Generalizations of the Springer correspondence and cuspidal Langlands parameters

TL;DR

This paper develops a new framework for understanding the local Langlands correspondence for reductive p-adic groups, introducing cuspidality notions, a support map, and revealing a disjoint union structure in enhanced L-parameters, extending Lusztig's Springer correspondence.

Contribution

It introduces a notion of cuspidality for enhanced Langlands parameters and extends Lusztig's Springer correspondence to disconnected complex reductive groups, linking these to the local Langlands program.

Findings

Defined cuspidality for enhanced L-parameters.

Established a cuspidal support map and Bernstein components for L-parameters.

Revealed a disjoint union of twisted extended quotients structure in enhanced L-parameters.

Abstract

Let H be any reductive p-adic group. We introduce a notion of cuspidality for enhanced Langlands parameters for H, which conjecturally puts supercuspidal H-representations in bijection with such L-parameters. We also define a cuspidal support map and Bernstein components for enhanced L-parameters, in analogy with Bernstein's theory of representations of p-adic groups. We check that for several well-known reductive groups these analogies are actually precise. Furthermore we reveal a new structure in the space of enhanced L-parameters for H, that of a disjoint union of twisted extended quotients. This is an analogue of the ABPS conjecture (about irreducible H-representations) on the Galois side of the local Langlands correspondence. Only, on the Galois side it is no longer conjectural. These results will be useful to reduce the problem of finding a local Langlands correspondence for H-representations to the corresponding problem for supercuspidal representations of Levi subgroups of H. The main machinery behind this comes from perverse sheaves on algebraic groups. We extend Lusztig's generalized Springer correspondence to disconnected complex reductive groups G. It provides a bijection between, on the one hand, pairs consisting of a unipotent element u in G and an irreducible representation of the component group of the centralizer of u in G, and, on the other hand, irreducible representations of a set of twisted group algebras of certain finite groups. Each of these twisted group algebras contains the group algebra of a Weyl group, which comes from the neutral component of G. In 2025 an erratum was added, to repair Theorem 3.1.a.