Foundation-theoretic Epistemology II: Conditional Justification and Non-circularity

Abstract

Abstract In this chapter, the ‘second half’ of the theory of first-order justification is elaborated: the theory of conditional first-order justification. The developed justification concept is not limited to the narrow boundaries of deductive-monotonic arguments, but it fully accounts for the probabilistic and defeasible nature of inductive and abductive arguments. A further crucial difference between these three kinds of inferences is that while deductions don’t allow the introduction of new relevant non-logical terms in the conclusion, inductive inferences introduce new relevant individual constants and theory-generating abductions introduce new relevant predicates. Contrary to a position that is criticized as ‘naive probabilism’, probabilistic support relations are not logically determined but have to be inductively supported by empirical evidence. A theory of justification nets is developed that allows us to reduce rule-circular justifications to premise-circular ones. It is shown that fully circular argument chains are epistemically worthless because they allow the justification of mutually contradictory beliefs. In contrast, justification circles that are additionally supported by independent evidence, so-called partial circles, are epistemically valuable. In the final section the notion of a conditional justification is formally explicated as similar to the notion of a ‘proof’ in logic as a sequence of propositions, with the difference that the involved rule steps need not be deductive but may also be inductive or abductive in nature, being annotated with associated conditional probabilities.