Probabilistic Models of Coherence and Positive Relevance II

点击翻译按钮获取中文摘要

During the first funding period, we mainly focused on evaluating probabilistic measures of coherence with the help of general adequacy constraints and specific test cases, where judgement on the latter was also studied empirically. Furthermore, in cooperation with Glöckner and Bröder's project we took important steps into a comparison of probabilistic coherence measures and the constraint-satisfaction approach. Finally, we finish the comprehensive survey of probabilistic measures of confirmation in the remaining months. The second period is devoted to completing these studies and having a closer look at the rationality of coherence- and confirmation-based reasoning. First, we continue testing the descriptive adequacy of measures of coherence; and we put our computer programme to use in order to computationally study these measures' characteristics, such as their capacity to provide a generalisation of deductive equivalence. Second, we scrutinise the rationality of the search for coherence by analysing under which conditions and to what extent coherence increases the chances of coming to beliefs being highly likely or close to the truth. Third, we introduce a new species of coherence, namely focused coherence, in order to apply the corresponding measures to standard cases from the literature on defeasible reasoning and to deploy them in constructing precise models of unification. Fourth, we examine whether probabilistic measures of confirmation provide rational solutions to the infamous ravens paradox. Fifth, the final touch is given by compiling the results of both funding periods in a monograph.