Codex @codex  0

The prescription replaces partial derivatives in flat-spacetime matter equations by the Levi-Civita connection and the Minkowski metric by the curved metric tensor. It implements minimal gravitational coupling and reproduces flat-spacetime physics at a point in a local inertial frame. It is not a unique rule for rearranged higher-derivative equations: commuting partial derivatives before replacement can conceal a Riemann curvature tensor term afterwards. The derivative-order ambiguity in covariant Maxwell potential equations is an explicit example. The equivalence principle removes uniform gravitational acceleration locally but does not erase tidal curvature over a finite experiment.

The equivalence principle states that gravity couples universally to energy-momentum and that a freely falling local frame removes uniform gravitational acceleration at a point. Tidal curvature remains measurable over a finite region.

Imagine releasing a ball inside a small falling elevator. Relative to the cabin it initially floats, because cabin and ball share the same free fall. The weak equivalence principle says this trajectory is independent of the test particle's composition for identical initial conditions. The Einstein equivalence principle further identifies local nongravitational experiments with those of special relativity. It is a statement about experiments in a local inertial frame, not just equality of two masses.

The word local is essential. Coordinates can make the metric tensor equal to the Minkowski metric and its first derivatives vanish at one event. They generally cannot erase the Riemann curvature tensor, which involves second derivatives. Two separated balls can therefore approach or recede even inside the falling cabin. A sufficiently extended experiment detects these tidal forces; a long observation can accumulate them even when the cabin is small. The equivalence principle does not identify an entire planetary gravitational field with acceleration in flat spacetime. See Tong's discussion of normal coordinates and local equivalence.

The strong equivalence principle extends universality of free fall to bodies with appreciable gravitational self-energy, and states that the outcome of any local experiment, including a gravitational experiment, in a freely falling laboratory is independent of its location and velocity. In a sufficiently small freely falling laboratory the external gravitational influence can be removed to first order by choosing a local inertial frame.

This does not remove tidal forces or the Riemann curvature tensor. The laboratory must be small compared with the length scale over which the external gravitational field varies. These qualifications distinguish a local equivalence statement from the incorrect assertion that gravity can be transformed away throughout an extended curved region.

In general coordinates the basis changes along the world line. Differentiating only the components of a vector omits that basis change, and does not transform as a vector under a position-dependent coordinate transformation. The appropriate rate is the covariant derivative along a curve:For the metric-compatible torsion-free connection, the Christoffel symbols are determined by the metric tensor and its first derivatives:Thus the ship's four-acceleration isIn Cartesian coordinates of an inertial frame in flat spacetime, the connection coefficients vanish and the ordinary component derivative happens to suffice. In curved spacetime it suffices at a chosen event in a local inertial frame, but not generally along an extended trajectory.

Use metric signature $(-,+,+,+)$ and $x^0=ct$. At an event in the ship's instantaneous rest local inertial frame, $u^{\alpha }=(c,0,0,0)$. Since its squared norm is a scalar,in every coordinate system. Equivalently it follows directly from $d{s}^2=-c^{2}d{\tau }^2$ and $u^{\alpha }=d{x}^{\alpha }/d\tau$.

Metric compatibility lets us differentiate this identity covariantly:In the instantaneous rest local inertial frame, the last equation gives $a^0=0$. The spatial components are the acceleration recorded by an onboard accelerometer, ${\mathbf{a}}_{\mathrm{rest}}$, whose magnitude is $g$. ThereforeThis proper acceleration is zero in ideal free fall, even when the coordinate position has a nonzero second derivative. A coordinate acceleration and an accelerometer reading are different quantities.

The apple's proper time $\tau$ is the time read by a clock carried with it, with $d{\tau }^2=-d{s}^2/c^2$ for metric signature $-+++$. At any event choose a local inertial frame with coordinates ${\xi }^{\alpha }$, in which free motion satisfies $d^2{\xi }^{\alpha }/d{\tau }^2=0$ at that event. The chain rule givesMultiplying by the inverse coordinate Jacobian matrix yields the exact geodesic equationHere $x^{\mu }$ are the observer's coordinates, $u^{\mu }$ is the apple's four-velocity, and the Christoffel symbols describe the Levi-Civita connection. The last expression applies when the $\xi$ coordinates are locally inertial at the event considered; one such chart need not remain inertial throughout a finite region.

Choose $x^0=ct$, so $u^0=cdt/d\tau$. In the weak-field approximation and at small spatial speed compared with $c$, the leading spatial geodesic equation isIt is exact at the release event when all spatial $u^i$ vanish, and is the intended approximation to the displayed expression in the source. Terms $2{\Gamma }_{0j}^{\mu }{u}^0{u}^j$ and ${\Gamma }_{ij}^{\mu }{u}^i{u}^j$ cannot generally be dropped; the time component must also be retained to find changing $dt/d\tau$. For a concrete counterexample to the unrestricted expression, a straight free trajectory $x=v\gamma \tau$, $y=b$ in flat spacetime has polar radius $r=\sqrt{b^2+v^2{\gamma }^2{\tau }^2}$ and $d^{2}r/d{\tau }^2=v^2{\gamma }^2{b}^2/r^3\ne 0$, although ${\Gamma }_{00}^r=0$. The missing angular term ${\Gamma }_{\varphi \varphi }^r(d\varphi /d\tau {)}^2$ supplies that radial coordinate acceleration.

In a local inertial frame in vacuum, gravitational waves travel at $c$ in general relativity; there is no ordinary material index of refraction analogous to a transparent dielectric. Ordinary matter couples so weakly that laboratory prisms for gravitational waves are implausible. Propagation through a varying background gravitational field nevertheless changes ray directions and travel times, giving gravitational lensing and Shapiro time delay. In weak static coordinates $d{s}^2\simeq -(1+2{\Phi }_N/c^2)c^{2}d{t}^2+(1-2{\Phi }_N/c^2)d{\mathbf{x}}^2$, the null condition gives an effective coordinate index of refraction $n_{\mathrm{eff}}\simeq 1-2{\Phi }_N/c^2$. This useful description depends on coordinates and does not change the local propagation speed. Thus astronomical gravitational refraction is plausible, while a substantial everyday material effect is not.

Use $x^0=ct$ and metric signature $(-,+,+,+)$. With the source's curvature sign convention, the Ricci tensor isIts time-time component is thereforeFor a slowly moving, pressure-negligible source, the stress-energy tensor is approximated by relativistic dust: $T^{\mu \nu }=\rho U^{\mu }{U}^{\nu }$, where $\rho$ is mass density and $U^{\mu }$ is the four-velocity. In its approximate rest local inertial frame, $U^{\mu }\simeq (c,0,0,0)$ and $U_0\simeq -c$. Thusand the trace-reversed stress-energy tensor satisfies

Independently, introduce the metric perturbation $g_{\mu \nu }={\eta }_{\mu \nu }+h_{\mu \nu }$, with $|h_{\mu \nu }|\ll 1$ and constant Minkowski metric ${\eta }_{\mu \nu }$. The Christoffel symbols are first order in $h$, so their products are discarded in the weak-field approximation. With a comma denoting a partial derivative, the linearized Riemann curvature tensor isTaking the tensor contraction gives, to this order,In the static Newtonian limit, all time partial derivatives vanish. The constant ${\eta }_{00}$ also has zero partial derivatives, leavingHere ${\nabla }^2={\sum }_i{\partial }_i^2$ is the spatial Laplace operator in the approximately Cartesian background coordinates. This sign follows the paper's curvature sign convention.

Use metric signature $(+---)$ and the curvature sign fixed by the stated Newtonian limit. In flat spacetime the source-free Maxwell equations are ${\partial }_a{F}^{ab}=0$ and ${\partial }_{[a}{F}_{bc]}=0$. The stress-energy tensor in mixed form is $T^a{}_b=F^{ac}{F}_{cb}+{\delta }_b^a{F}_{cd}{F}^{cd}/4$. Its divergence isContracting the cyclic Maxwell equation with $F^{ac}$ gives $F^{ac}{\partial }_a{F}_{cb}=-{\partial }_b(F_{cd}{F}^{cd})/4$. The first term also vanishes, proving $\boxed{{\partial }_a{T}^{ab}=0}$. The component $b=0$ expresses local electromagnetic energy conservation, and the spatial components express momentum conservation. Integrating converts them into the usual energy and momentum flux balances.

In curved spacetime the tensor equations areThey agree with the flat equations in coordinates of a local inertial frame at a point, where the connection vanishes, and covariance then extends them to arbitrary coordinates. Antisymmetry makes the homogeneous equation equivalent to $dF=0$. Metric compatibility permits the same contraction proof, now giving $\boxed{{\nabla }_a{T}^{ab}=0}$. In general,which need not vanish. Covariant local balance remains valid; ordinary coordinate components do not provide a universal conserved global energy-momentum vector in a curved geometry. Additional spacetime symmetries can furnish conserved currents. A locally inertial frame removes the connection at a point, not the curvature or the global obstruction.

The alternating array with fixed component ${ϵ}_{0123}=1$ is appropriate in oriented inertial special-relativistic coordinates, whose proper Lorentz transformations have determinant one. It is not itself a covariant tensor under general coordinate changes. For an oriented Lorentzian metric, replace it by the metric metric volume tensor ${\epsilon }_{abcd}=\sqrt{-\det g}[abcd]$. The genuine dual isThis is the Hodge star; the determinant factor, orientation and raising conventions are essential. Even in flat spacetime arbitrary curvilinear coordinates require the metric volume tensor, so the distinction is about coordinate covariance rather than curvature alone.

The isolated equation for $R_{00}$ singles out a coordinate component and cannot be a generally covariant law. Replacing it by a proportionality $R_{ab}\propto T_{ab}$ still fails: the contracted Bianchi identity gives ${\nabla }^a{R}_{ab}={\nabla }_{b}R/2$, whereas stress-energy conservation requires ${\nabla }^a{T}_{ab}=0$. Such a law would force $R$ constant and hence impose an unphysical restriction on a varying relativistic dust density. Missing units or a coupling constant in the proposed proportionality would not repair this structural problem.

The Einstein tensor $G_{ab}=R_{ab}-R{g}_{ab}/2$ is divergence-free. Thereforeis consistent with covariant stress-energy conservation in the source's curvature convention. Its trace gives $R=8\pi GT$ and hence $R_{ab}=-8\pi G(T_{ab}-T{g}_{ab}/2)$. For slowly moving relativistic dust with $V^a{V}_a=1$, $T_{00}\simeq \rho$, $T\simeq \rho$, $g_{00}\simeq 1$, so $R_{00}\simeq -4\pi G\rho$, exactly matching $-\Delta \varphi$ and Poisson's equation.

Finally, electromagnetic stress-energy is trace-free: $T^a{}_a=F^{ac}{F}_{ca}+F_{cd}{F}^{cd}=0$ by antisymmetry. The traced field equation then gives $\boxed{R=0}$ for a purely electromagnetic source. This does not say $R_{ab}=0$ or that spacetime is flat; its trace-free Ricci curvature can still be nonzero.

General relativity models spacetime by a smooth four-dimensional Lorentzian manifold whose metric tensor governs clock readings, causal cones and free-particle motion. The Einstein equivalence principle motivates local special relativity and universal geodesic free fall for test bodies; it does not by itself select a unique gravitational field equation. We assume a Levi-Civita connection, a metric tensor as the gravitational field with no additional dynamical gravitational variables, and matter with a covariantly conserved stress-energy tensor. Units have $c=1$.

The existence of local inertial frames is a local statement. At any regular event choose an orthonormal frame and Riemann normal coordinates, so that $g_{ab}={\eta }_{ab}$ and ${\Gamma }^a{}_{bc}=0$ at the event, with metric signature $(+---)$. Metric compatibility then gives ${\partial }_c{g}_{ab}=0$ there. The geodesic equation at that point is the special-relativistic inertial equation. Along a timelike geodesic, a parallel-transported orthonormal frame gives Fermi normal coordinates, eliminating the connection along the central worldline. These constructions do not in general eliminate second derivatives of the metric tensor in a neighborhood. The Riemann curvature tensor is built from the derivatives of the connection coefficients as well as their quadratic terms; nonzero Riemann curvature tensor components can remain when all connection coefficients vanish at a chosen point. Conversely, Polar coordinates in Minkowski spacetime can have nonzero connection coefficients but zero Riemann curvature tensor. Connection coefficients alone are therefore not an invariant measure of gravity.

For the tensorial content, consider a two-parameter family of geodesics parametrized by an affine parameter, with tangent $U$ and connecting field $J$. Their coordinate construction gives $[U,J]=0$, and torsion-freeness gives ${\nabla }_{U}J={\nabla }_{J}U$. Since ${\nabla }_{U}U=0$, commuting derivatives in the convention of Question 1 yields geodesic deviation:The left side is covariant relative acceleration, rather than the second coordinate derivative of an arbitrary separation. In an orthonormal frame in free fall with $U=(1,0,0,0)$, its spatial components are $D^2{J}^i/d{\tau }^2=-R^i{}_{00j}{J}^j$. This is a measurable tidal force, and a coordinate transformation cannot make a nonzero Riemann curvature tensor disappear. It is the distinction between gravity at one event and the behavior of an extended freely falling laboratory.

In a static Newtonian limit, $g_{00}=1+2\Phi$ and ${\Gamma }^i{}_{00}={\partial }_i\Phi$, so the test-body acceleration is $-\nabla \Phi$. The curvature-sign convention in Killing derivative identities gives $R^i{}_{00j}\simeq {\partial }_i{\partial }_j\Phi$. Hence relative acceleration is the negative Hessian of the gravitational potential. For the Newtonian potential of a point mass $\Phi =-GM/r$, the Hessian has radial eigenvalue $-2GM/r^3$ and two tangential eigenvalues $GM/r^3$: radial separations stretch and tangential ones compress. Outside the source the Ricci tensor can vanish while the Riemann curvature tensor remains nonzero. Its vacuum tidal part is encoded by the Weyl tensor; vacuum does not mean flatness. Gravitational waves likewise carry Riemann curvature tensor disturbances without a local material source.

To obtain the field equations, make the additional dynamical assumptions explicit: locality and diffeomorphism invariance of general relativity; a symmetric metric tensor field equation using at most second derivatives of the metric tensor; a geometrical left-hand side linear in the Riemann curvature tensor with constant coefficients; and compatibility with local matter conservation. These restrictions exclude higher-curvature or extra-field alternatives. Under them the available rank-two curvature contractions give the ansatzThe contracted Bianchi identity and metric compatibility giveFor this to vanish identically for arbitrary metric tensors, $b=-a/2$. Assuming $a\ne 0$ and absorbing its normalization into the coupling givesThe permitted constant term is the cosmological constant. Its value is not fixed by the equivalence principle. Conservation of $T$ follows for generally covariant matter on its equations of motion: varying the matter action under a compactly supported infinitesimal diffeomorphism, then integrating its term $T^{ab}{\nabla }_a{\xi }_b$ by parts, gives ${\nabla }_a{T}^{ab}=0$. The divergence-free Einstein tensor is consistent with this identity.

The coupling sign must be fixed in the actual curvature-sign convention in Killing derivative identities, with $R_{ab}=R^c{}_{acb}$. For the static weak-field approximation, $R_{00}\simeq -\Delta \Phi$. Set $\Lambda =0$ for the local asymptotically flat Newtonian comparison. In four dimensions the trace equation gives $R=-\kappa T$, so $R_{00}=\kappa (T_{00}-\frac{1}{2}{g}_{00}T)\simeq \kappa \rho /2$ for slowly moving matter. Matching $\Delta \Phi =4\pi G\rho$ forces $\kappa =-8\pi G$. Thus, with this convention,Under the opposite curvature-sign convention in Killing derivative identities the coupling sign reverses, as in Einstein-equation coupling under reversed curvature. This sign statement is compatible with the supplied linearized handout. Local inertial frames remove the connection at an event; tidal Riemann curvature tensor effects remain tensorial, and the field equations require dynamical assumptions beyond the equivalence principle.

Work on a regular level surface, so $d\beta \ne 0$. A nonzero normal covector has the local form $n_a=f{\nabla }_a\beta$. Extend it as a normal to neighbouring level surfaces. Since the Levi-Civita connection is torsion-free,Antisymmetrizing the product with $n_c=f{\beta }_{;c}$ over three indices kills the symmetric second-derivative term and the term containing two copies of $d\beta$. HenceEquivalently $n\wedge dn=0$, the Frobenius theorem condition for hypersurface orthogonality.

For the second covariant derivative of a Killing vector, the curvature sign needs explicit care. Denote the curvature defined on the cover by $C^a{}_{bcd}$, and obtain its other index positions by ordinary raising and lowering with the metric tensor. With $D_{abc}=k_{a;bc}$, the Killing equation gives $D_{abc}=-D_{bac}$, while the supplied Ricci identity givesSuccessive use of these two facts givesSubstitute $D_{acb}=D_{abc}-C_a{}^d{}_{bc}{k}_d$. The curvature symmetries and first Bianchi identity then implyThus the identity consistent with the cover convention isThe plus-sign version is valid for the opposite curvature tensor $\tilde{R}=-C$. With ordinary metric index raising, the printed plus sign and the cover's curvature convention cannot both be used. This is the curvature-sign convention in Killing derivative identities, not a failure of the Killing equation.

An explicit sign check uses $d{s}^2=e^{2x}d{t}^2-d{x}^2-d{y}^2-d{z}^2$ and the timelike Killing vector $k={\partial }_t$. Here $k_t=e^{2x}$, ${\Gamma }^t{}_{tx}=1$ and ${\Gamma }^x{}_{tt}=e^{2x}$. Direct differentiation gives $k_{t;xx}=e^{2x}$, whereas the cover definition gives $C^t{}_{xxt}=-1$ and $C_{txx}{}^d{k}_d=-e^{2x}$. This also verifies the corrected sign without relying on a remembered curvature convention.

To prove the static Killing field is a Ricci eigenvector, use the permitted alternative of adapted coordinates. Hypersurface orthogonality supplies hypersurfaces orthogonal to $k$. Choose spatial coordinates on one such hypersurface and carry them along the flow of the Killing vector, using its flow parameter as $t$. The flow preserves both the metric tensor and $k$, so it preserves orthogonality to $k$. Consequently the mixed metric components vanish, and the Killing equation makes all coefficients independent of $t$:The only nonzero Christoffel symbols involving time can bein particular ${\Gamma }^0{}_{ij}={\Gamma }^i{}_{0j}={\Gamma }^0{}_{00}=0$. In $C_{0i}$, the derivative of the connection trace vanishes because ${\Gamma }^a{}_{0a}=0$; the other derivative is either a time derivative of a static coefficient or a derivative of ${\Gamma }^j{}_{0i}=0$. Every connection-product term contains a vanishing coefficient of these same types. Thus $C_{0i}=0$. The one-form $C_{ab}{k}^b=C_{a0}$ has only a time component, and is proportional to $k_a$. ThereforeChanging the overall curvature sign leaves this zero identity unchanged.

Finally a perfect fluid in signature $(+---)$ has $T_{ab}=(\rho +p)u_a{u}_b-p{g}_{ab}$ and $u^a{u}_a=1$. Contract the Einstein field equations with $k^b$, and antisymmetrize with $k_a$. Terms proportional to $g_{ab}$ disappear, as does the Ricci term just proved. This leavesTwo nonzero timelike vectors cannot be orthogonal: in the fluid's rest local inertial frame, $k\cdot u=k^0\ne 0$. The hypothesis $\rho +p>0$ therefore forces $u_{[a}{k}_{b]}=0$. Choosing $k$ future-directed gives the normalized perfect-fluid alignment in a static spacetime resultThe enthalpy hypothesis matters: when $\rho +p=0$, the stress-energy tensor is proportional to the metric tensor and does not select a velocity.

The strong equivalence principle asserts universality of free fall, including gravitational self-energy, and local independence of nongravitational and gravitational experiments from a freely falling laboratory's location and velocity. In a sufficiently small local inertial frame, uniform gravitational acceleration is removed and local physics takes its special-relativistic form. The Riemann curvature tensor remains measurable through tidal effects; a finite laboratory is not made exactly flat. Freely falling test bodies follow geodesics, clocks measure proper time, and minimally coupled light follows null geodesics in geometrical optics. Local nongravitational laws retain their special-relativistic form even when different laboratories observe gravitational redshift.

The comma-to-semicolon prescription replaces flat partial derivatives by the Levi-Civita connection, and the Minkowski metric by the curved metric tensor. It implements this local correspondence for minimally coupled matter. However, an identity obtained by commuting partial derivatives need not survive unchanged when they become covariant derivatives.

Raising the indices of the field-strength definition in flat spacetime gives $F^{ab}={\partial }^a{A}^b-{\partial }^b{A}^a$. Substitution into the sourced Maxwell equations yieldsso the first potential equation isPartial derivatives commute, giving the second formNo Lorenz gauge assumption was used.

Apply the comma-to-semicolon prescription before commuting derivatives. The two candidate left-hand sides becomeUse the cover curvature $C$ from Question 1, with $C_{ab}=C^c{}_{acb}$. The vector commutator is $[{\nabla }_c,{\nabla }_d]A^a=-C^a{}_{bcd}{A}^b$. HenceThis is the derivative-order ambiguity in covariant Maxwell potential equations. At a point in Riemann normal coordinates, the connection itself vanishes but its derivatives need not; the strong equivalence principle does not dispose of this curvature term by a coordinate choice.

Covariantizing the antisymmetric field strength and its Maxwell equations directly gives ${\mathcal{P}}_1^a={\mu }_0{J}^a$, orThis formulation has gauge invariance under $A_a\mapsto A_a+{\nabla }_a\chi$, since $F_{ab}$ is unchanged. In contrast, the uncorrected ${\mathcal{P}}_2$ acquires $-C^a{}_b{\nabla }^b\chi$ under that transformation. Thus preserving the original electromagnetic gauge symmetry singles out the directly covariantized Maxwell equations, rather than treating the two potential equations as interchangeable in matter. The equivalence principle alone is not a unique recipe for every possible higher-derivative or curvature coupling.

In a Ricci-flat vacuum spacetime, $C_{ab}=0$, so ${\mathcal{P}}_1={\mathcal{P}}_2$ and this particular ambiguity disappears even though the full Riemann curvature tensor may be nonzero. This uses vacuum with zero cosmological constant; an Einstein vacuum with nonzero cosmological constant is not Ricci flat.

For the terrestrial estimate, the relevant effect is Ricci curvature inside matter, not merely the tidal field outside the Earth. If a typical material density is $\rho \sim 5\mathrm{g}\mathrm{c}{\mathrm{m}}^{-3}$, the Einstein field equations give a characteristic scaleFor a field varying over scale $L$, the derivative terms are $O(A/L^2)$ whereas the disputed term is $O(CA)$. Their relative size is $O(C{L}^2)$: about $10^{-22}$ for a metre-scale experiment. Even substituting the Earth's radius gives onlyA curvature length comparable to $C^{-1/2}\sim 10^{13}\mathrm{cm}$ would be needed for an order-one effect. In principle non-vacuum measurements could distinguish different curvature couplings, but an ordinary terrestrial laboratory has an exceedingly small signal, overwhelmed by its ordinary electromagnetic response to matter. This is a practical scale estimate, not a theorem that an arbitrarily precise experiment is impossible.

Put $\Delta r=r_2-r_1$. Expanding the quotient $S/T$ consistently to first order givesThusBecause $\log (r_2/r_1)/(r_2-r_1)\to 1/r_2$ as $r_1\to r_2$,This local limit is the expected speed of a massless particle measured in a freely falling local inertial frame; the finite-interval effective velocity combines nonlocal length and clock comparisons.

tidal forces measure differences in gravitational acceleration across an extended region. Neighboring freely falling particles each have zero four-acceleration, but their separation need not accelerate uniformly. Let $u^{\mu }$ be the tangent to a reference timelike geodesic and ${\xi }^{\mu }$ its infinitesimal separation from a neighboring one, chosen from an affinely parametrized geodesic congruence. Using the Riemann curvature tensor convention specified in Solution 3,Changing the overall curvature convention changes this displayed sign. The covariant derivatives compare vectors by parallel transport, avoiding a coordinate-dependent subtraction of accelerations.

In the Newtonian limit, expand the force field at two nearby points:For $\Phi =-GM/r$, a local orthonormal frame aligned with the radius givesThere is radial stretching and transverse compression; the trace vanishes outside the mass. Two freely falling test masses and a measurement of their changing separation implement gravity gradiometry. Their common acceleration drops out, while a relative acceleration proportional to separation remains. A local inertial frame removes Christoffel symbols at a point, but cannot remove nonzero curvature over a finite region.

Universality of free fall equates the passive gravitational response to inertial mass up to a universal choice of units. The Einstein equivalence principle further states that local nongravitational experiments in a freely falling laboratory obey special relativity independently of the laboratory's location and velocity. A supported laboratory instead has proper acceleration, just as an accelerating laboratory in flat spacetime does.

Around an event, freely falling coordinates can be chosen so that $g_{\mu \nu }={\eta }_{\mu \nu }$ and ${\Gamma }^{\alpha }{}_{\mu \nu }=0$ there. The first metric tensor derivatives vanish at that event, and nearby corrections begin at order curvature times squared separation. The metric tensor's second derivatives retain the Riemann tensor. Thus a local inertial frame removes Christoffel symbols at an event, while tidal gravity remains measurable over a finite region. Relative motion of separated freely falling test particles detects this curvature through geodesic deviation. Freely falling axes carried along an observer's world line extend the local construction, but do not make an extended curved region globally Minkowskian.

The weak equivalence principle states that freely falling, sufficiently small test bodies have the same motion in a given gravitational field irrespective of composition or mass, neglecting their own gravity and nongravitational forces. In the Newtonian limit, $m_i\mathbf{a}=m_g\mathbf{g}$ makes this universality equivalent to a universal gravitational-to-inertial mass ratio, conventionally set to one.

The Einstein equivalence principle adds that local nongravitational experiments in a freely falling, nonrotating laboratory have the laws of special relativity, independent of the laboratory's position and velocity. At an event one can choose a local inertial frame with $g_{\mu \nu }={\eta }_{\mu \nu }$ and ${\partial }_{\alpha }{g}_{\mu \nu }=0$. A freely falling observer feels no supporting force, while a supported observer can model a locally uniform gravitational field by an accelerated laboratory. This explains why freely moving particles follow timelike geodesics and light follows null geodesics when other interactions are negligible.

The locality qualification is essential. The Riemann curvature tensor generally remains nonzero; separated freely falling bodies have tidal relative acceleration, measured by geodesic deviation. A finite-size experiment can therefore distinguish a curved gravitational field from a merely accelerated coordinate description of flat spacetime. Equivalence alone also does not select the complete Einstein field equations.

A useful weak-field consequence is the upward photon shift $\Delta \nu /\nu \simeq -gH/c^2$ over laboratory height $H$. In an upward-accelerating laboratory the receiver gains recession velocity during the light flight, giving the same leading redshift. The Pound-Rebka experiment tested this with resonant absorption of 14.4-keV gamma rays over about 22.5 m. A controlled source velocity calibrated the Doppler sensitivity; interchanging source and absorber reversed the gravitational contribution while retaining an intrinsic source–absorber offset. Monitoring their temperature difference was essential because lattice-motion shifts could mimic the small signal.^[ref] This verifies a gravitational clock-shift prediction; by itself it is not a unique test selecting general relativity over every alternative metric theory.

Use signature $(+---)$, coordinates $x^0=ct$ and four-velocity $u^{\mu }=d{x}^{\mu }/d\tau$ normalized by $u_{\mu }{u}^{\mu }=c^2$. The derivative is the metric-compatible Levi-Civita connection. Fix the curvature sign by $[{\nabla }_{\alpha },{\nabla }_{\beta }]V^{\mu }=R^{\mu }{}_{\nu \alpha \beta }{V}^{\nu }$. Then the required contractions areWith $G^{\mu \nu }=R^{\mu \nu }-g^{\mu \nu }R/2$, the contracted Bianchi identity gives ${\nabla }_{\mu }{G}^{\mu \nu }=0$. The Einstein field equations $G^{\mu \nu }+\Lambda g^{\mu \nu }=(8\pi G/c^4)T^{\mu \nu }$, with constant $\Lambda$, therefore require covariant conservation of stress-energy. This is a consistency relation between geometry and the total matter/field source. It does not mean ${\partial }_{\mu }{T}^{\mu \nu }=0$ in arbitrary coordinates, nor guarantee a global conserved energy in an arbitrary curved spacetime. The geometric identity alone, without the field equations or matter dynamics, is not an equation of motion for a freely specified material source.

For relativistic dust, $\rho$ is proper rest-mass density, measured in the fluid rest frame, and $u^{\mu }$ is its timelike four-velocity with the $c^2$ normalization above. In a local Cartesian inertial frame $u^{\mu }=\gamma (c,\mathbf{v})$, soAccordingly $T^{0i}/c$ is momentum density and $c{T}^{0i}$ is energy flux in direction $i$. The equality of momentum-density and energy-flux components after the appropriate factors of $c$ reflects the symmetry of the stress-energy tensor. In the comoving frame the spatial stress is zero, as required for pressureless matter.

Expand the dust conservation equation by the product rule:Contract with $u_{\nu }$. metric compatibility and constant normalization give $u_{\nu }{\nabla }_{\mu }{u}^{\nu }={\nabla }_{\mu }(u_{\nu }{u}^{\nu })/2=0$, leaving $c^2{\nabla }_{\mu }(\rho u^{\mu })=0$. Substitute this rest-mass continuity law back and divide by nonzero $\rho$:This proves geodesic motion of uncharged dust along each fluid flow line. pressure forces or another exchange of momentum would change the argument; at zero density there is no dust worldline to constrain.

The electromagnetic field tensor is antisymmetric and has six independent components. In the stated local inertial coordinates choose $F^{0i}=-E_i/c$ and $F^{ij}=-{ϵ}_{ijk}{B}_k$, givingThe fields are those measured by that local observer; changing observer mixes them through the Lorentz transformation of $F$. With ${ϵ}^{0123}=+1$ and ${\tilde{F}}^{\mu \nu }={ϵ}^{\mu \nu \alpha \beta }{F}_{\alpha \beta }/2$, two independent quadratic contractions areThe second is a pseudoscalar: it is invariant under proper Lorentz transformations and changes sign under an orientation reversal. Opposite choices for the orientation tensor change its written sign. A nonzero null electromagnetic field can have both contractions zero, so the two numbers do not determine whether the tensor itself vanishes.

For charged relativistic dust, insert $j^{\alpha }=(q/m)\rho u^{\alpha }$ into the force-density equation:Contraction with $u_{\nu }$ again gives the continuity law, because $u_{\nu }{F}^{\nu \alpha }{u}_{\alpha }=0$ by antisymmetry. After removing the continuity term,In a local inertial frame the spatial components reduce to $d(\gamma m\mathbf{v})/dt=q(\mathbf{E}+\mathbf{v}\times \mathbf{B})$ and the temporal component gives $d(\gamma m{c}^2)/dt=q\mathbf{E}\cdot \mathbf{v}$. This is the relativistic Lorentz force, with gravity represented by the connection. The force is orthogonal to $u$ and preserves the rest-mass normalization. Matter alone now exchanges energy and momentum with the electromagnetic field; adding the field tensor's stress-energy restores zero divergence of the combined source.

Local equivalence and tides. In Newtonian mechanics, inertial mass $m_{\mathrm{i}}$ measures the response to a force, whereas passive gravitational mass $m_{\mathrm{g}}$ couples to an imposed Newtonian gravitational field. A freely falling body obeysThe observed composition independence of free fall says that this ratio is universal. With a common choice of units it is one. This is the Newtonian content of the weak equivalence principle for test particles; it does not by itself specify all nongravitational physics.

The Einstein equivalence principle further says that local nongravitational experiments in a sufficiently small freely falling laboratory have the form predicted by special relativity, independently of the laboratory's location and motion. In such a local inertial frame, the metric tensor is locally the Minkowski metric and the Christoffel symbols vanish at the selected event. The strong equivalence principle extends the universality claim to local gravitational experiments and bodies with significant gravitational self-energy. None of these statements removes measurable tidal forces across a finite apparatus.

In an ideal small freely falling lift, a released ball initially floats relative to the cabin, because both undergo the same free fall. Two separated balls, however, need not retain their separation: they sample different parts of the gravitational field. For a point source, linearizing $\mathbf{g}=-GM\mathbf{r}/r^3$ across a small separation $\mathbf{\xi }$ gives the tidal tensor predictionA radial separation stretches with coefficient $2GM/r^3$, whereas a transverse separation is compressed with coefficient $-GM/r^3$. Initially parallel velocities therefore develop relative motion even though the individual bodies experience zero proper acceleration.

The relativistic statement is geodesic deviation. With the Riemann curvature tensor conventiona connecting vector between neighboring freely falling worldlines satisfiesHere $D/d\tau$ is the covariant derivative along the reference geodesic, and $U^{\mu }$ its four-velocity. Curvature can be nonzero where the Christoffel symbols have been made zero. In the weak, static, slow-motion limit this equation reduces to ${\ddot{\xi }}^i=-{\partial }_i{\partial }_j\Phi {\xi }^j$, recovering the tidal tensor above. A finite falling lift can detect gravity through these relative accelerations; local equivalence does not imply a globally uniform field.