The action homomorphism, quasimorphisms and moment maps on the space of compatible almost complex structures (Q2448614)

In this paper, Weinstein's action homomorphism (see [\textit{A. Weinstein}, Math. Z. 201, No. 1, 75--82 (1989; Zbl 0644.57024)]) is extended to Hamiltonian actions with equivariant moment maps of Lie groups on symplectic manifolds, and then, under certain conditions, to a quasimorphism on the universal cover of the group. The author establishes a general principle for building bounded 2-cocycles, in terms of the action of a group \(\mathcal{G}\) on a space \(\mathfrak{X}\), using the integration on geodesic simplices. Then he gives a general method for constructing primitives to such cocycles in the group cohomology and primitives to wit-quasimorphisms-functions which satisfy the homomorphism property up to a uniform bounded error. The main ingredients of the construction are the notion of equivariant moment map for the Hamiltonian action of a group \(\mathcal{G}\) on the space \(\mathfrak{X}\) with a symplectic form \(\Omega\) and the notion of Weinstein's action homomorphism extended to general Hamiltonian actions with equivariant moment maps. The author formulates a general principle for constructing quasimorphisms, a principle which involves groups acting well enough on spaces of sufficiently negative curvature. Then, he proposes some ``negative enough curvature''-symplectic manifolds \((\mathfrak{X},\Omega)\) with bounded Gromov norm of \(\Omega\), by introducing the Domic-Toledo spaces. For ``act well enough'' -- by ``isometries'' with an equivariant moment map -- the author defines a Hamiltonian-Hermitian group \(\mathcal{G}\) as being a Lie group acting on a Domic-Toledo space \((\mathfrak{X},\Omega,\mathcal{K})\) and preserving \(\mathcal{K}\) and \(\Omega\), with an equivariant moment map. Theorem 1 asserts that on the universal cover of any Hamiltonian-Hermitian group \(\mathcal{G}\) acting with an equivariant moment map on the corresponding Domic-Toledo space \((\mathfrak{X},\Omega,\mathcal{K})\), there exists a real-valued quasimorphism, whose homogenization does not depend on the basepoint. Next, the author studies the invariance of the quasimorphism under conjugation with respect to a suitable normal extension of a given action of a Hamiltonian-Hermitian group on a Domic-Toledo space (see Proposition 1.5.1). In the Donaldson-Fujiki framework (where the scalar curvature is seen as a moment map for the action of the Hamiltonian group on the space of compatible almost complex structures), new quasimorphisms are constructed on the Hamiltonian group (and on its universal cover) of an arbitrary symplectic manifold of finite volume. In this sense, the author proves that there exists a nontrivial homogeneous quasimorphism on the universal cover \(\widetilde{\mathcal{G}}\) of the group \(\mathcal{G}\) of Hamiltonian diffeomorphisms of an arbitrary closed symplectic manifold (Corollary 1). The restriction of this quasimorphism to the fundamental group of \(\widetilde{\mathcal{G}}\) is equal (by Theorem 2) to a homomorphism \(I_{c_1}:\pi_1(\widetilde{\mathcal{G}})\rightarrow \mathbb{R}\), given by \[ I_{c_1}(\gamma)=\int_{P_\gamma}c_1^Vu^n, \] where \(\gamma\) is a loop in \(\mathcal{G}\) based at Id, \(P_\gamma\) is the Hamiltonian fibration corresponding to \(\gamma\), \(c_1^V\) is the vertical Chern class of a Hamiltonian fiber bundle \(P\) over the 2-sphere, and \(u\in H^2(P,\mathbb{R})\) is a natural characteristic class of the fiber bundle \(P\). In Theorem 3, the author shows that the homogeneous quasimorphism from Theorem 1 is non-trivial, by computing its local type. Another interesting result of the paper is Proposition 1.9.1 on the reconstruction of the Guichardet-Wigner quasimorphisms via moment maps. Theorem 4 states the proportionality between the general quasimorphism from Corollary 1 and the Py quasimorphism (see [\textit{P. Py}, Ann. Sci. Éc. Norm. Supér. (4) 39, No. 1, 177--195 (2006; Zbl 1098.57014)]) on a monotone symplectic manifold \((M,\omega)\), i.e., with \(c_1(TM,\omega)=k[\omega]\) for \(k\neq 0.\) Moreover, in the case where \(c_1(TM,\omega)=0,\) the general quasimorphism is proportional to the Entov quasimorphism from [\textit{M. Entov}, Comment. Math. Helv. 79, No. 1, 58--104 (2004; Zbl 1048.53056)].