Generic $W^{1,2}$-inextendibility of the dynamical Kerr Cauchy horizon: Christodoulou's formulation of strong cosmic censorship for the Einstein vacuum equations without symmetry

The theorem of Dafermos and Luk on the $C^0$-stability of the Kerr Cauchy horizon shows that generic vacuum black hole interiors admit a continuous metric extension across the Cauchy horizon, so Penrose's strong cosmic censorship conjecture is false in its classical $C^0$-formulation. The sharp surviving question, due to Christodoulou, asks whether these horizons are nevertheless singular at the lowest regularity at which the Einstein equations still make sense as weak equations: are the Christoffel symbols of any continuous extension generically non-square-integrable at the Cauchy horizon, so that the maximal development is inextendible even as a weak ($W^{1,2}$) solution? This problem is open without symmetry assumptions: the best current result (Gurriaran, 2026, conditional on a nonlinear Price's law ansatz) proves only the weaker Lipschitz-inextendibility near timelike infinity, and the full $W^{1,2}$-level, generic, global statement remains unproved. The problem asks for a proof or refutation of Christodoulou's reformulation for generic vacuum data.

Background

The Kerr family $g_{a,M}$ is the two-parameter family of stationary, axisymmetric, rotating solutions of the Einstein vacuum equations $\mathrm{Ric}(g)=0$ (zero cosmological constant), parametrized by mass $M$ and angular momentum parameter $a$ with $0\le|a|\le M$; the subextremal range is $0<|a|<M$. Inside a subextremal Kerr black hole, the interior region bounded by the event horizon $\mathcal{H}^+$ (the null boundary of the black-hole region, i.e. of the region from which no signal can reach infinity) also contains a Cauchy horizon $\mathcal{CH}^+$: a null boundary across which the Kerr geometry extends smoothly, but which is the boundary of the domain of dependence of any initial hypersurface (the maximal region whose evolution is uniquely fixed by the data prescribed there) — beyond it, the Einstein equations lose their predictive (deterministic) character, since additional data would have to be prescribed. Penrose's strong cosmic censorship conjecture asserts that such Cauchy horizons should not occur for generic initial data.

Physically, the expected mechanism destroying the Cauchy horizon is the blue-shift instability: outgoing radiation falling into the black hole is infinitely blueshift-amped as it approaches $\mathcal{CH}^+$, an effect that in charged, spherically symmetric models manifests as mass inflation — blow-up of the Hawking mass (a standard quasi-local mass functional in general relativity) at the horizon. Rigorously, Dafermos (2005) proved for the spherically symmetric Einstein–Maxwell–scalar field system that the development is extendible as a continuous ($C^0$) metric across the Cauchy horizon but is $C^2$-inextendible generically; Luk and Oh (2017–2019) completed the $C^2$-formulation of strong cosmic censorship for two-ended asymptotically flat data in this model, including the genericity (openness and density) of the instability. In contrast, for the exactly stationary spacetimes themselves, Sbierski proved that Schwarzschild is $C^0$-inextendible, so the $C^0$-formulation is nontrivial.

The landscape changed with Dafermos and Luk's study of the interior of dynamical (genuinely perturbed, non-stationary) vacuum black holes without any symmetry: they proved that for data posed on a hypersurface just inside a black hole interior close to subextremal Kerr, the maximal development extends across a non-empty piece of Cauchy horizon as a Lorentzian manifold with continuous metric. Hence, conditional on the expected nonlinear stability of the Kerr exterior, the $C^0$-formulation of strong cosmic censorship is false. The $C^0$ horizons so constructed are expected to be weak null singularities: continuous metric, blowing curvature.

This leaves the question of the correct regularity threshold. Below $C^2$ there is a natural floor: if a Lorentzian metric is continuous with Christoffel symbols locally square integrable — equivalently, the metric belongs to $W^{1,2}_{\mathrm{loc}}$, the Sobolev space of functions with square-integrable first derivatives — the Einstein equations can still be interpreted as equations for a weak (distributional) solution. Christodoulou's reformulation of strong cosmic censorship therefore predicts: for generic vacuum initial data, the maximal Cauchy development is inextendible as a Lorentzian manifold with continuous metric and locally square-integrable Christoffel symbols. Dafermos and Luk state this precisely as Conjecture 3 and identify proving this higher-regularity instability for the vacuum equations without symmetry as the central open problem left by their theorem.

The linear evidence is strong: generic scalar waves on Reissner–Nordström (Luk–Oh) and on subextremal Kerr (Luk–Sbierski, Ma–Zhang) fail to be in $W^{1,2}_{\mathrm{loc}}$ at the Cauchy horizon, and Gurriaran (2024–2025) proved precise oscillatory blow-up asymptotics for the spin $\pm2$ Teukolsky fields (the linearized curvature components of gravitational perturbations), giving a generic linearized curvature singularity at the Kerr Cauchy horizon. Sbierski (2024) proved a general criterion deriving Lipschitz ($C^{0,1}_{\mathrm{loc}}$, i.e. metric with bounded first derivatives) inextendibility of weak null singularities from curvature blow-up, without symmetry assumptions in the criterion.

Very recently, Gurriaran (2026) achieved the first nonlinear instability result without symmetry: for smooth data slightly inside a dynamical vacuum black hole settling down to subextremal Kerr, and assuming a precise nonlinear Price's law-type decay ansatz on the data (Price's law is the expected generic inverse-polynomial decay of perturbations along the event horizon; the ansatz is expected but not yet proven to hold generically), the development admits the Dafermos–Luk Cauchy horizon emanating from timelike infinity $i^+$ (the ideal endpoint of the spacetime at infinite future time), the Riemann curvature blows up in the constructed coordinates, and the spacetime is Lipschitz-inextendible there. Since a Lipschitz metric has bounded derivatives and hence lies in $W^{1,2}_{\mathrm{loc}}$, Lipschitz-inextendibility is strictly weaker than the required $W^{1,2}$-inextendibility; Gurriaran's paper itself leaves open the genericity of the ansatz, the region of the horizon away from $i^+$, and the coordinate-invariant (Kretschmann-scalar, i.e. the squared norm of the Riemann tensor) blow-up. (Two adjacent questions raised in the source paper are distinct from this core and remain open separately: whether one-ended generic vacuum collapse must also produce a spacelike portion of the singular boundary — proven only for charged spherical collapse by Van de Moortel — and the fate of censorship for positive cosmological constant $\Lambda>0$, where near-extremal charged de Sitter black holes are expected, and in spherical charged models provably are, extendible even in $H^1=W^{1,2}$.)

Problem Statement

Prove or refute Christodoulou's reformulation of strong cosmic censorship for the Einstein vacuum equations $\mathrm{Ric}(g)=0$ (with $\Lambda=0$) without symmetry assumptions, in the black-hole-interior setting made precise by the Dafermos–Luk $C^0$-stability theorem:

For smooth characteristic or Cauchy data for the vacuum equations that are close to subextremal Kerr $g_{a,M}$, $0<|a|<M$, in the black hole interior (in the Dafermos–Luk class, or more generally for data whose exterior region settles down to subextremal Kerr, as expected for generic asymptotically flat collapse), the maximal globally hyperbolic development (the largest spacetime uniquely determined by the data) admits a continuous extension across a non-empty Cauchy horizon (Dafermos–Luk). Christodoulou's conjecture in its full generality quantifies over all generic asymptotically flat vacuum initial data; the present problem targets its near-Kerr core above — the class in which the Cauchy horizon is proven (interior data) or expected (generic collapse) to arise — which is precisely the formulation Dafermos and Luk identify as the open problem left by their theorem. Show that, for a generic set of such data — genericity to be established in an appropriate topology (e.g. open and dense), not assumed via a Price's law ansatz — every continuous extension of the metric across the Cauchy horizon has Christoffel symbols that fail to lie in $L^2_{\mathrm{loc}}$ at the horizon. Equivalently: the development is inextendible as a Lorentzian manifold with continuous metric and locally square-integrable Christoffel symbols, i.e. inextendible even as a weak solution of the Einstein vacuum equations. A full solution should address the entire Cauchy horizon (including regions away from timelike infinity and the late-time blueshift region), and the expected mechanism is a coordinate-invariant curvature blow-up (Kretschmann scalar) of oscillatory weak-null-singularity type.

A valid refutation would exhibit a generic class of such vacuum data whose maximal developments admit $W^{1,2}_{\mathrm{loc}}$ extensions across the Cauchy horizon solving the vacuum equations weakly.

The statement must be resolved at exactly the $W^{1,2}$ threshold: results at weaker regularity (e.g. Lipschitz-inextendibility) or for restricted data classes (data assumed to satisfy a nonlinear Price's law ansatz) are partial progress, not a resolution. Two adjacent open questions raised in the same source passage — whether one-ended generic vacuum collapse must also produce a spacelike portion of the singular boundary, and the fate of censorship for $\Lambda>0$ — are not part of this problem's completion boundary.

References

1. Demetrios Christodoulou, The Formation of Black Holes in General Relativity, EMS Monographs in Mathematics, European Mathematical Society, Zürich, 2009. DOI: 10.4171/068, ISBN 978-3-03719-068-5; also arXiv:0805.3880. https://doi.org/10.4171/068
2. Mihalis Dafermos, The interior of charged black holes and the problem of uniqueness in general relativity, Comm. Pure Appl. Math. 58 (2005), 445–504. arXiv:gr-qc/0307013, DOI: 10.48550/arXiv.gr-qc/0307013. https://arxiv.org/abs/gr-qc/0307013
3. Mihalis Dafermos and Jonathan Luk, The interior of dynamical vacuum black holes I: The $C^0$-stability of the Kerr Cauchy horizon, Ann. of Math. (2) 202 (2025), no. 2, 309–630. DOI: 10.4007/annals.2025.202.2.1. Preprint arXiv:1710.01722 (2017). https://arxiv.org/abs/1710.01722
4. Jonathan Luk and Sung-Jin Oh, Proof of linear instability of the Reissner–Nordström Cauchy horizon under scalar perturbations, Duke Math. J. 166 (2017), no. 3, 437–493. DOI: 10.1215/00127094-3715189. arXiv:1501.04598. https://arxiv.org/abs/1501.04598
5. Jonathan Luk and Sung-Jin Oh, Strong cosmic censorship in spherical symmetry for two-ended asymptotically flat initial data I. The interior of the black hole region, Ann. of Math. (2) 190 (2019), no. 1, 1–111. DOI: 10.4007/annals.2019.190.1.1. arXiv:1702.05715. https://arxiv.org/abs/1702.05715
6. Jonathan Luk and Sung-Jin Oh, Strong cosmic censorship in spherical symmetry for two-ended asymptotically flat initial data II. The exterior of the black hole region, Annals of PDE 5 (2019), no. 1, article 6. DOI: 10.1007/s40818-019-0062-7. arXiv:1702.05716. https://arxiv.org/abs/1702.05716
7. Jonathan Luk and Jan Sbierski, Instability results for the wave equation in the interior of Kerr black holes, J. Funct. Anal. 271 (2016), no. 7, 1948–1995. DOI: 10.1016/j.jfa.2016.06.013. arXiv:1512.08259. https://arxiv.org/abs/1512.08259
8. Siyuan Ma and Lin Zhang, Precise late-time asymptotics of scalar field in the interior of a subextreme Kerr black hole and its application in Strong Cosmic Censorship conjecture, Trans. Amer. Math. Soc. (2023). DOI: 10.1090/tran/8957. arXiv:2208.08702. https://arxiv.org/abs/2208.08702
9. Sebastian Gurriaran, Precise asymptotics of the spin $+2$ Teukolsky field in the Kerr black hole interior, Commun. Math. Phys. 406 (2025), article 152. DOI: 10.1007/s00220-025-05332-3. arXiv:2409.02670. https://arxiv.org/abs/2409.02670
10. Sebastian Gurriaran, Generic linearized curvature singularity at the perturbed Kerr Cauchy horizon, arXiv:2503.24114 (2025), DOI: 10.48550/arXiv.2503.24114. https://arxiv.org/abs/2503.24114
11. Sebastian Gurriaran, Non-linear instability of the Kerr Cauchy horizon near $i^+$, arXiv:2603.17911 (2026), DOI: 10.48550/arXiv.2603.17911. https://arxiv.org/abs/2603.17911
12. Jan Sbierski, Lipschitz inextendibility of weak null singularities from curvature blow-up, arXiv:2409.18838 (2024, revised 2025), DOI: 10.48550/arXiv.2409.18838. https://arxiv.org/abs/2409.18838
13. Jan Sbierski, The $C^0$-inextendibility of the Schwarzschild spacetime and the spacelike diameter in Lorentzian geometry, J. Differential Geom. 108 (2018), no. 2. DOI: 10.4310/jdg/1518490820. arXiv:1507.00601. https://arxiv.org/abs/1507.00601
14. Maxime Van de Moortel, Mass inflation and the $C^2$-inextendibility of spherically symmetric charged scalar field dynamical black holes, Commun. Math. Phys. 382 (2021), no. 2, 1263–1341. DOI: 10.1007/s00220-020-03923-w. arXiv:2001.11156. https://arxiv.org/abs/2001.11156
15. Maxime Van de Moortel, The breakdown of weak null singularities inside black holes, Duke Math. J. 172 (2023), no. 15, 2957–3012. arXiv:1912.10890, DOI: 10.48550/arXiv.1912.10890. https://arxiv.org/abs/1912.10890
16. Maxime Van de Moortel, The coexistence of null and spacelike singularities inside spherically symmetric black holes, arXiv:2504.12370 (2025), DOI: 10.48550/arXiv.2504.12370. https://arxiv.org/abs/2504.12370
17. Vitor Cardoso, Joao L. Costa, Kyriakos Destounis, Peter Hintz, Aron Jansen, Quasinormal modes and Strong Cosmic Censorship, Phys. Rev. Lett. 120 (2018), 031103. DOI: 10.1103/PhysRevLett.120.031103. arXiv:1711.10502. https://arxiv.org/abs/1711.10502
18. Oscar J.C. Dias, Felicity C. Eperon, Harvey S. Reall, Jorge E. Santos, Strong cosmic censorship in de Sitter space, Phys. Rev. D 97 (2018), 104060. DOI: 10.1103/PhysRevD.97.104060. arXiv:1801.09694. https://arxiv.org/abs/1801.09694
19. Mihalis Dafermos and Yakov Shlapentokh-Rothman, Rough initial data and the strength of the blue-shift instability on cosmological black holes with $\Lambda>0$, Class. Quantum Grav. 35 (2018), no. 19, 195010. DOI: 10.1088/1361-6382/aadbcf. arXiv:1805.08764. https://arxiv.org/abs/1805.08764
20. Raimon Luna, Miguel Zilhão, Vitor Cardoso, João L. Costa, José Natário, Strong cosmic censorship: The nonlinear story, Phys. Rev. D 99 (2019), 064014. DOI: 10.1103/PhysRevD.99.064014. arXiv:1810.00886. https://arxiv.org/abs/1810.00886
21. Flavio Rossetti, Strong cosmic censorship for the spherically symmetric Einstein–Maxwell-charged-Klein-Gordon system with positive $\Lambda$: stability of the Cauchy horizon and $H^1$ extensions, Ann. Henri Poincaré 26 (2025), 675–753. DOI: 10.1007/s00023-024-01454-x. arXiv:2309.14420. https://arxiv.org/abs/2309.14420
22. Flavio Rossetti, Mass inflation from rough initial data for the spherically symmetric Einstein–Maxwell-scalar field system with $\Lambda$, arXiv:2506.08075 (2025), DOI: 10.48550/arXiv.2506.08075. https://arxiv.org/abs/2506.08075