Could the disagreement over "dynamical equivalence" stem from whether unobservable geometric degrees of freedom should count as physically distinct theories?

Your question gets at the heart of a deep philosophical disagreement in contemporary gravity research—and yes, the preprints confirm that much of the tension hinges on exactly this point.

The Core Disagreement

The "dynamical equivalence" dispute between General Relativity (GR) and its alternatives like teleparallel gravity (TEGR) centers on whether two theories that make *identical physical predictions* should be considered the same theory or distinct ones. The disagreement arises because:

GR and TEGR describe spacetime geometry differently. In GR, gravity arises from spacetime *curvature* (bent geometry). In TEGR, spacetime is flat but possesses *torsion* (a twisting structure) [AG-2024.06-418]. Both theories yield the same equations of motion and make identical observational predictions [AG-2025.04-820].

The question becomes: if the extra geometric structure in TEGR (the torsion) is unobservable—you can't design an experiment to detect it directly—does it count as a real physical difference?

What the Archive Shows

The "stronger structure" view: One camp argues these are genuinely distinct theories. TEGR postulates *more mathematical structure* than GR does—it specifies both a metric and a torsion tensor, whereas GR specifies only the metric. By the criterion of categorical equivalence (a rigorous mathematical standard), they are not equivalent [AG-2024.06-418]. This view treats unobservable geometric degrees of freedom as real parts of the theory's content.

The "gauge redundancy" view: The opposing position (implied in [AG-2025.01-442]) treats the choice between curvature vs. torsion as a *gauge choice*—a human decision about how to mathematize the same physics. Under this reading, the unobservable geometric structures are artifacts of representation, not features of reality.

Why This Matters

This disagreement has concrete consequences. If TEGR genuinely posits extra structure, it makes stronger metaphysical claims about spacetime than GR—claims we cannot empirically verify. If it's merely an alternative gauge choice, then realism about spacetime curvature (the intuitive view Einstein had) remains viable [AG-2025.04-820]. The question also illuminates a broader philosophical problem: how much mathematical freedom can we allow in formulating physical laws before equivalence stops being a meaningful concept?

A related warning appears in the literature: some researchers have conflated the equivalence of different formulations without always acknowledging it, creating confusion in modified-gravity research [AG-2025.08-434].

Deeper Framework

The disagreement also reflects a larger shift in how physicists think about theories themselves. Rather than asking "which geometry is real?", some propose viewing theories as geometric objects with structure—where dualities and alternative formulations emerge naturally as features of the space of all models, not as failures of equivalence [AG-2025.03-475, AG-2025.11-200]. Under this view, the question "are they the same theory?" becomes less sharp; what matters is understanding the *structure* relating them.

Notably, the preprints suggest a subtler position: GR, TEGR, and their cousins are equivalent *dynamically* but not in all aspects—particularly in how they relate to the Equivalence Principle and what assumptions about spacetime they embed [AG-2025.01-442]. So the answer may be: they are equivalent for predicting observations, but not equivalent as complete descriptions of spacetime structure.