Much of the reasoning we carry out on a daily basis is based on uncertain changing information,and takes us to uncertain interim conclusions. Trivalent and probabilistic logics have played a cen-tral role in developing ways of modelling uncertain reasoning with precision, allowing us to makeempirical predictions and develop theories about what types of uncertain inferences are more or lessjustified, and why. We review recent logical, philosophical and psychological literature on this topic,focusing on two Bayesian approaches to conditionals: A trivalent and a probabilistic one. Drawingon differences between the two conditionals in their representation of probabilistic dependencies,counterfactual scenarios, and their use in inferences, we argue that overall, probabilistic conditionalsare more expressive, general, and empirically compelling as models for reasoning under uncertainty.
Cruz et al. (Fri,) studied this question.