Foundation-theoretic Epistemology III: Optimality Justifications

Abstract

Abstract The major justification load for the foundation-theoretic account developed in this book lies in the higher-order justification of inferences. Most pressing is the justification of induction (Hume’s problem), and of the abductive inference to external reality (Descartes’ problem); but in the light of contemporary non-classical logics the higher-order justification of logical frameworks also becomes a challenge. Many contemporary epistemologists consider these problems unsolvable. The central innovation of this book consists in the attempt at solving the higher-order regress problem by the method of optimality justifications, a novel research programme within the paradigm of epistemic engineering. Chapter 5 presents the programme of optimality justifications at a general level. An optimality justification does not demonstrate that an epistemic method is reliable; it pursues the more modest goal of demonstrating that a given method is optimal with regards to a given epistemic goal, for example, predictive success. The breakthrough of the optimality programme lies in its application at the level of meta-methods. Meta-methods use the track records and outputs of accessible object-level methods as their input and try to construct from this input a universally optimal meta-method; one that is never worse but is possibly better than all accessible alternative methods. Another important cornerstone of the optimality programme is the optimality principle. It leads from justifiably optimal methods to the rational acceptance of the beliefs recommended by these methods, under the proviso that we are forced to act as if we believe in one of the competing belief candidates generated by the competing methods.