Key Principle
The book's central conceptual chain: empirical content = improbability = testability = corroborability. A theory's empirical content is the class of basic statements it forbids (its potential falsifiers). The more a theory forbids, the more it says about the world — and the less logically probable it is. Science should therefore seek bold, improbable theories that risk the most and tell us the most.
Simplicity is identified with degree of falsifiability: a linear equation is simpler than a parabola because fewer data points suffice to refute it. Simplicity is not aesthetic — it is epistemological, measuring how much a theory exposes itself to testing.
Why This Matters
This inverts the inductivist value system. Inductivists seek high-probability theories (well-supported by evidence); Popper shows this path leads to vacuous tautologies. A tautology has maximal probability (it is compatible with everything) and zero content (it tells us nothing). The ideal scientific theory sits at the opposite end: maximal content, zero probability, maximal testability.
Without this linkage, there is no principled way to prefer one unfalsified theory over another. Conventionalists (Poincaré) chose the simplest system as the most convenient convention; Popper shows they should be chosen because they are the most testable — they tell us the most about the world and are the easiest to refute if wrong.
Good Examples
"Not for nothing do we call the laws of nature 'laws': the more they prohibit the more they say" (Chapter 1). A law that forbids nothing permits everything and says nothing. Newton's law of gravitation forbids a vast range of possible planetary motions — that is precisely why it is informative.
Linear vs. quadratic curves: A linear law (y = ax + b) can be refuted by three non-collinear points; a quadratic law (y = ax² + bx + c) requires four points to refute. The linear law is simpler because it is more falsifiable — it tolerates less and therefore claims more (Chapter 7).
Degrees of falsifiability as theory comparison: When two theories are unfalsified, prefer the one whose class of potential falsifiers is larger. It has survived more stringent tests by exposure to more potential counterexamples (Chapter 6).
Counterpoints
"But we want true theories, not just bold ones": Popper agrees — but truth is never demonstrable. What we can do is test boldly and eliminate errors. A bold theory that survives severe tests is better corroborated precisely because it was easier to refute. A timid theory that survives mild tests tells us little (Chapter 10).
Conventional simplicity: Some argue simplicity is merely a matter of taste or convenience. Popper's identification with falsifiability converts this aesthetic intuition into a measurable logical property: count the potential falsifiers (Chapter 7).
Dimension as measure: The "dimension" of a theory (the minimum number of data points needed to determine its parameters) provides a rough quantitative handle on degrees of falsifiability. Higher dimension means lower falsifiability. But Popper acknowledges this measure does not capture all aspects of comparative testability (Chapter 6).
Key Quotes
"Not for nothing do we call the laws of nature 'laws': the more they prohibit the more they say." — Karl Popper, Chapter 1
"We should prefer the better-tested theory, the one which, by its logical character, can be most severely tested." — Karl Popper, Chapter 10
"The content of a theory — which is the same as its improbability — determines its testability and its corroborability." — Karl Popper, Appendix *ix
Rules of Thumb
- When comparing theories, ask: which one forbids more? That one has more empirical content
- Prefer theories that risk more — they are more informative when they survive testing
- Simplicity is not aesthetic preference: a simpler theory is one that can be refuted by fewer data points
- A theory that accommodates any possible observation has zero content — it is scientifically worthless
- High probability is a symptom of low informative content, not a mark of scientific quality
Related References
- Falsifiability and the Logic of Deductive Testing — The falsifiability criterion that grounds this chain
- Corroboration — How a Theory Stands Up to Tests — How content and testability determine corroboration
- The Formal Probability Calculus and Zero-Probability Proof — The formal proof that universal laws have zero probability