Introduction to moduli spaces and Dirac geometry

Eckhard Meinrenken · 2025

Read Introduction to moduli spaces and Dirac geometry by Eckhard Meinrenken (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2506-04150.txt.

Introduction

Eckhard Meinrenken gives an introduction to moduli spaces and Dirac geometry, based on lectures at a 2024 summer school in Naples, in a 2025 article. The setting starts with a Lie group and a surface with boundary. From these, a space of representations is built, and it carries a special two-form.

Meinrenken explains the finite-dimensional construction of that form and its basic traits. The tools come from quasi-Hamiltonian methods and Dirac geometry. The article is an extended version of the lectures, written to bring newcomers into the topic. Readers get both the main construction and the ideas needed to work with it further.

Catalog record

CREATOREckhard Meinrenken
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE138 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2506-04150

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Introduction to moduli spaces and Dirac geometry Eckhard Meinrenken Abstract . Let G G be a Lie group, with an invariant metric on its Lie algebra 𝔤 \mathfrak{g} . Given a surface Σ \Sigma with boundary, and a collection of base points 𝒱 ⊆ Σ \mathcal{V}\subseteq\Sigma meeting every boundary component, the moduli space (representation variety) ℳ G ​ ( Σ , 𝒱 ) \mathcal{M}_{G}(\Sigma,\mathcal{V}) carries a distinguished ‘quasi-symplectic’ 2-form. We shall explain the finite-dimensional construction of this 2-form and discuss its basic properties, using quasi-Hamiltonian techniques and Dirac geometry. This article is an extended version of lectures given at the summer school ’Poisson 2024’ at the Accademia Pontaniana in Napoli, July 2024. 1. Introduction This article is an extended version of lectures given at the summer school ’Poisson 2024’ at the Accademia Pontaniana in Napoli, July 2024. The moduli spaces in its title are the moduli spaces \mathcal{M}_{G}(\Sigma,\mathcal{V}) of flat G -bundles over compact oriented surfaces \Sigma , with framings at a finite collection \mathcal{V}\subseteq\Sigma of base points, meeting every component of the boundary. Equivalently, they are described as spaces of homomorphims from the fundamental groupoid into the Lie group G . If the Lie algebra of \mathfrak{g} carries an invariant metric, then the moduli space acquires a distinguished 2-form \omega . For the case without boundary (and \mathcal{V}=\emptyset ), this is the Atiyah-Bott symplectic form. When \Sigma has non-empty boundary, then this 2-form is neither closed or nondegenerate, but for \mathcal{V}\subseteq\partial\Sigma it is quasi-symplectic in the sense of quasi-Hamiltonian geometry. My original plan for the lectures had been to divide the material into two parts. The first part would present a direct ‘low-tech’ construction of the 2-form via cutting and gluing of the surface. Most properties of the 2-form are direct consequences from the construction, with the exception of ‘minimal degeneracy’ of the 2-form. This would then serve as a motivation for the second part, leading to an introduction to Dirac geometry as the proper framework for quasi-Hamiltonian spaces, with moduli spaces serving as the main examples. In reality, this plan was too ambitious, and the lectures never actually reached the second part. Accordingly, Sections 2 – 5 of the present article are written in the style of lecture notes, replete with exercises. In later sections the presentation gradually morphs into a survey style, discussing material that was only hinted at in the lectures. In more detail, the material is organized as follows. Section 2 begins with a rapid overview of moduli spaces of flat bundles over surfaces, and associated representation varieties. Section 3 defines the moduli space \mathcal{M}_{G}(\Sigma,\mathcal{V}) for any finite collection \mathcal{V}\subseteq\Sigma (meeting every component) as the space of homomorphisms from the fundamental groupoid \Pi(\Sigma,\mathcal{V}) into G . The space carries an action of G^{\mathcal{V}} , and taking holonomies along boundary edges \mathsf{e}\in\mathcal{E} defines an equivariant map \Phi\colon\mathcal{M}_{G}(\Sigma,\mathcal{V})\to G^{\mathcal{E}} . Starting in Section 4 , we assume that \mathfrak{g}=\operatorname{Lie}(G) carries an invariant metric. This determines a Cartan 3-form \eta\in\Omega^{3}(G) and a related 2-form \beta\in\Omega^{2}(G\times G) . The main result of this section is the existence of a distinguished 2-form \omega on the space \mathcal{M}_{G}(\Sigma,\mathcal{V}) , whose differential is the sum of pullbacks of \eta under the boundary holonomies, and whose contractions with the generators of the G^{\mathcal{V}} -action are described explicitly. The 2-form admits a direct description for the case that \Sigma is a polyhedral region (an n -gon); the general case is reduced to this case via gluing diagrams. Using cutting and re-gluing, we prove that the 2-form does not depend on the choice of gluing diagram. Section 5 continues the discussion of cutting and gluing operations, examining in particular the gluing of two boundary circles of a surface. Among other things, this leads to a description of the 2-form for surfaces without boundary. Section 6 is devoted to Hamiltonian dynamics on the moduli spaces: Every G^{\mathcal{V}} -invariant function f on the moduli space determines a Hamiltonian vector field X_{f} (despite the fact that \omega is degenerate). For special choices of f , this leads to the so-called Goldman flows which we compute explicitly. Section 7 shows that the moduli space for a cylinder is naturally a ‘quasi-symplectic groupoid’, where the groupoid multiplication is defined via gluing of cylinders. Finally, Section 8 gives an introduction to Dirac geometry, and supplies proofs of minimal degeneracy properties of the 2-form via a cross-section theorem for Dirac structures. It is a pleasure to thank the organizers of the Poisson 2024 summer school and the Accademia Pontaniana for the invitation to present these lectures in a wonderful setting, as well as INdAM for sponsoring the event. Special thanks to Luca Vitagliano for help during the meeting and Alfonso Tortorella, Antonio de Nicola, and Chiara Esposito for editing the proceedings. 2. Representation varieties Let G be a Lie group, and \Sigma a compact, connected surface, without boundary. The representation variety or character variety is the space \mathcal{M}_{G}(\Sigma)=\operatorname{Hom}(\pi_{1}(\Sigma,x_{0}),G)/G of homomorphisms from the fundamental group \pi_{1}(\Sigma,x_{0}) into the group G , up to conjugacy. The definition of fundamental group involves the choice of a base point x_{0}\in\Sigma , but the quotient space no longer depends on this choice. The space \mathcal{M}_{G}(\Sigma) may also be regarded as the moduli space of flat bundles , \mathcal{M}_{G}(\Sigma)=\{\mbox{flat principal $G$-bundles}\ \ P\to\Sigma\}\big/\sim where two such bundles are identified if they are related by a G -equivariant diffeomorphism intertwining the flat connections and inducing the identity on the base. The correspondence is induced by the map taking the homomorphism \kappa\colon\pi_{1}(\Sigma,x_{0})\to G to the associated principal bundle P=\widetilde{\Sigma}\times_{\pi_{1}(\Sigma,x_{0})}G. In the opposite direction, given a flat G -bundle P\to\Sigma , one obtains \kappa as the parallel transport (holonomy) of the connection, after choice of trivialization P|_{x_{0}}\cong G . Different choices of trivialization at x_{0} change \kappa by G -conjugacy. Remark 2.1 . If the group G is connected and simply connected, then every principal G -bundle P\to\Sigma is isomorphic to the trivial one, and \mathcal{M}_{G}(\Sigma) is the moduli space of flat connections on the trivial bundle. Equivalently, it is the moduli space of Maurer-Cartan elements \mathcal{M}_{G}(\Sigma)=\Big\{A\in\Omega^{1}(\Sigma,\mathfrak{g})|\ {\mbox{d}}A+{\frac{1}{2}}[A,A]=0=0\Big\}/C^{\infty}(\Sigma,G). where the group C^{\infty}(\Sigma,G) acts by g\cdot A=\operatorname{Ad}_{g}(A)-g^{*}\theta^{R} . Here \theta^{R}\in\Omega^{1}(G,\mathfrak{g}) is the right-invariant Maurer-Cartan form. For general groups G , the space \mathcal{M}_{G}(\Sigma) may also include non-trivial flat bundles. One obtains a concrete description of the moduli space by choosing generators of the fundamental group. For a surface of genus \mathsf{g} , there are the standard generators \mathsf{a}_{1},\mathsf{b}_{1},\ldots,\mathsf{a}_{g},\mathsf{b}_{g} with the relation \mathsf{a}_{1}\mathsf{b}_{1}\mathsf{a}_{1}^{-1}\mathsf{b}_{1}^{-1}\cdots\mathsf{a}_{g}\mathsf{b}_{g}\mathsf{a}_{g}^{-1}\mathsf{b}_{g}^{-1}=1 . It corresponds to a gluing diagram, depicted below for genus \mathsf{g}=2 . In terms of these generators, the moduli space is identified with \mathcal{M}_{G}(\Sigma)=\Big\{(a_{1},b_{1},\ldots,a_{\mathsf{g}},b_{\mathsf{g}})\in G^{2\mathsf{g}}|\ \prod_{i=1}^{\mathsf{g}}a_{i}b_{i}a_{i}^{-1}b_{i}^{-1}=e\Big\}\big/G where we interpret a_{i},b_{i}\in G as the holonomies along \mathsf{a}_{i},\mathsf{b}_{i}\in\pi_{1}(\Sigma,x_{0}) . One may also consider other gluing diagrams presenting the surface, corresponding to ‘non-standard’ generators. For example, gluing opposite sides of an 8-gon gives a surface of genus \mathsf{g}=2 ; this gluing pattern corresponds to generators \mathsf{c}_{1},\ldots,\mathsf{c}_{4} with a relation \mathsf{c}_{1}\cdots\mathsf{c}_{4}\mathsf{c}_{1}^{-1}\cdots\mathsf{c}_{4}^{-1}=1 . The space \mathcal{M}_{G}(\Sigma) appears in many areas of mathematics and physics – far too many to properly review them here. Let us just point out two major directions: • For G=\operatorname{SU}(n) , and a given choice of complex structure on the surface, \mathcal{M}_{G}(\Sigma) it is the moduli space of stable holomorphic

Read the complete text