Initial Data for Geometry and Quantum State

Initial data in semiclassical gravity are joint data for geometry and quantum matter. A spatial metric and extrinsic curvature must satisfy constraints sourced by the state’s renormalized stress tensor, while the state must be admissible and evolved on the geometry it helps determine. Because the stress expectation is nonlocal and higher-curvature terms can raise the formal differential order, there is no general theorem that every plausible pair of classical and quantum initial data produces a unique, globally controlled semiclassical spacetime.

Required background. Einstein Equations supplies the Hamiltonian and momentum constraints, and Gravitational Actions and Boundary Terms supplies the chosen variational and initial-boundary problem.

Helpful background. Quantum Fields in Curved Spacetime supplies Hadamard state data and field evolution, while Black-Hole Geometry and Mechanics supplies the horizon regularity conditions used in later examples.

Joint geometric and quantum data

On a spacelike Cauchy surface $\Sigma$ with future unit normal $n^\mu$, take

$$h_{\mu\nu}=g_{\mu\nu}+n_\mu n_\nu, \qquad K_{ij}=-h_i{}^\mu h_j{}^\nu\nabla_\mu n_\nu.$$

The sign of $K_{ij}$ is declared here because it affects the constraint formulas. For the two-derivative Einstein part, the state supplies

$$\rho_\omega=n^\mu n^\nu\langle T_{\mu\nu}\rangle_\omega, \qquad j_i=-h_i{}^\mu n^\nu\langle T_{\mu\nu}\rangle_\omega,$$

and the constraints are schematically

$${}^{(3)}R+K^2-K_{ij}K^{ij}-2\Lambda =16\pi G\rho_\omega+\rho_{\mathrm{hd}},$$ $$D_j(K^{ij}-h^{ij}K)=8\pi G j^i+j^i_{\mathrm{hd}}.$$

The higher-derivative terms are computed from the same renormalized action. They are not extra matter chosen independently. The quantum state may be encoded by its $n$-point functions or density operator; for a quasifree scalar state, equal-time two-point data and their normal derivatives must satisfy the commutation relations, positivity, field equation, and Hadamard short-distance condition.

Why the data are coupled

One cannot first solve arbitrary classical constraints and then attach an arbitrary state. The state’s energy and momentum appear in those constraints, while its notion of positive frequency and its future evolution depend on the metric. A self-consistent solution is a triple consisting of a metric, field operator, and admissible state that satisfies the renormalized equation together Hu and Verdaguer 2008, § 3.1, following Eq. (3.10).

Real-time expectation values also retain memory of the initial density matrix. In the closed-time-path construction, that density matrix weights and correlates the two initial field configurations; the forward and backward histories are sewn together at the later return surface. The resulting kernel is retarded, but not generally local in time. Anderson, Molina-París, and Mottola 2003, §§ II–III, PDF make the initial-state and causal response structure explicit for linearized semiclassical gravity.

If the four-derivative equation is treated as fundamental, extra time derivatives appear to demand extra geometric data. In an effective theory truncated at first quantum order, order reduction instead expresses those derivatives using lower-order equations, leaving the classical amount of physical initial data through the retained order. These are different formulations and must not be mixed.

Results and open questions as of 2026-08-19

Evidence or formulation	Result	Assumptions	Limitation or unresolved issue
Conserved-source semiclassical constraints, Hu and Verdaguer 2008, §3.1, Eqs. (3.8)–(3.10)	Initially satisfied gravitational constraints propagate with the evolution equations	Common metric, differentiable source, and $\nabla^\mu T_{\mu\nu}=0$	Does not construct the quantum state or guarantee source regularity
Hadamard admissibility on a prescribed background, Hu and Verdaguer 2008, §3.1, discussion around Eqs. (3.5)–(3.7)	Renormalized stress-tensor matrix elements are finite for the stated physically acceptable Hadamard states	Fixed smooth globally hyperbolic geometry and specified state data	Does not supply a coupled existence theorem or by itself prove propagation for state–geometry data
Closed-time-path linear response, Anderson, Molina-París, and Mottola 2003, §§ II–III, PDF	State-dependent response is causal and includes initial data	Perturbations of a self-consistent background, chosen contour, and state	Only linearized control; initial singularities and nonlinear feedback remain
Perturbative order reduction, Parker and Simon 1993, § II and Appendix A, PDF	Removes extra initial data and nonanalytic runaway branches at retained order	Low-energy derivative expansion and fixed truncation	Alternative contour prescriptions can differ beyond the retained order
Symmetry-reduced solutions, Hu and Verdaguer 2008, §§7–8	Demonstrate that joint state–geometry solutions can be constructed in examples	Strong symmetry, chosen state, finite couplings, and boundary data	No general global well-posedness or stability conclusion

Parker and Simon derive the perturbative reduction and show why physical data remain of classical order in Parker and Simon 1993, § 2 and Appendix A, PDF. That result supports an effective-theory initial-value prescription; it does not prove that an exact higher-derivative completion has no additional degrees of freedom.

Competing formulations

The effective-field-theory position admits only data analytic in the small parameter and uses order reduction or an equivalent low-energy contour. Its strength is a controlled remainder and the exclusion of cutoff-scale runaways. Its weakness is that it deliberately cannot decide whether a UV completion contains additional healthy modes.

The enlarged-data position treats the renormalized four-derivative equation as a differential equation in its own right, specifying the additional derivatives on $\Sigma$. It explores a larger solution space, but many branches have frequencies at the cutoff and are not continuously connected to classical solutions. Without a UV stability and energy prescription, formal existence is not evidence of semiclassical control.

A third, nonlocal formulation retains the full causal response kernel instead of replacing it by a finite derivative series. It can avoid some local runaway artifacts, but it needs the entire initial-state prescription and may introduce memory kernels whose nonlinear well-posedness is difficult. These positions coincide only through the perturbative order at which their matching is established.

Consistency tests and limitations

Three checks should precede evolution. Evaluate the Hamiltonian and momentum residuals using the renormalized state on the proposed metric. Verify the stress Ward identity to the retained order. Finally, perturb both geometric and state data within the admissible class and confirm that the target observable varies continuously below the cutoff.

A counterexample is flat initial geometry with $K_{ij}=0$, $\Lambda=0$, and a homogeneous thermal state of positive energy. The state is perfectly admissible on fixed Minkowski space, but the joint Hamiltonian constraint is not satisfied. Conversely, constraint-compatible expectation values do not guarantee small noise or stable response.

Fixed-background state construction remains with Quantum Fields in Curved Spacetime. Microscopic choices that resolve additional modes belong to Quantum Gravity, and fine-grained recovery belongs to Black-Hole Information. The semiclassical problem ends when no joint low-energy data set makes the constraints, state regularity, response, and truncation simultaneously controlled.

This page should be reassessed after a general local or global well-posedness theorem for coupled Hadamard state–geometry data; an independent nonlinear calculation that shows loss or restoration of constraint propagation; a demonstrated equivalence or inequivalence of order reduction and causal contour prescriptions below cutoff; or a counterexample in which admissible small data develop uncontrolled stress fluctuations. Corrections to the central response or order-reduction sources also trigger reassessment.

Exercise

Why do Minkowski initial geometry, vanishing $K_{ij}$, $\Lambda=0$, and a thermal state with $\rho\gt 0$ fail as joint data? Give three distinct repairs.

Solution

The left side of the Hamiltonian constraint vanishes, while the right side contains $16\pi G\rho\gt 0$, so the joint data violate the constraint even though both pieces make sense separately. One may choose curved or expanding geometric data that satisfy the sourced constraint, choose a zero-energy state, or change the cosmological parameter to $\Lambda=-8\pi G\rho$ so that the initial Hamiltonian constraint balances. The last option changes the physical theory or its measured parameters; it is not a state redefinition or a compensated scheme change. It repairs this constraint on the initial slice, not a static Minkowski solution: a thermal source also has nonvacuum pressure, and the spatial evolution equations generally make the geometry evolve.

References

• Anderson, P. R., Molina-París, C., and Mottola, E. 2003. “Linear Response, Validity of Semi-Classical Gravity, and the Stability of Flat Space.” Physical Review D 67, 024026. Open PDF.
• Hu, B. L., and Verdaguer, E. 2008. “Stochastic Gravity: Theory and Applications.” Living Reviews in Relativity 11, 3. DOI.
• Parker, L., and Simon, J. Z. 1993. “Einstein Equation with Quantum Corrections Reduced to Second Order.” Physical Review D 47, 1339–1355. Open PDF.