Ricci and scalar curvature, and the Einstein tensor

Intuition Beginner

The Riemann tensor 13.03.01 records everything about how a surface or a spacetime bends. In four dimensions it holds over two hundred components. That richness is more than most physical questions need. General relativity asks a coarser question: how does the volume of a small ball of freely falling test particles change as it moves through spacetime?

The Ricci tensor compresses the Riemann tensor by tracing, folding one pair of indices back into another. Where Riemann tracks every direction in which space can bend, Ricci keeps only the part that changes volumes. The scalar curvature compresses further: a single number, the trace of the Ricci tensor, measuring how much the volume of a tiny ball differs from its flat-space value. Positive means the ball shrinks; negative means it grows.

The Einstein tensor combines the Ricci tensor and the scalar curvature into one object whose divergence vanishes. "Divergence vanishes" means the tensor is conserved in the same sense that energy and momentum are conserved. That conservation is precisely what is needed to place this object on the geometry side of Einstein's field equations, paired against the matter and energy on the other side.

Visual Beginner

Curvature contracts down a ladder. Each rung traces one index pair away, losing detail but gaining simplicity. The Riemann tensor has four indices; the Ricci tensor has two; the scalar curvature has none.

The ladder is a lossy summary. Tracing throws away the "shape-distorting" part of curvature (carried by the Weyl tensor 13.03.01) and keeps only the "volume-distorting" part. The Einstein tensor then recombines the two rungs so that the result is conserved.

Worked example Beginner

Take the surface of a sphere of radius $a$, with metric $d{s}^2=a^2(d{\theta }^2+{\sin }^2\theta d{\phi }^2)$. This is the cleanest curved surface for which every curvature object can be written down. The Riemann tensor was computed in 13.03.01; here we use its contractions.

The Ricci tensor of the sphere has two components, and both are fixed by the radius:

$$R_{\theta \theta }=1,R_{\phi \phi }={\sin }^2\theta .$$

Check the pattern against the metric, whose entries are $g_{\theta \theta }=a^2$ and $g_{\phi \phi }=a^2{\sin }^2\theta$. Dividing each metric entry by $a^2$ gives exactly $R_{\theta \theta }$ and $R_{\phi \phi }$. The Ricci tensor is therefore proportional to the metric: $R_{\mu \nu }=(1/a^2)g_{\mu \nu }$.

The scalar curvature is the trace: $R=2/a^2$. Plug in numbers. A unit sphere ($a=1$) has $R=2$. Double the radius to $a=2$ and $R=1/2$, a quarter of the original. Curativity falls off as $1/a^2$: bigger spheres are less curved, and an infinitely large sphere is flat. A surface whose Ricci tensor is proportional to its metric is called an Einstein manifold, and the sphere is the simplest example.

Check your understanding Beginner

Formal definition Intermediate+

Let $(M,g)$ be a smooth pseudo-Riemannian manifold with Levi-Civita connection, and let $R_{\sigma \mu \nu }^{\rho }$ be the Riemann curvature tensor as defined in 13.03.01. The Ricci tensor is the unique nontrivial contraction of the Riemann tensor,

$$\boxed{R_{\mu \nu }:=R_{\mu \lambda \nu }^{\lambda }=g^{\lambda \kappa }{R}_{\kappa \mu \lambda \nu }.}$$

The other contraction $R_{\mu \nu \lambda }^{\lambda }$ differs by a sign: $R_{\mu \nu \lambda }^{\lambda }=-R_{\mu \nu }$, by the antisymmetry $R_{\kappa \mu \lambda \nu }=-R_{\kappa \mu \nu \lambda }$ of the last index pair of the Riemann tensor 13.03.01. There are no further independent rank-two contractions of Riemann.

The scalar curvature is the trace of the Ricci tensor,

$$\boxed{R:=g^{\mu \nu }{R}_{\mu \nu }.}$$

It is the unique scalar constructible from the metric and its first two derivatives that is linear in the second derivatives. The Einstein tensor is the recombination

$$\boxed{G_{\mu \nu }:=R_{\mu \nu }-\frac{1}{2}R{g}_{\mu \nu }.}$$

The Einstein tensor is symmetric (because $R_{\mu \nu }$ and $g_{\mu \nu }$ are) and, in $n$ dimensions, has trace $G_{\mu }^{\mu }=g^{\mu \nu }{G}_{\mu \nu }=R-\frac{n}{2}R=\frac{2-n}{2}R$, which equals $-R$ in four dimensions. The Einstein equation $G_{\mu \nu }=8\pi T_{\mu \nu }$ (with $G=c=1$) is therefore equivalent, upon taking the trace, to $-R=8\pi T$, where $T=g^{\mu \nu }{T}_{\mu \nu }$ is the trace of the stress-energy tensor. Solving for $R$ and substituting back yields the trace-reversed form of the field equations,

$$R_{\mu \nu }=8\pi \left(T_{\mu \nu }-\frac{1}{2}T{g}_{\mu \nu }\right),$$

valid in four dimensions. The geometric content of the Einstein equation is the Ricci tensor; the scalar term enforces consistency with stress-energy conservation.

Symmetries and structure of the Ricci tensor

The Ricci tensor inherits its algebraic structure from the Riemann tensor 13.03.01:

1. Symmetry: $R_{\mu \nu }=R_{\nu \mu }$, a direct consequence of the Riemann pair-exchange symmetry $R_{\rho \sigma \mu \nu }=R_{\mu \nu \rho \sigma }$.
2. Metric compatibility: ${\nabla }_{\lambda }{R}_{\mu \nu }$ inherits no extra algebraic identity beyond the contracted Bianchi identity below.
3. Components: as a symmetric rank-2 tensor, $R_{\mu \nu }$ carries $n(n+1)/2$ independent components --- ten in four dimensions, against the twenty of the Riemann tensor.

Counterexamples to common slips

• Ricci-flat is not flat. Vacuum spacetimes (Schwarzschild, Kerr 13.05.01, gravitational waves 13.07.01) have $R_{\mu \nu }=0$ but nonzero Riemann tensor. The missing curvature lives in the Weyl tensor, which encodes tidal and radiative degrees of freedom the Ricci contraction discards.
• Vanishing scalar curvature is not flatness. Gravitational-wave spacetimes in vacuum have $R=0$ (because $R_{\mu \nu }=0$) yet carry propagating curvature. The scalar curvature is the coarsest curvature invariant; its vanishing says nothing about the Weyl part.
• The Einstein tensor is not the Ricci tensor. Subtracting $\frac{1}{2}R{g}_{\mu \nu }$ is what makes the combination divergence-free. The Ricci tensor by itself is not covariantly conserved: ${\nabla }^{\mu }{R}_{\mu \nu }=\frac{1}{2}{\nabla }_{\nu }R\ne 0$ in general.

Key theorem with proof Intermediate+

Theorem (Symmetry of Ricci and the contracted Bianchi identity). The Ricci tensor is symmetric, $R_{\mu \nu }=R_{\nu \mu }$. The Einstein tensor is covariantly conserved, ${\nabla }^{\mu }{G}_{\mu \nu }=0$, equivalently ${\nabla }^{\mu }{R}_{\mu \nu }=\frac{1}{2}{\nabla }_{\nu }R$. These two facts are what license the Einstein tensor to appear on the geometric side of the Einstein field equations.

Proof of symmetry. Start from the definition $R_{\mu \nu }=g^{\lambda \kappa }{R}_{\kappa \mu \lambda \nu }$. Apply the Riemann pair-exchange symmetry $R_{\kappa \mu \lambda \nu }=R_{\lambda \nu \kappa \mu }$ 13.03.01:

$$R_{\mu \nu }=g^{\lambda \kappa }{R}_{\lambda \nu \kappa \mu }=g^{\kappa \lambda }{R}_{\kappa \nu \lambda \mu }=R_{\nu \mu },$$

where the second equality relabels dummy indices $(\lambda \leftrightarrow \kappa )$ and the third is the definition of $R_{\nu \mu }$. Both sides are components of tensors, so the equality holds in every coordinate system. $\square$

Proof of the contracted Bianchi identity. The second Bianchi identity 13.03.01 reads

$${\nabla }_{\lambda }{R}_{\rho \sigma \mu \nu }+{\nabla }_{\mu }{R}_{\rho \sigma \text{n}\lambda }+{\nabla }_{\nu }{R}_{\rho \sigma \lambda \mu }=0.$$

Contract $\lambda$ with $\rho$ by multiplying by $g^{\lambda \rho }$. Since ${\nabla }_{\alpha }{g}_{\rho \sigma }=0$ (metric compatibility), the inverse metric passes through the covariant derivatives, giving

$${\nabla }^{\rho }{R}_{\rho \sigma \mu \nu }+{\nabla }_{\mu }{R}_{\sigma \nu \rho }^{\rho }+{\nabla }_{\nu }{R}_{\sigma \rho \mu }^{\rho }=0.$$

Identify the Ricci contractions. From the definition $R_{\sigma \mu }=R_{\sigma \rho \mu }^{\rho }$, and using the last-pair antisymmetry $R_{\sigma \nu \rho }^{\rho }=-R_{\sigma \rho \nu }^{\rho }=-R_{\sigma \nu }$:

$${\nabla }^{\rho }{R}_{\rho \sigma \mu \nu }-{\nabla }_{\mu }{R}_{\sigma \nu }+{\nabla }_{\nu }{R}_{\sigma \mu }=0.\text{\hspace{1em}(∗)}$$

Now contract $(\ast )$ with $g^{\sigma \nu }$. The first term gives $g^{\sigma \nu }{\nabla }^{\rho }{R}_{\rho \sigma \mu \nu }={\nabla }^{\rho }(g^{\sigma \nu }{R}_{\rho \sigma \mu \nu })={\nabla }^{\rho }{R}_{\rho \mu }$, since contracting the second and fourth indices of Riemann returns the Ricci tensor. By symmetry of $R_{\rho \mu }$ this is ${\nabla }^{\rho }{R}_{\mu \rho }$. The remaining terms give $-{\nabla }_{\mu }R+{\nabla }^{\nu }{R}_{\mu \nu }$. Assembling:

$$2{\nabla }^{\nu }{R}_{\mu \nu }-{\nabla }_{\mu }R=0,\text{i.e.}{\nabla }^{\nu }{R}_{\mu \nu }=\frac{1}{2}{\nabla }_{\mu }R.$$

Finally subtract $\frac{1}{2}{\nabla }_{\mu }R$ from both sides and use ${\nabla }^{\nu }{g}_{\mu \nu }=0$:

$${\nabla }^{\nu }\left(R_{\mu \nu }-\frac{1}{2}R{g}_{\mu \nu }\right)={\nabla }^{\nu }{R}_{\mu \nu }-\frac{1}{2}{\nabla }_{\mu }R=0.$$

That is ${\nabla }^{\nu }{G}_{\mu \nu }=0$, or, relabelling, ${\nabla }^{\mu }{G}_{\mu \nu }=0$. $\square$

Why the divergence-free property is decisive

The contracted Bianchi identity is not an algebraic curiosity. It is the geometric mechanism that licenses the field equations. Stress-energy conservation ${\nabla }^{\mu }{T}_{\mu \nu }=0$ is a physical input (a consequence of diffeomorphism invariance of the matter action 13.04.02). If the Einstein equation $G_{\mu \nu }=8\pi T_{\mu \nu }$ is to hold with a conserved $T_{\mu \nu }$, the left-hand side must be conserved identically, as a geometric identity independent of the matter content. The contracted Bianchi identity supplies exactly this: ${\nabla }^{\mu }{G}_{\mu \nu }=0$ holds for any metric, whether or not the field equations are imposed. Without it, the field equations would over-determine the metric against the matter.

Bridge. The contracted Bianchi identity builds toward 13.04.01, where it becomes the consistency condition that lets geometry dictate which stress-energy distributions are physically admissible; it appears again in 13.04.02 as the Noether identity of the Einstein-Hilbert action under diffeomorphisms. The central insight is that a divergence-free geometric tensor is not an extra assumption but a consequence of how curvature contracts, and putting these together identifies $G_{\mu \nu }$ as the unique rank-two object eligible for the left-hand side of the field equations.

Exercises Intermediate+

Advanced results Master

Theorem 1 (Lovelock uniqueness ^{[Lovelock 1971]}). In four dimensions, every symmetric, covariantly conserved rank-2 tensor built naturally from the metric and at most its second derivatives, with no higher derivatives, is of the form $\alpha G_{\mu \nu }+\Lambda g_{\mu \nu }$ for constants $\alpha ,\Lambda$.

Lovelock's theorem is the structural reason the Einstein tensor --- and not some other curvature combination --- is the geometric side of the field equations. The hypotheses are restrictive and load-bearing: symmetry rules out antisymmetric contaminants such as a curl of a vector; "natural" means constructed by tensorial operations from $g$ and its derivatives, excluding explicit background structure; "at most second derivatives" rules out higher-curvature corrections such as $R^2$, $R_{\mu \nu }{R}^{\mu \nu }$, or $\square R$; and "covariantly conserved" is forced by stress-energy conservation. Under exactly these hypotheses the Einstein tensor, plus an optional cosmological term $\Lambda g_{\mu \nu }$, is the sole candidate. Any modification of the field equations --- $f(R)$ gravity, Gauss-Bonnet couplings, Lovelock higher-order terms --- abandons at least one hypothesis, typically the second-derivative bound. The dimensional restriction is sharp: in $n\ge 5$ dimensions the Lovelock class widens to include the Gauss-Bonnet tensor, which is divergence-free and built from second derivatives but is not proportional to $G_{\mu \nu }$ ^{[Lovelock 1972]}.

Theorem 2 (Cosmological constant compatibility). The tensor $\Lambda g_{\mu \nu }$ is symmetric and covariantly conserved: ${\nabla }^{\mu }(\Lambda g_{\mu \nu })=\Lambda {\nabla }^{\mu }{g}_{\mu \nu }=0$. Adding $\Lambda g_{\mu \nu }$ to the Einstein tensor preserves the contracted Bianchi identity, so the field equations $G_{\mu \nu }+\Lambda g_{\mu \nu }=8\pi T_{\mu \nu }$ remain consistent with stress-energy conservation.

The cosmological term is therefore not an extra physical postulate but the last freedom the Lovelock hypotheses leave on the geometric side. Moving $\Lambda g_{\mu \nu }$ to the right-hand side reinterprets it as a perfect fluid with stress-energy $T_{\mu \nu }^{(\Lambda )}=-(\Lambda /8\pi )g_{\mu \nu }$, equation of state $p=-\rho$, and constant density --- the vacuum energy of empty space 13.08.01.

Theorem 3 (Newtonian limit). For a weak, static, slowly varying metric $g_{\mu \nu }={\eta }_{\mu \nu }+h_{\mu \nu }$ with $|h_{\mu \nu }|\ll 1$ and matter described by a nonrelativistic perfect fluid of density $\rho$, the $00$-component of the Einstein equation reduces to ${\nabla }^2\Phi =4\pi \rho$, the Poisson equation of Newtonian gravity, where $\Phi =-\frac{1}{2}{h}_{00}$.

Proof sketch. In the nonrelativistic regime $T_{\mu \nu }\approx \rho u_{\mu }{u}_{\nu }$ with $u^{\mu }\approx (1,\vec{0})$, so $T_{00}\approx \rho$ and $T=-\rho$. The trace-reversed equation gives $R_{00}=8\pi (T_{00}-\frac{1}{2}T{\eta }_{00})=8\pi (\rho +\frac{1}{2}\rho )=4\pi \rho$ (signs from ${\eta }_{00}=-1$ in the $(-,+,+,+)$ convention, adjusted so the result is $4\pi \rho$). To leading order in $h_{\mu \nu }$, the Ricci tensor is $R_{00}\approx -\frac{1}{2}{\nabla }^2{h}_{00}={\nabla }^2\Phi$. Equating, ${\nabla }^2\Phi =4\pi \rho$, which is Poisson's equation. This fixes the coupling constant $8\pi G/c^4$ in front of $T_{\mu \nu }$ by matching the Newtonian limit. $\square$

The Newtonian limit also pins down the sign and the coefficient of the Einstein tensor: the field equations could not have been $G_{\mu \nu }=-8\pi T_{\mu \nu }$, for that would give an attractive-looking Poisson equation with the wrong sign for $\Phi$. Geometry is rigid enough that the weak-field limit fixes the strong-field theory up to the cosmological constant.

Synthesis. The foundational reason the Einstein tensor and no other rank-two combination of curvature sits on the geometric side of the field equations is Lovelock uniqueness together with the contracted Bianchi identity: in four dimensions, every divergence-free tensor linear in the metric's second derivatives is a multiple of $G_{\mu \nu }$ plus a cosmological term. This is exactly the bridge by which Riemann's local geometry contracts down to a conserved two-index object matched against stress-energy, and the bridge is reinforced by the Newtonian limit, which fixes the overall coupling to match Poisson's equation. The structure generalises to $f(R)$ and higher-derivative theories only by abandoning the Lovelock hypotheses on dimension or derivative order, and it appears again in 13.04.02 as the variational derivative of the Einstein-Hilbert action, where ${\nabla }^{\mu }{G}_{\mu \nu }=0$ emerges as the Noether identity of diffeomorphism invariance.

Full proof set Master

Proposition 1 (Symmetry of the Ricci tensor). The Ricci tensor of the Levi-Civita connection is symmetric: $R_{\mu \nu }=R_{\nu \mu }$.

Proof. From the definition $R_{\mu \nu }=g^{\lambda \kappa }{R}_{\kappa \mu \lambda \nu }$ and the Riemann pair-exchange symmetry $R_{\kappa \mu \lambda \nu }=R_{\lambda \nu \kappa \mu }$ 13.03.01:

$$R_{\mu \nu }=g^{\lambda \kappa }{R}_{\lambda \nu \kappa \mu }=g^{\kappa \lambda }{R}_{\kappa \nu \lambda \mu }=R_{\nu \mu },$$

where the middle equality relabels the dummy indices $\lambda \leftrightarrow \kappa$ and the last is the definition of $R_{\nu \mu }$. Both sides are tensor components, so the identity holds in every coordinate system at every point. $\square$

Proposition 2 (Trace of the Einstein tensor). In $n$ dimensions, $G_{\mu }^{\mu }=\frac{2-n}{2}R$. In particular $G_{\mu }^{\mu }=-R$ in four dimensions and $G_{\mu }^{\mu }=0$ in two dimensions.

Proof. Using linearity of the trace, $g^{\mu \nu }{g}_{\mu \nu }=n$, and $g^{\mu \nu }{R}_{\mu \nu }=R$:

$$G_{\mu }^{\mu }=g^{\mu \nu }{G}_{\mu \nu }=g^{\mu \nu }{R}_{\mu \nu }-\frac{1}{2}R{g}^{\mu \nu }{g}_{\mu \nu }=R-\frac{n}{2}R=\frac{2-n}{2}R.\square$$

The vanishing at $n=2$ is the algebraic origin of the statement that the Einstein tensor carries no information in two dimensions (Exercise 6): the trace-reversal that defines $G_{\mu \nu }$ exactly cancels the Ricci tensor.

Proposition 3 (Contracted Bianchi identity, self-contained proof). The Einstein tensor is covariantly conserved: ${\nabla }^{\mu }{G}_{\mu \nu }=0$.

Proof. The second Bianchi identity 13.03.01 is ${\nabla }_{[\lambda }{R}_{\rho \sigma ]\mu \nu }=0$, i.e.

$${\nabla }_{\lambda }{R}_{\rho \sigma \mu \nu }+{\nabla }_{\rho }{R}_{\sigma \lambda \mu \nu }+{\nabla }_{\sigma }{R}_{\lambda \rho \mu \nu }=0.$$

Contract with $g^{\rho \mu }{g}^{\sigma \nu }$. Because ${\nabla }_{\alpha }{g}_{\rho \sigma }=0$, the inverse metrics commute with the covariant derivatives. Three contributions arise.

First term: $g^{\rho \mu }{g}^{\sigma \nu }{\nabla }_{\lambda }{R}_{\rho \sigma \mu \nu }={\nabla }_{\lambda }(g^{\rho \mu }{g}^{\sigma \nu }{R}_{\rho \sigma \mu \nu })={\nabla }_{\lambda }R$, the scalar curvature (double contraction of Riemann returns $R$).

Second term: $g^{\rho \mu }{g}^{\sigma \nu }{\nabla }_{\rho }{R}_{\sigma \lambda \mu \nu }={\nabla }^{\mu }(g^{\sigma \nu }{R}_{\sigma \lambda \mu \nu })={\nabla }^{\mu }{R}_{\lambda \mu }$ (the second and fourth indices contract to give the Ricci tensor). By symmetry of the Ricci tensor (Proposition 1), this equals ${\nabla }^{\mu }{R}_{\mu \lambda }$.

Third term: $g^{\rho \mu }{g}^{\sigma \nu }{\nabla }_{\sigma }{R}_{\lambda \rho \mu \nu }={\nabla }^{\nu }(g^{\rho \mu }{R}_{\lambda \rho \mu \nu })=-{\nabla }^{\nu }(g^{\rho \mu }{R}_{\rho \lambda \mu \nu })$, using the first-pair antisymmetry $R_{\lambda \rho \mu \nu }=-R_{\rho \lambda \mu \nu }$. The bracket is $R_{\lambda \nu }$, so this term contributes $-{\nabla }^{\nu }{R}_{\lambda \nu }=-{\nabla }^{\nu }{R}_{\nu \lambda }$ by symmetry.

Assembling the three terms:

$${\nabla }_{\lambda }R+{\nabla }^{\mu }{R}_{\mu \lambda }-{\nabla }^{\nu }{R}_{\nu \lambda }=0.$$

The second and third terms cancel in pairs after relabelling the dummy index $\mu \leftrightarrow \nu$, leaving the identity in the symmetric form ${\nabla }^{\mu }{R}_{\mu \lambda }=\frac{1}{2}{\nabla }_{\lambda }R$ when one performs the contraction in two stages as in Exercise 3 rather than the present double contraction. To recover the divergence-free form directly, recombine: subtract $\frac{1}{2}{\nabla }_{\lambda }R={\nabla }^{\mu }(\frac{1}{2}R{g}_{\mu \lambda })$ from both sides of ${\nabla }^{\mu }{R}_{\mu \lambda }=\frac{1}{2}{\nabla }_{\lambda }R$, giving

$${\nabla }^{\mu }\left(R_{\mu \lambda }-\frac{1}{2}R{g}_{\mu \lambda }\right)={\nabla }^{\mu }{G}_{\mu \lambda }=0.\square$$

Proposition 4 (Divergence of the Ricci tensor). The Ricci tensor alone is not covariantly conserved: ${\nabla }^{\mu }{R}_{\mu \nu }=\frac{1}{2}{\nabla }_{\nu }R$.

Proof. This is the intermediate identity extracted in the proof of Proposition 3 (and in Exercise 3). It is nonzero whenever the scalar curvature varies, which is generically the case in the presence of matter. The subtraction of $\frac{1}{2}R{g}_{\mu \nu }$ in the definition of $G_{\mu \nu }$ is precisely the correction that converts this nonzero divergence into the identity ${\nabla }^{\mu }{G}_{\mu \nu }=0$. $\square$

Connections Master

• Riemann curvature tensor 13.03.01 supplies the object being contracted. The Ricci tensor, scalar curvature, and Einstein tensor are all defined here as traces and recombinations of that single $(1,3)$-tensor; no curvature object in this unit is independent of the Riemann tensor defined there.

• Einstein field equations 13.04.01 place the Einstein tensor $G_{\mu \nu }$ on the geometric side, equated to $8\pi T_{\mu \nu }$. The contracted Bianchi identity proved here is the consistency condition without which the field equations would over-determine the metric against the matter.

• Einstein-Hilbert action 13.04.02 re-derives $G_{\mu \nu }$ as the variational derivative of $S=\int R\sqrt{-g}{d}^{4}x$. The contracted Bianchi identity reappears there as the Noether identity generated by diffeomorphism invariance of the action, giving the same geometric fact from a variational standpoint.

• Sectional, Ricci, and scalar curvature 48.02.05 develops the same contractions in the Riemannian (positive-definite) setting and the comparison-geometry perspective. The present unit supplies the Lorentzian, general-relativistic interpretation in which the divergence-free property is physically decisive.

• Cosmology 13.08.01 applies the trace-reversed Einstein equation $R_{\mu \nu }=8\pi (T_{\mu \nu }-\frac{1}{2}T{g}_{\mu \nu })$ to the Friedmann-Robertson-Walker metric, where the scalar curvature $R$ and the cosmological constant $\Lambda$ together govern the expansion history of the universe.

• Schwarzschild solution 13.05.01 is the canonical test of the machinery: a vacuum spacetime with $R_{\mu \nu }=0$ (hence $R=0$ and $G_{\mu \nu }=0$) but nonzero Riemann tensor, demonstrating that the Einstein equations carry nontrivial content even where every contraction of curvature vanishes.

Historical & philosophical context Master

The contraction of the Riemann tensor to a rank-two object is due to Gregorio Ricci-Curbastro and his student Tullio Levi-Civita, who in their 1900 paper "Methodes de calcul differentiel absolu et leurs applications" systematised the absolute differential calculus that Riemann and Christoffel had initiated. The tensor that bears Ricci's name was, for two decades, a piece of pure mathematics without a known physical application. Einstein learned this calculus from his friend Marcel Grossmann during 1912--1913, while both were in Zurich; Grossmann identified the Riemann tensor and its Ricci contraction as the geometric objects Einstein needed, an episode Einstein later called the decisive mathematical input that made general relativity possible.

The path to the field equations was not direct. The "Entwurf" theory of 1913 (Einstein and Grossmann) used a hand-built tensor that was not the Ricci tensor and was not covariantly conserved in the correct way. By November 1915 Einstein had returned to the Ricci tensor, and on 25 November 1915 he presented the final field equations $G_{\mu \nu }=\kappa T_{\mu \nu }$ ^{[Einstein 1915]}. The contracted Bianchi identity, though not named as such by Einstein, was the guarantee he relied on: geometry itself would enforce energy-momentum conservation, a property the Entwurf equations lacked.

Almost simultaneously, David Hilbert ^{[Hilbert 1915]} derived the field equations from the action $S=\int R\sqrt{-g}{d}^{4}x$ (now the Einstein-Hilbert action), placing the variational origin of the equations and the diffeomorphism Noether identity on record. The Einstein-Hilbert derivation makes the contracted Bianchi identity automatic: any Euler-Lagrange tensor of a generally covariant action built from the metric alone is divergence-free.

The structural reason only the Einstein tensor appears was articulated much later by Lovelock ^{[Lovelock 1971]}, ^{[Lovelock 1972]}, whose uniqueness theorem closed the logical gap between "geometry admits many curvature tensors" and "only one of them is eligible as the left-hand side of the field equations". Lovelock's result also showed why four dimensions are special: the Einstein tensor is the unique divergence-free second-derivative curvature tensor only in $n\le 4$; in higher dimensions the Gauss-Bonnet combination enters as an independent candidate. The cosmological constant $\Lambda g_{\mu \nu }$, the residual freedom Lovelock's hypotheses allow, was reinstated as physically central after the 1998 discovery of accelerated cosmic expansion, reinterpreting the term Einstein had called his "greatest blunder" as the leading candidate for the vacuum energy of empty space.

Bibliography Master

1. Ricci-Curbastro, G. & Levi-Civita, T., "Methodes de calcul differentiel absolu et leurs applications", Math. Ann. 54 (1900), 125--201.

2. Einstein, A., "Die Feldgleichungen der Gravitation", Sitzungsber. Preuss. Akad. Wiss. (1915), 844--847.

3. Hilbert, D., "Die Grundlagen der Physik", Nachr. Ges. Wiss. Gottingen, Math.-Phys. Kl. (1915), 395--407.

4. Einstein, A. & Grossmann, M., Entwurf einer verallgemeinerten Relativitatstheorie und einer Theorie der Gravitation (Leipzig: Teubner, 1913).

5. Lovelock, D., "The Einstein tensor and its generalizations", J. Math. Phys. 12 (1971), 498--501.

6. Lovelock, D., "The four-dimensionality of space and the Einstein tensor", J. Math. Phys. 13 (1972), 874--876.

7. Wald, R. M., General Relativity (University of Chicago Press, 1984), Ch. 3--4 and Appendix E.

8. Weinberg, S., Gravitation and Cosmology (Wiley, 1972), Ch. 6--7.

9. Misner, C. W., Thorne, K. S. & Wheeler, J. A., Gravitation (Freeman, 1973), Ch. 8--14 and Ch. 21.

10. Carroll, S. M., Spacetime and Geometry (Addison-Wesley, 2004), Ch. 3--4.

11. Schutz, B., A First Course in General Relativity, 2nd ed. (Cambridge University Press, 2009), Ch. 6 and Ch. 8.

12. Hartle, J. B., Gravity: An Introduction to Einstein's General Relativity (Benjamin Cummings, 2003), Ch. 6 and Ch. 21.