Towards a Normative Theory of Scientific Evidence

A scientific reasoning system makes decisions using objective evidence in the form of inde­ pendent experimental trials, propositional ax­ ioms, and constraints on the probabilities of events. As a first step towards this goal, we propose a system that derives probability in­ tervals from objective evidence in those forms. Our reasoning system can manage uncertainty about data and rules in a rule based expert system. We expect that our system will be particularly applicable to diagnosis and anal­ ysis in domains with a wealth of experimental evidence such as medicine. We discuss limi­ tations of this solution and propose future di­ rections for this research. This work can be considered a generalization of Nilsson's "prob­ abilistic logic" [Nil86] to intervals and experi­ mental observations. *We gratefully acknowledge the work of Carolyn Sher in editing this work and the assistance of Moises Sudit regarding nonlinear programming.