Research Group
in Analytic Philosophy

Abstract

The best systems account (BSA) of laws of nature consists of two components (Dorst, 2019): the “Humean base” and the “nomic formula”. The former postulates a humean fundamental ontological structure of the world: a mosaic of fundamental nonmodal facts to which all other facts reduce. The latter consists of a certain operation that gets applied to the Humean base and generates the laws (according to Lewis, this formula involves a balance of simplicity and strength among all the true systems describing the mosaic; laws are then the regularities described by the system that best balances strength and simplicity). Lewis’ version of the BSA and other recent modifications pretend to satisfy two properties: humeanism (all modal facts reduce to some nonmodal facts) and objectivity (what are the fundamental laws is not to be explained by psychological facts). I will show that the “nomic formula” consists of two components: some variables and some operation (balancing) that combines them. New versions of the BSA have focused on modifying the variables postulated by Lewis, but no attention has been devoted to how the function (balancing) or “the best” (balanced) are to be defined. Appealing to multi-objective optimization problems (MOOPs), I will argue that the BSA can’t assume, without further argument, the existence of a unique, objective and a priori answer to which system is the “best” balanced. In MOOPs, the choice of a “best” solution is typically highly subjective, which would imply that the BSA can’t satisfy the objectivity property. Further, I will argue that even if the BSA could provide an objective way to define the balance function of the “nomic formula”, the possibility of “tied best systems” (cases of no laws) seems very likely to arise. The final conclusion would be that, according to the BSA, lawful worlds, if possible at all, are an extremely rare occurrence among all possible worlds.