Topological aspects of Chow quotients (Q814645)

The construction of moduli spaces is frequently done by expressing them as quotients of schemes by the action of reductive groups. The most known way to construct such quotients is by geometric invariant theory (GIT) by Mumford. These quotients depend on choices of linearized line bundles, and are not canonical. Besides this, the closures of the orbits parameterized by a GIT quotient almost always belong to different cohomology classes, and so make these quotients misbehaved compactifications. Despite the fact that the Hilbert quotient, derived from a Hilbert scheme, enjoys more functorial properties, the Chow quotient, as it parameterizes cycles, is more geometrically friendly and approachable. Thus in this article, the canonical quotient considered is the Chow quotient. The main purpose is to give some topological interpretations and characterization of the Chow quotient. This is to be done over the field of complex numbers and in the languages that are familiar to topologists and differential geometers. The author focuses on actions by a torus \(G\). A GIT quotient parameterizes orbits that are closed in the semistable locus. The Chow quotient parameterizes cycles of generic orbit closures and their toric degenerations which are certain sums of orbit closures of dimension \(\dim G\). These cycles are called Chow fibers. A main goal of the article is to answer when two arbitrary points belong to the same Chow fibre. The answer is a theorem telling that two points \(x\) and \(y\) of \(X\), with \(\dim G\cdot x=\dim G\cdot y=\dim G\), belong to the same Chow fiber if and only if \(x\) can be perturbed (to general position), translated along \(G\)-orbits (to positions close to \(y\)), and then specialized to the point \(y\). This is called the PTS-principle. The PTS-relation is computable and provides some much needed information on boundary cycles of the Chow quotient. In the introduction, these propositions on the Chow fibres are illustrated on an easy example. The example also illustrates a very general correspondence between GIT-quotients and symplectic structures, and the author proves that there is a correspondence between Chow quotients and symplectic structures as well. Let \(K\) be the compact part of the polar decomposition of \(G\). The moduli space \(\mathfrak{M}_{\gamma}\) of stable \(K\)-orbits with prescribed momentum charges \(\gamma\) is homeomorphic to the Chow quotient \(X//^{\text{ch}}G\), regardless of the choice of \(\gamma\). Under some assumptions, the Chow quotient is the least common refinement of all GIT quotients, in a strict sence. That is, \(X//^{\text{ch}}G\) is isomorphic to the limit quotient \(X//^{\text{lim}}G,\) the distinguished irreducible component of the inverse limit of all GIT quotients. Then the morphism from the Chow quotient to a GIT quotient, in terms of stable \(K\)-orbits, correspond to a map from \(\mathfrak{M}_{\gamma}\) to the symplectic quotient \(\Phi^{-1}(r)/K\) where \(r\) is the principal momentum charge in \(\gamma\). Every stable orbit contains a unique principal orbit, the one with the principal momentum charge, all the rest will be called ``bubble''orbits of the principal one. Then the map from \(\mathfrak{M}_{\gamma}\) to \(\Phi^{-1}(r)/K\) simply forgets all the bubble orbits and sends a stable orbit to its principal part. The space \(\mathfrak{M}_{\gamma}\) reflects the Hamiltonian aspects of the \(G\)-action on \(X\). The author relates this to another topological approach to the Chow quotient by introducing stable action manifolds. The author then proves that the moduli space \(\mathfrak{M}\) of stable action manifolds exists as a separated complex variety homeomorphic to the Chow quotient. In this article, a short treatment of the GIT quotient and the symplectic reduction is given. Also a very nice algebro-geometric treatment of the Chow quotient starting with the Chow variety is given together with its Chow family. Its relation with the limit quotient and its ample line bundles are extensively studied. This then ends with the PTS-principle and the proof of the main theorem given above. An example on point configurations on \(\mathbb{P}^n\) ends this part of the article. The rest of the article treats the symplectic approach of the Chow quotient, then the topological approach. This is a nice article that illustrates the existence and importance of quotients with, among other good properties, good compactifications.