General Covariance, Artificial Gauge Freedom and Empirical Equivalence

This dissertation updates the debate over the nontriviality of general covariance for Einstein's General Theory of Relativity (GTR) and considers particle physics in the debate over underdetermination and empirical equivalence. Both tasks are tied to the unexplored issue of artificial gauge freedom, a valuable form of descriptive redundancy. Whereas Einstein took general covariance to characterize GTR, Kretschmann thought it merely a formal feature that any theory could have. Anderson and Friedman analyzed substantive general covariance as the lack of absolute objects, fields the same in all models. Some extant counterexamples and a new one involving the electron spinor field are resolved. However, Geroch and Giulini diagnose an absolute object in GTR itself in the metric's volume element. One might instead analyze substantive general covariance as formal general covariance achieved without hiding preferred coordinates as scalar 'clock fields,' recalling Einstein's early views. Theories with no metric or multiple metrics make the age of the universe meaningless or ambiguous, respectively, so the ancient and medieval debate over the eternity of the world should be recast. Particle physics provides case studies for empirical equivalence. Proca's electromagnetism with some nonzero photon mass constitutes a family of rivals to Maxwell's theory. Whereas any Proca theory can be distinguished empirically from Maxwell's, the Proca family approaches Maxwell's for small masses, yielding permanent underdetermination with only approximate empirical equivalence. The weak nuclear force also displays a smooth massless limit classically, but not after quantization, recalling the instability of empirical equivalence under change of auxiliary hypotheses. The standard electroweak theory apparently permits a photon mass term and hence underdetermination, but possible further unification might not. The question of underdetermination regarding massive gravity is unresolved. Physicists often reformulate a theory to install additional fields and new gauge symmetries preserving empirical content. Post-positivist philosophers might judge the result a distinct and inferior theory, but physicists consider it a perhaps superior formulation of the same theory. Evidently a step back from naïve realism towards mathematical-empirical equivalence is appropriate for physical ontology and theory individuation. Artificial gauge freedom licenses a generalized Kretschmann objection, but the clock field case suggests a resolution.