On the Minimization Principle in the Boolean Approach to Causal Discovery

Abstract

I examine the role of minimization in the Boolean approach to causal discovery, focusing on a recent development of the approach, implemented in a method called Coincidence Analysis (CNA). I present some prima facie counterexamples to the soundness of the minimization steps in CNA, or to the validity of a minimization principle that is naturally suggested by the minimization steps. I discuss two possible responses to the challenge, and argue that while one (but not the other) of them is viable, it renders the role of minimization steps inessential in an important sense. I end by suggesting that the trouble with the minimization principle arises out of locally uninstantiated regularities that regularity theorists have no reason to dismiss.