How (not) to be a pragmatic Humean

Abstract Recently, philosophers have argued that the laws of Lewis’s Best System Analysis (BSA) are not pragmatically useful. In this paper, I argue that these criticisms are substantive only because the BSA laws are characterized independently of their pragmatic roles. It is important to maintain the distinction between what the laws are and what the laws should do if we are to engage in the traditional, metaphysical debates found in the laws literature. While it is plausible that usefulness is an important criterion by which we judge the success of a Humean best system, I argue that we should not take usefulness to constitute a Humean best system. In response to the shortcomings of Lewis’s BSA, many authors have developed new “pragmatic” Humean best system accounts. I argue that these accounts are not characterized independently of the useful roles they are supposed to play for us. This leaves them vulnerable to objections that they may be stipulating the virtues they claim to explain, or they may fail to be Humean best system accounts at all.