Weak Null Singularities

A weak null singularity is singular because curvature or connection-level quantities blow up on a null boundary, yet “weak” because the metric may remain continuous and the integrated tidal distortion of some infalling bodies may stay finite. The adjective does not mean negligible, removable, or known to admit a deterministic continuation. It names an integrability property that must be stated separately from metric extendibility.

Helpful background. Black-Hole Geometry and Mechanics supplies stationary horizons and freely falling frames; Black-Hole Perturbation Theory supplies tail decay; Formulations of General Relativity supplies characteristic equations; Causality and Global Structure supplies extensions and geodesic incompleteness; Computational Relativity supplies invariant numerical tests.

Curvature strength along a timelike geodesic

Let $\gamma(\tau)$ be a timelike geodesic reaching a null boundary at finite proper time $\tau_0$. Carry a parallel-propagated orthonormal tetrad with $e_{\hat0}=u$. The measured tidal matrix is

$$\mathcal R_{\hat i\hat j}(\tau) =R_{\hat0\hat i\hat0\hat j}.$$

An infinitesimal separation $\xi^{\hat i}$ obeys geodesic deviation,

$$\frac{d^2\xi^{\hat i}}{d\tau^2} =-\mathcal R^{\hat i}{}_{\hat j}\xi^{\hat j}.$$

Curvature blowup means some component or invariant becomes unbounded. Finite distortion asks a different question: whether the effect remains integrable twice along the observer. A useful sufficient diagnostic is the distortion accumulated from a regular proper time $\tau_i$ up to the current time $\tau\lt \tau_0$,

$$I(\tau)= \int_{\tau_i}^{\tau}d\tau' \int_{\tau_i}^{\tau'}d\tau''\, \left|\mathcal R_{\hat i\hat j}(\tau'')\right|, \qquad \lim_{\tau\to\tau_0^-} I(\tau)\lt \infty.$$

For a local behavior $|\mathcal R|\sim(\tau_0-\tau)^{-q}$, the double integral is finite for $q\lt 2$ and diverges for $q\ge2$, with a logarithmic borderline at $q=2$. This calculation illustrates why “curvature diverges” and “finite body is crushed to zero volume” are not synonyms. The strength criteria and their relation to mass-inflation singularities are discussed in Ori 1992, pp. 2117–2120 and Marolf and Ori 2012, §§II–IV, pp. 2–8, PDF.

A logarithmic borderline

The logarithmic borderline is particularly relevant to mass inflation. Write $x=\tau_0-\tau$ and let $x_*\gt x$ be a fixed proper-time scale. For

$$\lvert\mathcal R\rvert\sim \frac{1}{x^2[\log(x_*/x)]^p},$$

the limiting double integral has the same convergence as $\int_0 dx/[x\log^p(x_*/x)]$. It is finite precisely when $p\gt1$. Thus curvature almost as singular as $x^{-2}$ can still pass this distortion test. Inverse powers of advanced time naturally become logarithms in an exponentially compressed proper-time coordinate. This is an illustrative integrability criterion: a spacetime result must control the full transported tidal matrix and Jacobi fields, not one assumed component.

The theorem and its hypotheses

The nonsymmetric vacuum result of Dafermos and Luk starts with smooth geometric initial data in an interior region that are suitably close to data induced by a fixed subextremal Kerr solution. The analysis uses a double-null foliation, quantitative weighted control of the data, and a nonzero separation from extremality. It proves that the evolved vacuum metric extends continuously across a nontrivial Cauchy-horizon segment Dafermos and Luk 2025, §4.1, pp. 50–52, and §16, pp. 165–185, PDF.

That statement has a deliberately limited conclusion:

• it is an existence and stability result for a $C^0$ metric extension in the specified local interior problem;
• it does not assert a $C^1$ or $C^2$ extension;
• it does not by itself prove generic curvature blowup;
• it does not derive the required interior data from every asymptotically flat exterior evolution.

The 2026 exterior result now supplies a narrower rigorous bridge: Hintz reports that its decay estimates meet the Dafermos–Luk hypotheses, giving $C^0$ stability of the Cauchy horizon for the stated small vacuum perturbations of rotating subextremal Kerr Hintz 2026, v2, §1.1, consequence (1), p. 3, PDF. This current preprint does not prove generic curvature blowup, classify the complete singular boundary, or cover arbitrary collapse data.

The proof architecture renormalizes null connection and curvature quantities by subtracting the Kerr background, propagates weighted energy estimates through a blueshift region, closes a bootstrap for the difference variables, and finally changes to coordinates in which the metric coefficients converge continuously. Section 17 explains how additional lower-bound information is expected to turn the continuous horizon into a weak null singularity rather than a regular horizon Dafermos and Luk 2025, §§1.5.3–1.5.6, pp. 28–33, and §17, pp. 185–187, PDF.

Continuous metric, singular connection

A $C^0$ Lorentzian metric supplies continuous light cones and lengths, but its Christoffel symbols involve first derivatives:

$$\Gamma^\rho_{\mu\nu} =\frac12g^{\rho\sigma} \left( \partial_\mu g_{\nu\sigma} +\partial_\nu g_{\mu\sigma} -\partial_\sigma g_{\mu\nu} \right).$$

If $\partial g$ is not locally square integrable, the usual weak formulation of the Einstein equation may fail even though $g$ is continuous. Curvature adds another derivative and can blow up more strongly. Consequently a continuous extension may have no accepted evolution law, no unique continuation, and no regular freely falling frame across the boundary.

The converse failure is also important: a divergent scalar invariant is sufficient to obstruct $C^2$ extension, but some null singularities can have divergent parallel-propagated components without a polynomial scalar invariant detecting the entire effect. One must specify the diagnostic.

Proof boundary and counterexample

Remove the incoming or outgoing flux in an exactly stationary Reissner–Nordström or Kerr solution. The inner horizon is then smooth and analytic, so a Cauchy horizon does not automatically constitute a weak null singularity. Conversely, a spacelike crushing singularity can be $C^0$-inextendible and strong; it is not weak or null merely because it lies inside a black hole. These two cases show why causal character, curvature behavior, and extension regularity are independent axes.

Check the strength criterion

Assume one tidal eigenvalue behaves as $\mathcal R(\tau)=A(\tau_0-\tau)^{-3/2}$ with $A\gt 0$. Does it diverge, and is its double proper-time integral finite?

Solution

The curvature component diverges as $\tau\to\tau_0$. Integrating once gives a term proportional to $(\tau_0-\tau)^{-1/2}$; integrating a second time gives a finite limiting constant plus a term proportional to $(\tau_0-\tau)^{1/2}$. Thus $I(\tau)$ has a finite limit even though its first proper-time derivative diverges. The model is curvature-singular but weak according to this sufficient distortion test. It says nothing by itself about $C^0$, $C^1$, or Sobolev extendibility of the metric.

References

• Dafermos, Mihalis, and Jonathan Luk, “The Interior of Dynamical Vacuum Black Holes I: The C⁰-Stability of the Kerr Cauchy Horizon”, Annals of Mathematics 202, 309–630, 2025. Open PDF.
• Hintz, Peter. “Nonlinear Stability of Subextremal Kerr Black Holes.” arXiv:2606.28253v2 (3 August 2026). Open PDF.
• Marolf, Donald, and Amos Ori, “Outgoing Gravitational Shock-Wave at the Inner Horizon: The Late-Time Limit of Black Hole Interiors”, Physical Review D 86, 124026, 2012. Open PDF.
• Ori, Amos, “Structure of the Singularity Inside a Realistic Rotating Black Hole”, Physical Review Letters 68, 2117–2120, 1992.