A Localization Argument for Characters of Reductive Lie Groups

Matvei Libine
DOI: https://doi.org/10.48550/arXiv.math/0206019
2003-05-20
Abstract:This article provides a geometric bridge between two entirely different character formulas for reductive Lie groups and answers the question posed by W.Schmid in [Sch]. A corresponding problem in the compact group setting was solved by N.Berline, E.Getzler and M.Vergne in [BGV] by an application of the theory of equivariant forms and particularly the fixed point integral localization formula. This article (besides its representation-theoretical significance) provides a whole family of examples where it is possible to localize integrals to fixed points with respect to an action of a NONcompact group. Moreover, a localization argument given here is not specific to the particular setting considered in this article and can be extended to a more general situation. There is a broadly accessible article [L] which explains how the argument works in the SL(2,R) case, where the key ideas are not obstructed by technical details and where it becomes clear how it extends to the general case.
Representation Theory,Differential Geometry
What problem does this paper attempt to address?
The problem that this paper attempts to solve is about the geometric connection between two completely different character formulas in reductive Lie groups. Specifically, by introducing a localization technique, the author realizes the localization of the integral to the zeros of the vector field generated by the infinitesimal action of $\mathfrak{g}_R$ on $X$. This technique not only provides a result in representation theory, but also provides a class of examples in which the integral can be localized to the fixed points of non - compact group actions. Moreover, the localization argument mentioned in the paper is not limited to a specific setting and can be extended to more general situations. ### Background and Motivation of the Paper 1. **Compact Group Case**: - For a connected compact group $K$, consider its maximal torus $T\subset K$. Let $\mathfrak{k}_R$ and $\mathfrak{t}_R$ be the Lie algebras of $K$ and $T$ respectively, and $\mathfrak{k}$ and $\mathfrak{t}$ be their complexified Lie algebras. For a finite - dimensional representation $\pi$ of $K$, define the character function $\Theta_\pi(x)=\text{tr}(\pi(x))$. - When $\pi$ is an irreducible representation, there are two completely different character formulas: the Weyl character formula and the Kirillov character formula. 2. **Non - Compact Group Case**: - Consider a connected linear reductive Lie group $G_R$ with Lie algebra $\mathfrak{g}_R$. Most of the representations of interest are infinite - dimensional. Consider the representations on complete, locally convex Hausdorff topological vector spaces and require that the action of $G_R$ is continuous. - For the maximal compact subgroup $K$ of $G_R$, the reasonable representation category includes admissible representations of finite length. Although the trace of a linear operator cannot be defined in an infinite - dimensional space, an $\text{Ad}(G_R)$ - invariant distribution $\theta_\pi$ can still be defined as a character function. ### Main Contributions of the Paper 1. **Localization Technique**: - A localization technique is introduced, which makes it possible to localize the integral at the zeros of the vector field generated by the infinitesimal action of $\mathfrak{g}_R$ on $X$. This not only provides a result in representation theory, but also provides a class of examples in which the integral can be localized to the fixed points of non - compact group actions. 2. **Equivalence of Character Formulas**: - The equivalence of the fixed - point character formula and the integral character formula in this type of representations is proved without taking a detour through representation theory. This equivalence can be directly seen through the localization of the character sheaf $F$. 3. **Generalization**: - The localization argument is not only applicable to a specific setting, but can also be extended to more general situations, providing a new tool for studying the integral localization under non - compact group actions. ### Conclusion By introducing a new localization technique, this paper successfully establishes the geometric connection between two different character formulas in reductive Lie groups, providing a new perspective and method for the integral localization under non - compact group actions. This result is not only of great significance in representation theory, but also provides new insights in geometry and topology.