Causal Methodology. A Comment on Nancy Cartwright's Hunting Causes and Using Them

摘要

In the introduction of her book, Nancy Cartwright states that with respect to causation ‘[m]etaphysics, methods and use must march hand in hand’ (1). By this she means that philosophers of causation should try to develop an integrated account, which answers three interrelated questions about causation: What do causal claims mean? How do we confirm them? What use can we make of them? She complains that philosophy traditionally only tries to answer the first question. I want to start this article by stressing that I agree wholeheartedly.1 I have tried to develop such integrated accounts for the social sciences in Weber (2007a, 2007b). nIn this comment, I will focus on methods for hunting causes. First, I will argue that there is no philosophically interesting distinction between methods that ‘clinch the conclusion’ and methods that ‘merely vouch for a conclusion’ (a distinction introduced in Chapter 3 of the book). Second, I will argue that – in economics and elsewhere – scientists should combine different types of evidence in order to warrant causal claims. This seems to go against what I call Cartwright's ‘methodological situationism’ (I will explain later what this means). Finally, I will argue that simulations (laboratory experiments) are important for economics, especially because they have high policy relevance. Given the overall idea of the book (with its emphasis on policy and use) I had expected a chapter or part of a chapter on this type of evidence.