Statistical Defaults and Paraconsistency∗

This paper discusses the relationship between the property of paraconsistency and Statistical Default Logic, a nonmonotonic logic that models common inference forms found in classical statistical inference such as hypothesis testing and the estimation of a population’s mean, variance and proportions. Statistical Default Logic is paraconsistent in the sense that a set of default inference forms may induce an inconsistent set of conclusions yet not, by that fact alone, trivialize the statistical default consequence relation. The paper also clarifies a distinction between the studies of implication, reasoning and inference. This distinction is useful for understanding how to evaluate statistical defaults vis a vis probabilistic approaches to statistical argumentation and, it is suggested, useful for evaluating non-standard logics.