Geometric Langlands in positive characteristic from characteristic zero

Dennis Gaitsgory; Sam Raskin

REVIEW 1 major objections 2 minor 1 cited by

Geometric Langlands in positive characteristic from characteristic zero

T0 review · 1 major / 2 minor · reviewed 2026-08-06 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read Over a field of positive characteristic, automorphic l-adic sheaves with nilpotent singular support match ind-coherent sheaves on a union of some connected components of the stack of Langlands parameters.

desk verdict Partial positive-characteristic geometric Langlands result from two credible authors, but the abstract alone is too underspecified to judge the mathematics. read the letter →

arxiv 2508.02237 v1 pith:VR77SYJJ submitted 2025-08-04 math.AG

Dennis Gaitsgory , Sam Raskin This is my paper

classification math.AG MSC 14D2414F2014G17

keywords geometricLanglandsconjecturepositivecharacteristicl-adicsheavesnilpotentsingularsupportind-coherentstackofparameterscategoricalequivalencezeroreduction

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper establishes a substantial part of the geometric Langlands conjecture over fields of positive characteristic, in the l-adic sheaf setting. It proves an equivalence between automorphic sheaves whose singular support is nilpotent and a category of ind-coherent sheaves on a union of connected components of the stack of Langlands parameters. The title points to the proof route: the positive-characteristic statement is obtained from the characteristic-zero categorical form of geometric Langlands, so the structures known in characteristic zero serve as the template. If the argument is right, Langlands duality in its categorical form is not confined to characteristic zero; it also describes sheaves on the moduli of bundles over positive-characteristic curves.

What carries the argument

The machinery is the singular-support condition together with ind-coherent sheaves on the Langlands-parameter stack. Singular support is a microlocal invariant that records, roughly, in which direction along the cotangent bundle a sheaf spreads; requiring it to be nilpotent cuts the full automorphic-sheaf category down to a subcategory fine enough to match a subcategory of $\mathrm{IndCoh}$ on $\mathrm{LocSys}_{\check G}$, where $\check G$ is the Langlands dual group. The load-bearing move is showing that this cutting survives the trip from characteristic zero to positive characteristic, so that the two sides remain identified after specialization or reduction.

What would settle it

Work out both sides explicitly in the simplest nontrivial case, for example $G = GL_1$ on a curve of positive characteristic, where the automorphic side is sheaves on the Picard stack and the parameter side is local systems; if the claimed equivalence does not reproduce the known rank-one duality, the central claim is wrong.

Watch

Extended reading notes

Core claim

The paper's central claim is a categorical equivalence in the l-adic setting over a base field of positive characteristic, with $l$ invertible in the ground field. For a smooth projective curve and a reductive group $G$, the category of automorphic sheaves on $\mathrm{Bun}_G$ with nilpotent singular support is equivalent to the appropriately defined category of ind-coherent sheaves on the union of some of the connected components of the stack of Langlands parameters. The phrase “from characteristic zero” tells the reader that the proof proceeds by transferring the already available characteristic-zero equivalence to positive characteristic, rather than by constructing the correspondence directly there.

Load-bearing premise

The argument stands on the premise that the automorphic-sheaf category with nilpotent singular support and the ind-coherent category on the parameter stack exist and have the same good properties in positive characteristic that they have in characteristic zero, even though those definitions are not given in the abstract.

Editorial extensions

If this is right

The geometric Langlands equivalence is now a theorem for l-adic sheaves with nilpotent singular support over positive-characteristic fields, rather than only a conjecture.
Any categorical construction that holds in the characteristic-zero equivalence and is compatible with the reduction functors transfers to positive characteristic, so derived invariants on one side can be computed on the other.
The statement places the singular-support-restricted part of the Langlands correspondence for function fields on an equal footing with the characteristic-zero statement, giving a direct bridge between the two worlds.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

A natural next step is to ask whether the same method, applied without the singular-support cut, yields an equivalence on all components; the paper's selected union indicates that the nilpotent condition is doing essential work there.
The reduction-from-characteristic-zero route suggests the equivalence should also hold over arbitrary fields of positive characteristic, including finite fields, provided the categories are defined there; a concrete check would be to compare the action of Frobenius on both sides.
For $G = GL_1$ the claimed equivalence should reduce to classical rank-one duality for sheaves on the Picard stack and local systems; checking that reduction is a low-cost test of the whole machinery.

Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

The abstract states a partial geometric Langlands equivalence in positive characteristic: automorphic sheaves with nilpotent singular support match ind-coherent sheaves on a union of some Langlands-parameter components. If correct, this is a real step toward the full conjecture, and the abstract is honest that it is only part of the statement. That honesty is a plus.

What's genuinely new here is the claimed bridge from characteristic zero to positive characteristic, presumably via some form of deformation or specialization of categories. The authors' prior framework is the natural setting, and the novelty seems to be in extending the nilpotent-singular-support story across characteristics rather than in any single new definition. The paper is clearly aimed at experts and builds on a large body of established work.

The soft spots are real, but they are soft spots of the abstract, not necessarily of the paper. The stress-test note is correct: 'nilpotent singular support' for l-adic sheaves in positive characteristic is not a standard notion in the literature, and without a definition or a pointer to one, the theorem is not fully well-posed. Likewise, 'the union of some of the connected components' is not an intrinsic specification; the right-hand category depends on which components are chosen. If the full text defines these objects and proves the expected categorical properties, the concern evaporates. But from the abstract alone, the statement cannot be verified or even precisely parsed.

A second, softer concern is the paper's heavy reliance on the authors' own categorical machinery. That is not a flaw when the referenced prior work is solid, but it does mean the independence of the current result from earlier unverified claims cannot be assessed from the abstract alone. No circularity is visible, and the stated theorem is not tautological.

My overall take: this is a serious paper by serious people, and the result—if solid—will matter. But the abstract is not self-contained enough for a definitive technical judgment. A capable referee with access to the full text should be able to evaluate the definitions and the proof. I would not cite it based on the abstract alone, but I would put it on the reading group list once the full version is available, and I would definitely send it to peer review rather than desk-reject it.

Referee Report

1 major / 2 minor

Summary. The paper claims to establish part of the geometric Langlands conjecture for l-adic sheaves over a field of positive characteristic, namely an equivalence between the category of automorphic sheaves with nilpotent singular support and an appropriately defined category of ind-coherent sheaves on the union of some connected components of the stack of Langlands parameters. The abstract contains no proof details, definitions, or statements of lemmas, and the full text was not available for review.

Significance. If the claimed equivalence holds, it would be a substantial advance in the geometric Langlands program in positive characteristic, extending a categorical statement known or expected in characteristic zero to the l-adic setting. The claim is non-tautological and falsifiable, and it addresses an open problem of considerable interest. The abstract is too terse to allow verification, but the statement as given is not vacuous and appears to require genuine new input to handle singular support and ind-coherent sheaves in positive characteristic.

major comments (1)

[Abstract] The theorem as stated is not well-posed because its two central objects are not fixed. First, 'nilpotent singular support' for l-adic sheaves in positive characteristic is not a standard imported notion; the usual singular support formalism is developed in characteristic zero, and a positive-characteristic analogue requires either a construction or a proved property. Second, 'the union of some of the connected components of the stack of Langlands parameters' leaves unspecified which components are included; until that union is characterized intrinsically, the right-hand category is not uniquely determined. If the full text supplies these definitions, this is only a presentation issue in the abstract, but from the available material the central claim cannot be evaluated.

minor comments (2)

[Abstract] The abstract would be more informative if it indicated the source of the positive-characteristic singular support definition and the criterion for selecting the connected components of the Langlands parameter stack, even by reference to numbered definitions in the body.
[Abstract] The phrase 'appropriately defined category' is vague; specifying the category's definition or citing a section would help readers assess the scope of the claimed equivalence.

Circularity Check

0 steps flagged · score 0.0 of 10

Abstract-only review finds no circular derivation; definitional gaps are completeness concerns, not circularity.

full rationale

This is an abstract-only review, so no derivation chain, equations, or cited prior results are available to inspect. The two sides of the claimed equivalence—automorphic sheaves with nilpotent singular support and ind-coherent sheaves on a specified union of connected components of the Langlands parameter stack—are standard objects in the geometric Langlands program, and the asserted equivalence is the substance of the conjecture rather than a restatement of a definition. The phrase 'appropriately defined category' indicates that the right-hand side must be constructed with care, but it does not say that the category is defined as 'the category equivalent to the left-hand side,' which would be circular. Likewise, 'union of some of the connected components' is a stated restriction on the parameter stack and is part of the conjecture's content, not a fitted parameter. No self-citation appears in the abstract, and no reduction of the theorem to its own inputs can be exhibited from the text provided. Lack of definitions is a well-posedness and completeness concern, not a circularity finding under the hard rules, which require quoting a specific reduction rather than inferring circularity from vagueness.

Assumptions & free parameters 0 free parameters · 2 assumptions · 0 invented entities

No fitted numbers appear; this is a theorem proof. The axioms are background assumptions about the categorical framework used to state the result, which cannot be justified from the abstract.

assumptions (2)

domain assumption The framework of derived algebraic geometry, higher categories, and l-adic sheaves developed in prior literature is available and works as expected in positive characteristic.
The abstract states an equivalence between categories defined via this machinery; no details are provided.
domain assumption The 'appropriately defined category of ind-coherent sheaves' and the 'automorphic sheaves with nilpotent singular support' correspond to the standard geometric Langlands objects, so that the proven equivalence matches the intended conjecture.
The abstract does not specify the definitions; the result's meaning depends on these being the right objects.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Geometric Langlands in positive characteristic from characteristic zero." pith.science (2026). https://pith.science/paper/VR77SYJJ

@misc{pith2026250802237,
  author       = {Pith},
  title        = {Pith review of: Geometric Langlands in positive characteristic from characteristic zero},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/VR77SYJJ}},
  note         = {Machine review of arXiv:2508.02237}
}

read the original abstract

We establish part of the statement of the geometric Langlands conjecture for l-adic sheaves over a field of positive characteristic. Namely, we show that the category of automorphic sheaves with nilpotent singular support is equivalent to the appropriately defined category of ind-coherent sheaves on the union of some of the connected components of the stack of Langlands parameters.

Discussion (0). Sign in to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

On the excursion algebra
math.AG 2026-02 conditional novelty 6.0 of 10

The excursion algebra of a scheme over a finite field is canonically isomorphic to the ring of functions on the geometrically semisimple locus, yielding reduced/normal components, finite generation, GL_n surjectivity,...

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.