On the minimal parabolic induction

Xinyu Li · 2024

Read On the minimal parabolic induction by Xinyu Li (2024) 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-2412-19008.txt.

Introduction

Xinyu Li wrote a 2024 math paper defining a new induction functor for Kac-Moody algebras, the minimal parabolic induction. The idea is modeled on a classic proof of the Jantzen conjectures. Li sets out basic properties of the functor. Two uses follow: computing first extension groups between simple highest weight modules, and studying what kills certain simple modules in the weight module category.

Catalog record

CREATORXinyu Li
YEAR2024
COLLECTIONPAPERS
FORMATTXT
FILE SIZE86 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2412-19008

Browse the full library · About the Tower of Babel

Full text

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

On the minimal parabolic induction Xinyu Li Abstract Motivated by Beilinson–Bernstein’s proof of the Jantzen conjectures [ BB93 ] , we define the minimal parabolic induction functor for Kac–Moody algebras, and establish some basic properties. As applications of the formal theory, we examine first extension groups between simple highest weight modules in the category of weight modules, and analyze the annihilators of some simple highest weight modules. 0 Introduction 0.1 Motivation Let \mathfrak{g} be a Kac–Moody algebra. For two weights \lambda,\mu of \mathfrak{g} , it is a basic problem in representation theory to calculate the first extension groups between L(\lambda) and L(\mu) in the category of \mathfrak{g} -weight modules (or the BGG category \mathcal{O} ). In particular, such problems arise in recent studies on simple affine vertex algebras (for instance, [ KR22 ] and [ ACK23 ] ). However, the explicit result is not easy to calculate in general. Even for \mathfrak{g} being a finite dimensional semisimple Lie algebra, the answer depends on the Jantzen conjecture (cf. [ Hum08 ] Chapter 8.15.), which is a deep result in representation theory. In the finite case, the Jantzen conjecture was proved by Beilinson–Bernstein [ BB93 ] , using their localisation theorem [ BB81 ] , the weight filtration of \ell -adic mixed perverse sheaves [ BBD82 ] , and the monodromy weight filtration of nearby cycles [ Del80 ] . We do not have a Beilinson–Bernstein type localisation theorem for a general Kac–Moody algebra \mathfrak{g} . Therefore, some more tools need to be developed to calculate the extension groups. Instead of trying to obtain a direct formula, our idea is to reduce the calculation to some Levi subalgebra (which is usually of finite type, and hence we can apply the known results for finite dimensional reductive Lie algebras). This leads us to the construction of minimal parabolic induction . Let us briefly describe our main constructions and main results below. 0.2 Main results Let \mathfrak{g} be a Kac–Moody algebra with a fixed choice \Pi of simple roots. For any subset \Xi\subset\Pi , we have the corresponding standard (resp. opposite) parabolic subalgebra \mathfrak{p}^{+}_{\Xi} (resp. \mathfrak{p}^{-}_{\Xi} ), and the associated Levi subalgebra \mathfrak{l}_{\Xi} . The parabolic induction functor \operatorname{Ind}_{\Xi,!}\colon\mathfrak{l}_{\Xi}\text{-}\mathsf{Mod}\to\mathfrak{g}\text{-}\mathsf{Mod},N\mapsto U\mathfrak{g}\otimes_{U\mathfrak{p}^{+}_{\Xi}}N, which maps \mathfrak{l}_{\Xi} -Verma modules to \mathfrak{g} -Verma modules, is intensively studied in representation theory. Here we inflate an \mathfrak{l}_{\Xi} -module N to a \mathfrak{p}^{+}_{\Xi} -module through the projection \mathfrak{p}^{+}_{\Xi}\twoheadrightarrow\mathfrak{l}_{\Xi} . Equally important is the parabolic coinduction functor \operatorname{Ind}_{\Xi,*}\colon\mathfrak{l}_{\Xi}\text{-}\mathsf{Mod}\to\mathfrak{g}\text{-}\mathsf{Mod},N\mapsto\operatorname{Hom}_{U\mathfrak{p}^{-}_{\Xi}}(U\mathfrak{g},N), which maps completed \mathfrak{l}_{\Xi} -coVerma modules to completed \mathfrak{g} -coVerma modules. Here we inflate an \mathfrak{l}_{\Xi} -module N to a \mathfrak{p}^{-}_{\Xi} -module through the projection \mathfrak{p}^{-}_{\Xi}\twoheadrightarrow\mathfrak{l}_{\Xi} . In the other direction, there is the parabolic restriction functor \operatorname{Res}_{\Xi}^{!}\colon\mathfrak{g}\text{-}\mathsf{Mod}\to\mathfrak{l}_{\Xi}\text{-}\mathsf{Mod},M\mapsto\operatorname{Hom}_{U\mathfrak{u}^{+}_{\Xi}}(\mathbb{C},M), and the parabolic corestriction functor \operatorname{Res}_{\Xi}^{*}\colon\mathfrak{g}\text{-}\mathsf{Mod}\to\mathfrak{l}_{\Xi}\text{-}\mathsf{Mod},M\mapsto\mathbb{C}\otimes_{U\mathfrak{u}^{-}_{\Xi}}M, where \mathfrak{u}^{\pm}_{\Xi} is the nilradical of \mathfrak{p}^{\pm}_{\Xi} . We have adjoint pairs of functors (\operatorname{Ind}_{\Xi,!},\operatorname{Res}_{\Xi}^{!}),(\operatorname{Res}_{\Xi}^{*},\operatorname{Ind}_{\Xi,*}) . Moreover, \operatorname{Res}_{\Xi}^{!}\circ\operatorname{Ind}_{\Xi,*}=\operatorname{Res}_{\Xi}^{*}\circ\operatorname{Ind}_{\Xi,!}=\operatorname{id}. This induces a natural transformation (see Definition 2.3 ) \operatorname{Ind}_{\Xi,!}\to\operatorname{Ind}_{\Xi,*}. Let us denote by \operatorname{Ind}_{\Xi,!*} the image of the above transformation. We call \operatorname{Ind}_{\Xi,!*} the minimal parabolic induction , or the intermediate parabolic induction . In this paper, we establish some basic properties of minimal parabolic induction. For example, like the construction of IC sheaves, one may expect that \operatorname{Ind}_{\Xi,!*} sends “good” (lisse) simple objects to simple objects. We confirm this expectation in Proposition 3.7 by showing that \operatorname{Ind}_{\Xi,!*} maps a simple weight module to a simple weight module (see Section 3 for the definition of weight modules and the category \mathfrak{g}\text{-}\mathsf{wtMod} ). Main Theorem 1 . For any simple \mathfrak{l}_{\Xi} -weight module N , \operatorname{Ind}_{\Xi,!*}(N) is a simple \mathfrak{g} -weight module. We exhibit two applications of the general theory in Section 4 . First, we explore the behavior of first extension groups between some simple highest weight modules under minimal parabolic induction. Under some rather technical conditions on two weights \lambda,\mu , we are able to prove that there is an isomorphism (Proposition 4.4 ) \operatorname{Ext}^{1}_{\mathfrak{g}\text{-}\mathsf{wtMod}}(L(\lambda),L(\mu))=\operatorname{Ext}^{1}_{\mathfrak{l}_{\Xi}\text{-}\mathsf{wtMod}}(L_{\Xi}(\lambda),L_{\Xi}(\mu)). Main Theorem 2 . Let \mu,\lambda be two weights of \mathfrak{g} such that \mu-\lambda\in\mathbb{Z}\Xi , then we always have an inclusion \operatorname{Ext}^{1}_{\mathfrak{g}\text{-}\mathsf{wtMod}}(L(\lambda),L(\mu))\hookrightarrow\operatorname{Ext}^{1}_{\mathfrak{l}_{\Xi}\text{-}\mathsf{wtMod}}(L_{\Xi}(\lambda),L_{\Xi}(\mu)). Moreover, suppose that \mu-\lambda\notin\mathbb{Z}_{\geq 0}\Xi and \lambda is \Xi -joyful (a technical notion introduced in Definition 4.1 ), then the above inclusion is an isomorphism. Another application is about annihilators. Assume \mathfrak{l}_{\Xi} is a finite dimensional reductive Lie algebra, then we are able to prove that \operatorname{Ann}_{U\mathfrak{g}}(\operatorname{Ind}_{\Xi,!*}(M_{\Xi}(\lambda)))=\operatorname{Ann}_{U\mathfrak{g}}(\operatorname{Ind}_{\Xi,!*}(M_{\Xi}(w\cdot\lambda))) for \lambda being a \rho -anti-dominant integral weight for \mathfrak{l}_{\Xi} and w\in W_{\Xi} , the Weyl group of \mathfrak{l}_{\Xi} . As a corollary, we show that under the same assumption, \operatorname{Ann}_{U\mathfrak{g}}(L(\lambda))\subset\operatorname{Ann}_{U\mathfrak{g}}(L(w\cdot\lambda)) . Main Theorem 3 . Suppose \Xi is of finite type, i.e. \mathfrak{l}_{\Xi} is a finite dimensional reductive Lie algebra. Let \lambda\in\mathfrak{h}^{*} be a weight such that \langle\lambda+\rho,\alpha^{\vee}\rangle\in\mathbb{Z}_{\leq 0} for any \alpha^{\vee}\in\Xi^{\vee} , then we have \operatorname{Ann}_{U\mathfrak{g}}(L(\lambda))=\operatorname{Ann}_{U\mathfrak{g}}(\operatorname{Ind}_{\Xi,!*}(M_{\Xi}(w\cdot\lambda)))\subset\operatorname{Ann}_{U\mathfrak{g}}(L(w\cdot\lambda)) for any w\in W_{\Xi} , the Weyl group of \mathfrak{l}_{\Xi} . To the knowledge of the author, this result is new for arbitrary Kac--Moody algebra. 1 1 1 When \mathfrak{g} is a finite dimensional semisimple Lie algebra, this result was proved by Vogan [ Vog80 ] . When \mathfrak{g} is an affine Kac–Moody algebra, this result was claimed by Dhillon, based on his work [ CD21 ] . Both of them used Harish-Chandra bimodules, whose general theory has not been developed for arbitrary Kac–Moody algebras yet. 0.3 Convention Throughout the paper, we work over an algebraically closed field \mathbb{C} of characteristic 0 . Symbols \otimes and \operatorname{Hom} without subscripts mean the corresponding operations in the category of \mathbb{C} -vector spaces, that is, \otimes_{\mathbb{C}} and \operatorname{Hom}_{\mathbb{C}} . For a ring (or a Lie algebra) \Lambda , we use \Lambda\text{-}\mathsf{Mod} to denote the category of left \Lambda -modules. By a module, we mean a left module. 0.4 Acknowledgement The author would like to express his deepest gratitude to his advisors, Prof. Wenbin Yan and Prof. Peng Shan, without whose help the current work would have never been done. He especially thanks Peng Shan for pointing out many inaccuracies, due to the author’s arrogance and over-optimism, in an earlier draft of the paper. There is a similar construction in the theory of vertex algebras, namely the Zhu’s induction. The author thanks Tomoyuki Arakawa, Thomas Creutzig, and Yongchang Zhu for correspondences on this subject. The author thanks Gurbir Dhillon for teaching him parabolic induction in categorical representation theory, Dingxin Zhang for teaching him \ell -adic mixed perverse sheaves, and Qixian Zhao for teaching him Duflo’s theorem. 1 Kac–Moody algebra setup 1.1 Parabolic type Let \mathfrak{g} be a Kac–Moody algebra associated with the triple (\mathfrak{h},\Pi,\Pi^{\vee}) , where \mathfrak{h} is a fixed Cartan subalgebra, \Pi\subset\mathfrak{h}^{*} (resp. \Pi^{\vee}\subset\mathfrak{h} ) is the collection of simple roots (resp. coroots). Denoted by \Delta the set of roots of \mathfrak{g} , \Delta^{+} (resp. \Delta^{-} ) the set of positive (resp. negative) roots of \mathfrak{g} , we have the root decomposition \mathfrak{g}=\mathfrak{h}\oplus\bigoplus_{\alpha\in\Delta}\mathfrak{g}_{\alpha}=\mathfrak{n}^{-}\oplus\mathfrak{h}\oplus\mathfrak{n}^{+}, where \mathfrak{n}^{+}=\bigoplus_{\alpha\in\Delta^{+}}\mathfrak{g}_{\alpha},\mathfrak{n}^{-}=\bigoplus_{\alpha\in\Delta^{-}}\mathfrak{g}_{\alpha}. Any subset \Xi of \Pi (together with the corresponding subset \Xi^{\vee} of \Pi^{\vee} ) defines a parabolic type . More precisely, let \Delta_{\Xi}=\Delta\cap\mathbb{Z}\Xi,\Delta^{\pm}_{\Xi}=\Delta_{\Xi}\cap\Delta^{\pm} . Then we have the standard (opposite) parabolic subalgebra of type \Xi \mathfrak{p}^{+}_{\Xi}=\bigoplus_{\alpha\in\Delta^{-}_{\Xi}}\mathfrak{g}_{\alpha}\oplus\mathfrak{h}\oplus\mathfrak{n}^{+},\mathfrak{p}^{-}_{\Xi}=\bigoplus_{\alpha\in\Delta^{+}_{\Xi}}\mathfrak{g}_{\alpha}\oplus\mathfrak{h}\oplus\mathfrak{n}^{-}, the (opposite) nilradical \mathfrak{u}^{\pm}_{\Xi}=\bigoplus_{\alpha\in\Delta^{\pm}\backslash\Delta^{\pm}_{\Xi}}\mathfrak{g}_{\alpha}, and the Levi \mathfrak{l}_{\Xi}=\mathfrak{h}\oplus\bigoplus_{\alpha\in\Delta_{\Xi}}\mathfrak{g}_{\alpha}. 1.2 Transpose anti-involution Let us name the simple roots of \mathfrak{g} by \alpha_{1},\cdots,\alpha_{n} . We fix the Chevalley generators e_{i}\in\mathfrak{g}_{\alpha_{i}} , f_{i}\in\mathfrak{g}_{-\alpha_{i}} . From the structure theory of Kac–Moody algebra (the Serre relations),

Read the complete text