Epistemic Modeling with Justifications

Existing logical models do not fairly represent epistemic situations with fallible justifications, e.g., Russell’s Prime Minister example, though such scenarios have long been at the center of epistemic studies. We introduce justification epistemic models, JEM, which can handle such scenarios. JEM makes justifications prime objects and draws a distinction between accepted and knowledge-producing justifications; belief and knowledge become derived notions. Furthermore, Kripke models can be viewed as special cases of JEM s with additional assumptions of evidence insensitivity and common knowledge of the model. We argue that JEM can be applied to a range of epistemic scenarios in CS, AI, Game Theory, etc. 1 Context and motivations We make the case that formal epistemology should venture, when needed, beyond modal logic limits. This includes a formal treatment of justifications, which have been at the heart of epistemic discussions since Plato. Proof systems of justification logic and general purpose classes of models for these systems have been studied in [3, 4, 5, 7, 10, 11, 16, 17, 19] and many other sources. However, for formalizing epistemic scenarios, one needs specific domain-dependent models, with additional features that are not necessary for standard soundness and completeness analysis of proof systems. We introduce justification epistemic models, JEM, in which justifications are primary objects and a distinction is made between accepted and knowledge-producing justifications. In JEM s, belief and knowledge are derived notions which depend on the status