The local justification of beliefs and hypotheses has recently become a
major concern for epistemologists and philosophers of induction. As
such, the problem of local justification is not entirely new. Most
pragmatists had addressed themselves to it, and so did, to some extent,
many classical inductivists in the Bacon-Whewell-Mill tradition. In the
last few decades, however, the use of logic and semantics, probability
calculus, statistical methods, and decision-theoretic concepts in the
reconstruction of in- ductive inference has revealed some important
technical respects in which inductive justification can be local: the
choice of a language, with its syntactic and semantic features, the
relativity of probabilistic evalua- tions to an initial body of evidence
or background knowledge and to an agent's utilities and preferences,
etc. Some paradoxes and difficulties encountered by purely formal
accounts of inductive justification, the erosion of the once dominant
empiricist position, which most approaches to induction took for
granted, and the increasing challenge of noninduc- tivist epistemolgies
have underscored the need of accounting for the methodological problems
of applying inductive logic to real life contexts, particularly in
science. As a result, in the late fifties and sixties, several related
developments pointed to a new, local approach to inductive
justification.
