Weak Cosmic Censorship

Censorship is a live experimental and computational frontier. Numerical relativity is its principal laboratory: every simulated binary black-hole merger behind the LIGO–Virgo–KAGRA detections settles to a clean Kerr remnant with its singularity hidden — an enormous, if non-rigorous, body of four-dimensional evidence. Choptuik's critical-collapse experiments probe the razor's edge, where at the exact threshold of black-hole formation a fine-tuned, codimension-one self-similar solution produces an arbitrarily small naked singularity, consistent with WCC precisely because it demands infinite fine-tuning. The Event Horizon Telescope's images of M87\ and Sgr A\ assume implicitly that the exterior is determined by Einstein's equations and free of signals erupting from a singularity. Meanwhile the conjecture's one clear failure — Lehner and Pretorius's 2010 result that Gregory–Laflamme-unstable black strings pinch into a naked singularity in five dimensions — has become a springboard for understanding why four dimensions may be special. And gedanken over-spinning experiments, revisited with self-force and quantum corrections, remain a sharp analytic test of whether a near-extremal hole can ever be pushed over the edge.

Gravitational collapse is a race between two surfaces. As a star implodes, curvature climbs toward a singularity at the center; simultaneously the region of no escape — the event horizon — inflates outward from where the trapped region first appears. Weak Cosmic Censorship is the wager that, for generic matter, the horizon always wins: by the time the curvature diverges, the light that could carry news of it is already trapped, and a telescope at infinity sees only a featureless dark disk. The conformal diagram makes the whole causal story fit on one page, because null rays run at a fixed 45°.

The diagram encodes the definition literally: the black-hole region $\mathcal{B}=M\setminus J^-(\mathscr{I}^+)$ is everything that cannot signal infinity, and the event horizon $\mathcal{H}$ is its boundary — a null surface running at exactly $45°$. A photon emitted just inside $\mathcal{H}$ is doomed: even moving "outward" it is parallel to the horizon and terminates on the singularity; a photon just outside climbs to $\mathscr{I}^+$. Censorship is the claim that the singularity always sits above the horizon, never poking into the blue region. What makes a solution violate it is quantitative — and the cleanest place to see the boundary is the Kerr–Newman horizon function, which has real roots only when the hole is under-spun and under-charged.

Censorship is a statement about causal sets, so its equations are set-theoretic before they are differential. The first pair defines the objects a distant observer cares about — the visible universe, and the dark region it cannot see.

where: $M$ — the physical spacetime manifold (the maximal Cauchy development); $\mathscr{I}^+$ — future null infinity, the ideal endpoint of outgoing null geodesics that reach infinity; $J^-(\mathscr{I}^+)$ — causal past of $\mathscr{I}^+$, the events that can send a signal to infinity; $\mathcal{B}$ — the black-hole region, events that can never communicate with $\mathscr{I}^+$; $\mathcal{H}$ — the event horizon, the boundary of $\mathcal{B}$, a null hypersurface. Reading it: the horizon is defined teleologically, by reference to the entire future ($\mathscr{I}^+$), which is why its location depends on everything that will ever fall in. The conjecture itself is the demand that this structure exist and be predictable from initial data. It is phrased as a subset relation, not a differential equation.

where: $\Sigma$ — a partial Cauchy surface carrying regular, asymptotically flat initial data; $D(\Sigma)$ — the domain of dependence of $\Sigma$, the region determined by evolving that data; $\overline{J^-(\mathscr{I}^+)}$ — the closure of the causal past of $\mathscr{I}^+$, i.e. the visible universe and its boundary. Reading it: everything an observer at infinity can ever see is fixed by the data on $\Sigma$; no singularity leaks new, undetermined information into the visible region. WCC claims this holds for generic data — an open dense set — not for all data. The next relation is the physical fuel for both the singularity theorems and the area theorem: the null energy condition, which after using Einstein's equation becomes a statement about matter.

where: $R_{ab}$ — Ricci tensor; $T_{ab}$ — stress–energy tensor; $k^a$ — any null vector. Reading it: because the trace term $\tfrac12 R g_{ab}k^ak^b$ vanishes for null $k$, the purely geometric condition $R_{ab}k^ak^b\geq0$ equals the physical requirement that energy density be non-negative along every light ray — the mildest of the energy conditions, and the one that drives focusing. Focusing is quantified by the Raychaudhuri equation, whose sign structure forces horizons to grow — the Hawking area theorem, a pillar that presupposes censorship.

where: $\theta$ — expansion of the horizon's null generators; $\lambda$ — affine parameter along them; $\sigma_{ab}$ — shear; $A$ — cross-sectional area of the horizon. Reading it: every term on the right is $\leq0$ under (3), so once $\theta$ dips negative it reaches $-\infty$ (a caustic) in finite $\lambda$; but horizon generators have no future endpoints, so $\theta\geq0$ everywhere and the area cannot decrease. The "no future endpoints" step needs a smooth, complete horizon — which is exactly what censorship guarantees. Penrose's own falsifiable test packages area growth and the final-state uniqueness into a single inequality between mass and horizon area on one slice.

where: $M$ — ADM mass of an asymptotically flat slice; $A$ — area of the outermost apparent horizon (an outermost minimal surface) on that slice. Reading it: if censorship and the area theorem hold, this must be true; a single configuration violating it would refute WCC. Proven rigorously in the time-symmetric (Riemannian) case, it is the strongest evidence for censorship — details in Derivation B. Finally, the local seed of a black hole, and the boundary between censored and naked, are both diagnosed by the Kerr–Newman horizon function and the trapped-surface condition.

where: $a=J/M$ — spin parameter; $Q$ — charge; $r_\pm$ — outer/inner horizon radii; $\theta_\pm$ — expansions of the outgoing/ingoing null normals of a closed 2-surface. Reading it: horizons exist only when $M^2\geq a^2+Q^2$ (real roots), and a closed trapped surface — where even outgoing rays converge — is the local signature that collapse has passed the point of no return, guaranteeing a singularity by Penrose's theorem.

Censorship is stress-tested against explicit solutions and extreme regimes; the conjecture stands or falls on whether the naked cases below are generic.

1 · Reading the extremality boundary — can you over-spin a hole into nakedness?

A charged, rotating black hole has mass $M$, spin parameter $a=J/M$, and charge $Q$ (geometric units $G=c=1$). Take $M=1$. (a) For $a=0.6,\ Q=0.5$, locate the horizons. (b) You now try to spin it up to $a=0.9$ while keeping $Q=0.5$ — does a horizon survive? (c) What does WCC say about actually reaching case (b) by dropping matter in?

Step 1. Compute the extremality combination for (a): $a^2+Q^2 = 0.6^2+0.5^2 = 0.36+0.25 = \mathbf{0.61}$. Since $M^2=1>0.61$, the discriminant $M^2-a^2-Q^2 = 0.39>0$ is positive — two horizons exist. This is the sub-extremal, censored regime.

Step 2. Their radii: $r_\pm = M\pm\sqrt{M^2-a^2-Q^2} = 1\pm\sqrt{0.39} = 1\pm 0.6245$, giving an outer horizon $\mathbf{r_+=1.6245\,M}$ and an inner (Cauchy) horizon $\mathbf{r_-=0.3755\,M}$. The singularity is safely inside.

Step 3. Now case (b): $a^2+Q^2 = 0.9^2+0.5^2 = 0.81+0.25 = \mathbf{1.06}$. This exceeds $M^2=1$, so the discriminant $M^2-a^2-Q^2 = -0.06<0$ is negative — $\Delta(r)>0$ for all $r$ and no horizon exists. Naively, the hole has been over-spun and the ring singularity is naked.

Step 4. WCC's resolution: you cannot get there. To raise $a$ from $0.6$ to $0.9$ at fixed $M$ you must feed the hole angular momentum, and any particle carrying enough $\delta J$ to over-extremalize is centrifugally repelled before it crosses $r_+$ — the first-order gedanken bound $\delta M\geq\Omega_H\,\delta J+\Phi_H\,\delta Q$ is exactly saturated at the extremal boundary. The censored endpoint (a) is reachable; the naked endpoint (b) is not, for generic infall. Later self-force analyses (Barausse–Cardoso–Khanna) confirm that radiative and backreaction corrections restore the protection where the naive first-order argument leaves a loophole.

2 · The Penrose inequality is saturated by Schwarzschild

Show that a Schwarzschild black hole saturates $M\geq\sqrt{A/16\pi}$, and evaluate the horizon radius and area for a $10\,M_\odot$ stellar-mass hole to confirm the bound is an equality, not a slack inequality.

Step 1. For Schwarzschild the event horizon sits at the areal radius $r_s=2GM/c^2$, and its area is that of a sphere of that radius: $A = 4\pi r_s^2 = 4\pi(2GM/c^2)^2 = 16\pi\,(GM/c^2)^2$. In geometric units $(G=c=1)$ this is $A = 16\pi M^2$. The horizon is a round 2-sphere, so its area is simply $4\pi r_s^2$.

Step 2. Insert into the Penrose bound: $\sqrt{A/16\pi} = \sqrt{16\pi M^2/16\pi} = \sqrt{M^2} = M$. Hence $M\geq\sqrt{A/16\pi}$ holds with exact equality — Schwarzschild is the minimum-mass configuration for a given horizon area, the extremal case of the inequality. Any spin or additional matter at fixed area would only raise $M$ above this floor.

Step 3. Numbers for $M=10\,M_\odot$: with $GM_\odot/c^2 = 1.477$ km, $r_s = 2\times10\times1.477 = \mathbf{29.5\ km}$. The area is $A = 4\pi r_s^2 = 4\pi(2.95\times10^4\ \text{m})^2 = \mathbf{1.09\times10^{10}\ m^2}$.

Step 4. Convert the mass to geometric length, $M_{\text{geo}} = GM/c^2 = 10\times1.477 = 14.77$ km, and check: $\sqrt{A/16\pi}$ in the same geometric units returns $r_s/2 = 14.77$ km $= M_{\text{geo}}$. The inequality is saturated to the digit — mass equals $\sqrt{A/16\pi}$ exactly for Schwarzschild, which is why a violation of the bound, by even a hair, would be a clean death certificate for censorship.

Weak Cosmic Censorship is one man's conjecture in the most literal sense: it was posed, named, and championed by Roger Penrose, whose singularity theorem had just made the problem unavoidable by proving that the singularities were really there.

References. Misner, Thorne & Wheeler, Gravitation (1973), Ch. 31–34 · Wald, General Relativity (1984), Ch. 12 (§12.1–12.2) and Ch. 9 · Hawking & Ellis, The Large Scale Structure of Space-Time (1973), Ch. 8–9 · Penrose, "Gravitational Collapse: the Role of General Relativity", Riv. Nuovo Cim. 1, 252 (1969) · Carroll, Spacetime and Geometry (2004), Ch. 5–6 · Hartle, Gravity (2003), Ch. 12 · Wald, "Gravitational Collapse and Cosmic Censorship", gr-qc/9710068 (1997). Key results used throughout this article: horizon radii $r_\pm=M\pm\sqrt{M^2-a^2-Q^2}$ and the extremality bound $M^2\geq a^2+Q^2$, confirmed numerically in the worked example ($a{=}0.6,Q{=}0.5\Rightarrow r_\pm=1.6245,0.3755$); Kerr area $A=8\pi(M^2+\sqrt{M^4-J^2})\leq16\pi M^2$, saturating at Schwarzschild $A=16\pi M^2$; Penrose inequality $M\geq\sqrt{A/16\pi}$, with equality at Schwarzschild checked explicitly at $r_s=29.5$ km and $A=1.09\times10^{10}$ m² for $10\,M_\odot$; area theorem from the Raychaudhuri equation $\mathrm{d}\theta/\mathrm{d}\lambda=-\tfrac12\theta^2-\sigma^2-R_{ab}k^ak^b$ under the NEC; and the causal definitions $\mathcal{B}=M\setminus J^-(\mathscr{I}^+)$, $\mathcal{H}=\partial J^-(\mathscr{I}^+)$.

Builds on: Causal Structure of Spacetime · Chronological and Causal Futures and Pasts · Domain of Dependence and Cauchy Horizons · Global Hyperbolicity · Conformal Infinity and Asymptotic Structure · Asymptotic Flatness. Leads to: Penrose–Carter Conformal Diagrams · Cauchy Surfaces · Causality Conditions Hierarchy.

Ask the tutor: Ask me to walk through why the Penrose singularity theorem stops short of proving censorship, or to work a fresh over-spinning gedanken experiment and show where the naked singularity is forbidden.