Priors

Every Inquiry Starts Somewhere

A prior is the confidence assigned before considering the newest evidence. Priors may come from base rates, causal knowledge, track records, known mechanisms, or the frequency of similar events. They are unavoidable because no mind receives evidence from nowhere.

The danger is not having priors. The danger is hiding them, freezing them, or smuggling prejudice into them.

Good and Bad Priors

A good prior is relevant, explicit, proportionate, and revisable. It says, "Before this new evidence, here is what similar cases led me to expect." A bad prior is rigid, identity-protective, or selected only because it favors the desired conclusion.

Base rates are often the most neglected priors. People prefer stories about this case, but this case belongs to some class, and the history of that class matters.

Priors and Updating

Bayesian updating does not let the prior dominate forever. Strong evidence can overwhelm a skeptical prior; weak evidence may barely move a strong one. The art is in letting prior and evidence interact without allowing either to become tyrannical.

Adjacent References

These are starting points rather than exhaustive citations. They connect this page to nearby work in probability, calibration, cognitive bias, and social epistemology.

• |Thomas Bayes, "An Essay towards solving a Problem in the Doctrine of Chances" (1763).
• |E. T. Jaynes, Probability Theory: The Logic of Science (2003).