Proof of the $p$-adic Kazhdan-Lusztig hypothesis for $\mathrm{GL}(n)$
Kristaps John Balodis
TL;DR
This work proves the $p$-adic Kazhdan-Lusztig hypothesis for $GL_n(F)$ by fusing Zelinski–Langlands multisegment theory with Chriss–Ginzburg’s geometric framework on affine Hecke algebras. It establishes that standard representations $S_Q(a)$ are the Langlands-standard models, and that multiplicities $[S_Z(a):Z(b)]$ and $[S_Q(a):Q(b)]$ coincide via Aubert duality, allowing a reduction to simple inertial support and a geometric interpretation through Vogan varieties. The paper then aligns Zel’s geometric setup with CG’s, transferring results from the semi-simple simply connected setting to $GL_n(F)$ (via a PGL$_n$ bridge) and proving the proposition for simple support before assembling the general case. As a result, it provides a formal, comprehensive proof of the $p$-adic KL hypothesis for the general linear group over a $p$-adic field, clarifying gaps and offering a robust framework for potential generalizations using Solleveld’s graded Hecke algebras. The work also records several technical lemmas and proofs that experts had taken for granted, contributing to a rigorous, publishable foundation for these connections.
Abstract
In this article, we prove the $p$-adic Kazhdan-Lusztig hypothesis for $\mathrm{GL}_n(F)$. While the approach via graded affine Hecke algebras due to recent work of Solleveld leads to more general results, this article serves to completes and clarifies the approach via affine Hecke algebras of Chriss and Ginzburg. In particular, this article serves as an opportunity to articulate several results which are undoubtedly known to experts, but have not been formally recorded in the literature.
