What mathematical framework distinguishes modified gravity from general relativity predictions?

# Mathematical Frameworks Distinguishing Modified Gravity from GR

The key distinction lies in which core principles of general relativity are preserved or violated. General Relativity rests on several mathematical foundations—local Lorentz invariance, gauge invariance, the equivalence principle, and Riemannian geometry—and modified theories diverge by relaxing one or more of these.

Three Main Categories of Deviation

[AG-2025.02-126] organizes modified gravity into three broad classes:

1. Metric theories preserving fundamental symmetries: These keep local Lorentz invariance and gauge invariance intact but modify the field equations themselves. The most famous example is $f(R)$ gravity, where the Einstein–Hilbert action is generalized from the Ricci scalar $R$ to an arbitrary function $f(R)$. This sounds like a small change, but it dramatically alters how spacetime curves in response to matter [AG-2024.11-168].

2. Theories breaking gauge invariance or Lorentz symmetry: These allow spacetime to have preferred directions or introduce additional scalar or vector fields that couple to gravity asymmetrically.

3. Beyond-metric theories: These abandon the Einstein equivalence principle itself—the assumption that locally, gravity disappears and physics becomes special-relativistic. Instead, they use non-Riemannian geometries. For instance, teleparallel gravity encodes gravity in spacetime *torsion* rather than curvature, and symmetric teleparallel gravity uses *nonmetricity* [AG-2025.01-442]. Remarkably, these three frameworks—curvature-based GR, torsion-based teleparallel gravity, and nonmetricity-based gravity—are dynamically equivalent in how they predict bulk motion, yet they differ in their conceptual relationship to the equivalence principle [AG-2025.01-442].

How to Tell Them Apart Observationally

The practical way to distinguish modified gravity from GR is through gravitational lensing and growth of cosmic structure. In modified gravity theories like Hu-Sawicki $f(R)$ and Dvali-Gabadadze-Porrati (nDGP) models, the effective gravitational "strength" changes with scale and cosmological era, producing measurably different Einstein radii, lensing probabilities, and time delays compared to GR predictions [AG-2024.11-168]. However, this creates a subtle problem: dark matter models with additional fifth-force interactions can mimic modified-gravity signatures. Only precision measurements of *gravitational redshift*—the frequency shift of light climbing out of a gravitational well—can disentangle true modifications of gravity from exotic dark matter [AG-2024.05-211].

A Unifying Example: Fractional Gravity

One emerging approach uses fractional calculus—a generalization of ordinary derivatives—to deform Newton's gravity with a single parameter $α$. When $α = 1$, you recover standard gravity; when $|α - 1| \ll 1$, the same modified potential describes Mercury's orbit, light deflection, and the universe's accelerating expansion with consistent precision [AG-2026.03-379]. This suggests modified gravity may not be exotic after all, but a natural scaling phenomenon.

---

Note: These preprints focus on classification and observational tests but don't dive deeply into the algebraic machinery (Lagrangian formalism, field equations) that would appear in a technical review; the optical lensing framework [AG-2026.04-190] offers an accessible entry point to those details.