Causality in Statistical Theories as an Example for the Idealization of Physical Laws

Abstract

When two physical theories describe the physical “facts” equally well, Poincaré has proposed to always adopt the “simpler” theory. While there may be doubts about the applicability of this proposal as a general criterion, it often describes the behaviour of scientists even in the case when the two theories do not describe the facts “equally well”: In practical work physicists usually tolerate a larger margin of inaccuracy if they can only use a “simple” theory. Simplicity may concern the mathematical part of the theory as well as its physical content. The precise meaning of “simplicity” and “approximation” is generally quite clear for theories which are “simplifications” of more a comprehensive theory. But in most cases no such theory is available as a standard of comparison. Yet even in the framework of one given theory physicists often introduce “simplifying assumptions”, called “idealizations” in the work of Ludwig. The purpose of this paper is to discuss some problems related to idealizations and the role of physical imprecision and approximation in this context.