Karl Popper

Works

One book on method, one on politics, and one collection that states the programme best.

• The Logic of Scientific Discovery (1934 in German, 1959 in English) sets out falsifiability, the asymmetry between confirmation and refutation, and the rejection of induction.
• The Open Society and Its Enemies (1945) applies the same anti-authoritarianism to politics: piecemeal reform against utopian design, and the critique of historical inevitability.
• Conjectures and Refutations (1963) is where the programme is stated most usably — bold conjectures, severe tests, and the claim that a theory forbidding more is better because it is more exposed.
• Objective Knowledge (1972) develops the account of knowledge without a knowing subject.

Central ideas

Four, and the first two are what these placements borrow.

• Falsifiability as demarcation: a claim belongs to empirical science if it forbids something observable. A claim compatible with every possible outcome is not thereby meaningless or worthless — it is simply not doing that job.
• The asymmetry: confirmations are cheap and can be found for almost anything, while a refutation is informative. So the test of a position is what it rules out, not what it accommodates.
• Against induction: corroboration is not probability of truth. A theory that has survived severe testing is not thereby likely, only not yet defeated — which is a weaker and more honest status than most methods claim.
• Piecemeal against utopian engineering: prefer reforms whose failures are visible and correctable to designs whose failure can only be total.

Distinctive vocabulary

Three terms whose casual use inverts them.

• Falsifiable: capable of being refuted by an observation. Not false, and not doubtful.
• Corroboration: having survived attempts at refutation so far. Emphatically not confirmation, support, or evidence of probable truth.
• Demarcation: the line between empirical science and everything else. Not a judgement of value — calling something unfalsifiable is not calling it worthless, though it is almost always used that way.

Debates and disagreements

The objections are strong enough that falsificationism does not survive them intact.

• The Duhem and Quine problem: a hypothesis is never tested alone. A failed prediction can always be absorbed by adjusting an auxiliary assumption, so a decisive falsification is rarely available in practice. Quine is a reference in this work's epistemological lineage, and the two positions genuinely conflict.
• Kuhn's objection that working scientists do not try to refute their paradigms and that anomalies are absorbed rather than treated as refutations; Lakatos's research programmes were an attempt to rescue the position from this.
• Feyerabend's objection that no methodology, falsificationism included, survives contact with the actual history of science.
• Whether his own demarcation excludes sciences it should not — evolutionary biology, cosmology and historical geology all make claims that are not straightforwardly falsifiable.

Intellectual relationships

He is defined as much by what he rejected as by what he built.

• The Vienna Circle is the adjacent movement he rejected: verification as a criterion of meaning is precisely what falsifiability replaces, and the difference is the whole argument.
• Hume supplies the problem of induction, which Popper claims to dissolve rather than solve.
• Kuhn, Lakatos and Feyerabend are the succession and the critique.
• Quine's holism, in this work's epistemological lineage, is the technical reason decisive falsification is hard to come by.
• Adorno was his opponent in the German positivism dispute; Adorno is a reference in this work's critical-theory lineage.

How it changes this reading

Popper carries ten placements here — two claims, two frameworks and four through-lines — and every one of them is an explicitly bounded analogy.

• The borrowing is a discipline rather than a criterion: a reading should be able to say what would change it. A framework that can absorb any evidence has stopped being answerable to what it reads, whatever its fluency, and that is the same structural complaint Popper made about theories that explain everything.
• The application to descriptions of people is the sharpest use, and it belongs to the calibration through-lines. A characterological description forbids nothing — whatever the person does next confirms the type — while a description at the level of an exchange names something a next instance could contradict.
• The bound is not decoration. Interpretations are not scientific hypotheses, readings of a life are not testable, and Popper would not have called any of this falsifiable. What is borrowed is the demand that a claim expose its own defeater, and that demand can be honoured without any of the epistemology that motivated it.
• They are not evidence about any reading's correctness or about anyone's conduct.

Useful comparisons

In the lineages he shares, he is the member who supplies a test rather than a way of reading.

• Popper supplies the demand that a claim name what would defeat it.
• Gadamer, in the same two lineages, supplies the account of why a reading is situated and answerable to its source rather than to a neutral standard.
• Levinas and Foucault, in the claim about machine interpretation, supply the ethical demand and the analysis of how fluent description acquires authority.
• The traditions named alongside them — cybernetics, testimony theory, discourse ethics, historical source criticism — supply procedures rather than philosophies.

Where the ideas meet

They agree on what makes an interpretation trustworthy.

• None treats fluency or coherence as evidence that a reading is right.
• All hold that a claim's answerability to something outside itself is what gives it standing.
• Each treats correction as a normal condition rather than as a failure state.

Where they part

The difference between a test and a relation is not small.

• Popper's answerability runs to evidence that could refute a claim; Gadamer's runs to a source that can refuse a reading. The first is a logical relation and the second a hermeneutic one, and only the second applies where there is no possible observation.
• Levinas's demand is ethical and asymmetrical: the other's claim on me does not depend on my being able to test it.
• Quine's holism, elsewhere in this work, is a direct objection to Popper rather than a complement — if no hypothesis faces the tribunal of experience alone, the decisive refutation the method requires is not generally available.

Limits

What these placements can carry here is narrow, and the boundary is the point.

• An interpretation is not a scientific hypothesis. Nothing in this work is falsifiable in his sense, and describing it that way would misuse the term.
• Naming what would change a reading is a discipline borrowed from him, not an application of his criterion.

Criticisms

The objections are serious and some are conceded even by sympathisers.

• Decisive falsification is rarely available, because auxiliary assumptions can always absorb a failed prediction — the single most damaging technical objection.
• Verisimilitude, his attempt to define closeness to truth and make scientific progress meaningful without induction, was shown not to work formally, and the programme has no accepted replacement for it.
• The historical claim that scientists proceed by attempted refutation is not supported by the record.
• His readings of Plato, Hegel and Marx in The Open Society are polemical and are widely judged unfair to the texts, whatever the force of the political argument.
• Treating unfalsifiable as a term of dismissal is his own habit as much as his readers', despite the demarcation criterion carrying no such judgement.

Common misreadings

Five uses these placements do not license.

• Using unfalsifiable to mean worthless or meaningless, which the demarcation criterion explicitly does not entail.
• Calling an interpretation falsifiable. Interpretations are not the kind of thing the criterion applies to, and this work borrows the discipline rather than the test.
• Treating a well-corroborated position as probably true, which is the inductive conclusion the programme exists to refuse.
• Citing falsifiability as a settled criterion of science without the Duhem and Quine objection that decisive refutation is rarely available.
• Using the demarcation to dismiss a discipline rather than to describe what kind of claim it is making.

What remains outside this idea

• It cannot establish that any reading here is correct or incorrect.
• It cannot convert a methodological discipline into evidence about a life.