The Einstein Field Equations

The Einstein field equations are the dynamical heart of general relativity: they replace Newton's instantaneous force law with a local, geometric statement that the curvature of four-dimensional spacetime is fixed, point by point, by its energy and momentum content. Written compactly as $G_{\mu\nu}+\Lambda g_{\mu\nu}=8\pi G\,T_{\mu\nu}$ , they weld the kinematics of curved spacetime — carried by the metric $g_{\mu\nu}$ and the Einstein tensor $G_{\mu\nu}$ built from it — to the physics of matter carried by the stress-energy tensor $T_{\mu\nu}$ . Wheeler's aphorism names the dialogue: spacetime tells matter how to move; matter tells spacetime how to curve. Mathematically this is a system of ten coupled, nonlinear, second-order partial differential equations for the ten independent components of the symmetric metric. The nonlinearity has a physical name — gravity gravitates, because the field carries energy that itself curves spacetime — and it is what separates GR from linear theories like electromagnetism, giving rise to black holes, self-interacting gravitational waves, and the impossibility of superposing solutions. Because of the contracted Bianchi identity the ten equations are not ten independent evolutions: six are dynamical, four are constraints on the initial data, and the residual freedom is the four-fold choice of coordinates. The left-hand side is not a lucky guess. Lovelock's theorem states that in four dimensions the only symmetric, divergence-free, two-index tensor built locally from the metric and at most its second derivatives, linear in those second derivatives, is $a\,G_{\mu\nu}+b\,g_{\mu\nu}$ . The two constants are handed to experiment: the Newtonian limit fixes the coupling $8\pi G/c^{4}$ , and the cosmological constant $\Lambda$ survives as a number measured by cosmology. That uniqueness turns GR from one model among many into the essentially forced theory of a metric — and the same equations fall out of the Einstein-Hilbert action.

Key takeaways

The Einstein field equations $G_{\mu\nu}=8\pi G\,T_{\mu\nu}$ (units $c=1$ ) equate the Einstein curvature tensor of spacetime to the stress-energy tensor of matter: energy and momentum source the geometry that dictates how matter moves.
They are ten coupled, nonlinear, second-order PDEs for the metric $g_{\mu\nu}$ , of which four are constraints and six dynamical — the mismatch is the fingerprint of general covariance.
The geometric side is the unique symmetric, divergence-free tensor from the metric and its first two derivatives, linear in the second, up to two constants — Lovelock's theorem in four dimensions.
Their consistency rides on the contracted Bianchi identity $\nabla^{\mu}G_{\mu\nu}=0$ , which forces conservation $\nabla^{\mu}T_{\mu\nu}=0$ as a theorem, not a separate postulate.
The Newtonian limit fixes the coupling to $8\pi G/c^{4}\approx2.08\times10^{-43}\ \mathrm{N^{-1}}$ by matching Poisson's equation $\nabla^{2}\Phi=4\pi G\rho$ ; a term $+\Lambda g_{\mu\nu}$ is the only other consistent addition.
Key idea: demand a local tensor equation, second order in the metric, sourced by a conserved symmetric stress-energy tensor — divergence-freedom (Bianchi) plus the correct Newtonian limit fixes the left side to $G_{\mu\nu}+\Lambda g_{\mu\nu}$ and the coupling to $8\pi G$ , uniquely (Lovelock).

The core insight

Demand a local tensor equation, second order in the metric, whose source is a conserved symmetric stress-energy tensor: divergence-freedom (the contracted Bianchi identity) together with the correct Newtonian limit fixes the geometric side to $G_{\mu\nu}+\Lambda g_{\mu\nu}$ and the coupling to $8\pi G/c^{4}$ , and Lovelock's theorem makes that choice unique.

Where this lives today

The field equations are the working engine of modern gravitational physics. Numerical relativity solves them on supercomputers by splitting them into four constraints and six evolutions (the ADM and BSSN formulations), generating the waveform templates that let LIGO, Virgo, and KAGRA pull a merger out of detector noise — the first detection, GW150914 in 2015, matched a full nonlinear simulation of two merging holes. The Event Horizon Telescope's images of M87 and Sagittarius A are the vacuum solution $R_{\mu\nu}=0$ photographed at the horizon scale. In cosmology the same equations, fed an FLRW metric, become the Friedmann equations of $\Lambda$ CDM, and the measured $\Lambda$ (worked example 2) is the heart of the cosmological-constant problem — the worst quantitative mismatch between theory and experiment in physics. Every attempt at quantum gravity must reduce to $G_{\mu\nu}=8\pi G\,T_{\mu\nu}$ in the classical limit. A century on, the equation still sets the questions.

The physical picture

Picture spacetime as a deformable medium whose very shape is the gravitational field. Concentrations of energy and momentum — mass, pressure, radiation, fields — warp the geometry, and freely falling bodies follow the straightest available paths, the geodesics, through that warp. The field equation is the local converter turning "how much energy is here, and how is it flowing" into "how sharply is spacetime curved here." There is no force reaching across space; only geometry. This locality is the deep break from Newton: where Newtonian gravity is instantaneous action at a distance, the field equations are causal, and a change in the matter distribution ripples outward as curvature at the speed of light. The self-interaction — the nonlinearity — means even a region of pure field carries energy that curves spacetime further, which is why black holes can merge and radiate.

Figure

The Wheeler dialogue made local: matter and energy Tμν source curvature through the coupling κ = 8πG/c⁴ ≈ 2.08×10⁻⁴³ N⁻¹, and the resulting geometry Gμν tells freely-falling matter how to move along geodesics

A counting heuristic clears up why "ten equations for ten unknowns" is misleading. Four of the ten carry no second time-derivative and are constraints on any initial slice, leaving six evolutions for the spatial geometry; strip the four-fold coordinate freedom and those four constraints, and exactly two physical polarizations survive — the count of a massless spin-2 graviton.

Figure

Counting the degrees of freedom: the 10 components of the symmetric metric split as 10 = 4 constraints + 4 gauge choices + 2 physical polarizations, the four constraints enforced by the contracted Bianchi identity ∇^μ Gμν = 0

Trace structure closes the picture. Contracting the field equation gives $R=-8\pi G\,T$ , so in vacuum $R=0$ and $R_{\mu\nu}=0$ — Ricci-flat, but not necessarily Riemann-flat. Curvature persists in empty space through the Weyl tensor: tidal forces and gravitational radiation live exactly where there is no matter, which is why a distant detector still rings when a wave passes.

The core equations

The central statement equates the Einstein tensor — a divergence-free combination of Ricci curvature — to the stress-energy tensor. In units $c=1$ the coupling is bare $8\pi G$ , keeping geometry and physics adjacent; this is the form to reason with, and the constants are restored afterward.

G_{\mu\nu}\equiv R_{\mu\nu}-\tfrac{1}{2}R\,g_{\mu\nu}=8\pi G\,T_{\mu\nu} \qquad (1)

where: $G_{\mu\nu}$ — Einstein tensor, the divergence-free curvature combination; $R_{\mu\nu}$ — Ricci tensor, $R_{\mu\nu}=R^{\alpha}{}_{\mu\alpha\nu}$ ; $R$ — Ricci scalar, $R=g^{\mu\nu}R_{\mu\nu}$ ; $g_{\mu\nu}$ — the dynamical metric; $T_{\mu\nu}$ — stress-energy of matter and non-gravitational fields; $G$ — Newton's constant. The curvature on the left is completely fixed by the energy-momentum on the right. Restoring $c$ and admitting the one extra allowed term gives the full equation as Einstein wrote it. The cosmological constant $\Lambda$ multiplies the metric — the only other divergence-free two-index tensor available — and the coupling $\kappa=8\pi G/c^{4}$ is so small that macroscopic masses barely dent spacetime.

G_{\mu\nu}+\Lambda g_{\mu\nu}=\frac{8\pi G}{c^{4}}\,T_{\mu\nu} \qquad (2)

where: $\Lambda$ — cosmological constant, dimensions of inverse length squared; $c$ — speed of light; $\kappa=8\pi G/c^{4}\approx2.08\times10^{-43}\ \mathrm{N^{-1}}$ — the Einstein gravitational coupling, whose tininess is why spacetime is so stiff. Algebraically solving for the Ricci tensor gives the trace-reversed form — the practical shape of the equation, because the Ricci tensor, not $G_{\mu\nu}$ , is what appears in the geodesic-deviation and Raychaudhuri equations that govern the focusing of matter.

R_{\mu\nu}=8\pi G\left(T_{\mu\nu}-\tfrac{1}{2}T\,g_{\mu\nu}\right) \qquad (3)

where: $T=g^{\mu\nu}T_{\mu\nu}$ — the trace of the stress-energy tensor. In vacuum the right-hand side vanishes, giving the compact vacuum equation $R_{\mu\nu}=0$ . The equation's internal consistency rests on a pure geometric identity: the Einstein tensor is divergence-free for any metric whatsoever, an identity true off-shell. Taking the divergence of the field equation then forces the source to be conserved, making local energy-momentum conservation an output of the theory rather than a separate postulate.

\nabla^{\mu}G_{\mu\nu}=0 \quad\Longrightarrow\quad \nabla^{\mu}T_{\mu\nu}=0 \qquad (4)

where: $\nabla^{\mu}$ — covariant divergence with the Levi-Civita connection. The left identity holds for every metric; applied to the field equation it compels the conservation on the right. Linearizing about flat space in Lorenz gauge collapses the full nonlinear system to a flat-space wave equation with a source — the gateway to gravitational-wave physics.

\Box\bar h_{\mu\nu}=-\frac{16\pi G}{c^{4}}\,T_{\mu\nu},\qquad \bar h_{\mu\nu}=h_{\mu\nu}-\tfrac{1}{2}\eta_{\mu\nu}h \qquad (5)

where: $h_{\mu\nu}$ — metric perturbation, $g_{\mu\nu}=\eta_{\mu\nu}+h_{\mu\nu}$ ; $\bar h_{\mu\nu}$ — its trace-reversed form; $\Box=\eta^{\alpha\beta}\partial_\alpha\partial_\beta$ — flat d'Alembertian; $h=\eta^{\mu\nu}h_{\mu\nu}$ . Radiative solutions propagate at $c$ with two polarizations. Finally, the same equations descend from a single scalar action; varying it with respect to the metric returns $G_{\mu\nu}=8\pi G\,T_{\mu\nu}$ , with the stress-energy tensor emerging as the metric-derivative of the matter action.

S=\frac{c^{4}}{16\pi G}\int R\,\sqrt{-g}\;\mathrm{d}^{4}x+S_{\mathrm{matter}},\qquad T_{\mu\nu}=-\frac{2}{\sqrt{-g}}\frac{\delta S_{\mathrm{matter}}}{\delta g^{\mu\nu}} \qquad (6)

where: $\sqrt{-g}\,\mathrm{d}^{4}x$ — invariant volume element, $g=\det g_{\mu\nu}$ ; $S_{\mathrm{matter}}$ — action of all non-gravitational fields; $\delta/\delta g^{\mu\nu}$ — functional derivative in the inverse metric.

Derivation AThe contracted Bianchi identity forces energy-momentum conservation

The goal is to show that $\nabla^{\mu}T_{\mu\nu}=0$ is not a postulate bolted onto general relativity but an unavoidable consequence of the differential geometry of curvature. We begin from the second Bianchi identity — true of the Riemann tensor of any metric connection — contract it twice to prove that the Einstein tensor is identically divergence-free, then feed that into the field equation, where it compels the source to be conserved. The payoff is conceptual: energy-momentum conservation in GR is a theorem, and the equations would be inconsistent with a non-conserved source.

Step 1. Write the second Bianchi identity $\nabla_{[\lambda}R_{\mu\nu]\rho\sigma}=0$ , the antisymmetrization over the first three lower indices. Why: this is an identity for the curvature of any Levi-Civita connection, so what we derive from it constrains every spacetime without assuming dynamics.

Step 2. Contract once with $g^{\mu\lambda}$ and use the Riemann symmetries to get $\nabla^{\mu}R_{\mu\nu\rho\sigma}=\nabla_{\rho}R_{\nu\sigma}-\nabla_{\sigma}R_{\nu\rho}$ . Why: contracting one pair of indices turns a statement about full Riemann into one about its Ricci trace.

Step 3. Contract a second time to reach $\nabla^{\mu}R_{\mu\nu}=\tfrac{1}{2}\nabla_{\nu}R$ , the once-contracted Bianchi identity. Why: this is the exact combination that makes the divergence of the Einstein tensor cancel.

Step 4. Compute $\nabla^{\mu}(R_{\mu\nu}-\tfrac{1}{2}Rg_{\mu\nu})=\tfrac{1}{2}\nabla_{\nu}R-\tfrac{1}{2}\nabla_{\nu}R=0$ , using metric compatibility $\nabla_\alpha g_{\mu\nu}=0$ . Why: the two half-gradients of $R$ cancel exactly, proving $\nabla^{\mu}G_{\mu\nu}=0$ identically — a purely geometric fact.

Step 5. Take the divergence of $G_{\mu\nu}=8\pi G\,T_{\mu\nu}$ : the left side vanishes by Step 4, so $8\pi G\,\nabla^{\mu}T_{\mu\nu}=0$ and hence $\nabla^{\mu}T_{\mu\nu}=0$ . Why: the geometry can only be sourced by a conserved tensor, so local energy-momentum conservation is forced by the structure of the equation rather than assumed. *The cancellation in Step 4 is exact and coordinate-independent.*

Go deeper · Derivation B · 1 minFixing the coupling constant from the Newtonian limit

Derivation BFixing the coupling constant from the Newtonian limit

The purpose is to pin the one free proportionality constant in the field equation to a laboratory number by demanding that GR reproduce Newtonian gravity where Newton is known to work: weak fields, slow motion, static sources. We take a nearly flat static metric, identify its time-time component with the Newtonian potential, evaluate the trace-reversed equation to leading order, and match to Poisson's equation. The outcome is that the coupling must be exactly $8\pi G/c^{4}$ — the $8\pi$ that looks arbitrary in isolation is fixed once Newton's constant is required to come out right.

Step 1. Write the weak static metric $g_{00}=-(1+2\Phi/c^{2})$ with $|\Phi|\ll c^{2}$ and a nearly Euclidean spatial part, identifying $\Phi$ as the Newtonian potential. Why: this weak, static geometry embeds the one function $\Phi$ that Newton's theory uses.

Step 2. Model the source as slow dust dominated by rest energy, so $T_{00}\approx\rho c^{2}$ and $T\approx-\rho c^{2}$ , pressure negligible. Why: for non-relativistic matter the energy density swamps momentum flux and pressure, isolating $\rho$ .

Step 3. Evaluate the trace-reversed time-time equation $R_{00}=\kappa(T_{00}-\tfrac{1}{2}Tg_{00})\approx\kappa\cdot\tfrac{1}{2}\rho c^{2}$ , using $g_{00}\approx-1$ . Why: the $-\tfrac{1}{2}Tg_{00}$ term contributes a second $\tfrac{1}{2}\rho c^{2}$ , and this doubling produces the $8\pi$ rather than $4\pi$ .

Step 4. Show that to leading order the linearized Ricci component reduces to $R_{00}\approx\nabla^{2}\Phi/c^{2}$ . Why: only the second spatial derivatives of $g_{00}$ survive at this order, assembling into the Laplacian of Poisson's equation.

Step 5. Equate the two expressions, $\nabla^{2}\Phi/c^{2}=\tfrac{1}{2}\kappa\rho c^{2}$ , and impose Poisson $\nabla^{2}\Phi=4\pi G\rho$ to solve $\kappa=8\pi G/c^{4}$ ( $=8\pi G$ when $c=1$ ). Why: matching the empirical Newtonian law leaves no freedom — the coupling is fixed to $\kappa\approx2.08\times10^{-43}\ \mathrm{N^{-1}}$ . *The factor of two from the trace term is what converts $4\pi$ into $8\pi$ , reproducing $\nabla^{2}\Phi=4\pi G\rho$ with $\kappa=8\pi G/c^{4}$ .*

Go deeper · Derivation C · 1 minThe field equations from the Einstein-Hilbert action

Derivation CThe field equations from the Einstein-Hilbert action

The aim is to show that the field equations are not merely postulated but follow from an extremum principle applied to the simplest scalar built from curvature — the Ricci scalar integrated over spacetime. We vary the Einstein-Hilbert action with respect to the inverse metric, handle the three pieces that arise (Ricci variation, volume-element variation, and a boundary term), add the matter action, and demand stationarity. The reward is threefold: the geometric side emerges as $G_{\mu\nu}$ , the stress-energy tensor is defined intrinsically as the metric-response of the matter action, and the structure becomes the launch pad for modified-gravity and quantization programs.

Step 1. Vary $S=\tfrac{1}{16\pi G}\int R\sqrt{-g}\,\mathrm{d}^{4}x$ (units $c=1$ ), splitting $\delta(R\sqrt{-g})$ into $R_{\mu\nu}\delta g^{\mu\nu}\sqrt{-g}$ , the volume term $R\,\delta\sqrt{-g}$ , and $g^{\mu\nu}\delta R_{\mu\nu}\sqrt{-g}$ . Why: writing $R=g^{\mu\nu}R_{\mu\nu}$ shows the scalar depends on the metric in three ways, each of which must be varied.

Step 2. Use $\delta\sqrt{-g}=-\tfrac{1}{2}\sqrt{-g}\,g_{\mu\nu}\delta g^{\mu\nu}$ from $\delta g=-g\,g_{\mu\nu}\delta g^{\mu\nu}$ . Why: the volume-element variation supplies exactly the $-\tfrac{1}{2}Rg_{\mu\nu}$ term that turns bare Ricci into the Einstein tensor.

Step 3. Show by the Palatini identity that $g^{\mu\nu}\delta R_{\mu\nu}=\nabla_{\sigma}v^{\sigma}$ is a total divergence. Why: as a boundary term it is cancelled by the Gibbons-Hawking-York term and drops when the metric is fixed on the boundary, contributing nothing to the bulk.

Step 4. Collect the bulk terms and add the matter variation: $\tfrac{1}{16\pi G}(R_{\mu\nu}-\tfrac{1}{2}Rg_{\mu\nu})\delta g^{\mu\nu}\sqrt{-g}+\delta S_{\mathrm{matter}}=0$ . Why: demanding this vanish for arbitrary $\delta g^{\mu\nu}$ promotes an integral condition into a local field equation.

Step 5. Define $T_{\mu\nu}=-\tfrac{2}{\sqrt{-g}}\,\delta S_{\mathrm{matter}}/\delta g^{\mu\nu}$ , whereupon the bulk condition is exactly $G_{\mu\nu}=8\pi G\,T_{\mu\nu}$ . Why: this definition makes $T_{\mu\nu}$ symmetric and, via diffeomorphism invariance of $S_{\mathrm{matter}}$ , automatically conserved. *The GHY boundary term is required for a well-posed variational problem with a boundary.*

Limiting cases

Every simpler gravitational theory is a corner of the field equations, reached by switching off a source, a strength, or a velocity.

Case	Result	Meaning
Vacuum	$R_{\mu\nu}=0$	With $T_{\mu\nu}=0$ , $\Lambda=0$ the trace-reversed form gives Ricci flatness; Schwarzschild and waves are solutions, curvature surviving in the Weyl tensor.
Weak field / linearized	$\Box\bar h_{\mu\nu}=-16\pi G\,T_{\mu\nu}$	Writing $g_{\mu\nu}=\eta_{\mu\nu}+h_{\mu\nu}$ with $\|h\|\ll1$ in Lorenz gauge yields a flat-space wave equation for the trace-reversed perturbation.
Newtonian limit	$\nabla^{2}\Phi=4\pi G\rho$	Static, slow, weak: the $00$ -component reduces to Poisson's equation with $g_{00}\approx-(1+2\Phi/c^{2})$ , fixing the coupling to $8\pi G/c^{4}$ .
Cosmological constant	$G_{\mu\nu}+\Lambda g_{\mu\nu}=8\pi G\,T_{\mu\nu}$	The only other divergence-free term; drives cosmic acceleration, equivalent to a vacuum energy $\rho_\Lambda=\Lambda c^{2}/8\pi G$ .

Common trapThe ten equations are ten independent evolution equations

They are not. The contracted Bianchi identity $\nabla^{\mu}G_{\mu\nu}=0$ imposes four relations among the ten, so only six are genuine second-order evolutions; the remaining four are constraints on the initial data — the Hamiltonian and momentum constraints — automatically preserved by the evolution. Treating all ten as dynamical over-determines the system, and forgetting that $\nabla^{\mu}T_{\mu\nu}=0$ is implied rather than imposed leads to inconsistent sources. A quieter trap: $G_{\mu\nu}=8\pi G\,T_{\mu\nu}$ uses $c=1$ ; the physical coupling is $8\pi G/c^{4}\approx2.08\times10^{-43}\ \mathrm{N^{-1}}$ , and dropping the $c^{4}$ silently rescales every mass. And $R_{\mu\nu}=0$ does not mean flat: the Weyl curvature can be large, which is why vacuum carries tidal forces and radiation.

Worked examples

1 · The perihelion precession of Mercury (fully worked)

The general-relativistic correction advances the perihelion of an orbit by $\Delta\varphi=6\pi GM/[c^{2}a(1-e^{2})]$ per revolution. For Mercury, $M_\odot=1.989\times10^{30}$ kg, semi-major axis $a=5.79\times10^{10}$ m, eccentricity $e=0.2056$ , and orbital period $0.2409$ yr. Find the precession per orbit and per century.

Step 1. Compute the gravitational length $GM_\odot/c^{2}=(6.674\times10^{-11})(1.989\times10^{30})/(2.998\times10^{8})^{2}=\mathbf{1477\ m}$ — half the Sun's Schwarzschild radius, the natural scale of every solar relativistic effect.

Step 2. Form the semi-latus rectum $a(1-e^{2})=5.79\times10^{10}\times(1-0.0423)=5.545\times10^{10}$ m, then $\Delta\varphi=6\pi\,(1477)/(5.545\times10^{10})=\mathbf{5.02\times10^{-7}\ rad}$ per orbit — about $0.10$ arcsecond each time around.

Step 3. Mercury completes $100/0.2409=415$ orbits per century, so the advance accumulates to $5.02\times10^{-7}\times415=2.08\times10^{-4}$ rad; converting with $206265$ arcsec/rad gives $\mathbf{43''\ per\ century}$ .

Step 4. This is the residual Newtonian perturbation theory could not explain — the $43''$ /century anomaly Le Verrier catalogued in 1859 — and GR produces it with no free parameters. The GR precession is 5.0×10⁻⁷ rad per orbit, totalling 43″ per century — matching observation exactly.

2 · The vacuum energy behind the cosmological constant (fully worked)

Moved to the source side, the cosmological constant is a uniform vacuum energy of density $\rho_\Lambda=\Lambda c^{2}/8\pi G$ . For the measured $\Lambda\approx1.1\times10^{-52}\ \mathrm{m^{-2}}$ , find $\rho_\Lambda$ as protons per cubic metre.

Step 1. Insert the numbers: $\rho_\Lambda=(1.1\times10^{-52})(2.998\times10^{8})^{2}/[8\pi(6.674\times10^{-11})]$ , numerator $9.9\times10^{-36}$ over denominator $1.677\times10^{-9}$ .

Step 2. Divide to get $\rho_\Lambda=\mathbf{5.9\times10^{-27}\ kg\,m^{-3}}$ — a staggeringly small density, yet because it fills all of space it dominates the cosmic energy budget today.

Step 3. Divide by the proton mass $1.67\times10^{-27}$ kg: $\rho_\Lambda\approx\mathbf{3.5\ protons\ per\ cubic\ metre}$ — the accelerating universe is driven by an energy density of a few hydrogen atoms per cubic metre. Λ corresponds to ρ_Λ ≈ 5.9×10⁻²⁷ kg/m³ ≈ 3.5 protons per cubic metre — the empirical scale of dark energy.

3 · How weakly does matter curve spacetime?

Estimate the curvature produced by ordinary water. For static dust the trace-reversed equation gives a Ricci-scalar magnitude $R\sim(8\pi G/c^{2})\rho$ ; take $\rho=1000\ \mathrm{kg\,m^{-3}}$ .

Step 1. Compute $R\sim8\pi(6.674\times10^{-11})(1000)/(2.998\times10^{8})^{2}=\mathbf{1.9\times10^{-23}\ m^{-2}}$ ; the associated radius of curvature is $R^{-1/2}\approx\mathbf{2.3\times10^{11}\ m}$ .

Step 2. That is about $1.5$ astronomical units, larger than the orbit of Mars — a tank of water bends spacetime, but only on the scale of the inner solar system. The lesson is the smallness of $\kappa=8\pi G/c^{4}$ : it takes planetary masses or merger velocities to curve spacetime visibly. Everyday matter curves spacetime on a ~10¹¹ m radius — the field equations are nearly rigid, which is why gravity looked linear to Newton for centuries.

Physics of the portrait

The field equations are the summit of a decade Einstein spent converting a physical intuition — that free fall is indistinguishable from inertial motion — into the mathematics of curved spacetime. The story is one of a physicist reaching, at times unsteadily, for a geometry he had to learn as he went.

Origins · 1 minAlbert Einstein and how this was discovered

Albert Einstein

1879–1955 · the geometric theory of gravitation

Albert Einstein's road to the field equations began in 1907 with what he called the happiest thought of his life: a person in free fall feels no gravity, so gravitation and acceleration are locally the same. Turning that equivalence principle into a theory took eight years and a crash course in mathematics he did not command. By 1912 he had concluded that gravity must be the curvature of spacetime, and with Marcel Grossmann he adopted Riemannian geometry and the tensor calculus of Ricci and Levi-Civita. The middle years were a struggle — he wrote down, then abandoned, equations that were not generally covariant. Through the autumn of 1915, racing a correspondence with David Hilbert, he found the missing ingredient: the trace term. The trace-reversed source $T_{\mu\nu}-\tfrac{1}{2}Tg_{\mu\nu}$ made the equations both generally covariant and consistent with $\nabla^{\mu}T_{\mu\nu}=0$ , and on 25 November 1915 he presented $G_{\mu\nu}=8\pi G\,T_{\mu\nu}$ to the Prussian Academy. Within weeks it gave Mercury's $43''$ precession with no adjustable constants. In 1917 he added the $\Lambda$ term to force a static cosmos — later his self-described greatest blunder when the universe proved to be expanding, though it has since returned as dark energy. He spent his final decades chasing a unified field theory that never came, but the 1915 equations have withstood a century of sharpening tests without a confirmed deviation.

History

1859

— Le Verrier reports Mercury's anomalous perihelion advance of about 43 arcseconds per century

1907

— Einstein's equivalence-principle insight: free fall is locally indistinguishable from inertial motion

1915

— Einstein presents the final generally covariant field equations to the Prussian Academy in November

1916

— Schwarzschild solves the vacuum equations for the first exact black-hole geometry

1919

— Eddington's eclipse expeditions confirm the 1.75-arcsecond deflection of starlight by the Sun

Going further

Texts. Misner, Thorne & Wheeler, Gravitation, Ch. 17 · Wald, General Relativity, Ch. 4 and App. E · Carroll, Spacetime and Geometry, Ch. 4 · Hartle, Gravity, Ch. 22 · Schutz, A First Course in General Relativity, Ch. 8 · Lovelock, J. Math. Phys. 12, 498 (1971). Key results to keep in mind: $\nabla^{\mu}G_{\mu\nu}=0$ follows from double-contracting the second Bianchi identity; $\kappa=8\pi G/c^{4}\approx2.08\times10^{-43}\ \mathrm{N^{-1}}$ from matching the Newtonian limit to $\nabla^{2}\Phi=4\pi G\rho$ ; and the field equations emerge from the Hilbert variation with the GHY boundary term. Anchor numbers: Mercury $\Delta\varphi=6\pi GM/[c^{2}a(1-e^{2})]=5.0\times10^{-7}$ rad/orbit $=43''$ /century; $\rho_\Lambda=\Lambda c^{2}/8\pi G\approx5.9\times10^{-27}$ kg/m³ for $\Lambda=1.1\times10^{-52}\ \mathrm{m^{-2}}$ ; light deflection $1.75''$ at the solar limb.

Prerequisites. The Einstein Tensor · The Stress-Energy Tensor · Local Conservation of Energy-Momentum · The Newtonian Limit. Leads to. The Einstein-Hilbert Action · Variational Derivation of the Field Equations · The Cosmological Constant · The Weak-Field Approximation.

Ask the tutor: Ask me to walk through any single derivation — the Bianchi-to-conservation argument, the Newtonian-limit fixing of $8\pi G/c^{4}$ , or the Hilbert variation — step by step with the index gymnastics spelled out.

Why is there a factor of 8π rather than 4π in the field equations?

The $4\pi$ of Poisson's equation $\nabla^{2}\Phi=4\pi G\rho$ does appear, but the matching to GR runs through the trace-reversed form. In the Newtonian limit the source is not $T_{00}\approx\rho c^{2}$ alone but the combination $T_{00}-\tfrac{1}{2}Tg_{00}$ , and the trace term contributes a second equal piece that doubles the source. Working the $00$ -component (derivation 2) shows the two halves add, so the coupling that reproduces $4\pi G\rho$ must be $8\pi G/c^{4}$ . The $8\pi$ is not arbitrary — it is exactly what makes the weak-field theory agree with Newton.

In what sense is energy-momentum conservation derived rather than assumed?

In Newtonian physics and special relativity $\partial^{\mu}T_{\mu\nu}=0$ is imposed as a separate law; in general relativity it is a theorem. The Einstein tensor satisfies the geometric identity $\nabla^{\mu}G_{\mu\nu}=0$ for every metric, with no equations of motion invoked. So the moment you set $G_{\mu\nu}=8\pi G\,T_{\mu\nu}$ and take the covariant divergence of both sides, the left vanishes identically and forces $\nabla^{\mu}T_{\mu\nu}=0$ . Conservation is an integrability condition — the geometry cannot be sourced by a non-conserved tensor — so the same structure that guarantees consistency delivers a conservation law for free.

Does Rμν = 0 mean spacetime is empty and flat?

It means empty, not flat. The vacuum equation $R_{\mu\nu}=0$ says there is no local matter source, but the full Riemann tensor decomposes into a Ricci part and a Weyl part, and only the Ricci part is switched off. The Weyl tensor — the trace-free tidal curvature — can be arbitrarily large in vacuum: the Schwarzschild geometry outside a star is Ricci-flat yet strongly curved near the horizon, and a gravitational wave is a ripple of pure Weyl curvature in empty space. Flatness would require the entire Riemann tensor to vanish, a far stronger condition holding only in genuinely gravitation-free regions.

What makes the field equations nonlinear, and why does it matter?

The Einstein tensor is built from the metric, its inverse, and the Christoffel symbols, which contain products of metric derivatives, so $G_{\mu\nu}$ depends nonlinearly on $g_{\mu\nu}$ . Physically this is gravity gravitating: the field carries energy, that energy sources curvature, and so the field feeds back on itself. The consequences are large — solutions cannot be superposed, so you cannot add two black holes to get a two-hole spacetime; strong fields collapse to horizons; and mergers radiate through a regime no perturbation expansion reaches, which is why numerical relativity exists. Electromagnetism, being linear, has none of this.

Why can the ten equations be split into constraints and evolutions?

Because general covariance means the coordinates carry no physics, and four of the ten equations reflect that freedom rather than dynamics. Slicing spacetime into space-plus-time (the ADM decomposition), the six equations with two time derivatives evolve the spatial geometry forward, while the four without a second time derivative are constraints each slice must satisfy — the Hamiltonian and three momentum constraints. The contracted Bianchi identity guarantees that if the constraints hold on one slice, the evolution preserves them on every later one. This split makes GR a well-posed Cauchy problem and is the foundation of every numerical-relativity simulation.

Is the cosmological constant part of the field equations or an add-on?

It is a legitimate, in fact mandatory-to-consider, part. Lovelock's theorem says the most general geometric side is $a\,G_{\mu\nu}+b\,g_{\mu\nu}$ , because the metric is the only other symmetric, divergence-free tensor of the allowed form (divergence-free precisely because $\nabla_\alpha g_{\mu\nu}=0$ ). Setting $b=\Lambda$ gives the $\Lambda g_{\mu\nu}$ term, so $\Lambda$ is not a modification of GR but a constant the theory always permitted. Einstein introduced it in 1917 for a static universe, discarded it after the expansion was found, and it returned in 1998 when supernova surveys revealed cosmic acceleration. Today $\Lambda$ is measured, and its tiny non-zero value is one of the deepest puzzles in physics.