Does TsT preserve physics because it's a true duality, or just an effective mathematical reparametrization?

Your question cuts to the heart of a deep debate in modern physics: are dualities genuine physical equivalences, or are they just clever mathematical tricks that describe the same physics in different languages?

The short answer from the archive: it depends entirely on what you mean by "preserve physics," and this distinction matters enormously for understanding gravity.

The Equivalence Problem

A duality is a true equivalence only if the two formulations are theoretically equivalent — meaning they make identical predictions about every observable and contain no extra, unobservable structure [AG-2025.08-036]. But here's the catch: many apparent dualities in gravity fail this test [AG-2024.06-418].

Consider teleparallel gravity, which describes gravity using spacetime *torsion* (twisting) rather than *curvature* (bending). At first glance, teleparallel reformulations of general relativity look like a duality: same equations, different geometric language. But [AG-2024.06-418] argues they are not equivalent — teleparallel gravity posits strictly more mathematical structure (a flat parallel transport) than general relativity requires. If that extra structure is unobservable, it's excess baggage; if observable, it's a different theory.

What Makes a Real Duality?

The philosophical consensus [AG-2025.08-036] distinguishes three cases:

1. True duality: Two formulations that encode identical physics and make identical predictions. Example: the electric-magnetic duality in electromagnetism.

2. Effective mathematical reparametrization: Same physical content, but one formulation introduces unobservable auxiliary structures. This is what [AG-2024.11-413] calls "a new language for describing the same physics, in terms of absolutely unobservable and unpredictable geometrical inventions."

3. Distinct theories: Different physical content dressed in similar mathematical clothes.

Why This Matters

The distinction is not academic. In modified gravity theories (like $f(T)$ gravity, where you add functions of torsion), the "extra" structure of teleparallel formulations becomes *physical* — it changes predictions [AG-2024.05-450]. So whether you call it a duality or a reparametrization determines whether you've discovered a new theory or merely relabeled an old one.

The archive does not provide analysis of TsT transformations specifically — those appear to be a string theory or AdS/CFT construction not covered by these preprints.