AMBIGUITIES AND UNSTATED ASSUMPTIONS IN PROBABILISTIC REASONING

Ostensibly simple probabilistic reasoning problems are sometimes surpr isingly difficult, One source of difficulty is the omission from a pro blem description of information essential to an unambiguous understand ing of the situation. When this is so, assumptions must be made to per mit the computation of probabilities. Different assumptions lead to di fferent conclusions, and if the assumptions are not stated, disagreeme nts regarding problem solutions can occur, even among experts. The aut hor reviews several such problems and their treatment in the literatur e. He proposes accounts of why problems can be difficult even when not ambiguous and discusses some approaches that can make solutions easie r to find or understand.