ProbXiv
Explorersign in
Problem archiveProblem record

Statement

Let GG be a quasi-split connected reductive group over a local or global field FF. For a parameter psi\mathfrak{psi}, let mathitPipsi(G)Amathit{Pi}_{\mathfrak{psi}}(G)^A and mathitPipsi(G)Dmathit{Pi}_{\mathfrak{psi}}(G)^D denote the Arthur packets defined using Artin's and Deligne's conventions, respectively; let w\mathfrak{w} be a Whittaker datum, Spsi\mathcal{S}_{\mathfrak{psi}} the component group, and widehatCwidehat C the relevant duality automorphism. Artin–Deligne duality conjecture. Locally,

mathitPipsi(G)D=(mathitPipsi(G)A)∨=mathitPipsi∨(G)A,mathit{Pi}_{\mathfrak{psi}}(G)^D=(mathit{Pi}_{\mathfrak{psi}}(G)^A)^\vee=mathit{Pi}_{\mathfrak{psi}^\vee}(G)^A,

and, for mathitpi\inmathitPipsi(G)Dmathit{pi}\inmathit{Pi}_{\mathfrak{psi}}(G)^D and x\inmathcalSpsix\inmathcal{S}_{\mathfrak{psi}},

⟨pi,x⟩w,psiD=⟨pi∨,x−1⟩w−1,psiA=⟨pi,C^−1(x)⟩w,psi∨A.\langle\mathit{pi},x\rangle^D_{\mathfrak{w},\mathfrak{psi}}=\langle\mathit{pi}^\vee,x^{-1}\rangle^A_{\mathfrak{w}^{-1},\mathfrak{psi}}=\langle\mathit{pi},\widehat C^{-1}(x)\rangle^A_{\mathfrak{w},\mathfrak{psi}^\vee}.

Globally,

Lpsi2(G)D=(Lpsi2(G)A)∨=Lpsi∨2(G)A,m(pi∨,psi∨)=m(pi,psi).L^2_{\mathfrak{psi}}(G)^D=(L^2_{\mathfrak{psi}}(G)^A)^\vee=L^2_{\mathfrak{psi}^\vee}(G)^A, \qquad m(\mathit{pi}^\vee,\mathfrak{psi}^\vee)=m(\mathit{pi},\mathfrak{psi}).

This conjecture predicts compatibility between Artin and Deligne normalizations, contragredient representations, dual parameters, packet pairings, and global multiplicities. The source gives no resolution status.

Record

Comments

Nothing has been published against this problem yet, and nothing has been said about it.

Sign in with an institutional address to take part in the discussion. Reading every thread stays open to everyone.

Sign in

Solve with an agent

Open the statement in a chat, with the problem and the ground rules already written into the prompt.

This opens a third-party site. Nothing is posted back to ProbXiv and nothing you write there is recorded here — what a model gives you is an attempt, which a person still has to check.