Legendrian varieties (Q2477148): Difference between revisions

Let \(X\) be complex projective manifold \(X\) of dimension \(n\) embedded in a projective space \(\mathbb{P} V\) of dimension \(2n+1\). We say that the embedded manifold \(X\) is Legendrian if \(V\) can be endowed with a symplectic form such that in every point of \(X\) the affine tangent space to \(X\) is a maximal isotropic subspace. One reason to study such varieties is that any projective manifold of dimension \(n\) embeds in a projective space of dimension \(2n+1\), so it is natural to ask if there are some embeddings with special properties. By a theorem of \textit{R. Bryant} [in: The mathematical heritage of Hermann Weyl, Proc. Symp., Durham/NC 1987, Proc. Symp. Pure Math. 48, 227--240 (1988; Zbl 0654.53010)] every smooth curve admits a Legendrian embedding in \(\mathbb{P}^3\), but in higher dimension the authors establish a series of Chern class identities that yield strong restrictions on Legendrian varieties. This implies a series of corollaries, for example an abelian variety of dimension at least two is never Legendrian. Another motivation is the relation with contact Fano manifolds: \textit{S. Kebekus} has shown [J. Reine Angew. Math. 539, 167--177 (2001; Zbl 0983.53031)] that if \(Y\) is a smooth contact Fano manifold with Picard number one that is not the projective space, the set of tangent directions to contact lines passing through a general point is a smooth Legendrian variety \(X\) in its linear span. In the paper under review the authors show that if \(X\) is rational homogeneous (i.e. a quotient of a semi-simple algebraic group \(G\) by some parabolic subgroup \(P\)) with Picard number one and Legendrian, then it is the set of lines passing through a point of a contact Fano manifold \(Y\), the adjoint variety to \(G\). The second half of the paper is devoted to the construction of an example of a non-homogeneous smooth Legendrian surface: Let \(Z\) be a Kummer quartic surface in \(\mathbb{P}^3\), that is a quartic surface with sixteen double points as singularities. Its dual variety \(Z^*\) is again a Kummer quartic surface and there exists a \(K3\) surface \(Z^\#\) that resolves the singularities of \(Z\) and \(Z^*\). The pull-back of general hyperplane sections from \(Z\) and \(Z^*\) to the \(K3\) surface \(Z^\#\) meet transversely in twelve points. The authors show that the blow-up of these twelve points is Legendrian. We refer to the paper for the ideas behind this beautiful construction. Note that since the publication of this paper, \textit{J. Buczyński} has constructed further examples of non-homogeneous Legendrian varieties [Hyperplane sections of Legendrian subvarieties, 2007, \url{arXiv:math/0701590}].