Agnostic Endpoint Data for Quantum Gravity

Before predicting an evaporation endpoint, identify what can be calculated on the last controlled hypersurface: conserved charges and their fluxes, intrinsic and extrinsic geometry with gauge information, the renormalized state data needed to continue observables, correlations already established, the causal boundary, and explicit validity inequalities with uncertainties. Any microscopic completion must match these quantities where both descriptions overlap.

Helpful background. Quantum Fields in Curved Spacetime supplies state and correlator data; Einstein Equations supplies constraints; Black-Hole Geometry and Mechanics supplies invariant horizon data; Semiclassical Gravity supplies the last controlled mean geometry; Black-Hole Information assesses fine-grained correlations; and Quantum Gravity receives the microscopic endpoint problem.

Choose the last controlled surface

Let $\Sigma_*$ be a spacelike or characteristic hypersurface before the first declared control parameter reaches its tolerance. It should not be placed at a guessed singularity or endpoint. Record

$$\epsilon_{\rm ctrl} =\max\{\epsilon_R,\epsilon_\omega,\epsilon_{\rm ad}, \epsilon_{\rm fluc},\epsilon_{\rm trunc}\} \le\epsilon_*, \qquad \epsilon_*\ll1,$$

and state how each quantity was measured. The distance from $\Sigma_*$ to the estimated breakdown region must exceed the numerical and asymptotic error in its location.

For a spacelike surface, the geometric data are $(h_{ij},K_{ij})$ modulo spatial diffeomorphisms, with $K_{ij}=h_i{}^ah_j{}^b\nabla_an_b$ for the future unit normal; lapse and shift are gauge choices. This is the metric-data pair for an Einstein or consistently order-reduced formulation. If extra low-energy fields or higher-derivative degrees of freedom are deliberately retained, their independent data must also be supplied. A characteristic formulation instead uses null metric, shear, connection, and normalization data, typically on intersecting null surfaces with corner data or with a complementary boundary condition. One null sheet is not automatically a complete initial-value problem.

Charges, fluxes, and constraints

At minimum, provide Bondi or asymptotic mass, angular momentum, gauge charges, and any exactly conserved global or topological charges, together with flux integrals and uncertainty:

$$Q_A[\Sigma_*]-Q_A[\Sigma_0] =-\int_{\mathcal B}F_A\,d\lambda+\delta_A,$$

where $\mathcal B$ is the intervening boundary and $\delta_A$ includes discretization, subtraction, and omitted-order errors. The final record should quote $\delta_A$, not merely assert conservation.

The Hamiltonian and momentum constraints, or their null counterparts, must close on $\Sigma_*$. A microscopic model cannot be asked to repair a constraint violation inherited from an inconsistent semiclassical solution. Renormalized couplings and their scheme must accompany the geometry because changing them changes the split between local gravitational terms and state-dependent stress.

The four-dimensional construction claimed by del Río and Marañón-González illustrates the kind of stress-tensor family, conservation, and asymptotic flux data one would test del Río and Marañón-González 2026, Eqs. (6)–(12) and the quasi-stationary and Bondi-mass formulas on p. 4, PDF. Its local coincidence-data construction still requires a positive Hadamard two-point bisolution with the specified global boundary conditions. Consequently the candidate flux and geometry are conditional inputs, rather than established endpoint data. The arXiv record checked on 2026-09-26 remains v1 without a journal reference.

Quantum state and correlation data

A one-point stress tensor does not determine a quantum state. The record should specify the algebra or mode basis, state prescription, renormalization subtraction, and the domain on which Hadamard or other regularity holds. For a quasifree state, the one- and two-point functions determine the higher correlators; for a general interacting or non-Gaussian state they do not. A reduced covariance is sufficient only for a declared Gaussian approximation and the observables it controls. If stochastic fluctuations are relevant, include the smeared noise kernel, the retarded response, and any intrinsic metric covariance.

Fine-grained correlations already computed between outgoing radiation and retained degrees of freedom can constrain a completion, but their interpretation belongs to Black-Hole Information. Report which radiation algebra and time interval were used, whether entropy is von Neumann, generalized, or coarse grained, and the error from tracing out unmodeled sectors. A thermal spectrum alone does not fix purification data.

Causal information includes the domain of dependence of $\Sigma_*$, outer and inner trapping surfaces with slicing labels, event-horizon claims only when a future completion has been specified, and any Cauchy or chronology horizon approached by the controlled region. Do not draw a future event horizon using the endpoint model and then present it as input to that model.

Results and open questions through 2026-08-19

Evidence through 2026-08-19 supports conservation records and explicit breakdown surfaces in several models. Which additional state data a quantum-gravity completion must receive remains unresolved: proposed minimal records are conjectural, and no universal verdict follows from the available models.

Result or model	Reliably determined quantities	Limitation near semiclassical breakdown
del Río and Marañón-González 2026, p. 4, quasi-stationary solution and Bondi-mass discussion, PDF	Candidate Bondi mass and flux history, conditional on the proposed stress family and retained order	Global state realizability and fluctuations are not supplied; deep endpoint geometry is outside the approximation
Russo, Susskind, and Thorlacius 1992, §§ 1–4, PDF	Exact modified-model variables, apparent-horizon/singularity meeting, and a boundary-conditioned endpoint in two dimensions	Counterterm, state, boundary prescription, dimension, and field content are model specific
del Río and Marañón-González 2026, Eqs. (1)–(14) and p. 4, PDF	Constructed conserved stress family, flux, and claimed backreacted geometry	Recent unreviewed v1 without independent reproduction in the acquired corpus
Giddings 1995, §§ I–IV, PDF	Under an effective low-energy description with infinitely many internal remnant states, the relevant thermal trace and pair-production estimate can diverge	The conclusion depends on how internal states couple; it is an adverse consistency test, not an endpoint calculation or unconditional no-go theorem

The conservative position exports only gauge-invariant observables, constraint-satisfying surface data, and quantified error bars. The competing richer-record position also exports a chosen reduced state or effective interior variables to make matching practical. The latter can be useful, but it depends on factorization, gauge dressing, state reconstruction, and cutoff choices. A completion should be tested on both the invariant core and the declared auxiliary structure.

Data needed to compare endpoint models

A practical record contains:

• the location and type of $\Sigma_*$ and its coordinate-to-invariant map;
• geometric data, constraints, and numerical or asymptotic covariance;
• charges, flux integrals, balance residuals, and normalization conventions;
• field theory, quantum state, two-point or reduced-state data, and entropy type;
• renormalized couplings, regulator translation, and first omitted terms;
• causal domains and horizon definitions without endpoint-dependent circularity;
• the list of failed or soon-to-fail control parameters; and
• alternative records produced by credible branch or state choices.

Microscopic singularity resolution, topology change, and the final Hilbert space belong to Quantum Gravity. The quantities that semiclassical theory can determine remain useful: generalized entropy, mean geometry, backreaction, conserved fluxes, and a diagnosed breakdown.

The conclusions could change after a UV-complete model publishes an explicit matching map, an independent reproduction changes the last controlled four-dimensional data, or a theorem identifies a gauge-invariant minimal state record; a new numerical covariance analysis or a corrected endpoint geometry could also change the conclusion.

Exercise

A calculation reports $M_*=1.00\pm0.01$ and radiated energy $E_{\rm rad}=9.05\pm0.02$ from initial mass $M_0=10.00\pm0.01$, with no other flux. Is the balance consistent at one standard uncertainty?

Solution

The balance residual is $M_0-M_*-E_{\rm rad}=-0.05$. Assuming independent errors, its standard uncertainty is $\sqrt{0.01^2+0.01^2+0.02^2}\simeq0.0245$. The residual is about $2.04$ standard deviations from zero and hence outside the one-standard-uncertainty interval. This flags a discrepancy to investigate, not a proof that the equations or measurements are wrong. Correlated estimates require the full covariance: with $c=(1,-1,-1)$, $\operatorname{Var}(M_0-M_*-E_{\rm rad})=c^T\Sigma c$. Report the residual and error model honestly; enlarge a systematic uncertainty only for an identified reason.

References

• del Río, A., and Marañón-González, F. J. (2026). “Black-Hole Evaporation from the Semiclassical Einstein Equations in 3+1 Dimensions.” arXiv:2607.11737v1. Open PDF.
• Giddings, S. B. (1995). “Why Aren’t Black Holes Infinitely Produced?” Physical Review D 51, 6860–6869. DOI. Open PDF.
• Russo, J. G., Susskind, L., and Thorlacius, L. (1992). “End Point of Hawking Radiation.” Physical Review D 46, 3444–3449. DOI. Open PDF.