Singularity Theorems

Core Idea

The Penrose (1965) and Hawking-Penrose (1970) singularity theorems prove that singularities — defined as geodesic incompleteness (the existence of geodesics that cannot be extended to arbitrary parameter values) — are generic features of general relativity under physically reasonable conditions, not artifacts of special symmetry. Penrose's theorem requires: (1) the existence of a trapped surface (a closed surface from which all outgoing light rays converge), (2) a reasonable energy condition (null energy condition: T_μν k^μ k^ν ≥ 0 for all null vectors k^μ), and (3) global hyperbolicity. Under these conditions, at least one geodesic must be incomplete — a singularity exists. Hawking extended this to cosmological singularities, proving that an expanding universe satisfying the strong energy condition must have begun from a singularity (the Big Bang). These theorems do not describe the nature of the singularity — they only prove its existence.

Explainer

Before the singularity theorems, the singularity in the Schwarzschild and Friedmann solutions was widely regarded as an artifact of their perfect symmetry. Perhaps a slightly asymmetric collapse would produce a "bounce" rather than a singularity, and perhaps the Big Bang singularity would be avoided in a slightly inhomogeneous universe. The Penrose singularity theorem of 1965 demolished this hope. Using global geometric methods rather than explicit solutions, Penrose proved that once a trapped surface forms — a surface from which even outgoing light converges — a singularity is inevitable, regardless of any symmetry assumptions.

The key concept is geodesic incompleteness. A spacetime is called geodesically incomplete if there exists at least one geodesic (timelike or null) that cannot be extended to all values of its affine parameter. Physically, this means a freely falling particle or light ray reaches the "edge" of spacetime in finite proper time or affine parameter — its worldline simply ends. This is the mathematically precise definition of a singularity used in the theorems. The theorems do not say what happens at the singularity (infinite curvature, infinite density, etc.) — they only prove that geodesics terminate. In all known exact solutions, the termination is accompanied by divergent curvature, but this is not guaranteed in general.

Penrose's theorem requires three ingredients: (1) a trapped surface exists, (2) the null energy condition holds (T_μν k^μ k^ν ≥ 0 for all null vectors k — roughly, energy density is non-negative as seen by any light ray), and (3) the spacetime is globally hyperbolic (well-posed initial-value problem). The proof uses the Raychaudhuri equation, which governs the expansion of a congruence of geodesics: the energy condition ensures that the expansion of null geodesics from the trapped surface cannot stop decreasing, and the trapped-surface condition means the expansion starts negative. The geodesics must therefore reach zero expansion (a caustic or focal point) in finite affine parameter, and global hyperbolicity prevents them from simply exiting the spacetime. The conclusion: at least one geodesic is incomplete.

Hawking adapted Penrose's methods to cosmology. His theorem (and the later Hawking-Penrose theorem of 1970) proved that an expanding universe satisfying the strong energy condition — (ρ + 3p/c²) ≥ 0 for a perfect fluid — must have a past singularity: the Big Bang. The expanding universe plays the role of the trapped surface (run time backward and the expansion becomes convergence). The discovery of accelerating cosmic expansion (1998) means the strong energy condition is violated by dark energy, which technically invalidates the theorem's applicability to the far future. However, the past singularity (Big Bang) remains robust under weaker conditions. The profound lesson of the singularity theorems is that general relativity, under generic conditions, predicts its own breakdown — signaling the need for a quantum theory of gravity to describe the physics of extreme curvature. Penrose received the 2020 Nobel Prize in Physics for this work.