Causal Knowledge as a Resource for Selective Scientific Realism

The main challenge for selective scientific realism is to find reliable and prospectively applicable criteria on which parts of scientific theories one should be realist about. Many of the criteria proposed by selective realists have been shown by antirealists not to yield the correct results when applied to cases from the history of science, or at least to depend in their application on the benefit of hindsight. Taking account of these past failures, I propose a new version of selective scientific realism, inspired by Anjan Chakravartty’s idea of restricting realism to those posits of which we have detailed causal knowledge. Since this proposal bears some resemblance to earlier realist ideas, such as Stathis Psillos’s attempt to identify the “core causal description” of theoretical entities or Philip Kitcher’s distinction between “working posits” and “idle wheels”, I will first show that my proposal does not repeat the shortcomings of these earlier strategies. I will then seek to make precise the relevant notion of causal knowledge by elucidating three (individually necessary and jointly sufficient) criteria for those instances of inference to the best explanation generating the kind of knowledge that warrants realism. The virtues of this kind of selective realism will become apparent in my discussion of how the realist should respond to Kyle Stanford’s argument from unconceived alternatives. In the final part of my talk, I will briefly present two case studies that illustrate and support my proposal. The first one concerns Jean Perrin’s experimental work on the atomic hypothesis. Drawing on Stanford’s criticism of earlier realist treatments of this historical episode, I will show that the criteria previously introduced furnish a better account of the importance of Perrin’s work for establishing realism about atoms and molecules. The second case study demonstrates the strength of selective scientific realism in making sense of what particle physicists call the “direct detection” of a theoretically predicted particle. Analyzing the discovery of the neutrino, I show that my version of selective realism precisely captures what differentiates the direct detection of a particle from other (less convincing) means of confirming its existence. Reference: Egg, Matthias, “Expanding Our Grasp: Causal Knowledge and the Problem of Unconceived Alternatives”. British Journal for the Philosophy of Science 67 (2016), 115-141