works is a problem for philosophy, not evidence against the method. Science remains the most reliable instrument we have for finding out about the world; philosophy of science is the attempt to understand the instrument, not to dull it.

People also askQuick answers

What is philosophy of science?

The study of the nature, methods and limits of scientific knowledge — what distinguishes science from non-science, how evidence supports theories, how science changes, and whether theories describe reality.

What is the demarcation problem?

The question of what separates science from non-science or pseudoscience. Popper's proposal was falsifiability: a theory is scientific only if it forbids some observable outcome and could therefore be shown wrong.

What is a paradigm shift?

Kuhn's term for a fundamental change in the framework of assumptions, methods and exemplary problems that guides a scientific field. Anomalies accumulate until the framework itself gives way, changing what counts as a legitimate question.

Does philosophy of science undermine science?

No. These debates concern why the method works and what its results mean, not whether it works. Being unable to fully articulate a method's justification is a problem for philosophy, not evidence against the method.

Sources & further reading

A summary following standard scholarly accounts. These debates concern the interpretation of scientific method, not the reliability of science itself.