Falsifiability
If you can't prove it wrong, you can't really prove it right
Falsifiability is the principle that a theory belongs to empirical science only if it conflicts with possible experiences and is therefore capable of being proven wrong. Karl Popper argued that a good theory must have an element of risk: it must be able to be proven wrong under stated conditions. If a theory cannot be tested and potentially falsified, the best we can do is estimate its probability of being true.
Falsification works opposite to verification: instead of trying to confirm a theory by finding supporting evidence, you actively try to show the theory is incorrect. If you fail to disprove it despite rigorous testing, you have actually strengthened it. This mirrors natural selection in evolution: mutations that don't work are eliminated, strengthening the fitness of the surviving population. Theories that survive repeated attempts at falsification are more robust than those that have only been confirmed.
Popper applied this framework to critique pseudoscience and 'historicism,' the idea that history has fixed laws or trends leading to certain outcomes. He argued that trends are not destiny and that postulated 'laws' of history frequently become immune to falsifying evidence, with new evidence reinterpreted through the lens of the existing theory. Bertrand Russell's chicken parable illustrates the danger: daily feedings appear as a law until the day the chicken gets its head chopped off.
- A theory is part of empirical science only if it can be proven wrong by experience.
- Trying to disprove a theory and failing actually strengthens it more than finding confirming evidence.
- If observation shows the predicted effect is absent, the theory is simply refuted.
- Trend is not destiny; what appears to be a law may be revealed as merely a trend when conditions change.
- Theories that become immune to falsifying evidence have gained authority they haven't earned.
- State the theory in testable termsReformulate any claim, belief, or theory so that it makes specific predictions that can be tested against reality. If a theory cannot be stated in a way that allows for testing, note this limitation.Pro tipA good test: can you say 'If X happened, it would demonstrate that this theory is not true'?
- Design tests that could disprove itRather than looking for confirming evidence, actively design experiments or observations that would prove the theory wrong if they succeeded.Pro tipFalsification is the opposite of verification: try to show the theory is incorrect, and if you fail, you've actually strengthened it.
- Run the tests rigorouslyExecute your tests with intellectual honesty. Be willing to accept the results even if they contradict what you want to believe.Pro tipLike natural selection, eliminating what doesn't work strengthens what remains.WarningWatch for the temptation to reinterpret disconfirming evidence through the lens of your existing theory rather than updating the theory.
- Distinguish laws from trendsWhen evaluating historical patterns or predictions about the future, ask whether what appears to be a law is actually just a trend dependent on conditions. Conditions change.Pro tipThe worst events in history once smashed previous understandings of what was the worst. Prepare for extremes allowable by physics, not just what has happened so far.WarningWe tend to assume the worst that has happened is the worst that can happen. This is almost always wrong.
Popper argued that Freud's theory about repressed childhood memories influencing the unconscious is not possible to prove true or false. It does not make specific, testable predictions. The theory would need to be restated in a way that allows experience to refute it.
A chicken gets fed every day for its entire life. The daily feedings appear as a guaranteed law of nature. The chicken assumes feedings will continue in perpetuity based on consistent past experience.
Many claim that humans are subject to a law of ever-increasing technological complexity. We can certainly find confirmation for this trend. But Popper challenges whether this is an inviolable law or merely a trend dependent on conditions.
Karl Popper developed the concept of falsifiability as a demarcation criterion for distinguishing science from pseudoscience, most notably in his book The Logic of Scientific Discovery. He was particularly critical of theories like Freud's psychoanalysis, which he argued could not be proven true or false because they did not make specific, testable predictions.
Popper's attack on historicism, developed in The Poverty of Historicism, extended the concept to social and political theory. He challenged the idea that history follows fixed laws, arguing that claims like a 'Law of Increasing Technological Complexity' are not testable hypotheses but rather interpretations given undeserved authority. Bertrand Russell's famous chicken analogy from The Problems of Philosophy provided an accessible illustration: what appears to be an unchanging law (daily feeding) is revealed to be merely a trend when conditions change dramatically.