Classical Singularity Theorems

Classical singularity theorems are global incompleteness results, not calculations of a singular point. They combine local focusing with a global causal obstruction: curvature and initial geometry force focal or conjugate points, while causality demands geodesics that continue to maximize or generate an achronal boundary without such points. This chapter develops that shared mechanism, then distinguishes the exact hypotheses and conclusions of the Penrose, Hawking, and Hawking–Penrose theorems.

Enter this chapter

The central question is not whether gravity can focus geodesics; flat-space congruences already focus. It is whether specified initial geometry, convergence conditions, and causal topology make indefinite continuation impossible. Every theorem therefore has several independent inputs:

• a differentiability class in which geodesics, curvature, Jacobi fields, and causal boundaries have the required properties;
• a Ricci-convergence condition that keeps a twist-free congruence focusing;
• an initial condition that starts focusing, or a genericity condition that prevents permanent tidal degeneracy;
• a causal and topological condition that supplies global maximizers or an incompatible trapped set;
• an exactly typed conclusion: timelike, null, or disjunctive causal geodesic incompleteness, in a stated direction when available.

No entry can be replaced by the phrase “gravity is attractive.” The local Raychaudhuri inequality and the global contradiction perform different jobs. The common pattern and all three smooth classical statements are organized in Steinbauer 2023, §§3.1–3.5, pp. 9–20.

The chapter sequence

Read the pages in order for a theorem-building route.

1. Anatomy of a Singularity Theorem separates regularity, convergence, initial geometry, causality, topology, and conclusion. It shows why no single hypothesis produces incompleteness alone.
2. Focusing and Conjugate-Point Lemmas integrates Raychaudhuri and translates a zero of the Jacobi determinant into lost maximizing or boundary-generating behavior.
3. The Penrose Singularity Theorem uses a closed future trapped surface, null convergence, global hyperbolicity, and a noncompact Cauchy surface to prove future null incompleteness.
4. Hawking Cosmological Singularity Theorems uses uniform expansion or contraction of a compact Cauchy slice and timelike convergence to prove past or future timelike incompleteness.
5. The Hawking–Penrose Theorem replaces global hyperbolicity by chronology, adds generic tidal curvature on every causal geodesic, and permits three initial alternatives. Its basic conclusion is timelike-or-null incompleteness without a fixed direction.
6. What Singularity Theorems Conclude separates those results from curvature blowup, singular endpoints, inextendibility, horizon formation, cosmic censorship, and quantum resolution.

The final page is part of the theorem sequence, not an optional disclaimer. Correctly limiting a conclusion is as important as verifying its assumptions.

One analytic engine, different global engines

For a twist-free causal congruence with transverse dimension $d$, pointwise Ricci convergence gives

$$\dot\theta \leq-\frac{\theta^2}{d}.$$

If $\theta_0<0$, a focal or conjugate point occurs within parameter at most $d/\lvert\theta_0\rvert$, provided the geodesic exists that long. Timelike congruences use $d=n-1$; null screens use $d=n-2$. This is the shared analytic engine.

The theorems differ after that step.

Theorem	Focusing seed	Global causal input	Topological input	Exact basic conclusion
Penrose	both future null expansions negative on a closed trapped surface	global hyperbolicity	a noncompact Cauchy surface	a future-incomplete null geodesic
Hawking, expanding form	future expansion uniformly positive on a spacelike Cauchy slice, hence past-normal expansion negative	global hyperbolicity from the Cauchy slice	compactness gives a uniform expansion bound	a past-incomplete timelike geodesic
Hawking–Penrose	convergence plus generic transverse tidal curvature on every causal geodesic	chronology and the causal Hawking–Penrose lemma	one of three compact or trapped initial alternatives	an incomplete timelike or null geodesic, direction not fixed

Penrose confines boundary generators by focal points, makes the horismos compact, and projects it onto a Cauchy surface. Hawking bounds Lorentzian distance from a Cauchy slice and contradicts a complete timelike normal. Hawking–Penrose proves that chronology, a trapped set, and a conjugate pair on every inextendible causal geodesic cannot coexist. Calling all three “the Raychaudhuri proof” hides their most important differences.

Check your preparation

You should be able to distinguish an arbitrary causal curve from an affinely parametrized geodesic and explain why conjugate points affect maximizing behavior. Repair that material through Affine and Variational Geodesics and Raychaudhuri Equations for Timelike and Null Congruences.

For the global half, review Equivalent Criteria for Global Hyperbolicity and Trapped and Marginal Surfaces. Three distinctions should already be secure:

• a focal point is not an endpoint of the geodesic;
• a trapped surface is not an event horizon;
• an incomplete geodesic is not a singular point in the manifold.

Readers interested mainly in gravitational collapse can take the anatomy, focusing, Penrose, and conclusion pages before returning to Hawking–Penrose. Readers interested in cosmology can take the anatomy, focusing, Hawking, and conclusion pages. A mathematical-relativity route should read all six and track metric regularity at every theorem boundary.

Energy and convergence conditions

The proofs use geometric inequalities such as

$$R_{ab}k^ak^b\geq0$$

for null $k$ or $R_{ab}u^au^b\geq0$ for timelike $u$. Einstein’s equation translates the null condition directly into the null energy condition because terms proportional to the metric vanish on $k$. The timelike translation involves the stress–energy trace and the cosmological constant. A positive cosmological constant can violate timelike convergence even when ordinary matter satisfies its own inequality.

This distinction explains two important controls. Vacuum Schwarzschild has $R_{ab}=0$, so it satisfies null convergence while its Weyl curvature supplies nontrivial tidal geometry. Global de Sitter is geodesically complete with compact expanding slices but violates timelike convergence. Energy conditions enter only through the curvature terms used by the chosen theorem; they do not by themselves encode causality, trappedness, or completeness.

Counterexamples organize the hypotheses

Each familiar example removes a different overstatement.

Punctured Minkowski spacetime is incomplete, flat, and smoothly extendible. It separates incompleteness from curvature blowup and inextendibility.

Global de Sitter spacetime is complete and can contain trapped surfaces in a contracting region, while its Cauchy slices are compact. It shows why Penrose needs a noncompact Cauchy surface and why Hawking needs timelike convergence.

The Einstein static universe has compact Cauchy slices and is complete; its balanced congruence lacks the strict expansion seed, and special geodesics expose the role of genericity in the combined theorem.

Maximally extended Schwarzschild supplies trapped spheres, null convergence, noncompact Cauchy topology, and future null incompleteness. Curvature divergence at $r=0$ is established by the exact metric, not by the Penrose theorem.

These examples should be used one hypothesis at a time. A counterexample that violates three assumptions cannot identify which one a proof actually uses.

Smooth and low-regularity layers

The pages first state the classical $C^2$ or smooth theory, where pointwise curvature, unique geodesics, Jacobi fields, normal exponential maps, and familiar causal variation arguments are available. They then mark—not merge—the lower-regularity layer.

At $C^{1,1}$, curvature is locally bounded and the exponential map retains enough control for regularization methods. At $C^1$, curvature is distributional and the geodesic equation may lose uniqueness. Modern theorems formulate weak convergence and genericity conditions, approximate the metric smoothly, and in the Hawking–Penrose case may require maximal causal nonbranching. Steinbauer surveys these changes in Steinbauer 2023, §4, pp. 21–40.

A low-regularity theorem is not obtained by deleting “smooth” from a classical hypothesis list. Its curvature inequality, completeness notion, causal comparison, and nonbranching assumptions must all be stated afresh.

Review the chapter

Match each data set below to the theorem it directly supports and give only the guaranteed conclusion.

1. A smooth globally hyperbolic spacetime with noncompact Cauchy surface contains a closed future trapped surface and obeys null convergence.
2. A smooth spacetime has a compact spacelike Cauchy surface whose future expansion is uniformly positive and obeys timelike convergence.
3. A smooth chronological spacetime obeys causal convergence and genericity on every causal geodesic and contains a future trapped point.

Solution

Case 1 is the future-directed Penrose theorem: at least one future null geodesic is incomplete. It does not by itself give timelike incompleteness or an event horizon.

Case 2 is the past-directed Hawking cosmological theorem: at least one past timelike geodesic is incomplete. It does not identify a common Big Bang point or prove null incompleteness.

Case 3 meets a Hawking–Penrose initial alternative. The basic combined theorem guarantees that at least one causal geodesic is incomplete—timelike or null—with no direction fixed by the statement used here. A stronger directional claim would require a correspondingly strengthened theorem.

Onward use

The immediate next conceptual step is Timelike, Null, and Geodesic Completeness together with Regularity Classes and Inextendibility. Those pages explain what an incomplete curve does and does not say about extending the metric.

For gravitational collapse, continue to Weak Cosmic Censorship and Strong Cosmic Censorship. Censorship adds data classes, genericity topologies, visibility, maximal developments, and extension regularity. The singularity theorems supply indispensable global constraints, but not those verdicts.

References

• Hawking, Stephen W., and Roger Penrose, “The Singularities of Gravitational Collapse and Cosmology”, Proceedings of the Royal Society A 314, pp. 529–548, 1970.
• Penrose, Roger, “Gravitational Collapse and Space-Time Singularities”, Physical Review Letters 14, pp. 57–59, 1965.
• Steinbauer, Roland, “The Singularity Theorems of General Relativity and Their Low Regularity Extensions”, Jahresbericht der Deutschen Mathematiker-Vereinigung 125, §§3–4, pp. 9–40, 2023.