My PhD dissertation and a paper defending pragmatism about explanation in Philosophy of Science Matters explore the nature and role of chance in explanation in evolutionary biology. Philosophers of science, most notably Carl Hempel, claim that explanations require laws, and that they answer why questions. Although he provides clear and vivid examples drawn from the sciences, Hempel does not elaborate at length on his deeper understanding of why questions. Nonetheless, he is clear that he believes that explanations show that the event to be explained ought to have been expected. This informs his famous thesis of the structural symmetry of explanation and prediction. According to this thesis, all explanations would have been adequate before the fact for predicting the event to be explained, and all good predictions serve, after the fact, as good explanations. Accordingly, Hempel’s view is that improbable events cannot be explained. Just because they are improbable, their occurrence cannot be predicted; and after the fact, it is not possible to show that the event’s occurrence ought to have been expected.
This seems to me to be at odds with the explanatory strategies used by evolutionary biologists, who contrast explanations invoking natural selection with those invoking random drift, a chance process. “Is the evolution of [a trait, genetic difference, new species, etc.] due to natural selection, or drift” is a question asked frequently by evolutionists. Evolutionists see this question as a choice between competing explanations—selection or drift—-not as a choice between an explanation (selection) and none (drift).
The central work of my dissertation, which remains the focus of my work in the history and philosophy of science, is to describe and justify strategies of explanation invoking random drift. I take the contextualist, pragmatist position that explanations need not answer why questions, and that they sometimes answer how questions, for instance, “how did the human species come to diverge from other species in our genus?” The aim of how questions is to place events in sequence, formulating a narrative linking causes and effects, terminating with the event to be explained. It is not necessary to show that the event to be explained ought to be expected or ought to have been expected, and no laws of nature need to be cited. Indeed, explanations invoking drift do fit this narrative model. I indicate some of these explanations in the dissertation. The idea that narratives can be explanatory was introduced by William Dray in his book Law and Explanation, which concerns the problem of explanation in history. Michael Scriven has also elaborated similar views. The central proponent of the contextualist view today is Peter Achinstein, who supervised the thesis. Alexander Rosenberg, who also supervised the thesis, is its staunchest and most articulate enemy.
The viability of the contextualist-pragmatist position about explanation is the most serious issue facing my position. As befits an idea developed by someone as astute as Hempel and extended by other excellent philosophers over the course of five decades, Hempel’s position is especially resilient. I believe that undermining it requires consideration of fundamental issues about the nature and aims of scientific explanation.