Dynamical Tests of Cosmic Censorship

Cosmic censorship is tested by evolution, not by a superextremal arithmetic result or a singular exact metric alone. Weak censorship asks whether generic admissible data keep singular behavior from distant observers and preserve complete future null infinity. Strong censorship asks whether the maximal globally hyperbolic development is generically inextendible at a specified regularity. This chapter turns those statements into hypotheses, observables, and failure tests.

Helpful background. The chapter draws on Black-Hole Geometry and Mechanics, Black-Hole Perturbation Theory, Formulations of General Relativity, and Causality and Global Structure. Numerical claims should use the validation framework in Computational Relativity.

Three distinctions that control the subject

First, kinematics is not dynamics. A test particle can appear to move final charges across an extremality bound only because its self-force, radiation, and second-order energy have been omitted. A divergent local collision energy does not imply divergent energy reaching infinity. The same stress-energy must be used consistently in capture, flux, and final-state calculations.

Second, existence is not genericity. Symmetric dust, fluid, and scalar models contain naked-singularity candidates, but their failure set must contradict the conjecture’s specified meaning of “generic.” An open set of counterexamples rules out an open-dense validity claim; one fine-tuned example does not. Christodoulou proves black-hole-forming instability under a rough bounded-variation perturbation class Christodoulou 1999, Theorems 1.1–1.3, whereas Zheng’s 2026 preprint proves localized same-regularity Hölder stability for a restricted finite-regularity, continuously self-similar spherical scalar family Zheng 2026, Theorems 1.1 and 1.3, pp. 2–3. The conclusions concern different function-space topologies; neither is an unrestricted nonspherical genericity theorem.

Third, singularity is not one regularity threshold. A black-hole interior may admit a continuous extension across a Cauchy horizon even though curvature diverges and no $C^2$ extension exists. Strong censorship must therefore name the extension class, field equations, asymptotics, and genericity topology.

Route through the chapter

Overspinning and Overcharging Thought Experiments begins with capture and extremality inequalities and shows where second-order canonical energy enters. High-Energy Collisions and Horizon Destruction separates local center-of-mass energy, escaping energy, and self-gravitating backreaction. Naked-Singularity Candidates in Collapse compares dust, fluid, scalar, and critical-collapse examples under regularity and stability tests.

Weak Cosmic Censorship as a Dynamical Claim states a representative asymptotically flat formulation in terms of maximal development and complete null infinity. Strong Cosmic Censorship and Interior Regularity organizes $C^0$, $C^2$, and weak-solution thresholds through the competition between exterior decay and inner-horizon blueshift. Numerical Evidence and Failure Modes then sets the convergence and causal-tracing standard for any proposed violation.

From proposal to censorship test

A disciplined analysis follows a fixed chain. Specify the initial-data and matter class. State which maximal development and which asymptotics are used. Define genericity and the relevant topology. Identify the invariant failure criterion: incomplete null infinity, globally visible curvature blow-up, or inextendibility at a named regularity. Only then choose perturbative, numerical, or exact-solution tools.

For weak censorship, locating one outgoing ray from a singularity establishes at most local nakedness; the ray must reach the relevant asymptotic observers. For strong censorship, showing curvature blow-up can obstruct $C^2$ extension while leaving $C^0$ extension possible. In spherical Einstein–Maxwell–scalar theory, Luk and Oh’s Part I derives $C^2$ inextendibility from an event-horizon lower bound Luk and Oh 2019a, Theorem 5.7, p. 29, PDF, while Part II proves that lower bound for a generic data set Luk and Oh 2019b, Theorems 4.1–4.3, pp. 17–18, PDF.

The extremality experiments add another lesson: when the proposed effect is second order, every omitted second-order term is potentially decisive. The covariant canonical-energy inequality closes the near-extremal test-body window under its perturbative settlement and energy hypotheses Sorce and Wald 2017, §IV, Eqs. (115)–(122). This is a powerful obstruction, not a proof for arbitrary large data.

Evidence and limits

Exact models establish mechanisms and possible causal structures. Perturbation theory can control infinite time near a background. Numerical relativity can enter the nonlinear regime but only for finite resolution and duration. No one method supplies all three of existence, genericity, and global causal reach automatically.

A persuasive numerical violation would need convergent constraints and curvature, invariant horizon and ray diagnostics, controlled boundaries or null infinity, and robustness under gauge and formulation changes. Code termination is evidence about the algorithm until shown otherwise. Conversely, continued numerical horizon formation across a survey is evidence within that survey, not a theorem for an infinite-dimensional data space.

Review the chapter

A near-extremal simulation reports the dimensionless final spin $cJ_f/(G M_f^2)=1.0002$, no apparent horizon on its final slice, and a rapidly growing curvature scalar. List four questions that must be answered before claiming a weak-censorship violation.

Solution

One must show that the extremality excess exceeds converged errors in mass, spin, radiation, and constraints; verify charges and capture with a self-consistent second-order evolution; demonstrate that the horizon finder did not miss a distorted or later-forming surface; and establish global visibility of the singular behavior or failure of future-null-infinity completeness in the maximal development. The admissible data class and robustness under perturbations must also match the chosen genericity claim.

References

• Christodoulou, Demetrios. “The Instability of Naked Singularities in the Gravitational Collapse of a Scalar Field.” Annals of Mathematics 149 (1999): 183–217. doi:10.2307/121023.
• Luk, Jonathan, and Sung-Jin Oh. “Strong Cosmic Censorship in Spherical Symmetry for Two-Ended Asymptotically Flat Initial Data I.” Annals of Mathematics 190 (2019): 1–111. doi:10.4007/annals.2019.190.1.1.
• Luk, Jonathan, and Sung-Jin Oh. “Strong Cosmic Censorship in Spherical Symmetry for Two-Ended Asymptotically Flat Initial Data II.” Annals of PDE 5 (2019): 6. doi:10.1007/s40818-019-0062-7.
• Sorce, Jonathan, and Robert M. Wald. “Gedanken Experiments to Destroy a Black Hole II: Kerr–Newman Black Holes Cannot Be Overcharged or Overspun.” Physical Review D 96 (2017): 104014. doi:10.1103/PhysRevD.96.104014.
• Zheng, Weihao. “Nonlinear Stability of Continuously Self-Similar Naked Singularities for the Einstein–Scalar Field Equations I: Main Results.” arXiv:2605.16235v1 (15 May 2026). arXiv. Open PDF.

Dated Research assessment. For open questions and recent results, follow the link below. Each assessment states the date through which it considers results and the limits of its conclusions.

• Overspinning, Overcharging, and Censorship Tests