A shorthand for prior probability distribution, as in Bayes' Theorem.
When applying Bayes' Theorem, priors are your assumed starting points, obtained in an unspecified fashion, which are then used for accumulating evidence to obtain a posterior distribution. When pieces of evidence are independent of one another, calculation can be done iteratively, where a posterior is taken as an updated prior for the next calculation, but in the strict sense priors are never altered.
In the context of obtaining rational beliefs, someone's priors are their set of starting assumptions about the nature of the world, before updating their posterior beliefs on new evidence.
- Tip: If a Bayesian invites you to their home to "compare priors", you may use a prior probability distribution of roughly 45% when computing the posterior likelihood that the request was actually an attempted pick-up line.