Pseudo-isotopies of simply connected 4-manifolds
David Gabai, David T. Gay, Daniel Hartman, Vyacheslav Krushkal, Mark Powell
TL;DR
The paper completes Quinn's program for the 4-dimensional pseudo-isotopy theorems by supplying a replacement-criterion-free argument that yields both the topological and stable smooth pseud-isotopy results for simply connected 4-manifolds. It preserves the core Cerf-theoretic framework (nested eyes, Morse data) but replaces the problematic replacement step with a factorisation and dual-sphere stabilization strategy, coupled with a refined sum-square move to control intersections. A decomposition theorem for topological 4-manifolds with boundary is established, enabling Perron-type arguments in the topological setting and yielding a robust proof of the topological pseudo-isotopy theorem, along with a stable smooth analogue. The work also clarifies the status of the disc replacement criterion as an open problem and discusses its implications via an explicit S^4 diffeomorphism scenario, linking smoothing theory to deep questions about isotopy classes of homeomorphisms.
Abstract
Perron and Quinn gave independent proofs in 1986 that every topological pseudo-isotopy of a simply-connected, compact topological 4-manifold is isotopic to the identity. Another result of Quinn is that every smooth pseudo-isotopy of a simply-connected, compact, smooth 4-manifold is smoothly stably isotopic to the identity. From this he deduced that $π_4(\operatorname{TOP}(4)/\operatorname{O}(4)) =0$. A replacement criterion is used at a key juncture in Quinn's proofs, but the justification given for it is incorrect. We provide different arguments that bypass the replacement criterion, thus completing Quinn's proofs of both the topological and the stable smooth pseudo-isotopy theorems. We discuss the replacement criterion and state it as an open problem.
