Conceptual Evidence and Consistency Pressures

Quantum gravity is motivated by several sharply different facts. Some are theorems that classical general relativity becomes geodesically incomplete under stated hypotheses; some are controlled perturbative calculations showing that the Einstein action is not a complete ultraviolet quantum theory; and some are thermodynamic or information-theoretic clues whose microscopic interpretation remains open. Calling all of them “evidence” obscures what has actually been established. This page separates theorem, calculation, observation, and research judgement.

Required background. Einstein Equations supplies the classical field equation and its energy conditions. Quantum Mechanics supplies states, unitarity, and measurement.

Helpful background. Flat-Spacetime Quantum Field Theory supplies perturbative renormalization, while Quantum Information sharpens the entropy and recoverability questions.

Kinds of scientific support

The evidence cutoff for this assessment is 2026-08-20. “Consistency pressure” will mean a demonstrable tension between a domain of an established framework and a question posed outside that domain. It is not automatically a logical contradiction. A singularity theorem, for example, concludes causal-geodesic incompleteness—not divergent curvature everywhere, not the existence of a point belonging to spacetime, and not a unique quantum resolution. Hawking and Penrose prove incompleteness from causal, focusing, and genericity assumptions in a broad class of spacetimes Hawking and Penrose 1970, theorem and discussion, pp. 529–548.

Likewise, perturbative nonrenormalizability is a statement about a chosen low-energy field content and an expansion about a background. The two-loop divergence found by Goroff and Sagnotti requires a new cubic-curvature counterterm on shell Goroff and Sagnotti 1985, pp. 81–86. That establishes that pure Einstein gravity is not closed under perturbative renormalization as a fundamental theory. It does not establish that low-energy quantum GR is inconsistent: effective field theory absorbs the divergence into a higher-derivative coupling and remains predictive order by order below its cutoff.

Black-hole thermodynamics supplies a different kind of clue. Bekenstein’s generalized-entropy argument makes horizon area behave as entropy and fixes a strong target for microscopic state counting Bekenstein 1973, §§II–V, pp. 2333–2346. The area law is a robust semiclassical result, but agreement with it is a correspondence test rather than a unique discriminator: inequivalent microscopic constructions can reproduce the same leading thermodynamics.

Results and limitations

The table records what each pressure actually bears on. “Unresolved” means that the established result underdetermines the microscopic completion, not that the result itself is doubtful.

Pressure or programme	Degrees of freedom and observable	Assumptions	Established result at 2026-08-20	Limitation or live alternative
Classical singularity theorems	Lorentzian metric; causal-geodesic extendibility	Focusing condition, causal hypotheses, genericity, suitable trapped or cosmological data	Incompleteness follows in the theorem’s domain	Violating an assumption can evade the conclusion; no microscopic dynamics is selected
Perturbative Einstein gravity	Weak metric fluctuations; on-shell amplitudes and ultraviolet poles	Four dimensions, local covariant perturbation theory, Einstein field content	A nonzero two-loop counterterm is required	The low-energy EFT remains consistent; nonlocal or new-field ultraviolet completions are not excluded
Black-hole thermodynamics	Semiclassical geometry and quantum fields; entropy, temperature, flux	Stationarity or adiabaticity, controlled horizon geometry, semiclassical state	Universal thermodynamic relations constrain any correspondence limit	Leading entropy does not uniquely identify microscopic degrees of freedom
Information-loss problem	Quantum state and exterior radiation algebra; recoverability and entropy	A causal diagram, factorization or algebra choice, evaporation model, semiclassical validity	Standard semiclassical reasoning creates a sharp conflict among familiar assumptions	Which assumption fails is unresolved; a toy purification alone is not a gravitational derivation
Unification and ultraviolet completion	Candidate microscopic variables; low-energy spectrum and observables	Programme-dependent kinematics, dynamics, regulator, and continuum limit	Several frameworks supply partial constructions and nontrivial limits	Mathematical coherence is not empirical confirmation, and cross-programme comparisons remain assumption-sensitive

The conservative position is therefore that these are compelling reasons to study the limits of classical spacetime and perturbative Einstein gravity, but none is direct empirical evidence for one microscopic programme. The more permissive position treats the convergence of independent pressures—ultraviolet sensitivity, horizon thermodynamics, and quantum information—as abductive evidence that geometry itself must enter the quantum description. The latter can guide research, but it must not be reported with the status of a theorem or measurement.

A solvable focusing check

Consider a spatially flat, expanding Friedmann universe filled with pressureless matter. With scale factor $a(t)\propto t^{2/3}$ for $t\gt0$, a comoving timelike geodesic has proper time $t$. Running it to the past reaches $t=0$ after finite proper time; the Ricci scalar is

$$R=6\left(\frac{\ddot a}{a}+\frac{\dot a^2}{a^2}\right)=\frac{4}{3t^2}.$$

This explicit solution exhibits both geodesic incompleteness and curvature blow-up. Now change a named assumption: replace the dust solution by the flat slicing of de Sitter space, $a(t)=e^{Ht}$. The curvature becomes $R=12H^2$, finite at every $t$, while the flat patch is not the entire maximally extended spacetime. The comparison exposes two robust lessons. First, incompleteness and curvature divergence are logically distinct. Second, conclusions about a coordinate patch cannot replace a statement about maximal extension. A proposed quantum resolution must say which observable, state, and extension criterion it changes.

The same discipline applies to perturbative calculations. A regulator-dependent pole is not itself an observable, but the need for a counterterm allowed by the symmetries is. Changing gauge may rearrange off-shell terms; it cannot remove an on-shell divergence that survives field redefinitions. Conversely, adding a higher-curvature coupling changes the theory being tested and therefore changes the inference from the calculation.

What the pressures do and do not imply

Three boundaries keep the reasoning honest.

First, a breakdown of an approximation need not be a breakdown of quantum theory. Semiclassical backreaction may fail because stress-tensor fluctuations are large even when all curvatures are sub-Planckian. Second, a failure of perturbative renormalizability need not be a loss of low-energy predictivity. The EFT expansion simply asks for more measured coefficients at progressively higher order. Third, recovering a famous result is necessary but rarely sufficient. A state count that gives $S=A/(4G)$ has passed one hard correspondence test; it has not thereby derived evaporation, locality, or an observable signal.

These distinctions also prevent false symmetry between programmes. A kinematical Hilbert space, a finite truncation, a controlled continuum limit, and a discriminating prediction occupy different evidential levels. Parallel rows in a comparison table do not make their support equal.

Exercises

Exercise. In a spatially flat Friedmann spacetime with a perfect fluid $p=w\rho$ and constant $w\ne-1$, use $a(t)\propto t^{2/[3(1+w)]}$ to decide when the comoving past proper time to $a=0$ is finite. Explain why the result is a consistency pressure rather than evidence for a particular microscopic theory.

Solution

Let $q=2/[3(1+w)]$. A comoving observer has $d\tau=dt$. For $w\gt-1$, $q\gt0$, so the displayed branch expands from $a=0$ at $t=0$; its past duration from $t_0\gt0$ is the finite proper time $t_0$. For $w\lt-1$, $q\lt0$, and the displayed $t\gt0$ branch contracts from infinite scale factor: it does not reach $a=0$ to the past. Its expanding time-reversed version, $a\propto(t_*-t)^q$, instead reaches infinite scale factor at a finite future time $t_*$. The excluded case $w=-1$ has an exponential solution. Thus the branch and time orientation matter. The calculation uses the homogeneous perfect-fluid Einstein model and identifies no particular microscopic replacement; many theories could alter its assumptions near an endpoint.

What could change these conclusions

The conclusions could change after a theorem weakens or changes the focusing hypotheses, an independently reproduced calculation alters the ultraviolet divergence, a controlled microscopic construction derives additional observables beyond a correspondence check, or evidence obtained after 2026-08-20 changes the empirical standing of any pressure.

References

• Bekenstein, Jacob D. “Black Holes and Entropy.” Physical Review D 7 (1973): 2333–2346. DOI.
• Goroff, Marc H., and Augusto Sagnotti. “Quantum Gravity at Two Loops.” Physics Letters B 160 (1985): 81–86. DOI.
• Hawking, Stephen W., and Roger Penrose. “The Singularities of Gravitational Collapse and Cosmology.” Proceedings of the Royal Society A 314 (1970): 529–548. DOI.