Low-Regularity Spacetimes and Weak Solutions

Smooth spacetime geometry bundles together light cones, geodesics, curvature, Einstein’s equation, and compactness. At low regularity those objects separate. A continuous metric defines causal curves but not a classical connection; a weakly differentiable metric may possess distributional curvature but no unique geodesic flow; a weak limit may satisfy an effective matter equation even when every approximating metric is vacuum. This chapter states the hypotheses that reconnect those layers.

Helpful background. Analysis for Mathematical Relativity supplies Sobolev spaces, distributions, and weak compactness; The Einstein Equation supplies the nonlinear system; PDEs and Green Functions for Gravitation supplies weak formulations; Causality and Global Structure supplies the smooth causal theory used for comparison.

Enter low-regularity spacetime geometry

The guiding question is not simply “how rough may the metric be?” Different conclusions require different derivatives and different modes of convergence:

• the cone field and Lorentzian length require only continuity and nondegeneracy;
• classical Christoffel symbols require one derivative, while unique geodesic flow typically requires locally Lipschitz Christoffels;
• distributional curvature requires a class in which both $\partial\Gamma$ and $\Gamma\Gamma$ exist;
• a weak Einstein equation requires a declared test space and gauge treatment;
• passage to a nonlinear limit requires strong convergence or a defect measure.

A lower regularity index is therefore not automatically a stronger theorem. A geometric $L^2$-curvature result, a wave-coordinate $H^s$ result, and a continuous-metric causal theorem control different objects.

The six-page route

1. Sobolev Thresholds for Einstein Evolution compares classical $s>5/2$ energy theory, rough wave-coordinate results, and the bounded-$L^2$-curvature theorem through the evidence cutoff 2026-08-12.
2. Rough Lorentzian Metrics and Causality defines causal Lipschitz curves and time separation for continuous metrics, then identifies bubbling and geodesic nonuniqueness.
3. Distributional Curvature derives the Geroch–Traschen class and explains why hypersurface delta curvature is meaningful but arbitrary distributional metrics are not.
4. Weak Formulations of Einstein’s Equations compares distributional, reduced-hyperbolic, variational, compensated-compactness, and measure-valued equations.
5. Impulsive Waves and Thin Sources applies the framework to Israel shells, null shells, and impulsive pp-waves.
6. Compactness and Limits of Spacetimes develops defect measures and the two directions of the Burnett conjecture through the evidence cutoff 2026-08-12.

Two reading routes

For local evolution, read pages 1, 3, 4, and 6. This route follows the nonlinear products from a Sobolev iteration through distributional curvature to high-frequency limits.

For singular spacetime geometry, read pages 2, 3, and 5. It begins with the cone field, adds just enough differentiability for curvature, and tests the framework on shells and impulses.

Both routes meet at the same warning: weak convergence does not commute with nonlinear geometry. If $\partial g_n$ converges only weakly, then $\partial g_n\partial g_n$ can retain a nonzero limit even when $g_n-g$ tends uniformly to zero.

The resulting high-frequency defects and their effective-matter interpretation are surveyed in Huneau–Luk 2024, §§1.1–1.3.

The sufficient metric class that makes this nonlinear curvature expression distributional, together with its codimension restriction, is established in Geroch–Traschen 1987, §§II–VI.

A regularity ladder

The following implications are useful but intentionally one-way:

$$C^{1,1} \Longrightarrow \text{unique local geodesic flow},$$ $$W^{1,\infty}_{\mathrm{loc}} \Longrightarrow \text{a.e. connection and robust cone causality},$$ $$C^0 \Longrightarrow \text{causal Lipschitz curves and Lorentzian length}.$$

The Geroch–Traschen curvature condition is not simply another rung in this pointwise chain: it combines local boundedness and nondegeneracy with $H^1_{\mathrm{loc}}$ control. Likewise, the bounded-$L^2$-curvature theorem controls spatial $L^2$ norms of spacetime curvature on evolving slices, together with associated curvature fluxes. This geometric evolution control is not merely a Hölder class for metric components.

Three explicit examples test the distinctions. The causal-bubble metric has continuous cones but fails the smooth push-up property, as constructed in Chruściel–Grant 2012, Example 1.11. A continuous impulsive-wave metric has a finite jump in its first derivative and a well-defined curvature delta. An oscillating scalar sequence converges uniformly while the square of its first derivative retains a nonzero weak limit. Work through these examples before trying to compare theorem thresholds: they show which hypothesis performs which job.

The dated assessment Low-Regularity Einstein Evolution tracks threshold-improving and ill-posedness results; the present chapter supplies the definitions and established theorem boundaries it presupposes.

Review the chapter

1. Diagnose the object. A metric is continuous but nowhere differentiable. Which of the following are immediately defined: causal Lipschitz curves, Lorentzian length, classical Christoffel symbols, distributional curvature by the Geroch–Traschen construction, and unique geodesics?

Solution

The continuous nondegenerate metric defines its cone field, causal Lipschitz curves, and their Lorentzian lengths. Classical Christoffel symbols need first derivatives. The Geroch–Traschen construction needs weak first derivatives in $L^2$ plus bounded inverse and nondegeneracy; continuity alone does not supply them. Unique geodesics require still more regularity, typically $C^{1,1}$ for the standard ODE theorem.

2. Test a shell. Why does a continuous metric with a jump in its first normal derivative produce a delta curvature but not a delta squared?

Solution

The weak first derivative is a bounded step function, so the connection is piecewise bounded. Differentiating the connection produces a delta supported on the matching hypersurface, while the quadratic connection term is only a product of bounded functions. If the metric itself jumped, its first derivative would already contain a delta and the quadratic term could become undefined.

3. Transfer task. A sequence of vacuum metrics converges uniformly, while its first derivatives converge only weakly in $L^2$. Why may the limit fail to be vacuum?

Solution

Einstein curvature contains quadratic products of first derivatives. Weak convergence does not determine the limit of those products. Their difference from the product of the weak limits is a defect measure, which can act as an effective stress tensor. Stronger convergence would eliminate the defect; a high-frequency theorem identifies its allowed structure under additional gauge and compactness hypotheses.

The next chapter asks how weighted decay and continuation estimates prevent loss of control for selected small-data solutions: Global Existence and Nonlinear Stability.

References

• Chruściel, Piotr T., and James D. E. Grant. “On Lorentzian Causality with Continuous Metrics.” Classical and Quantum Gravity 29 (2012): 145001. arXiv:1111.0400.
• Geroch, Robert, and Jennie Traschen. “Strings and Other Distributional Sources in General Relativity.” Physical Review D 36 (1987): 1017–1031. DOI.
• Huneau, Cécile, and Jonathan Luk. “High-Frequency Solutions to the Einstein Equations.” arXiv:2404.07659, 2024.

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.

• Low-Regularity Einstein Evolution