Accession Number : ADA255506


Title :   Full Belief,


Corporate Author : ROCHESTER UNIV NY DEPT OF PHILOSOPHY


Personal Author(s) : Kyburg, Jr, Henry E


Full Text : https://apps.dtic.mil/dtic/tr/fulltext/u2/a255506.pdf


Report Date : Jan 1983


Pagination or Media Count : 25


Abstract : It may be questioned whether or not there is any such thing as full belief. One epistemological option is to suppose that, for any agent, every statement in his language bears a number between 1 and 0 reflecting the degree of belief the agent has in that statement, and (presumably, for ideal agents) satisfying the axioms of the probability calculus. One prima facie difficulty with this view is that it is conventional to look on changes in the epistemic status of statements as stemming, often if not always, from conditionalization. But conditionalization requires that the evidence on which conditionalization is done be given probability 1: P'(H) = P9H and E)/ P (E). But now P' (E) = 1, and on the view being discussed no statement, other than a priori truths, should receive probability 1. However, this is not an insuperable problem: there are ways of representing shifts in probability that do not require that any statement in our language be given probability. But as Diaconis and Zabell showed, every reasonable way of representing probability shifts can also be represented by conditionalization in an enriched language.


Descriptors :   *ARTIFICIAL INTELLIGENCE , INTELLIGENCE , PROBABILITY , CALCULUS , STEMMING , NUMBERS , LANGUAGE


Subject Categories : Cybernetics


Distribution Statement : APPROVED FOR PUBLIC RELEASE