On the quality of auxiliary assumptions

Abstract

In a recent contribution to this journal, David Trafimow and Joshua Uhalt (2015) argue against the tradeoff between predictive power and explanatory breadth. In their view, it is the quality of the auxiliary assumptions that allows one to make testable predictions; hence it is possible that theories of considerable explanatory breadth, when combined with the right quality auxiliary assumptions, lead to testable predictions. Unfortunately, they leave the notion “quality” unspecified. In this review, I consider three possible properties that might render auxiliary assumptions capable of yielding testable predictions, namely observability, detail, and precision. All three proposals are rejected. I end with an appeal to further reflect on the role of auxiliary assumptions in deriving testable predictions.