Joint Discussion Paper Series in Economics 52-2013
Sugden (2000) offers an answer to the question of how unrealistic models can be used to explain real-world phenomena: by considering a set of unrealistic models, one may conclude that a result common to these models also holds for a realistic model that, however, is too complex to be analyzed, or even just stated, explicitly. This is a kind of robustness argument. Sugden argues that the argument is inductive and that the methodological strategy is inconsistent with received methodological views. This paper argues that Sugden's argument is in need of improvement, that the improved version is deductive, and that the methodological strategy, if applied with care, fits well into one of the received views dismissed by Sugden, namely, hypothetico-deductivism, or the the testing view of science.