The problem
An ordinary sphere’s surface needs two coordinates near each point. Its four-dimensional counterpart needs four. Local coordinates let us take derivatives, but their overlaps must fit together smoothly for calculus to be consistent. Could a space that closes up without a boundary, and has the global topology of a four-sphere, support a way of doing calculus that cannot be smoothly identified with the usual one?[1] [2] [4]
A manifold is locally described by ordinary coordinates; closed means compact with no boundary. The standard four-sphere is , with its usual smooth coordinates. The target is every closed smooth four-manifold homotopy equivalent to . Must there be a diffeomorphism ? An exotic four-sphere would have the same underlying topology but no such smooth equivalence.[2] [4]
Scope
Closed smooth four-manifolds homotopy equivalent to S⁴; the smooth, not merely topological, classification.
What remains open
Continuously reversible need not mean smoothly reversible. Cubing sends 2 to 8 and is continuous in both directions, but its cube-root inverse has no finite slope at zero. This one-dimensional illustration does not produce an exotic line: the identity is still a smooth equivalence. Four-spheres are known to agree continuously. The unknown is whether some smooth equivalence always exists.[1] [2] [3] [4]
A space can be assembled from pieces called handles; simplifying it often involves canceling pairs of pieces. The surfaces used in these arguments can meet at isolated points in four dimensions. Removing paired crossings by a Whitney move requires an embedded disk with compatible normal directions, and the higher-dimensional argument does not provide these requirements automatically in dimension four. These obstacles concern proof methods; they do not establish that an exotic sphere exists.[2]
Why it matters
Calculus describes velocities, flows and differential equations on a space. It requires compatible coordinates, beyond knowing how points are connected. Smooth classification determines when these tools can be transferred consistently between spaces, so a sphere’s topology alone may leave an essential part of its geometry unspecified.[4]
Knots can encode how pieces of a four-manifold are attached. In studied candidate four-spheres, questions about whether associated knots bound smooth disks test whether the punctured space is a standard four-ball. This connects global classification to concrete knot invariants and handle manipulations, even when the particular candidates eventually prove standard.[5]
Definitions
3 terms
smooth structure
Compatible local coordinate charts used to define derivatives on a manifold. Two different chart collections need not represent inequivalent smooth manifolds: equivalence is tested by a diffeomorphism.[4] [1]
diffeomorphism
A one-to-one map onto the other smooth manifold, with both the map and its inverse smooth. Smooth here means having derivatives of every order in local coordinates.[4] [3]
homotopy equivalent
Two spaces admit continuous maps in both directions such that each round trip can be continuously deformed into the corresponding identity map. This is the precise sphere-like assumption in the problem.[2]
Sources
- Small Triangulations of 4-Manifolds and the 4-Manifold Census
Notes on Small Triangulations of 4-Manifolds and the 4-Manifold Census
Refereed 2026 research; Conjecture 1 explicitly gives the smooth four-sphere target.
- The 4 Dimensional Poincaré Conjecture
Notes on The 4 Dimensional Poincaré Conjecture
Author-hosted notes dated March 27, 2019. Sections 1.1–1.2 set up homotopy spheres, section 1.7 discusses handle cancellation and Whitney moves, and Corollary 4.10 / Theorem 4.12 give the topological four-sphere conclusion. Used for established mathematics, not current status claims.
- Introduction to Dynamical Systems — lecture notes
Notes on Introduction to Dynamical Systems — lecture notes
Author-hosted course notes dated April 11, 2007. Section 3.1, printed page 20, distinguishes homeomorphisms from diffeomorphisms and works out the cubing / cube-root example.
- Lecture Notes — MATH 215B (Winter 2021)
Notes on Lecture Notes — MATH 215B (Winter 2021)
Author-hosted Winter 2021 notes. January 11–13 sections explain compatible charts, calculus and tangent vectors; Question 12 distinguishes different atlases on the line from inequivalence up to diffeomorphism.
- Man and machine thinking about the smooth 4-dimensional Poincaré conjecture
Notes on Man and machine thinking about the smooth 4-dimensional Poincaré conjecture
Publisher-hosted research paper, Quantum Topology 1 (2010), 171–208, DOI 10.4171/QT/5. Sections 1–2 explain how handle descriptions and associated knot disk-bounding tests were used on candidate four-spheres; the studied Cappell–Shaneson family was ultimately proved standard.