Popper and Falsifiability

Karl Popper wanted to solve two problems at once.

First:

How can science grow if induction cannot logically prove universal laws?

Second:

How can science be distinguished from systems that explain everything and therefore risk nothing?

His answer centered on falsifiability.

A scientific theory should make claims that could, in principle, conflict with observation.

Science advances not by proving theories true, but by subjecting them to severe tests.

The Asymmetry of Universal Claims

Consider:

All swans are white.

No finite number of white swans proves the statement true.

A million white swans still leave open the possibility of a black one elsewhere.

But one genuine black swan is logically enough to refute the universal claim.

This is the asymmetry Popper emphasized.

Verification is difficult.

Falsification can be logically decisive in ideal cases.

Conjectures and Refutations

Popper described science as a process of:

1. conjecture,
2. attempted refutation,
3. replacement or revision.

Scientists propose bold theories.

Then they try to expose weaknesses.

A theory that survives severe tests earns corroboration.

It is not proven true.

It has simply survived attempts to show it false.

Falsifiable Does Not Mean False

A common misunderstanding is that a falsifiable theory is one already shown to be wrong.

No.

Falsifiable means:

there are conceivable observations that would count against it.

“The orbit of this planet follows equation X within specified tolerance” is falsifiable.

“Everything happens because of invisible forces that always adjust to produce whatever we observe” is not meaningfully falsifiable.

Risky Predictions

Popper admired theories that make risky predictions.

The more a theory rules out, the more informative it is.

A vague theory says:

“Something unusual may happen.”

A precise theory says:

“At this time and place, the instrument should measure a value within this range.”

If the precise prediction succeeds, the survival is impressive because failure was genuinely possible.

Einstein as Popper’s Model

Popper often contrasted Einstein’s relativity with systems he saw as too flexible.

General relativity made precise predictions about phenomena such as light deflection.

The theory risked failure.

If observations had strongly contradicted those predictions under reliable conditions, the theory would have faced serious trouble.

For Popper, this vulnerability was a mark of science.

Psychoanalysis and Marxism in Popper’s Critique

Popper criticized some forms of psychoanalysis and Marxist historical theory because he believed practitioners could reinterpret nearly any outcome as confirmation.

If behavior A occurred, the theory explained it.

If opposite behavior B occurred, the theory explained that too.

A framework compatible with every outcome cannot be tested sharply.

His historical assessments remain debated, but the methodological point is clear.

Confirmation Is Cheap

Humans are good at finding confirming examples.

Astrology can appear accurate if people remember hits and forget misses.

A political theory can explain every event after it happens.

A personality description can feel true because it is vague.

Popper therefore distrusted confirmation obtained too easily.

The important question is:

What observation would have counted as failure?

Severe Testing

A strong test gives the theory a real chance to fail.

Suppose a model predicts temperatures only between 20 and 21 degrees.

A measurement of 20.5 is significant support.

A model predicting anything between -100 and 500 degrees is not tested strongly by the same result.

Precision creates severity.

Corroboration

Popper used corroboration instead of confirmation in a strong inductive sense.

A theory that survives difficult tests becomes more worthy of provisional acceptance.

But survival does not logically make it more probable in the simple inductive way Popper wanted to avoid.

This distinction is philosophically subtle.

In practice, scientists often do increase confidence after predictive success.

Science Is Provisional

Popper’s philosophy fits a fallibilist picture.

No empirical theory becomes absolutely certain.

Even extremely successful theories remain revisable.

Newtonian mechanics worked brilliantly for centuries.

Relativity later revealed its domain limits.

Science progresses through increasingly powerful approximations.

Falsification Is Harder in Practice

The textbook picture says:

Theory predicts P.

Observation gives not-P.

Therefore theory is false.

Real science is messier.

A prediction depends on:

• auxiliary hypotheses,
• calibration,
• initial conditions,
• statistical assumptions,
• data processing.

When observation disagrees, several components may be responsible.

The Duhem-Quine Problem

Pierre Duhem and later W. V. O. Quine emphasized that hypotheses are rarely tested in isolation.

Suppose an astronomical prediction fails.

Maybe gravity theory is wrong.

Or telescope calibration.

Or atmospheric correction.

Or orbital parameters.

Logic alone does not tell us which assumption to reject.

This weakens naive falsificationism.

Neptune Again

Uranus did not move exactly as expected under known planetary models.

Scientists did not immediately falsify Newtonian gravity.

They considered another auxiliary hypothesis:

an unseen planet.

Neptune was found.

This was scientifically rational.

A failed prediction does not always mean the central theory should be abandoned.

Mercury Was Different

Mercury’s anomalous perihelion precession resisted explanation through an unseen planet.

Eventually general relativity explained the discrepancy.

Why keep Newton for Uranus but replace it for Mercury?

Because scientific judgment considers the entire network of evidence.

Falsification is historically and theoretically contextual.

Statistical Theories

Many theories do not predict exact outcomes.

They predict probabilities.

Suppose a coin model predicts 50% heads.

Ten heads in a row do not logically falsify the model.

The event is unlikely but possible.

Statistical theories require thresholds, likelihoods, repeated tests, and model comparison.

Falsification becomes probabilistic rather than absolute.

Measurement Error

No instrument has infinite precision.

Suppose a theory predicts 10.000 units and the measurement gives 10.003.

Is the theory falsified?

That depends on uncertainty.

Scientific tests must specify tolerances.

A theory cannot be compared with reality without a measurement model.

Theory-Ladenness

Observations themselves depend partly on theory.

A detector event is interpreted using physical models.

A fossil is dated using geological assumptions.

A spectral line is identified through atomic physics.

There is no perfectly theory-free observational tribunal.

This does not make testing impossible.

It makes testing interconnected.

Kuhn’s Critique

Thomas Kuhn argued that scientists normally work within paradigms and tolerate anomalies.

No major theory fits every observation perfectly.

If one anomaly forced abandonment, science would become unstable.

Researchers often treat anomalies as puzzles rather than refutations until an alternative framework becomes available.

Scientific revolution is not one failed test.

Lakatos’s Development

Imre Lakatos proposed evaluating research programmes rather than isolated hypotheses.

A programme is progressive when theoretical changes lead to successful new predictions.

It is degenerating when changes merely explain away failures after the fact.

This preserves Popper’s emphasis on empirical risk while fitting actual scientific history better.

Falsifiability and Pseudoscience

Despite limitations, falsifiability remains a useful diagnostic.

Ask of a claim:

• What observation would count against it?
• Are failure conditions stated in advance?
• Does the theory forbid anything?
• Are failed predictions acknowledged?

A claim with no possible failure condition is scientifically weak.

But Falsifiability Is Not Enough

A random prediction is falsifiable.

So is a badly designed hypothesis.

Science also requires quality.

A strong scientific framework should ideally have:

• evidential support,
• explanatory coherence,
• reproducible methods,
• integration with other knowledge,
• progressive problem solving.

Falsifiability is one virtue among several.

String Theory and Falsifiability

Modern theoretical physics has renewed debate about Popper.

Some theories involve scales inaccessible to current experiments.

Does lack of present testability make them unscientific?

Not necessarily.

A research programme can be scientific if it seeks empirical consequences and remains constrained by established physics.

But indefinitely escaping test creates tension with empirical science.

The boundary can be temporal rather than absolute.

Evolution and Falsifiability

Evolution is sometimes falsely claimed to be unfalsifiable.

In reality, evolutionary theory makes many risky predictions.

For example, particular fossil sequences, genetic nested hierarchies, and biogeographic patterns are expected.

Certain findings would create serious problems.

A framework can explain many observations without being compatible with every imaginable observation.

General Relativity and Falsifiability

General relativity has faced repeated tests:

• Mercury’s orbit,
• light deflection,
• gravitational redshift,
• binary pulsars,
• gravitational waves,
• black-hole observations.

Its survival does not prove it final.

It means the theory has repeatedly survived opportunities to fail.

This is exactly the spirit Popper valued.

Science Does Not Seek Refutation Only

Actual scientists do not spend all their time trying to destroy theories.

They also:

• measure constants,
• extend models,
• solve puzzles,
• build instruments,
• explore implications.

Popper’s philosophy captures the logic of critical testing more than the full sociology of research.

Science is broader than falsification.

Negative Results Matter

One of Popper’s lasting contributions is cultural.

A failed result is scientifically valuable.

A null result can eliminate a hypothesis.

An experiment that does not confirm expectations can still advance knowledge.

This contrasts with publication cultures that reward only positive findings.

Science needs failures to remain visible.

Intellectual Courage

Falsifiability demands intellectual courage.

A scientist must specify conditions under which a favored theory would be abandoned or revised.

This is difficult.

Humans become attached to ideas.

Institutions reward success.

Careers can depend on research programmes.

Methodological openness to failure is therefore both logical and psychological.

A Better Popperian Principle

A practical interpretation of Popper is:

Prefer theories that expose themselves to severe empirical tests and make it possible for reality to prove them wrong.

This avoids pretending one failed observation always destroys one theory.

It preserves the deeper insight:

knowledge grows when ideas take risks.

The Next Question

Popper reacted partly against verificationism.

But before falsifiability became the dominant popular slogan, philosophers associated with logical positivism had tried to build science around verification, observation, and meaning.

Their project was enormously influential and ultimately difficult to sustain.

The next question is:

What was verificationism, and why did it fail as a complete philosophy of science?