Searcharxiv⌕ Search

arXiv · 2610.23737

Plancherel Decompositions and Distinctions for $\mathrm{Sp}_{2n}\backslash \mathrm{Gl}_{2n+1}$ and $P(\mathrm{Sp}_{2n})\backslash \mathrm{Gl}_{2n}$

Abstract

Let $F$ be a non-Archimedean local field of characteristic zero. We determine the Plancherel decompositions of $\mathrm{Sp}_{2n}(F)\backslash\mathrm{GL}_{2n+1}(F)$ and $P(\mathrm{Sp}_{2n})(F)\backslash\mathrm{GL}_{2n}(F)$, where $P(\mathrm{Sp}_{2n})$ denotes the mirabolic subgroup of $\mathrm{Sp}_{2n}$, and study smooth distinction for both spaces. We carry out the local unfolding procedure of arXiv:1203.0039, Section 9.5, constructing unitary isomorphisms with degenerate Whittaker models by successive Fourier transforms. The isomorphism for the even model is already stated there in Example 9.5.17. The resulting Plancherel formulas are direct integrals of normalized parabolic inductions from the tempered duals of $\mathrm{GL}_n(F)\times\mathrm{GL}_{n+1}(F)$ and $\mathrm{GL}_n(F)\times\mathrm{GL}_n(F)$, respectively, with product Whittaker-Plancherel measures. In the odd case, the parameter map agrees with the prediction of arXiv:2409.04677. Using closed parabolic orbits and local Rankin-Selberg integrals, we prove that all representations induced from these tempered data are distinguished for the corresponding model. We also prove that essentially square-integrable representations are not distinguished for the odd model, and that supercuspidal representations are not distinguished for the even model. For the even model with $n=1,2$, we construct distinguished square-integrable representations outside the Plancherel support. Thus, the smooth tempered distinguished spectrum can be strictly larger than the corresponding $L^2$-spectrum. For $\mathrm{GL}_5(F)$, however, we prove that an irreducible tempered representation is $\mathrm{Sp}_4(F)$-distinguished if and only if it is induced from tempered representations of $\mathrm{GL}_2(F)$ and $\mathrm{GL}_3(F)$.

Explore related subjects

Keep this discovery

Explore connections, maps & timelines

BibTeXRIS

Sagnik Mukherjee. 2026-10-16. Plancherel Decompositions and Distinctions for $\mathrm{Sp}_{2n}\backslash \mathrm{Gl}_{2n+1}$ and $P(\mathrm{Sp}_{2n})\backslash \mathrm{Gl}_{2n}$. https://arxiv.org/abs/2610.23737

Cite the original work for its findings. Save a collection to share your selection of sources.

KEEP EXPLORING

Related papers

Totally elliptic surface group representations

A surface group representation into a Lie group is called totally elliptic if every simple closed curve on the surface is mapped to an elliptic element of the target group. In this note, we characterize all totally elliptic surface group representations into $\mathrm{PSL}_2\mathbb{R}$ and $\mathrm{PSL}_2\mathbb{C}$ by showing that they are either representations into a compact subgroup or Deroin--Tholozan representations.

math.RT↗

Stable Brauer--Thrall II$'$ conjecture for finite-dimensional Jacobian algebras

We prove that finite-dimensional Jacobian algebras associated with non-degenerate quivers with potentials satisfy the stable Brauer--Thrall II$'$ conjecture. In particular, this implies that the brick Brauer--Thrall II$'$ conjecture (also known as the $τ$-Brauer--Thrall II$'$ conjecture) holds for finite-dimensional Jacobian algebras.

math.RT↗

Wide Regular Subalgebras of Symmetrizable Kac-Moody Algebras and an Extension of Schur's Lemma

The behavior of representations under restriction is a central theme in Lie theory. We study wide regular subalgebras of symmetrizable Kac--Moody algebras, extending work of Douglas and Repka on semisimple Lie algebras. A subalgebra is \emph{wide} if every irreducible integrable highest weight module remains indecomposable upon restriction. Let $\mathfrak{g}$ be a symmetrizable Kac--Moody algebra with Cartan subalgebra $\mathfrak{h}$, root system $Φ$, simple roots $Π$, and root space decomposition $\mathfrak{g}=\mathfrak{h}\oplus\bigoplus_{α\inΦ}\mathfrak{g}_α$. Denote by $Φ_{\operatorname{re}}$ the set of real roots. To a regular subalgebra $\mathfrak{s}$ normalized by $\mathfrak{h}$, we associate a subset $T\subseteq Φ$ by declaring $α\in T$ if $\mathfrak{s}\cap \mathfrak{g}_α\ne \{0\}$. One of our main results is an extension of Schur's lemma: if $\mathfrak{h}\subseteq \mathfrak{s}$ and the real closure of $(T\cup(-T))\cap Φ_{\operatorname{re}}$ contains $Π$, then $(\operatorname{End} V)^{\mathfrak{s}}=\mathbb{C}\operatorname{Id}_V$ for every irreducible integrable highest weight module $V$. As a consequence, this real-root closure condition yields a sufficient condition for wideness. In the affine case, we establish a necessary condition for wideness: if $\mathfrak{s}$ is wide, then the closure of $T\cup(-T)$ in $Φ$ is all of $Φ$, and this implication holds without assuming that $\mathfrak{h}\subseteq \mathfrak{s}$. A key ingredient is a structural result showing that closed subsets of affine root systems are closed under arbitrary finite root sums that remain roots.

math.RT↗