Skip to content

Documentary state: EDITORIALLY_ADMITTED. The manuscript has a protected editorial admission, not a theorem certification.

OAI-26 family 069: Global quantum geometric Langlands at irrational level

Editorial status: DRAFT_UNACCEPTED pending protected admission. Mathematical status: NO_CERTIFICATION. This source-pinned account describes a manuscript claim; it does not independently establish the asserted equivalence.

The question in plain language

The geometric Langlands programme relates mathematical descriptions of bundles on a complex curve to descriptions associated with a dual algebraic group. Its quantum deformation asks whether suitable categories of twisted differential equations for a group and its Langlands dual agree after an inverse transformation of the twisting parameter. A category here records both objects and structure-preserving maps between them. An equivalence of categories is stronger than a mere correspondence between selected examples: it must preserve the full mathematical structure.

A small example

For a simply laced group, the lacing number is one, so the shifted parameter in the manuscript is sent from \(c\) to \(-1/c\). For a nonsimply laced root system the lacing number \(r\) is two or three, and the partner parameter is \(-1/(rc)\). The hypothesis \(c\in\mathbb C\setminus\mathbb Q\) excludes rational values but includes non-real complex values. These expressions describe the manuscript's parameter convention, not a numerical algorithm for constructing the correspondence.

What is known

The October 4, 2026 manuscript Global quantum geometric Langlands at irrational level states an unramified de Rham quantum geometric Langlands equivalence for every connected simple complex algebraic group \(G\), every smooth projective connected complex curve \(X\), and every shifted level \(c\in\mathbb C\setminus\mathbb Q\). Its main theorem asserts an equivalence of presentable complex-linear derived DG categories

\[ D_c(\mathrm{Bun}_G(X)) \simeq D_{-1/(rc)}(\mathrm{Bun}_{G^\vee}(X)). \]

where \(G^\vee\) is formed using the dual full root datum and \(r\) is the lacing number. The source specifies determinant-line twisting, uses the given global forms rather than quietly passing to adjoint or simply connected forms, and includes all connected components of the bundle stacks. A theta characteristic and dual pinnings enter the normalization of the construction.

This is a claimed equivalence of full twisted D-module categories, not a theorem merely about their finite-dimensional objects, a comparison of Grothendieck groups, or a Betti equivalence. The claimed range is irrational shifted levels. It does not include a statement covering all rational or critical levels.

The argument described by the manuscript

The manuscript's introduction presents two principal comparison mechanisms. First, an irrational-level Whittaker–Poincaré or Plancherel identity \(J_sR_s\to\mathrm{Id}\) is claimed on the entire twisted D-module category. Its proof invokes a finite Whittaker test, a filtered deformation argument and uniform weight bounds to transfer a vanishing statement from the ordinary-level setting. Second, local Kac–Moody and Whittaker pairings are compared over families of marked points, including collision diagonals. The source proposes extension through collision strata, reconstruction of maps of hearts, and derived realization.

The final section uses arbitrary-level localization, a global unipotent-period/Ran comparison and two Whittaker identities to construct inverse functors. The central-torsor and component compatibility is part of the claimed construction. These are source-reported proof steps, not independently replayed GCL lemmas.

GCL verification state

Source lock: openai/math@1581d4cbdd69a0faf34b2089614fe2a814316a6a. Primary PDF path: preprints/Global-quantum-geometric-Langlands-at-irrational-level-October-4-2026/quantum-langlands.pdf, Git blob cfae27b2f72fc043c0f4358d75fbc9a8d91790b3. The PDF's exact binary identity was obtained from the pinned GitHub source; the accompanying pinned build/main.tex, introduction, Plancherel, local-comparison and equivalence section sources were inspected. This verifies wording and scope against source text, not the mathematical validity of the manuscript.

No independent replay of the categorical localizations, finiteness assertions, weight argument, collision extension, global period comparison, or inverse-functor coherence has been completed here. No formalization, Comparator result or MATHCERT certification is asserted.

What remains

A substantive mathematical review must check the cited localization and geometric Langlands dependencies at the exact hypotheses; the determinant and shifted-level sign conventions; the uniform weight bound and colimit passage in the Whittaker test; the claimed conservativity at every irrational complex parameter; extension over collision diagonals; the Ran period comparison; and naturality and two-sided inverse identities for the final functors. It must separately verify central-torsor and connected-component compatibility for non-simply-connected global forms.

Editorial admission requires a bounded source critic, exact article-hash registration and source-bound regression, Markdown and MkDocs replay, followed by protected PR review and merge/readback. None of these documentary gates promotes the mathematical theorem to certification.

Sources and claim boundaries

Boundary: NO_CERTIFICATION.


Indexed primary manuscript paths: