Review of Philosophy of Science in Practice

Philosophy of Science in Practice: Nancy Cartwright and the Nature of Scientific Reasoning Chao, Hsiang-Ke and Julian Reiss, Editors Springer International, Switzerland, 2017.

Nancy Cartwright is a thinker who has been a major force in turning the philosophy of science towards a realist notion of causation. Her work has stressed empirical appreciation for how scientists actually achieve the results they do, and turned away from the scientism of the "unity of science" project. She has also been instrumental in clarifying the nature of models in scientific reasoning, both in the natural and social sciences. Her recent focus has been on the role of scientific studies in policy debates.

This volume is a collection of essays inspired by her work, and it ranges across all of the above topics. As such, it is difficult to give a unified thematic summary of the book, so instead, I will focus on a few highlights of its contents. And although primarily aimed at philosophers of science, the work is not beyond the grasp of non-specialists.

On the topic of the application of scientific studies to real-world problems, in particular to the practice of medicine, Otávio Bueno and Robin Neiman address the vexed issue of when, if ever, clinical trials should be halted early because of accumulating evidence that the treatment being tested is either causing such harm that it should no longer be continued for the treatment group, or producing such benefits that denying it to the control group is unethical. But their effort suffers from vagueness and tautology: they complain, for instance, that stopping the trial early only constitutes "partial evidence" for the hypothesis being tested, as though any "full" trial could ever provide "full evidence." They worry that in such circumstances "the emerging pattern in the data may turn out to be misleading: additional data could overthrow that pattern" (41). But obviously this worry also applies to "full" trials! They present such vacuous truisms as "one will eventually get full evidence [for a hypothesis, or] otherwise one won't" as genuine insights. They claim that after a partial trial, "additional information needs to be gathered before any epistemic decision is made" (41), not noticing that "additional information needs to be gathered" is an "epistemic decision." They contend that evidence does not and should not depend on the beliefs of the person evaluating the evidence. But evidence is always evidence to someone, and whether something counts as evidence will depend upon their beliefs.

In the most frustrating paper in the book, which addresses Cartwright's work on the unity of science, Michael Strevens defines 'fundamentalism' as the notion that "Everything is made up of a single kind of stuff and everything... is directed solely by fundamental laws of physics " (69). He goes on to claim that fundamentalism implies that all sciences really should operate by showing how, say, mate selection in bower birds, or the nature of parliamentary institutions in Medieval Europe, can be derived from the laws of physics alone.

Strevens backs his fundamentalist faith with the claim that "the empirical evidence for fundamentalism has accumulated swiftly" (69). But he presents no such evidence, and instead he admits, "Real science is not only largely disunified; it is largely content to be disunified" (71). He confesses that Nancy Cartwright has shown that not only are other sciences not being reduced to physics, but even within physics, a plethora of sub-studies are not being reduced to "fundamental" physics. But like a good Biblical fundamentalist arguing away the dating of dinosaur bones, Strevens argues that this evidence against fundamentalism is actually what we should expect if fundamentalism is true.

Roman Frigg and James Nguyen tackle yet another theme from Cartwright, that of how scientific models "represent" the phenomenon they model. Their answer is that a scientific model points to a class of events by declaration ("this is a model of the economy") but that it is successful as a scientific model to the extent it exemplifies particular aspects of the thing modeled. Thus, a model of the economy built from pipes and water is a good model if, say, the flow of cash through the real economy is exemplified by the flow of water through the pipes.

Space does not permit a more thorough examination of the many interesting papers in this volume (of particular note among the neglected is Julian Reiss's excellent paper on causality, and Cartwright's own paper on the validity of singular causes), but for anyone interested in the impact of Cartwright's work on the philosophy and the practice of science, this is an essential work.