Proof of the Arnold chord conjecture in three dimensions. I (Q2428793)

The paper under review proves the Arnold chord conjecture in dimension three: If \(Y\) is a closed oriented 3-manifold with contact form \(\lambda\), then every Legendrian knot \(K\) in \((Y,\lambda)\) has a Reeb chord. Recall that a contact form on an oriented 3-manifold is a 1-form \(\lambda\) satisfying \(\lambda \wedge d\lambda > 0\) everywhere. Such a contact form determines both a contact structure \(\xi = \ker(\lambda)\) and a Reeb vector field \(R\), characterized by \(d\lambda(R, \cdot) = 0\) and \(\lambda(R) = 1\). If \(K \subset (Y,\lambda)\) is a Legendrian knot, then a Reeb chord is a trajectory of the Reeb vector field which begins and ends on the knot \(K\). Arnold's chord conjecture asserts that, for any choice of contact form \(\lambda\) and Legendrian knot \(K\), such Reeb chords always exist. The proof makes extensive use of the isomorphism between embedded contact homology (ECH) \textit{M. Hutchings} [Proceedings of the international congress of mathematicians (ICM 2010), Hyderabad, India, August 19--27, 2010. Vol. II: Invited lectures. Hackensack, NJ: World Scientific; New Delhi: Hindustan Book Agency. 1022--1041 (2011; Zbl 1233.57020)] and monopole Floer homology [\textit{P. Kronheimer} and \textit{T. Mrowka}, Monopoles and three-manifolds. New Mathematical Monographs 10. Cambridge: Cambridge University Press. (2007; Zbl 1158.57002)] established by \textit{C. H. Taubes} [Geom. Topol. 14, No. 5, 2497--2581 (2010; Zbl 1275.57037)] and can be summarized as follows: Beginning with a Legendrian knot \(K \subset (Y,\lambda)\), perform Legendrian surgery along \(K\) to obtain a new contact 3-manifold \((Y_1,\lambda_1)\). The resulting 4-manifold cobordism \(X\) induces a map on monopole Floer homology \(\phi_X : \widehat{\mathrm{HM}}(Y_1) \to \widehat{\mathrm{HM}}(Y)\), which fits into an exact triangle \[ \dots \to \widehat{\mathrm{HM}}(Y_2) \to \widehat{\mathrm{HM}}(Y_1) \to \widehat{\mathrm{HM}}(Y) \to \widehat{\mathrm{HM}}(Y_2) \to \dots. \] see [{\textit{P. Kronheimer} and \textit{T. Mrowka}, loc. cit.]. Since the cobordism \(X\) comes from a Legendrian surgery, it can be endowed with an exact symplectic structure with positive end \((Y_1,\lambda_1)\) and negative end \((Y,\lambda)\). According to [\textit{F. Bourgeois}, \textit{T. Ekholm} and \textit{Y. Eliashberg}, Geom. Topol. 16, No. 1, 301--389 (2012; Zbl 1322.53080)], the Reeb orbits of \((Y_1,\lambda_1)\) are (essentially) in one-to-one correspondence with those of \((Y,\lambda)\) together with cyclic words in the Reeb chords of \(K\). If one assumes that \(K\) has no Reeb chords, then tools from [\textit{C. H. Taubes}, [loc. cit.] and [\textit{M. Hutchings} and \textit{C. H. Taubes}, ``Proof of the Arnold chord conjecture in three dimensions. II'', Preprint, \url{arXiv:1111.3324}] can be used to show that the map \(\phi_X\) above must be an isomorphism. This, in turn, implies that \(\widehat{\mathrm{HM}}(Y_2) = 0\), contradicting a result from [\textit{P. Kronheimer} and \textit{T. Mrowka}, [loc. cit.] asserting that this same group must be infinitely generated. In the case of a degenerate contact form, a more delicate argument involving a careful analysis of the filtration on ECH coming from the symplectic action is required. In particular, the authors establish upper bounds on the symplectic action of Reeb orbits in the non-degenerate case which depend ``continuously'' on the contact form. This, in turn, establishes the chord conjecture in the degenerate case via a compactness argument.}