ProbXiv
Explorersign in
Problem archiveProblem record

Statement

For a partition λ\lambda of nn, let CλC_\lambda be the unipotent conjugacy class of GL⁡n(C)\operatorname{GL}_n(\mathbb{C}) whose Jordan-block sizes are given by the dual partition λ′\lambda'. For a multi-partition μ=(μ1,…,μk)\bm{\mu}=(\mu^1,\dots,\mu^k) of nn, define

C‾μ:=GL⁡n2g×C‾μ1×⋯×C‾μk.\overline{C}_{\bm{\mu}}:=\operatorname{GL}_n^{2g}\times \overline{C}_{\mu^1}\times\cdots\times \overline{C}_{\mu^k}.

Fix a primitive nn-th root of unity ζ\zeta, and let Zμ{\mathcal{Z}}_{\bm{\mu}} be the space of tuples

(A1,…,Ag,B1,…,Bg,X1,…,Xk)∈C‾μ(A_1,\dots,A_g,B_1,\dots,B_g,X_1,\dots,X_k)\in\overline{C}_{\bm{\mu}}

satisfying

∏i=1g(Ai,Bi)∏j=1kXj=ζ⋅In,\prod_{i=1}^g(A_i,B_i)\prod_{j=1}^kX_j=\zeta\cdot I_n,

where (A,B)=ABA−1B−1(A,B)=ABA^{-1}B^{-1}. Put

Mμ:=Zμ/ ⁣/GL⁡n=Spec⁡(C[Zμ]GL⁡n).{\mathcal{M}}_{\bm{\mu}}:={\mathcal{Z}}_{\bm{\mu}}/\!/\operatorname{GL}_n=\operatorname{Spec}\left(\mathbb{C}[{\mathcal{Z}}_{\bm{\mu}}]^{\operatorname{GL}_n}\right).

Let PPc(Mμ,t):=∑sihcs,s;2s(Mμ)tsPP_c({\mathcal{M}}_{\bm{\mu}},t):=\sum_s ih_c^{s,s;2s}({\mathcal{M}}_{\bm{\mu}})t^s be the pure part of the compactly supported mixed Poincaré polynomial, and let Vμ(t)V_{\bm{\mu}}(t) and dμd_{\bm{\mu}} be as above. Letellier's conjecture.

Vμ(t)=t−dμ/2PPc(Mμ,t).V_{\bm{\mu}}(t)=t^{-d_{\bm{\mu}}/2}PP_c({\mathcal{M}}_{\bm{\mu}},t).

This conjecture proposes the interpretation of the polynomials Vμ(t)V_{\bm{\mu}}(t) as pure compactly supported intersection-cohomology polynomials of character varieties. It is attributed in the source to Letellier, and the status is not determined by the provided text.

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.