Stable map quotients (and orbifold log resolutions) of Richardson varieties
Allen Knutson · 2025
Read Stable map quotients (and orbifold log resolutions) of Richardson varieties by Allen Knutson (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-2505-09905.txt.
Introduction
Allen Knutson smooths the singularities of Richardson varieties by way of stable maps. A Richardson variety is the overlap of two Schubert varieties inside a flag space, a shape with corners where geometry snags. His paper builds a standard resolution from spaces of symmetric stable maps. The price is orbifold structure: the smooth stand-in carries finite symmetry scars. Its nodal-curve edge forms a simple normal crossings divisor, thought to be anticanonical.
The boundary's combinatorics mirror the order complex of the open Bruhat interval, known to be a sphere or a ball. In the Grassmannian case the resolution is a GKM space, whose fixed points are indexed by rim-hook tableaux. The field gains a fix with readable combinatorics.
Catalog record
| CREATOR | Allen Knutson |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 131 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2505-09905
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Stable map quotients (and orbifold log resolutions) of Richardson varieties Allen Knutson Abstract . Let X λ μ := X λ ∩ X μ ⊆ G / P X_{\lambda}^{\mu}:=X_{\lambda}\cap X^{\mu}\subseteq G/P be a Richardson variety in a generalized partial flag manifold. We use equivariant stable map spaces to define a canonical resolution X λ μ ~ \widetilde{X_{\lambda}^{\mu}} of singularities, albeit obtaining an orbifold not a manifold. The “nodal curves” boundary is an (orbifold) simple normal crossings divisor, and is conjecturally anticanonical. Its dual simplicial complex is the order complex of the open Bruhat interval ( λ , μ ) ⊆ W / W P (\lambda,\mu)\subseteq W/W_{P} , shown in [Björner-Wachs ’82] to be a sphere or ball. In the case of G / P G/P a Grassmannian, the resolution X λ μ ~ \widetilde{X_{\lambda}^{\mu}} is a GKM space, whose T T -fixed points are indexed by rim-hook tableaux. 1. Introduction Fix a pinning (G,B,B_{-},T,W) of a connected reductive complex algebraic group, and a parabolic subgroup P\leq G , P\geq B . For \lambda,\mu\in W/W_{P} , define the closed subvarieties of G/P X_{\lambda}:=\overline{B_{-}\,\lambda P}/P,\quad X^{\mu}:=\overline{B\mu P}/P,\quad X_{\lambda}^{\mu}:=X_{\lambda}\cap X^{\mu} and call them a Schubert variety , opposite Schubert variety , and Richardson variety respectively. They are known to be normal with rational singularities. One way (from [ Bri05 ] ; see also [ E16 ] ) to resolve those singularities involves choosing Bott-Samelson resolutions of X_{\lambda} and X^{\mu} , whose fiber product is a brick manifold resolution. In particular, this involves choosing reduced words for some elements of W . In this paper we produce another resolution \widetilde{X_{\lambda}^{\mu}} , with no choices involved, suffering the slight drawback that the resolution is not by a manifold but by a smooth orbifold (or, Deligne-Mumford stack). It is a couple of steps removed from a space of genus zero stable maps, and we relate it in § 3 to Chow quotients. 1 1 1 Several years after we embarked on this study [ Kn20 ] we learned that Mateusz Michalek, Leonid Monin, and Botong Wang were investigating the corresponding Chow quotients. In upcoming work, they make fascinating use of them to define a “kernel” in the poset W/W_{P} (see [ F24 ] for the definition), where the palindromicity of the rational Betti numbers of \widetilde{X_{\lambda}^{\mu}} plays a key role. If P=B , where we instead use Roman letters u,v\in W , the boundary \partial X_{u}^{v}:=\bigcup_{u"\gtrdot u}X_{u^{\prime}}^{v}\cup\bigcup_{v"\lessdot v}X_{u}^{v^{\prime}} (where \gtrdot,\lessdot indicate covering relations) is an anticanonical divisor [ KnLamSp14 , lemma 5.4] . Some properties shared by both the brick manifold resolution and \widetilde{X_{u}^{v}} are that • while not “strict resolutions” (ones that blow up no smooth points), they only blow up points on that boundary, • the preimage of \partial X_{u}^{v} is a simple normal crossings divisor (albeit orbifold, in our case), and • its “dual simplicial complex” (in the sense of [ KoX16 , Pay13 ] ; see § 4 ) to that sncd is homeomorphic to a sphere. The dual simplicial sphere to \partial\widetilde{X_{u}^{v}} is the suspension of the order complex of the open interval (u,v)\subseteq W in Bruhat order, shown combinatorially to be a sphere in [ BjWa82 ] . Analogously, the dual simplicial sphere to the boundary of a brick manifold was shown in [ E16 ] to be a spherical subword complex. Most of our results concern a simpler space we introduce, the “stable map quotient” X//_{StMap}\,S of a variety X by the action of a circle S\mathrel{\cong}\mathbb{C}^{\times} . The resolution \widetilde{X_{\lambda}^{\mu}} then arises via a “graph space” construction (X_{\lambda}^{\mu}\times{\mathbb{P}}^{1})//_{StMap}\,S_{\Delta} using the diagonal action, where the action of S on X_{\lambda}^{\mu} comes from S\leq T being the first regular dominant coweight {\check{\rho}} . After defining these spaces in § 2 , our first main result is the orbifold-smoothness of \widetilde{X_{\lambda}^{\mu}} . We comment in § 3 on the close relation of the stable map quotient to the Chow quotient X_{\lambda}^{\mu}//_{Chow}\,{\check{\rho}} . In § 4 we compute the stratification of X_{\lambda}^{\mu}//_{StMap}\,{\check{\rho}} induced from the sncd. The resolution \widetilde{X_{\lambda}^{\mu}} bears an action of the torus B_{-}\cap B=T , and we study this action in § 5 . While \widetilde{X_{\lambda}^{\mu}} has isolated T -fixed points and every point stratum is of course T -fixed, there are additional T -fixed points making it tricky to compute Betti numbers, which we do in the Grassmannian and GL(n)/B cases. In § 6 we further study the Grassmannian case, in which \widetilde{X_{\lambda}^{\mu}} has not only isolated fixed points but isolated fixed curves (the “Goresky-Kottwitz-MacPherson property”). In § 7 we define some T -equivariant classes on the T -fixed points, satisfying these GKM conditions, which seem likely to be the first Chern classes of T -equivariant (orbi-)line bundles. In § 8 we address the sense in which the boundary divisor \partial\widetilde{X_{\lambda}^{\mu}} is anticanonical, shedding new light on the difference between the ball and sphere cases of [ BjWa82 ] . 1.1. Acknowledgments I express my gratitude to Daniel Halpern-Leistner, Adeel Khan, Angelo Vistoli, and Rachel Webb for helping me with the basics of stacks and stable curves; any remaining mistakes, or even just gauchely stated arguments, are mine. I thank Christian Gaetz and Shiliang Gao for directing me to Dyer’s result, and Brian Hwang for directing me to Chen and Satriano’s result. Thanks also to Mateusz Michalek and Botong Wang for generously explaining their interests in and applications of the Chow quotients, and to the organizers of the Workshop on Combinatorics of Enumerative Geometry at the Insitute for Advanced Study in February 2025 for serendipitously bringing us together. 2. Stable map quotients Our reference for stable map spaces is [ FuPan97 ] . 2.1. Stable map spaces Given a curve class \beta\in A_{1}(M) in the first Chow group of a projective variety M , one defines the space StMap_{\beta}({\mathbb{P}}^{1},M) of genus zero stable maps \gamma\colon\Sigma\to M (with no “marked points”), where \Sigma is an at-worst-nodal (and probably reducible) connected curve of genus zero (so, a tree of {\mathbb{P}}^{1} s) and \gamma_{*}([\Sigma])=\beta . In addition, the group of automorphisms of \gamma (maps \Sigma\to\Sigma inducing a commuting triangle) is required to be finite; equivalently, each component of \Sigma that \gamma collapses to a point in M should meet at least three other components of \Sigma . So far this defines a set \{[\Sigma,\gamma]\} of isomorphism classes [\Sigma,\gamma] of pairs (\Sigma,\gamma) ; there is more work involved to define a moduli space , and the finite groups of automorphisms mean this space will generally be a Deligne-Mumford stack or “orbifold” in the differential-topological language. Theorem 1 . [ FuPan97 , part of Theorem 2] If M is a generalized flag variety G/P (or more generally is “convex”), then StMap_{\beta}({\mathbb{P}}^{1},M) can be given the structure of a smooth orbifold. 2 2 2 To be precise, in [ FuPan97 ] they punt discussion of stacks. In their §4 they construct their moduli spaces by gluing together quotients of manifolds by finite groups, leaving the reader to check that the stabilizer groups match across the gluing. At each node p of a genus zero stable curve \Sigma , the two components C,D of \Sigma that meet at p are distinct (i.e. unlike on a nodal elliptic curve). We relate the two tangent lines T_{p}C