Abstract

Books in this series have a first part providing an introduction to and history of a problem and a second part which builds up to the author's own views. For his historical treatment of the problems of explanation Ruben selects Plato, Aristotle, Mill, and Hempel. The space devoted to Plato and Aristotle is rather unexpected, but as with the other figures considered Ruben focuses on themes of relevance to current debates. Thus the discussion of Plato is largely on whether two "opposites" can explain the same thing and whether the same thing can explain two opposites. On Ruben's intricate analysis, the issue has a bearing on probabilistic dependency theories of explanation and on whether causes need be either necessary or sufficient for their effects. The discussion of Aristotle introduces among other things the idea that an adequate theory of explanation will have metaphysical underpinnings; it should, in other words, fit "what we think the world is like". Ruben is sympathetic to this idea and pursues it more fully later. Mill figures largely as a precursor of Hempel, but there is an interesting discussion of the relation between Mill's view that explanations are deductive arguments and his rather peculiar thesis that deduction is not "real" inference. Not surprisingly, the Hempelian account forms the main backdrop to Ruben's constructive analysis. Among the main topics discussed are the ontology of explanation, the standard counterexamples to the Hempelian account, the introduction of a causal condition into the analysis of explanation, the role of laws in relation to explanation, whether explanations are arguments, and whether all explanation is causal. The discussion is clear and thorough all the way through. In dealing with Ruben skillfully works up to the view that "facts explain facts only when the features and the individuals the facts are about, are appropriately conceptualized or named". This brings out the sense in which explanation is "epistemological." is crucial to the ensuing discussion, since if, as Ruben suggests, we "put the 'cause' back into 'because'" to deal with, the idea that laws must figure in explanantia, and the idea that explanations must be arguments, seem less compelling. Still, a further question is whether all singular explanations are causal. Ruben delivers a negative answer. There are explanatory identities. "A particular's having a property, described or conceptualized in one way, can explain the same particular's having the same property, described or conceptualized in another way". There is no self-explanation here given the favored ontology of explanation. There would be self-causation if such explanations were causal. So they are not causal. But even noncausal explanations explain something by "showing what is responsible for it or what makes it as it is". That, Ruben concludes, is what explanation is all about.