Is There an Ellsberg-Fellner Paradox? A Note on its Resolution

Abstract

The Ellsberg-Fellner Paradox can be shown to be nonparadoxical if Keynes' concept of weight-of-evidence is incorporated in decision making under risk and/or uncertainty. Keynes' concept is a more advanced version of C. S. Peirce's “second number” approach and is identical to R. Carnap's reliability-unreliability index. The Ellsberg-Fellner Paradox is paradoxical only to decision theorists who accept the Von Neumann-Morgenstern-Savage axioms. Decision theorists who accept the logical approach to probability can bypass this problem.