Historical Argumentation and Statistical Inference: Towards More Appropriate Logics for Historical Research