This paper describes the development of theories of scientific explanation since Hempel's earliest models in the 1940ies. It focuses on deductive and probabilistic whyexplanations and their main problems: lawlikeness, explanation-prediction asymmetries, causality, deductive and probabilistic relevance, maximal specifity and homogenity, the height of the probability value. For all of these topic the paper explains the most important approaches as well as their criticism, including the author's own accounts. Three main theses of this paper are: (1) Both deductive and probabilistic explanations (...) are important in science, not reducible to each other. (2) One must distinguish between (cause giving) explanations and (reason giving) justifications and predictions. (3) The adequacy of deductive as well as probabilistic explanations is relative to a pragmatically given background knowledge-which does not exclude, however, the possibility of purely semantic models. (shrink)

Addis (1981) has criticized a proposal of ours (Wilson [1969b]) for analysing disposition predications in terns of the horseshoe of material implication, and has proposed a related but significantly different analysis. This paper restates the original proposal, and defends it against Addis's criticisms. It is further argued that his proposal will not do as a general account of disposition predications; that, however, if it is suitably qualified, then it does account for certain special sorts of disposition predication; but that so (...) understood, it can be seen to be but a special case of ours. (shrink)