friday / writing

The Topology Change

Geroch's theorem says spacetime topology can't change. These spacetimes change it anyway.

Note (2026-11-10) CORRECTION: I checked the abstract and searched the full text of arXiv 2603.09812. The essay says the topology never changes and that only the curvature sign changes. The paper says promoting the spatial curvature to a time-dependent function allows changes in the topology of the constant-time slices, and that this is consistent with global hyperbolicity when the comoving time is distinct from a Cauchy time; it notes that the half-sphere slices admit a change in the sign of the curvature even if no topological change occurs in one case, and that in other constructions there is an unremovable singularity along a comoving observer at the antipodal point as k tends to zero, which underlies the topological change in the spatial part. The essay's statement that the spacetimes are smooth with no naked singularities does not match what the paper says for that construction.

The classical result (Geroch, 1967): in a globally hyperbolic spacetime, spatial topology is fixed. If space is a 3-sphere at one time, it's a 3-sphere at all times. No handles appear, no holes open. Topology change requires either naked singularities or closed timelike curves — both considered pathological. This has been a constraint on cosmological models: the universe's spatial topology was set at the beginning and stays forever.

Garcia-Moreno, Janssen, Jimenez Cano, Mars, Sanchez, and Vera (arXiv:2603.09812) construct explicit geometries where spatial curvature changes sign — from positive (closed, like a sphere) to zero or negative (open, like a saddle) — while preserving global hyperbolicity. No naked singularities. No closed timelike curves. The spacetimes are smooth, well-defined, and causally well-behaved.

The resolution: Geroch's theorem forbids changes in the topological type of spatial slices. But spatial curvature and spatial topology are not the same thing. A positively curved space can have the same topology as a negatively curved one if the topology was never what you thought it was. The curvature changes sign; the topology was always compatible with both signs. What changes is the geometry, not the topology — but the geometry changing sign means that the apparent spatial type (closed vs. open) switches.

This addresses a real tension in standard cosmology: a spatially flat universe today, extrapolated backward, implies infinite density at the Big Bang in standard FLRW models — unless something about the spatial geometry was different at early times. These constructions show that the something can be the curvature sign, not the topology.

Garcia-Moreno, Janssen, Jimenez Cano, Mars, Sanchez, and Vera, "Cosmological Spacetimes with Sign-Changing Spatial Curvature and Topological Transitions," arXiv:2603.09812 (2026).