Summary

Henri Poincaré argues that mathematical physics is not a mere tautology or a dangerous auxiliary to experiment, but a necessary tool for generalizing observations into hypotheses. He insists that experiment is the sole source of truth, yet without generalization—which always involves hypothesis—we cannot use our observations. Poincaré contends that a hypothesis that fails verification is not sterile; it reveals something unexpected and new, often serving science better than a true hypothesis by prompting decisive experiments. He warns that the most dangerous hypotheses are tacit and unconscious ones, which we cannot discard because we do not know we hold them. Mathematical physics, by its precision, forces us to formulate all our hypotheses explicitly. Poincaré also cautions against multiplying hypotheses indefinitely, as a theory built on many premises cannot be tested cleanly—if experiment condemns it, we cannot tell which premise to change. The reader takes away a nuanced view of hypothesis as both necessary and risky, and a method for using mathematical physics to discipline scientific reasoning.

Key concepts

  • Tacit and unconscious hypothesesHypotheses made without awareness, which are the most dangerous because they cannot be deliberately abandoned or tested.
  • Generalisation as hypothesisEvery generalization from observed facts is a hypothesis, and must be submitted to verification as soon as possible.
  • Sterile vs. fertile hypothesisA hypothesis that fails verification is fertile if it leads to a decisive experiment and reveals something unknown, whereas a true hypothesis may only confirm the expected.
  • Multi-hypothesis theory problemA theory built on multiple hypotheses cannot be tested cleanly—if experiment condemns it, it is impossible to tell which premise must be changed.
  • Mathematical physics as hypothesis-formulatorBy its precision, mathematical physics compels scientists to formulate all hypotheses explicitly, countering the danger of tacit assumptions.
  • Approximate simplicity of natureThe simplicity of laws like Kepler's is only apparent and approximate, not rigorous, and this approximate simplicity allows them to apply to analogous systems.

From the book

Title: Science and Hypothesis by Henri Poincaré
Title: Science and Hypothesis by Henri Poincaré

Popular questions readers ask