The_Unexpected_Hanging_as_an_Epistemic_Causal_Fixed_Point_Problem_is_excellent

This paper presents a unified treatment of the Unexpected Hanging Paradox by distinguishing between two interpretations: a standard non-causal reading and a richer causal reading where the prisoner's knowledge functions as a condition within the law of execution. The paper argues that the classical backward-induction argument relies on a hidden epistemic conditional that is incompatible with the judge's announcement. To resolve the paradox in the causal case, the paper models the scenario as a self-referential epistemic-causal network. It introduces the concept of Reflective Stability as a necessary constraint on knowledge, arguing that a belief cannot constitute knowledge if its endorsement would undermine its own justification. Using this framework, the paper distinguishes between "loop-trapped rationality" and "hyper-rationality", showing how a rational agent can suspend judgment to avoid epistemic fixed-point failures. This approach successfully blocks the "revenge" version of the paradox without abandoning classical logic.