MathOverflow is a question and answer site for professional mathematicians. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

In Symplectic field theory ( Hofer-Eliashberg....) and considering moduli of J-holomorphic curves asymptotic to Reeb orbits at punctures (J-holomorphic curve into a symplectic cobordism), The authors and every body else who works on SFT, fix markers both on Reeb orbit and at the puncture. Does any body clearly understand why they do that ? What is that for ? Why is that needed?


--An introduction to symplectic field theory and also --Coherent orientations in SFT

share|cite|improve this question

I'll talk about cylindrical contact homology, which is a relatively well-established part of SFT. It's a variant of Hamiltonian Floer homology for contact manifolds. Let $\alpha$ be a contact 1-form on a closed manifold $V$, and let $A\colon LV\to \mathbb{R}$, $\gamma\mapsto \int_{S^1}{\gamma^*\alpha}$ be the action functional on its loopspace. It's invariant under rotation of loops, hence not a Morse function. The critical points are the 1-periodic orbits of the Reeb vector field.

Here are three things we could try to construct:

(i) The Floer cohomology of $A$. Each geometric Reeb orbit with multiplicity contributes $H^*(S^1)$ to the cochain complex.

(ii) The $S^1$-equivariant Floer cohomology of $A$. Each geometric Reeb orbit with multiplicity contributes $H^\ast_{S^1}(S^1)$ to the complex (the $S^1$ action on itself depends on the multiplicity).

(iii) The Floer homology of $A$ on $LV/S^1$ (over $\mathbb{Q}$). Each geometric Reeb orbit with multiplicity contributes $\mathbb{Q}$ to the complex.

Which of these things work? None of them, without substantial modification. All of them, with modifications. They are called (i) symplectic cohomology; (ii) circle-equivariant symplectic cohomology; (iii) cylindrical contact homology. In each case, the differential essentially counts pseudo-holomorphic maps $S^1\times \mathbb{R}\to V\times \mathbb{R}$, asymptotic to periodic Reeb orbits, but with subtle differences.

Since (iii) is a quotient construction we have to allow loop-rotation; so we mark a standard point on $S^1\times -\infty$, an arbitrary point on $S^1\times +\infty$, and insist that these markers map to chosen points on the Reeb orbits. In (i), we would use the standard marker also on $S^1\times +\infty$; allowing it to vary defines a loop-rotation (BV) operator on symplectic cohomology.

See Bourgeois-Oancea's recent Inventiones paper for info on the relationship between these constructions.

share|cite|improve this answer
I don't think that this the reason. If there is a reason, that should be hided in the orientation problem, or definition of CZ index. None of the things you said force us to put markers, and in all cases you can simply consider moduli spaces which asymptote to Reeb orbits. – Mohammad F. Tehrani Apr 20 '10 at 15:52
Look at last paragraph of page 4 of "Coherent orientations in..." – Mohammad F. Tehrani Apr 20 '10 at 15:54
The differences become apparent when one looks at the compactified moduli spaces. For a visualization, see Bourgeois-Oancea, section 6 (p. 39 in the linked ArXiv version). – Tim Perutz Apr 20 '10 at 16:42
I agree that it makes some differences. But what will happen if we forget the marking? Does it make problem for : -orienting moduli space -compactifying moduli space -defining CZ index ? – Mohammad F. Tehrani Apr 20 '10 at 17:25
OK, thanks for pushing me to make this clear. The definition of the contact homology differential (e.g. as in B-O section 3) involves not just cylinders, but compactified spaces of rational curves with multiple punctures, each with an asymptotic marker. I believe that the SFT compactification is not quite the same as "D-M space with asymptotic markers", so forgetting markers would radically change the spaces. So far as I can see, the index and orientation (or bad orbit) issues are not particularly relevant here. – Tim Perutz Apr 20 '10 at 18:00

The reason is related to orientations, and more specifically to bad orbits (see the end of paper :Coherent orientation in SFT). Assume that a holomorphic curve is asymptotic to a bad orbit ( means $n-1+CZ_{index}$ is even) of multiplicity 2m (a bad orbit always has even multiplicity). Theorientation depends on the choice of an asymptotic marker near the corresponding puncture; if it is rotated by \pi/m, then the orientation is reversed. Hence, the algebraic count of curves asympttic to a bad orbit always vanishes. Of course, as soon as we restrict to good orbits only (so that the orienttion does not depend on the marker anymore), then the asymptotic markers are not needed anymore for the orientations.

share|cite|improve this answer

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

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