On dual pairs in Dirac geometry (Q1650172)

A driving idea in Poisson geometry advocates Poisson structures on manifolds are particular instances of more general objects, called \textit{Dirac structures}, which have become central objects in the area. One advantage obtained under this perspective is more flexibility; for example, Dirac structures can be pulled back under smooth maps satisfying a transversality condition, whereas Poisson structures can be pulled back only in very special cases. The article under review can be framed into this philosophy since the authors prove a general Dirac version of Libermann's theorem. Recall that this classical result states that, given a surjective submersion \(s\,: \Sigma\to M\), a non-degenerate Poisson structure \(\omega^{-1}\in\mathfrak{X}^2(\Sigma)\), corresponding to a symplectic structure \(\omega\in\Omega^2(\Sigma)\), can be pushed forward through \(s\) to a Poisson structure on \(M\) if and only if the symplectic orthogonal to the vertical foliation \(\ker s_*\) is involutive. In this article, the authors prove in Proposition 2 that a Dirac structure \( L\) on \(\Sigma\) can be pushed forward via \(s\) to a Dirac structure on \(M\) if and only if the Lagrangian family \(L^{s}\) (defined in page 177) is a Dirac structure on \(\Sigma\) (i.e., it forms a smooth, involutive subbundle of \(T\Sigma\oplus T^*\Sigma\)). Remarkably, in Examples 1 and 2, the authors show that under the hypotheses of this result, neither smoothness nor involutivity of \(L^s\) are ensured. So, some sufficient criteria are given in Propositions 3--5 to assure that \(L^s\) is a Dirac structure. Moreover, in Section 4, the authors discuss this result in the case when \(L=\text{Gr}(\omega)\) is the graph of a closed 2-form \(\omega\in\Omega^2(\Sigma)\). Let \((M,\pi)\) be a Poisson manifold. A \textit{symplectic realization} \(s\,: (\Sigma,\omega^{-1})\to (M,\pi)\) is a surjective and submersive Poisson map \(s\) from a symplectic manifold \((M,\omega)\). These objects are very natural in Poisson geometry from the point of view of integration and quantization theory of Poisson manifolds. However, although the existence of symplectic realizations is a fundamental result in Poisson geometry, the proofs are complicated; a direct, global, finite-dimensional proof was presented in [\textit{M. Crainic} and \textit{I. Mǎrcuţ}, J. Symplectic Geom. 9, No. 4, 435--444 (2011; Zbl 1244.53091)], by using contravariant geometry. Enframed in the perspective described at the beginning, a very natural question, which is the main motivation for the article under review, is the introduction and study of one of the possible Dirac-theoretic incarnations of the notion of symplectic realizations. More precisely, given a Dirac structure \(L\subset TM\oplus T^*M\) on a manifold \(M\), the authors consider the problem of finding a surjective submersion \(s\,: \Sigma\to M\) and a closed 2-form \(\omega\) on \(\Sigma\) such that \(s\,: (\Sigma,\text{Gr}(\omega))\to (M,L)\) is a forward Dirac map. The solution of this problem is presented in the Main Theorem in Section 6, where it is proved that any Dirac manifold \((M,L)\) fits into a self-dual pair \((M,L)\leftarrow (\Sigma,\text{Gr}(\omega))\rightarrow (M,-L)\). The main feature is the notion of a (weak) dual pair, which is introduced in this paper (Definitions 2 and 3) as natural generalizations of the corresponding notions in Poisson geometry, and whose natural operations (composition, pullback through a surjective submersion, reduction and pullback via transverse maps) are discussed in Section 5. We believe that the interesting notion of (weak) dual pairs will deserve some attention in future works in this area. Remarkably, as the authors point out, the proof of the main result is completely conceptual, relying on natural Dirac geometric operators (like pullback or gauge transformations) that are not generally available in the Poisson context. As the authors stress, the main motivation of their article is to deal with the Dirac counterpart of a symplectic realization. Nevertheless, a secondary motivation was the proof of a Dirac version of the normal form theorem, originally proved in [\textit{H. Bursztyn} et al., ``Splitting theorems for Poisson and related structures'', Preprint, \url{arXiv:1605.05386}]. This is achieved in Section 8, where the authors use a result on linearization of vector fields around submanifolds. The article ends with an insightful section where connections with other results as well as further directions are presented.