Philosophy of Science

Quick Facts

• Name: Philosophy of Science
• Kind: A branch of philosophy about science, not one single doctrine
• Time period: Ancient roots, major modern form from the 1600s onward
• Main questions: What makes something scientific? What counts as evidence? What is a good explanation? Do scientific theories describe reality?
• Main regions: Europe, Britain, United States, and global science studies
• Neighboring fields: Analytic philosophy, epistemology, history of science, sociology of science

In One Minute

Philosophy of science asks what science is doing when it works. It studies evidence, experiments, models, laws, explanation, objectivity, and theory change.

A scientist might say, "This drug works because the trial showed a real effect." Philosophy of science asks what "showed" means here. Was the trial large enough? Did the control group rule out other causes? Does the result explain the disease, or only predict improvement?

It also asks whether science tells us what the world is really like. Scientific realism says successful theories are usually getting at real things, even when those things are invisible, like electrons or genes. Scientific anti-realism says a theory can be useful, accurate, and worth trusting without proving that every invisible thing it talks about is literally real. For an anti-realist, a model of electrons may be like a map: excellent for navigation, but not a photograph of the territory.

Main Ideas

Induction means reasoning from past cases to new cases. If every tested sample of a metal expands when heated, we expect the next sample to expand too. Hume made the problem famous: the past has always guided us, but that alone does not logically prove the future must behave the same way.

Confirmation is support from evidence. A prediction confirms a theory when the result fits what the theory led us to expect. If a theory predicts where a planet should appear and the telescope finds it there, the theory gains support. The hard part is that many different theories can sometimes fit the same result.

Explanation means showing why something happens, not just that it happens. Saying "the patient recovered after taking the pill" reports a sequence. A stronger explanation says how the pill blocked a virus, changed blood pressure, or triggered an immune response.

Demarcation is the problem of separating science from non-science or pseudoscience. Astronomy predicts eclipses and risks being wrong. Astrology also makes claims about the stars, but often bends its claims so loosely that failures do not count against it.

Falsifiability is Popper's answer to demarcation. A claim is scientific only if some possible observation could show it is false. "All swans are white" is falsifiable because one black swan would refute it. "This crystal has invisible healing energy that cannot be measured" is not falsifiable in the same way.

Underdetermination means the evidence may not force one theory by itself. If a spaceship misses its target, maybe the gravity theory is wrong, maybe the engine data were wrong, maybe the instruments were misread. Duhem and Quine both pressed this point: tests usually check a whole package of theory plus background assumptions.

Theory-ladenness means observation is shaped by what observers already know, expect, and know how to measure. A trained radiologist sees a tumor pattern in an image where a beginner sees gray patches. This does not make observation useless. It means observation is skilled, instrument-guided, and interpreted.

Paradigm is Kuhn's word for a shared scientific framework: accepted examples, tools, standards, problems, and background assumptions. Newtonian mechanics was a paradigm for much of physics. It told scientists what counted as a good problem and what a good solution looked like.

Normal science is Kuhn's name for everyday research inside a paradigm. Scientists do not question everything at once. They solve puzzles, refine measurements, extend models, and explain awkward cases. A chemist using atomic theory to explain a reaction is doing normal science.

Scientific realism says mature, successful science often tells us roughly what the world is like. Realists think electrons, viruses, tectonic plates, and black holes are not just useful bookkeeping devices.

Scientific anti-realism says science can be reliable without requiring that its hidden entities be taken as literally true. The point may be to save the observable facts, predict well, and organize inquiry, while staying cautious about what lies beyond observation.

How It Works

Philosophy of science usually starts from real scientific practice. It asks how experiments, measurements, models, and arguments earn trust.

One question is method. Bacon pictured science as disciplined inquiry that resists wishful thinking and authority. Galileo showed how experiment, mathematics, and idealization can work together. Idealization means simplifying reality on purpose. A physics problem may treat a falling object as if air resistance does not exist, not because air is unreal, but because the simplified model reveals a pattern.

Another question is evidence. Scientific claims are not checked one at a time in isolation. Suppose a vaccine trial fails. That might show the vaccine does not work. It might also show the dose was too low, the storage temperature was wrong, the sample was too small, or the endpoint was badly chosen. This is why underdetermination matters. Evidence constrains theory, but it often leaves room for judgment.

A third question is explanation. Carl Hempel defended a "covering law" model: to explain an event is to show that it follows from general laws plus facts about the case. If a copper wire expanded, we explain it by saying copper expands when heated and this wire was heated. Later philosophers argued that this misses many real explanations. In biology, explaining a trait may involve function, history, and mechanism, not just one clean law.

A fourth question is theory change. Popper described science as bold guesses followed by severe attempts to refute them. Kuhn replied that science is not usually a constant attempt to overthrow its own framework. Most science is normal science: careful puzzle-solving inside a paradigm. Revolutions happen when problems pile up and a new framework starts solving them better, as when oxygen chemistry replaced phlogiston chemistry.

Lakatos tried to split the difference. He said scientists often work in research programmes. A research programme has a hard core protected by auxiliary assumptions. Newtonian physics, for example, could survive a failed planetary prediction by checking the telescope, the calculations, or the possibility of another planet. For Lakatos, the issue is whether the programme keeps producing new successes or only makes excuses.

Feyerabend pushed harder. He argued that no single rule captures all good science. Real history includes rule-breaking, competing methods, rhetoric, politics, instruments, and accidents. His slogan "anything goes" is easy to overread. The better point is that strict method rules can misdescribe how discovery actually happens.

Recent science studies added another layer. Scholars such as Bruno Latour studied laboratories, instruments, funding, authorship, and trust networks. They asked how facts become stable enough that later scientists can treat them as settled. This does not have to mean facts are fake. It means facts become usable through material work: samples, machines, inscriptions, peer review, calibration, and replication.

Key People

• Francis Bacon: defended experiment, organized observation, and methods for escaping inherited prejudice.
• Galileo Galilei: made mathematical experiment a model for modern science.
• Isaac Newton: gave later philosophers the classic example of mathematically powerful law-based physics.
• David Hume: made induction and causation philosophically difficult.
• Immanuel Kant: asked how objective natural science is possible for minds like ours.
• Rudolf Carnap: central logical empiricist who tried to clarify science with logic, probability, and formal languages.
• Carl Hempel: analyzed confirmation and scientific explanation, especially the covering-law model.
• Pierre Duhem: argued that physical theories are tested as groups, not one sentence at a time.
• W. V. O. Quine: extended that point into a wider picture of belief as a web revised under pressure from experience.
• Karl Popper: made falsifiability, criticism, and risky prediction central.
• Thomas Kuhn: explained science through paradigms, normal science, anomalies, and revolutions.
• Imre Lakatos: described science as competing research programmes that can be progressive or degenerating.
• Paul Feyerabend: attacked the idea that science follows one universal method.
• Bruno Latour: studied how laboratories, instruments, inscriptions, and networks help make facts durable.

Important Works

• Novum Organum by Francis Bacon: argues that inquiry should be organized around experiment and careful collection of evidence rather than inherited authority. It is an early modern manifesto for method.
• An Enquiry Concerning Human Understanding by David Hume: gives the classic problem of induction. Hume argues that experience teaches us patterns, but reason alone cannot prove nature must keep following them.
• The Aim and Structure of Physical Theory by Pierre Duhem: argues that a failed prediction does not refute one hypothesis by itself, because tests involve instruments, mathematics, and background assumptions.
• The Logical Structure of the World and later works by Rudolf Carnap: try to make science clearer through logical reconstruction, formal language, and careful accounts of confirmation.
• Studies in the Logic of Explanation by Carl Hempel and Paul Oppenheim: presents the covering-law model, where explanation shows how an event follows from laws plus initial conditions.
• "Two Dogmas of Empiricism" by W. V. O. Quine: attacks the idea that some truths are purely analytic and says our beliefs face experience as an interconnected web.
• The Logic of Scientific Discovery by Karl Popper: argues that science advances by conjectures and refutations. Theories should make risky claims that could fail.
• The Structure of Scientific Revolutions by Thomas Kuhn: argues that science alternates between normal puzzle-solving and revolutionary paradigm change. It made "paradigm shift" part of ordinary speech.
• "Falsification and the Methodology of Scientific Research Programmes" by Imre Lakatos: says theories live inside research programmes. A programme is healthy when it predicts new facts and unhealthy when it only patches failures after the fact.
• Against Method by Paul Feyerabend: attacks strict scientific method rules and argues that real scientific progress has often required pluralism, disorder, and rule-breaking.
• Laboratory Life by Bruno Latour and Steve Woolgar: studies a working laboratory to show how scientific facts are built through instruments, writing, negotiation, repetition, and trust.
• The Mangle of Practice by Andrew Pickering: describes science as an active struggle between human plans, instruments, materials, and unexpected resistance from the world.

Why It Matters

Philosophy of science matters because scientific authority is powerful. Courts, hospitals, schools, governments, and ordinary people rely on claims about evidence. This field helps explain when that trust is earned.

It also protects against two bad extremes. One extreme treats "science says" as magic words that end all argument. The other treats every scientific claim as just politics or fashion. Philosophy of science gives better tools: ask what the evidence is, what alternatives were ruled out, what assumptions were used, how the model works, and how the claim could be corrected.

The field is also useful because science itself is not one thing. Particle physics, climate modeling, evolutionary biology, psychology, medicine, and archaeology do not all explain in the same way. A climate model, a double-blind drug trial, a fossil record, and a particle detector need different kinds of trust.

Critics And Pushback

Some critics say Popper's falsifiability is too simple. Scientists rarely throw away a theory after one bad result. They check the equipment, the statistics, the background assumptions, and the auxiliary hypotheses. This is the Duhem-Quine problem in practice.

Some critics say Kuhn made science look too tribal, as if paradigms were closed worlds. Kuhn did not think scientists simply choose frameworks at random, but his view raised hard questions about objectivity, comparison, and progress across revolutions.

Realists and anti-realists argue over what success proves. Realists say it would be a miracle if mature scientific theories predicted so well while being deeply wrong about electrons, viruses, or genes. Anti-realists answer that past theories also worked well before being replaced, so success alone does not guarantee truth.

Science studies has its own pushback. If it stresses social practice too much, critics worry it makes facts sound like mere inventions. Its defenders answer that studying laboratories, instruments, money, and institutions explains how objectivity is produced, tested, and maintained.

The strongest lesson is modest: science is powerful, but not automatic. It depends on methods, tools, communities, criticism, and correction.

Related Pages

• Francis Bacon
• Galileo Galilei
• Isaac Newton
• David Hume
• Immanuel Kant
• Rudolf Carnap
• W. V. O. Quine
• Karl Popper
• Thomas Kuhn
• Paul Feyerabend
• Bruno Latour
• Analytic Philosophy