Limits of Four-Dimensional Uniqueness

Black-hole uniqueness is a theorem about a sharply delimited class, not a universal ban on hair or topology. Black rings appear when dimension changes; scalar or Yang–Mills hair appears when the matter model changes; multiple extremal horizons appear when connectedness and nondegeneracy change together. The right question is therefore not “Does no hair hold?” but “Which field equations, asymptotics, topology, symmetry, regularity, and boundary data are being held fixed?” This page compares those formulations using literature checked through 2026-08-12.

Required background. Analysis for Mathematical Relativity supplies the distinction between analytic continuation and smooth unique continuation, while Spacetime Symmetries and Killing Fields supplies rigidity, symmetry inheritance, and horizon generators.

Helpful background. Einstein Equations fixes the theory being classified, Causality and Global Structure fixes the exterior and asymptotics, Partial Differential Equations and Green Functions clarifies the smooth continuation gap, and Black-Hole Geometry and Mechanics supplies horizon topology, degeneracy, and charges.

Which hypotheses distinguish uniqueness theorems?

Each entry below is compared on the same axes: spacetime dimension; field equations and matter; asymptotics; horizon connectedness, topology, and degeneracy; assumed versus derived symmetry; analytic or smooth regularity; gauge and boundary normalization; and conclusion. These are stationary classification results, so an initial-data class and an evolution norm are not part of their claims. Numerical families are not promoted to theorem status, and a theorem about intrinsic horizon data is not promoted to uniqueness of a whole exterior.

Evidence cutoff: 2026-08-12. Versions, corrections, and publication status were checked through that date. Later theorems, revisions, counterexamples, or independent reviews lie outside this evidence window. The central unresolved formulation is the fully smooth, global, four-dimensional, asymptotically flat stationary vacuum classification without analyticity, a near-Kerr condition, or axisymmetry assumed in advance.

Uniqueness results through 2026-08-12

Four-dimensional analytic vacuum. Dimension four; vacuum Einstein; asymptotically flat and globally regular exterior; connected compact horizon; stationarity assumed and axisymmetry derived; analytic metric. The nondegenerate connected theorem is peer reviewed. A 2023 preprint claims the completion for all connected horizons, including the mean-degenerate branch: its Theorem 1.1 concludes Kerr under stationary, analytic, asymptotically flat, regular hypotheses. Its own §3 still says that global extension of the second Killing field in the smooth class remains open beyond special regimes; see Chruściel 2023, Theorem 1.1 and §3. Because that article remained an unpublished preprint at the cutoff, its extension is labeled a preprint theorem claim, not a peer-reviewed upgrade.

Four-dimensional smooth vacuum near Kerr. The dimension, theory, asymptotics, connected nondegenerate horizon, and regular exterior remain fixed, but analyticity is replaced by smoothness plus quantitative smallness of the Mars–Simon tensor. This is a peer-reviewed conditional theorem: the exterior is Kerr. Carleman estimates extend the local Hawking field through conditionally pseudoconvex level sets. It does not cover an arbitrary smooth stationary black hole; see Alexakis, Ionescu, and Klainerman 2010, main theorem and §§4–6, pp. 4–35.

Connected degenerate electrovacuum. Dimension four; Einstein–Maxwell; asymptotically flat, analytic, stationary, and globally regular; connected, rotating, degenerate horizon. This is a peer-reviewed theorem identifying the exterior with extremal Kerr–Newman. Degeneracy is therefore not by itself a counterexample. The theorem retains connectedness, analyticity, and electrovacuum, as stated in Chruściel and Nguyen 2010, theorem and §§2–4, pp. 585–609.

Extremal intrinsic geometry. In four-dimensional vacuum, every compact extremal horizon cross-section carries a Killing field; when the cross-section is a two-sphere, the nonstatic intrinsic data are those of extremal Kerr. This conclusion does not assume axial symmetry at the start, but the two-dimensional spherical-cross-section hypothesis is essential to the Kerr classification; see Dunajski and Lucietti 2026, Theorem 1.1 and Corollary 1.2, pp. 179–182. For four-dimensional Einstein–Maxwell, compact extremal cross-sections likewise admit a Killing field or belong to the static branch, and the nonstatic spherical data are Kerr–Newman; see Colling, Katona, and Lucietti 2025, Theorems 1.1–1.2. These are peer-reviewed horizon theorems, not full exterior classifications. The distinction, and the remaining question whether intrinsic extremal data fix a vacuum exterior, are reviewed in Lucietti 2026, Theorem 1, Corollary 2, and the discussion after Theorem 3.

Disconnected degenerate electrovacuum. Dimension four; Einstein–Maxwell; asymptotically flat; static; several degenerate spherical components. Majumdar–Papapetrou solutions are exact peer-reviewed solution families and show that the connected Kerr–Newman conclusion cannot simply be copied to multiple extremal components. Their balance uses charge and degeneracy. They do not violate connected-horizon uniqueness. A general stationary classification with several degenerate components is more delicate and is not settled merely by listing the known family; the distinct cases are reviewed in Chruściel, Costa, and Heusler 2012, §§3.4.1–3.4.3, pp. 17–19.

Four-dimensional scalar hair. Dimension four; Einstein gravity coupled to a complex massive scalar; asymptotically flat; connected rotating horizon; the metric is stationary and axisymmetric, while the scalar has harmonic time–azimuth dependence and is invariant only under the horizon’s helical generator. The published families are numerical constructions, with regularity and field-equation residuals checked in their numerical ansatz. They overlap Kerr in the $(M,J)$ plane but carry an additional Noether charge; see Herdeiro and Radu 2014, construction and Figs. 1–3, pp. 221101-1–5. This changes the matter model and symmetry-inheritance condition, so it is not a counterexample to electrovacuum uniqueness.

Five-dimensional vacuum rings. Dimension five; vacuum Einstein; asymptotically flat; regular connected horizon of topology $S^1\times S^2$; stationary and axisymmetric exact metric. This is a peer-reviewed exact counterexample to a dimension-independent charge-only classification: for a range of mass and angular momentum there are two rings and a spherical black hole with the same conserved pair. The solution and parameter overlap are stated in Emparan and Reall 2002, solution and charge comparison, pp. 101101-1–4. The relaxed assumption is dimension, which also changes allowed topology and the orbit-space data.

Non-flat asymptotics. With positive or negative cosmological constant, the conformal boundary, conserved charges, horizon topology, and elliptic boundary conditions all change. There are important conditional theorems and exact families, but no direct analogue of the full asymptotically flat classification. As of the cutoff, even spherical Schwarzschild–anti-de Sitter uniqueness is identified as an open question in Lucietti 2026, discussion following Corollary 2. Changing asymptotics is therefore a new boundary-value problem, not a small correction to Kerr uniqueness.

What remains open and what is already false

The scalar example makes symmetry inheritance concrete. Write $\Psi=\psi(r,\theta)e^{-i\omega t+im\varphi}$, with integer $m$ and real $\omega$. Its quadratic stress tensor can be independent of $t,\varphi$ even though $\Psi$ itself is not. For the horizon generator $\chi=\partial_t+\Omega_H\partial_\varphi$,

$$\mathcal L_\chi\Psi=-i(\omega-m\Omega_H)\Psi.$$

Thus the synchronization condition $\omega=m\Omega_H$ makes the scalar constant along the horizon generators and eliminates its flux through that stationary horizon. It does not restore invariance under $\partial_t$ and $\partial_\varphi$ separately. This is the precise symmetry distinction used in the numerical construction Herdeiro and Radu 2014, equations (2)–(4) and horizon boundary conditions, pp. 221101-1–3.

It is already false that mass and angular momentum classify regular vacuum black holes in every dimension: the black ring is an exact counterexample. It is also false that asymptotic gravitational and electromagnetic charges classify black holes in arbitrary matter theories: scalar and non-Abelian fields add degrees of freedom and can change the reduced target geometry. Those failures are sharply located outside the four-dimensional vacuum/electrovacuum theorem.

What remains open is narrower and more mathematical. Can local smooth horizon rigidity be globalized through an arbitrary ergoregion without analytic or near-Kerr input? Which disconnected stationary configurations are completely regular once all axis and balance conditions are imposed? How far can intrinsic extremal-horizon rigidity determine transverse derivatives and the global exterior? What is the correct uniqueness data under AdS or de Sitter boundary conditions? A universal verdict would erase precisely the formulation choices on which these questions depend.

Diagnose the changed hypothesis

Classify each case by the first standard assumption it changes: (i) a regular vacuum black ring with the same mass and angular momentum as a spherical black hole; (ii) a rotating black hole with a complex massive scalar and an extra Noether charge; (iii) a static configuration with several extremal charged horizons; (iv) a theorem that fixes an extremal horizon’s intrinsic geometry but not its exterior.

Solution

Case (i) changes dimension from four to five, allowing ring topology and charge nonuniqueness. Case (ii) changes the matter model from vacuum/electrovacuum and weakens symmetry inheritance for the scalar, so the Kerr–Newman harmonic-map proof no longer applies. Case (iii) changes connectedness and nondegeneracy together; the Majumdar–Papapetrou balance is electromagnetic and extremal. Case (iv) changes the conclusion rather than a hypothesis: near-horizon or intrinsic rigidity is local data and does not yet prove global exterior uniqueness. None of the four refutes the analytic, connected, four-dimensional asymptotically flat theorem in its stated field theory.

References

• Alexakis, Spyros, Alexandru D. Ionescu, and Sergiu Klainerman. “Uniqueness of Smooth Stationary Black Holes in Vacuum: Small Perturbations of the Kerr Spaces.” Communications in Mathematical Physics 299 (2010): 89–127. DOI. Open PDF.
• Chruściel, Piotr T. “Remarks on Stationary Vacuum Black Holes.” arXiv:2305.07329 (2023), unpublished preprint at the evidence cutoff. arXiv.
• Chruściel, Piotr T., and Luc Nguyen. “A Uniqueness Theorem for Degenerate Kerr–Newman Black Holes.” Annales Henri Poincaré 11 (2010): 585–609. DOI. Open PDF.
• Chruściel, Piotr T., João Lopes Costa, and Markus Heusler. “Stationary Black Holes: Uniqueness and Beyond.” Living Reviews in Relativity 15 (2012): 7. DOI. Open PDF.
• Colling, David, Dániel Katona, and James Lucietti. “Rigidity of Extremal Kerr–Newman Horizon.” Letters in Mathematical Physics 115 (2025): 19. DOI.
• Dunajski, Maciej, and James Lucietti. “Intrinsic Rigidity of Extremal Horizons.” Journal of Differential Geometry 132 (2026): 179–201. DOI.
• Emparan, Roberto, and Harvey S. Reall. “A Rotating Black Ring in Five Dimensions.” Physical Review Letters 88 (2002): 101101. DOI. Open PDF.
• Herdeiro, Carlos A. R., and Eugen Radu. “Kerr Black Holes with Scalar Hair.” Physical Review Letters 112 (2014): 221101. DOI. Open PDF.
• Lucietti, James. “Intrinsic Rigidity of Extremal Horizons and Black Hole Uniqueness.” General Relativity and Gravitation 58 (2026): 79. DOI.