What the Equivalence Principles Do and Do Not Imply

Equivalence principles constrain how gravity is experienced locally; they do not uniquely determine how the gravitational field is generated. Universal test-body motion motivates a common geometry, and local Lorentz and position invariance motivate universal metric coupling of nongravitational matter. Additional assumptions are required to choose Einstein’s equation, its derivative order, its field content, its boundary data, or its global solutions.

Required background. We use special-relativistic local physics, observer-defined measurements, and the Newtonian gravity limit to identify what the principles actually test.

Helpful background. Tensor fields and natural constructions clarify the distinction between covariance and dynamics. Accelerated-observer physics supplies useful local counterexamples to overly global readings of equivalence.

What follows from the local principles

Within their test domains, the weak equivalence principle supports a universal family of free-fall trajectories. Adding local Lorentz invariance and local position invariance for nongravitational experiments yields the Einstein equivalence principle. A natural synthesis is metric gravity: a Lorentzian metric determines proper time and null cones, freely falling test bodies follow its geodesics, and nongravitational fields couple universally so that their local laws reduce to special relativity in freely falling frames.

Even this implication is partly constructive rather than deductive. Under Ehlers–Pirani–Schild-type regularity assumptions, universal unparametrized free-fall trajectories identify a projective structure and light propagation identifies a conformal structure. Light-cone compatibility determines a Weyl-compatible connection, not by itself the Levi-Civita connection of a unique Lorentzian metric; additional physical conditions are needed for that reduction and for a preferred clock parametrization. Matveev and Scholz prove the central compatibility step while keeping Weyl and Riemann compatibility distinct Matveev and Scholz 2020, §§ 2–4, pp. 3–6, PDF. The empirical principles strongly motivate this construction, but they do not erase its regularity and universality assumptions.

What is robust is the local hierarchy. Connection coefficients can vanish at one event, while curvature produces corrections of order $|R|L^2$ across a laboratory of size $L$. A freely falling central observer already has zero proper acceleration; an adapted coordinate system makes that geodesic motion simple to describe. Coordinates cannot remove a nonzero accelerometer reading, and they cannot remove tidal relative acceleration. Reall’s sequence from universality through local inertial frames to tidal gravity makes this boundary explicit Reall 2013, §§ 1.2–1.4, pp. 2–5, PDF.

What does not follow

EEP does not determine an equation for the metric. It does not select the Einstein tensor, fix Newton’s constant or the cosmological constant, prohibit extra gravitational fields, or choose boundary and initial data. It makes no global assertion about topology, geodesic completeness, horizons, or asymptotic structure. Nor does it say that curvature vanishes: curvature is precisely the locally irreducible tidal remainder.

EEP also concerns nongravitational experiments and negligibly self-gravitating test bodies. The strong equivalence principle extends the universality claim to local gravitational experiments and gravitational binding energy; that extension is substantially stronger and is not smuggled into the elementary statement. Will separates metric coupling inferred from EEP from the further restrictions placed on gravitational dynamics Will 2014, §§ 2.1 and 3.1, pp. 6–16 and 26–29, PDF.

General covariance is not the missing dynamical principle. Almost any local field theory can be expressed in generally covariant notation by introducing appropriate geometric fields. Covariance controls transformation behavior; it does not select which covariant equation nature obeys.

A discriminating comparison of theories

Einstein gravity with universally coupled matter can be written schematically as

$$S_{\mathrm E}[g,\psi] =\frac{1}{16\pi G}\int d^4x\,\sqrt{-g}\,(R-2\Lambda) +S_{\mathrm m}[g,\psi].$$

Now compare a scalar–tensor theory in a frame where ordinary matter still couples to the same metric:

$$S_{\mathrm{ST}}[g,\phi,\psi] =\frac{1}{16\pi G}\int d^4x\,\sqrt{-g} \left[F(\phi)R-Z(\phi)(\nabla\phi)^2-2U(\phi)\right] +S_{\mathrm m}[g,\psi].$$

If $S_{\mathrm m}$ is universal and the fields are smooth over a sufficiently small laboratory, neutral test matter can satisfy EEP in both theories. Yet variation of the second action produces additional scalar and metric equations, derivative terms involving $F(\phi)$, and gravitational behavior different from Einstein gravity. The local test-matter principle therefore cannot distinguish the two actions by logic alone.

Higher-curvature metric actions supply another comparison. Terms such as $R^2$ can preserve diffeomorphism covariance and universal matter coupling while changing the field equation and adding degrees of freedom. Will’s scalar–tensor and metric $f(R)$ examples make this separation between universal matter coupling and gravitational dynamics concrete Will 2014, §§ 3.1 and 3.3.2–3.3.3, pp. 26–29 and 34–36, PDF. A uniqueness theorem can recover the Einstein tensor only after imposing hypotheses about dimension, locality, field content, divergence freedom, and derivative order. Those hypotheses are independent of EEP.

Boundary cases and failure modes

Pointwise nonminimal couplings of ordinary matter can violate EEP even when their coefficients are universal. For example, a term $\xi R\phi^2$ leaves $R(p)$ in the local scalar equation after normal coordinates have removed the connection, so the law at $p$ is not the corresponding special-relativistic law Okon and Callender 2011, § 5, p. 142, Eq. (6). Species dependence introduces an additional failure: a composition-dependent scalar charge can make test-body trajectories differ and thereby violate WEP. By contrast, higher-curvature terms confined to the gravitational action can alter metric dynamics without altering the universal local matter law, while finite-size tidal operators lie outside the point-particle limit. A preferred tensor background can preserve universality of fall while violating local Lorentz invariance, and a varying dimensionless coupling can preserve local Lorentz symmetry while violating local position invariance. The components and domains must therefore be tested separately.

Global acceleration–gravity analogies also have limits. Uniform acceleration in flat spacetime can imitate a homogeneous gravitational acceleration locally, but it cannot reproduce arbitrary Riemann curvature. Conversely, a small freely falling laboratory in curved spacetime looks inertial at its center even though a larger laboratory detects tides. No extension from the pointwise statement to the entire spacetime is justified without additional geometry.

Exercises

Two theories share the same universal matter action $S_{\mathrm m}[g,\psi]$, but one has the Einstein–Hilbert gravitational action and the other adds $\alpha R^2$. Explain why observing EEP for test matter does not force $\alpha=0$.

Solution

EEP constrains the local coupling and motion of nongravitational test matter, represented here by the common $S_{\mathrm m}[g,\psi]$. The $R^2$ term changes the equation determining $g_{\mu\nu}$ but need not change the universal local matter coupling. Setting $\alpha=0$ requires an additional dynamical or empirical restriction—such as a derivative-order hypothesis plus a suitable uniqueness theorem—not EEP alone.

References

• Matveev, Vladimir S., and Erhard Scholz. “Light Cone and Weyl Compatibility of Conformal and Projective Structures.” General Relativity and Gravitation 52, 66 (2020). DOI; arXiv.
• Okon, Elias, and Craig Callender. “Does Quantum Mechanics Clash with the Equivalence Principle—and Does It Matter?” European Journal for Philosophy of Science 1 (2011): 133–145. DOI.
• Reall, Harvey S. General Relativity. Cambridge Part III lecture notes (2013). PDF.
• Will, Clifford M. “The Confrontation between General Relativity and Experiment.” Living Reviews in Relativity 17, 4 (2014). DOI; arXiv.