A new approach to assessing the accuracy of beliefs.

In Guesswork, Sophie Horowitz develops a new theory of accuracy for degrees of belief. She proposes that we can understand accuracy in terms of all-or-nothing guesses in response to forced-choice questions. Horowitz argues that your credences, or degrees of belief, are accurate insofar as they license true guesses. (If you are asked whether it will be rainy or sunny tomorrow, for example, and you are more confident of rain, you will be licensed to guess “rain.”) The book explores applications of this theory of accuracy, as well as perennial questions about the connection between rationality and truth.

Horowitz shows that guessing can be used in service of the “accuracy first” approach to formal epistemology and suggests connections between the degreed-belief framework and topics in traditional epistemology, such as reliability and (full) belief. Throughout, she contrasts the guessing approach with epistemic utility theory, the most dominant recent movement in formal epistemology. She shows that although guessing is less powerful than epistemic utility theory, it is simpler, requires fewer controversial commitments, and can often be discussed in accessible, nontechnical language. Therefore, she argues, many popular principles in formal epistemology can be justified on simpler grounds than has been assumed.