2
$\begingroup$

What is an example of a connected symplectic manifold $(M,\omega)$, with a Hamiltonian action of $G = U(1) =S^{1}$ with infinitely many stabiliser types?

Infinitely many stabiliser types means that infinitely many sub-groups of $G$ appear as stabilisers as points in $M$.

I am aware that $M$ is necessarily non-compact.

$\endgroup$
12
  • $\begingroup$ Could you define "stabilizer type"? $\endgroup$
    – YCor
    Commented Oct 29, 2017 at 12:42
  • $\begingroup$ Thanks for the comment. I will clarify in the question, maybe there is a more standard terminology for this? $\endgroup$
    – Nick L
    Commented Oct 29, 2017 at 12:47
  • $\begingroup$ I guess it's standard as I found it on Google but it didn't come with a definition (as it's natural to guess it). Here in an abelian group it can be restated as "infinitely many distincts point stabilizers". For non-abelian group actions, probably it means infinitely many up to conjugation. $\endgroup$
    – YCor
    Commented Oct 29, 2017 at 12:54
  • $\begingroup$ You want $M$ to be connected as well I presume. Otherwise you can take a disjoint copy of $\mathbb{CP^1}$ for every $n\in \mathbb{N}$. The action on this sphere will be the $n$-fold rotation over this sphere. $\endgroup$
    – Thomas Rot
    Commented Oct 29, 2017 at 12:58
  • 7
    $\begingroup$ The question really has nothing to do with symplectic geometry: Given any connected $n$-manifold $N^n$ with a smooth $S^1$-action with infinitely many stabilizer types, then the canonical extension of this action to $T^*N$ will be an example of a Hamiltonian $S^1$-action with infinitely many stabilizer types. $\endgroup$ Commented Oct 29, 2017 at 14:30

1 Answer 1

0
$\begingroup$

As an simple example with infinitely many different stabilizer subgroups you may take countably many disks. Since each disk may be rotated independently with a different speed, we obtain an action of $ U(1) $ such that all the subgroups $ \mathbb Z / n \mathbb Z$ with $ n \in \mathbb N $ occur as stabilizer.

Of course the example obtained in this way is disconnected and even has infinitely many connected components. However, one can use the same strategy to obtain a connected manifold with infinitely many distinct stabilizer groups. The idea is to embed the collection of disks in a bigger space and "smear" the action between them. Details can be found in an old paper "On a problem of Montgomery" by C.T. Yang.

$\endgroup$

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .