13
$\begingroup$

Is it true that any simply connected closed manifold possesses a Morse functions that does not have critical points of index one?

If the dimension is at least 5, this is a consequence of the results from Milnor's "Lectures on the $h$-cobordism theorem", but what about dimensions 3 and 4?

If such a function does not exist in general in low dimensions, are there reasonable topological conditions for the existence of a Morse function without critical points of index one?

$\endgroup$

2 Answers 2

15
$\begingroup$

It is still an open and very interesting question in dimension 4. Akbulut (The Dolgachev surface. Disproving the Harer-Kas-Kirby conjecture. Comment. Math. Helv. 87 (2012), no. 1, 187–241) showed that the Dolgachev surface (and subsequently other elliptic surfaces in the same homotopy type) has a handle decomposition with no 1 or 3 handles.

In the opposite direction, Rasmussen gave an argument that an exotic homotopy $S^2 \times S^2$ detected by an Ozsvath-Szabo (or presumably Seiberg-Witten) invariant would require either 1 or 3-handles. That paper is currently withdrawn from the arxiv; quoting the abstract, `I am withdrawing the paper because it is unclear to me if such a manifold exists.'

$\endgroup$
1
  • $\begingroup$ Thanks a lot! I wouldn't have guessed that this question is so delicate... $\endgroup$ Oct 3, 2013 at 13:43
7
$\begingroup$

In dimension $3$ we know from Perelman's work that a simply connected $3$-manifold is a sphere. In dimension $4$ I quote Kirby in 1989

"It is not known if a simply connected $4$-manifold needs $1$-handles and/or $3$-handles but the Dolgachev surface is a good candidate for needing them." page 8, R. Kirby: The Topology of $4$-Manifolds, Lect. Notes in Math. vol. 1374.

Many things have happened from the time Kirby made the above statement and I have not kept up with all the developments. In any case, the answer to your question is not straightforward.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.