Semiclassical Einstein Equation: Existence and Stability Results

The semiclassical Einstein equation is a coupled state–geometry problem, not Einstein’s equation with a precomputed number inserted on the right. Its source is a conserved renormalized expectation value in a state defined on the same metric being solved for. General local well-posedness is not established. Rigorous existence results apply to restricted symmetry classes and state families; linear stability results apply to declared backgrounds, gauges, response kernels, length scales, and prescriptions for higher-derivative modes.

Required background. The renormalized stress tensor supplies the conserved source and curvature counterterms. Existence and gluing of Hadamard states supplies admissible states. Perturbative agreement and background independence supplies the comparison of nearby backgrounds.

Helpful background. Trace anomalies and scaling supplies the local trace. Stress–energy response supplies relative Cauchy variation. The curved-spacetime sequence covers the semiclassical Einstein equation, state–geometry initial data, in-in effective actions, constraints and the Bianchi identity, self-consistent solutions, state evolution on backreacted backgrounds, the higher-derivative initial-value problem, order reduction and runaways, linear-response stability, cosmological benchmarks, black-hole evaporation, and numerical error budgets. Causal, thermodynamic, and stability constraints gives the adjacent theorem domains.

The coupled equation and its data

In four dimensions a renormalized equation may be written

$$G_{ab}+\Lambda g_{ab}+\alpha I_{ab}+\beta J_{ab} =8\pi G\,\langle T_{ab}\rangle_{\omega[g]},$$

where $I_{ab}$ and $J_{ab}$ are the conserved variations of curvature-squared actions. The constants $G,\Lambda,\alpha,\beta$ and the stress prescription must be matched at one renormalization scale. The state $\omega[g]$ must satisfy the field equation and Hadamard condition on the unknown geometry. Conservation of the right side is required by the Bianchi identity; it does not follow from a truncated mode sum unless the subtraction and state evolution are consistent.

Initial data therefore include more than a spatial metric and extrinsic curvature. They include suitable two-point-function data satisfying positivity, the commutator, field equation, constraints, and a surface form of the Hadamard condition. Curvature-squared terms raise the differential order unless treated perturbatively. These facts block a blanket appeal to the classical Einstein Cauchy theorem.

There are nevertheless controlled theorems in symmetric settings. For a spatially flat FLRW geometry and specified homogeneous quasifree scalar states, the equations can be reduced to a trace evolution plus an initial constraint and a causal integral equation. Pinamonti proves existence in a restricted cosmological construction in Pinamonti 2011, §§ 3–4, pp. 581–600; later global results impose additional state and parameter hypotheses. These do not establish well-posedness for arbitrary inhomogeneous four-dimensional data.

Causal linear response about Minkowski space

Let $g_{ab}=\eta_{ab}+h_{ab}$ and take the Minkowski vacuum of a free massive scalar as the background state. After gauge fixing and renormalizing local contact terms, the first variation has the form

$$\delta\langle T_{ab}(x)\rangle =\delta T^{\mathrm{local}}_{ab}[h](x) +\int\mathrm d^4x'\,Pi^{\mathrm{ret}}_{ab}{}^{cd}(x,x')h_{cd}(x'),$$

with

$$\Pi^{\mathrm{ret}}_{ab}{}^{cd}(x,x') =-i\,\theta(x^0-x'^0) \langle[T_{ab}(x),T^{cd}(x')]\rangle_0$$

up to the declared local contact terms. The commutator makes the kernel retarded; an in-out effective action would instead give the wrong boundary condition for causal evolution. Stress conservation makes the response transverse after local terms are included. Decomposing $h_{ab}$ into gauge, scalar, and transverse-traceless channels then reduces the stability question to the poles and branch cuts of renormalized response functions.

Anderson, Molina-París, and Mottola use this spectral representation to test flat space and find no growing gauge-invariant perturbations on length scales much larger than the Planck scale for the free-field model, while Planck-scale roots lie outside the controlled semiclassical regime Anderson, Molina-París, and Mottola 2003, §§ III–V, pp. 024026-7–024026-17. This is a linear-response criterion about one background, not nonlinear stability of semiclassical gravity.

The concrete calculation on linear-response stability and runaway solutions builds the massive-scalar retarded kernel and states the gauge and scale cutoff before classifying modes.

Higher derivatives and order reduction

The tensors $I_{ab}$ and $J_{ab}$ contain four metric derivatives. If their coefficients are treated exactly, additional modes appear, often with frequencies of order the inverse semiclassical expansion scale. Order reduction instead uses the lower-order Einstein equation inside the higher-derivative correction and retains only terms through the declared power of $G$ or curvature. It discards solutions nonanalytic in that expansion. Flanagan and Wald use a closely controlled perturbative treatment through second order about flat space in Flanagan and Wald 1996, §§ II–IV, pp. 6241–6267.

Order reduction is a prescription defining the approximation’s solution space. It is not proof that the full fourth-order equation lacks additional solutions. Conversely, keeping all exact roots while trusting an effective equation at their Planckian frequencies exceeds its domain.

Failure boundary: count the runaway consistently

An adversarial stability test solves the fourth-order truncated equation exactly, retains an exponentially growing high-frequency root, and declares the low-energy background unstable. Apply the declared order-reduction condition: the root is nonanalytic in the small higher-curvature coefficient and is removed, while the low-frequency causal response remains. Under that prescription, the instability claim fails.

The opposite overclaim also fails. Removing every inconvenient mode without showing that it lies beyond the expansion or violates admissible initial data cannot prove stability. A valid conclusion must name the retained function space, gauge quotient, response boundary condition, renormalization constants, and maximum frequency. No result here proves global existence or nonlinear stability for generic spacetime.

Exercises

Why must the linear-response kernel be transverse?

Solution

The renormalized stress tensor is conserved for every metric in the family. Differentiating $\nabla^a\langle T_{ab}\rangle=0$ produces a Ward identity relating the divergence of the nonlocal kernel to local variations of the connection and background stress. Once the required contact terms are included, the complete linearized source is conserved. Therefore pure-gauge metric perturbations do not generate an independent physical response.

References

• Anderson, Paul R., Carmen Molina-París, and Emil Mottola. “Linear Response, Validity of Semiclassical Gravity, and the Stability of Flat Space.” Physical Review D 67 (2003): 024026. DOI; Open PDF.
• Flanagan, Éanna É., and Robert M. Wald. “Does Back Reaction Enforce the Averaged Null Energy Condition in Semiclassical Gravity?” Physical Review D 54 (1996): 6233–6283. DOI; Open PDF.
• Pinamonti, Nicola. “On the Initial Conditions and Solutions of the Semiclassical Einstein Equations in a Cosmological Scenario.” Communications in Mathematical Physics 305 (2011): 563–604. DOI; Open PDF.