We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

math.RT

Change to browse by:

References & Citations

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Mathematics > Representation Theory

Title: Covers of reductive groups and functoriality

Authors: Tasho Kaletha
Abstract: For a quasi-split connected reductive group $G$ over a local field $F$ we define a compact abelian group $\tilde\pi_1(G)$ and an extension $1 \to \tilde\pi_1(G) \to G(F)_\infty \to G(F) \to 1$ of topological groups equipped with a splitting over $G_\textrm{sc}(F)$. Any character $x : \tilde\pi_1(G) \to \mu_n(\mathbb{C})$ leads to an $n$-fold cover $G(F)_x$ of $G(F)$ via pushout. We define an $L$-group $^LG_x$ for this cover that is generally a non-split extension of $\textrm{Gal}(F^s/F)$ by $\hat G$. We prove a refined local Langlands correspondence for $G(F)_x$, assuming it is known for connected reductive groups with the same adjoint group as $G$.
Motivation for this construction comes from considerations of Langlands' functoriality conjecture, where subgroups $\mathcal{H} \subset {^LG}$ of the $L$-group of $G$ arise that need not be $L$-groups of other reductive groups. If such a subgroup is full and intersects $\hat G$ in a connected reductive subgroup of maximal rank, we construct a natural triple $(H,x,\xi)$ consisting of a quasi-split connected reductive group $H$, a double cover $H(F)_x$, and an $L$-embedding $\xi : {^LH}_x \to {^LG}$ that is an isomorphism onto $\mathcal{H}$. We expect that genuine representations of $H(F)_x$ transfer functorially to representations of $G(F)$.
In the special case of endoscopy, we show that the construction of transfer factors simplifies when the natural double cover $H(F)_x$ of the endoscopic group is used. The transfer factor becomes the product of two natural invariants that do not depend on auxiliary choices. One of them is closely related to Kottwitz's work on transfer factors for Lie algebras. The other one is not specific to the case of endoscopy, and will likely play a role in general functoriality questions. Our work is motivated by work of Adams and Vogan over the real numbers.
Subjects: Representation Theory (math.RT); Number Theory (math.NT)
Cite as: arXiv:2209.14357 [math.RT]
  (or arXiv:2209.14357v1 [math.RT] for this version)

Submission history

From: Tasho Kaletha [view email]
[v1] Wed, 28 Sep 2022 18:39:42 GMT (65kb)

Link back to: arXiv, form interface, contact.