On the foundations and necessity of classical gauge invariance

Abstract

We argue that, ideally, the ways to measure magnitudes in non-quantum theories of physics (spacetime, field theory), limit drastically their possible mathematical models. In particular, gauge invariance in the Yang-Mills framework, is a necessity of our way of measuring rather than an a priori imposition on symmetry. A general postulational basis for the geometric aspects of classical field theories is introduced, and the permitted models are studied. Some of them (for example, compatible with signature-changing metrics or variations of the speed of interactions) are new, and require a generalization of the concept of principal fiber bundle, which may be of interest both, physically and mathematically.