Knots in homology spheres and Levine's theorem #
Knots in integral homology 3-spheres are compared by homology concordance: two knots
K₀ ⊆ Y₀ and K₁ ⊆ Y₁ are homology concordant when they cobound a smoothly and properly embedded
annulus in some smooth integral homology cobordism between Y₀ and Y₁
(TauCeti.SmoothEmbedding.HomologyConcordant, with the circle as source). Every knot in S³ is
homology concordant to a knot in S³, namely itself
(TauCeti.not_forall_not_homologyConcordant_sphere).
Matsumoto asked (Kirby's problem 1.31) whether the knots in homology spheres bounding acyclic
4-manifolds are, up to homology concordance, just the knots in S³. Levine answered this
negatively: there is a smooth compact contractible 4-manifold X and a knot K in its boundary
Y = ∂X which is not homology concordant to any knot in S³. Since X is contractible, Y is an
integral homology sphere, by Lefschetz duality. Levine states the result as: K bounds no
piecewise-linear disc in any homology ball with boundary Y. This implies the statement here. A
homology concordance W from K to a knot K' in S³, capped off by the four-ball along S³,
would be a homology ball with boundary Y in which the annulus, together with the cone on K',
is a piecewise-linear disc bounded by K.
TauCeti.LevineHomologyConcordanceTheorem records the statement in the smooth category, with
ℤ coefficients. It is stated, not proved. The four-manifolds X and W are modelled on the
half-space ℝ³ × [0, ∞), the model (𝓡 3).prod (𝓡∂ 1) of the cylinders over 3-manifolds; this
is the half-space of Lee's Introduction to Smooth Manifolds. No orientations are imposed; by the
discussion in TauCeti/Geometry/Manifold/SmoothEmbedding/HomologyConcordance.lean this only allows
K' to be replaced by its mirror or reverse, which are again knots in S³.
Main definitions #
TauCeti.LevineHomologyConcordanceTheorem: some knot in the boundary of a smooth compact contractible4-manifold is not homology concordant to any knot inS³.
Main results #
TauCeti.not_forall_not_homologyConcordant_sphere: every knot inS³is homology concordant to some knot inS³, so the boundary in Levine's theorem is notS³.TauCeti.LevineHomologyConcordanceTheorem.exists_isEmpty_diffeomorph_sphere: by Levine's theorem, some smooth compact contractible4-manifold has boundary not diffeomorphic toS³.
References #
- A. S. Levine, Nonsurjective satellite operators and piecewise-linear concordance, Forum Math. Sigma 4 (2016), e34, Theorem 1.1 and Remark 1.5.
- R. Kirby (ed.), Problems in Low-Dimensional Topology, Problem 1.31, in Geometric Topology, AMS/IP Stud. Adv. Math. 2.2 (1997).
- J. M. Lee, Introduction to Smooth Manifolds, 2nd ed., Graduate Texts in Mathematics 218, Springer (2013), Chapter 1, for the half-space model.
Levine's theorem (an answer to Kirby's problem 1.31): there are a smooth compact
contractible 4-manifold X, whose boundary is a closed 3-manifold Y, and a knot K in Y
that is not integrally homology concordant to any knot in S³.
The boundary of X is presented by a smooth embedding κ : Y → X onto it. The four-manifolds are
modelled on the half-space ℝ³ × [0, ∞). It is recorded as a proposition and not proved.
Instances For
Levine's theorem spelled out.
Every knot in S³ is homology concordant to a knot in S³, namely itself, through the
cylinder S³ × [0, 1]. So the conclusion of Levine's theorem fails for knots in S³: the boundary
Y there is not S³.
Levine's theorem gives a boundary that is not S³. Assuming Levine's theorem, some smooth
compact contractible 4-manifold X has as boundary a closed 3-manifold Y not diffeomorphic to
S³. A diffeomorphism e : Y ≃ₘ S³ would make the knot K of Levine's theorem homology concordant
to the knot e ∘ K in S³ (TauCeti.SmoothEmbedding.homologyConcordant_transDiffeomorph).