Artificial reasoning with subjective logic

Abstract. Judges and jurors must make decisions in an environment of ignorance and uncertainty for example by hearing statements of possibly unreliable or dishonest witnesses, assessing possibly doubtful or irrelevant evidence, and enduring attempts by the opponents to manipulate the judge’s and the jurors ’ perceptions and feelings. Three important aspects of decision making in this en-vironment are the quantification of sufficient proof, the weighing of pieces of evidence, and the relevancy of evidence. This paper proposes a mathematical framework for dealing with the two first aspects, namely the quantification of proof and weighing of evidence. Our approach is based on subjective logic, which is an extension of standard logic and probability theory, in which the notion of probability is extended by including degrees of uncertainty. Subjective Logic is a framework for modelling human reasoning and we show how it can be applied to legal reasoning