Three Problems of Direct Inference

Abstract

Direct inference consists in inference from a premise describing the incidence of a property among a given population to a conclusion about the likelihood of a particular element of the population having the property in question. For example, from the premise that 2% of American males are doctors one may, in appropriate circumstances, draw the conclusion that the probability is 0.02 that Joe, a particular American male, is a doctor. Despite the apparent centrality of direct inference to human belief formation, the manner in which direct inference is to be justified is not well understood. Similarly, no one has succeeded (or even claimed to succeed) in articulating adequate criteria that specify the conditions under which respective instances of direct inference are correct. My dissertation addresses the three most well know problems of direct inference.