Unification and Abductive Confirmation

Citation data:

THEORIA. An International Journal for Theory, History and Foundations of Science, ISSN: 0495-4548, Vol: 31, Issue: 1, Page: 107-123

Publication Year:
Usage 244
Downloads 96
Full Text Views 86
Abstract Views 58
Link-outs 4
Captures 3
Readers 3
Repository URL:
Niiniluoto, Ilkka
UPV/EHU Press; Euskal Herriko Unibertsitatea / Universidad del PaĆ­s Vasco
Arts and Humanities
article description
According to the traditional requirement, formulated by William Whewell in his account of the "consilience of inductions" in 1840, a scientific hypothesis should have unifying power in the sense that it explains and predicts several mutually independent phenomena. Variants of this notion of consilience or unification include deductive, inductive, and approximate systematization. Inference from surprising phenomena to their theoretical explanations was called abduction by Charles Peirce. As a unifying theory is independently testable by new kinds of phenomena, it should also receive confirmation from its empirical success. The study of the prospects of probabilistic Bayesianism to motivate this kind of criterion for abductive confirmation is shown to lead to two quite distinct conceptions of unification, linking up and screening off, and in both cases the unifying theory can be seen to receive probabilistic support from empirical phenomena.