Langlands parameters for reductive groups over finite fields

David A. Vogan Jr; Naoki Imai

Langlands parameters for reductive groups over finite fields

T0 review · 0 major / 6 minor · reviewed 2026-08-07 · deepseek-v4-flash

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

Pith's one-line read Finite fields get a Langlands correspondence: each L-packet is indexed by a component group's irreducible representations.

desk verdict A clean, honest reformulation of Deligne–Lusztig in Langlands-parameter language; the new Weil group for finite fields is the real contribution, and the packet conjecture is explicitly conjectural, so no red flags. read the letter →

arxiv 2506.06961 v1 pith:3EQYBDRY submitted 2025-06-08 math.NT math.RT

Naoki Imai , David A. Vogan Jr This is my paper · ORCID

classification math.NTmath.RT MSC 20G4011F70

keywords LanglandscorrespondencefinitefieldsreductivegroupsWeilgroupWeil-DeligneL-packetsDeligne-Lusztigtheorycomponent

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 tries to bring finite fields into the Langlands framework that already organizes representations of real and p-adic groups. It defines a Weil group $W_k = I_k \rtimes \langle \sigma_q\rangle$ for $k = \mathbb{F}_q$, where $I_k$ is the inverse limit of the norm maps among the groups $\mathbb{F}_{q^m}^\times$, and then defines Weil L-parameters from $W_k$ to the L-group ${}^L G$. Its main proved result is that equivalence classes of rigid Weil L-parameters match $G(\mathbb{F}_q)$-conjugacy classes of pairs $(T, \theta)$ consisting of a maximal torus and a character; read through Deligne\textendash Lusztig theory, this is a Langlands classification of the irreducible representations of $G(\mathbb{F}_q)$ over $\mathbb{Q}_\ell$, partitioned into L-packets. The paper then enlarges to a Weil\textendash Deligne group and states Conjecture 4.3: irreducible representations should be indexed by special Frobenius-semisimple parameters, with each fiber a packet in bijection with the irreducible representations of a finite component group $A_\varphi$. If true, the finite-group case would sit on the same conceptual footing as local Langlands, with packet sizes and internal structure controlled by the same kind of component-group bookkeeping.

What carries the argument

The central object is the Weil group of a finite field, $W_k = I_k \rtimes \langle \sigma_q\rangle$, where $I_k = \varprojlim_m \mathbb{F}_{q^m}^\times$ with transition maps the norm maps; its Weil\textendash Deligne extension is $WD_k = \mathbb{G}_a \rtimes W_k$. The identity that carries the argument is the torus character computation (3.1)\textendash(3.2): $\widehat{T(\mathbb{F}_{q^m})} = \operatorname{Hom}(\mathbb{F}_{q^m}^\times, T^\vee(K)) \cong \operatorname{Hom}(I_k, T^\vee(K))^{\sigma_q^m}$, identifying characters of a rational torus with Frobenius-fixed homomorphisms from the Weil inertia group to the dual torus. A rigid Weil L-parameter is a pair $(\varphi, T^\vee)$ with $\varphi(I_k)$ inside a maximal torus $T^\vee$ and the image of the whole parameter inside its normalizer; these correspond exactly to the torus-with-character data. The packet machinery is the component group $A_\varphi = Z_{A(\varphi_0)}(\varphi(\sigma_q))$ formed from the stabilizer of $\varphi_0 = \varphi|_{\mathbb{G}_a \times I_k}$, using Lusztig's canonical quotient to remove a central kernel.

What would settle it

Enumerate the rigid Weil L-parameters and the $G(\mathbb{F}_q)$-conjugacy classes of pairs $(T,\theta)$ for a small group such as $PGL_2(\mathbb{F}_q)$ or $GL_3(\mathbb{F}_q)$; any mismatch would refute Proposition 3.11(4). For Conjecture 4.3, compute one special Frobenius-semisimple parameter $\varphi$, form $A_\varphi$, and compare the number of irreducible constituents in the corresponding Deligne\textendash Lusztig packet with $|\operatorname{Irr}_{\mathbb{Q}_\ell}(A_\varphi)|$; a single packet whose constituent count differs from $|\operatorname{Irr}(A_\varphi)|$ would refute the conjecture.

Watch

Extended reading notes

Core claim

The paper's central claim is that a Langlands parametrization for finite reductive groups is not only possible but natural, once the right Weil group is used. Taking $W_k = I_k \rtimes \langle \sigma_q\rangle$ with $I_k = \varprojlim_m \mathbb{F}_{q^m}^\times$ under norm maps, a Weil L-parameter is a homomorphism $\varphi: W_k \to {}^L G$ compatible with the Galois quotient, with semisimple image and with finite image on $I_k$. Proposition 3.11(4) establishes a natural bijection between equivalence classes of rigid Weil L-parameters and $G(k)$-conjugacy classes of pairs $(T, \theta)$ with $T$ a rational maximal torus and $\theta$ a character of $T(\mathbb{F}_q)$; Proposition 3.8 is the torus case $\widehat{T(\mathbb{F}_q)} \cong \operatorname{Hom}(I_k, T^\vee(K))^{\sigma_q}$. Reinterpreting Deligne\textendash Lusztig's virtual characters $R_T(\theta)$ through this bijection, the paper shows that L-packets partition the irreducible representations and that packet size is governed by the Weyl group $(W(G,T)^F)_\theta$, equivalently $W((G^\vee)_{\varphi_0}, T^\vee)_x$. The final formulation is Conjecture 4.3: there should be a natural map from $\operatorname{Irr}_{\mathbb{Q}_\ell}(G(k))$ to special Frobenius-semisimple Weil\textendash Deligne parameters, and for each parameter $\varphi$ the fiber $L_G^{-1}(\varphi)$ should be in bijection with $\operatorname{Irr}_{\mathbb{Q}_\ell}(A_\varphi)$, where $A_\varphi$ is the component group built from the stabilizer of the inertial parameter and the image of $\varphi(\sigma_q)$.

Load-bearing premise

The load-bearing premise is that for every rational torus the norm maps are surjective, which makes characters of $T(\mathbb{F}_q)$ exactly the Frobenius-fixed homomorphisms from $I_k$ to $T^\vee(K)$; the full conjecture adds the packet-size requirement that each fiber $L_G^{-1}(\varphi)$ has $|\operatorname{Irr}(A_\varphi)|$ elements.

Editorial extensions

If this is right

The Deligne\textendash Lusztig virtual representations $R_T(\theta)$ become a Langlands classification: every irreducible $\mathbb{Q}_\ell$-representation of $G(\mathbb{F}_q)$ appears as a summand of some $R_T(\theta)$, and two such virtual representations have common irreducible summands only when their rigid parameters have equivalent inertial restrictions.
Equivalence classes of rigid Weil L-parameters are naturally counted by $G(\mathbb{F}_q)$-conjugacy classes of pairs $(T,\theta)$, so the combinatorial data of rational maximal tori with characters is exactly the data of Langlands parameters, as stated in Proposition 3.11.
If Conjecture 4.3 is correct, each L-packet $\Pi_{\varphi_0}$ is refined into smaller packets $\Pi_\varphi$ whose sizes are $|\operatorname{Irr}_{\mathbb{Q}_\ell}(A_\varphi)|$, making packet size computable from a finite group attached to the parameter.
The Weil\textendash Deligne group over a finite field gives finite-field analogues of the special and Frobenius-semisimple conditions familiar from local Langlands, allowing the finite-field correspondence to be compared with local correspondences at least at the level of parameter shapes.

Reading between the lines

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

Because the torus bijection is stated over any algebraically closed coefficient field $K$ (with possible shrinkage when $\operatorname{char} K$ divides $|T(\mathbb{F}_q)|$), the parameter side should support Galois-descent statements about Deligne\textendash Lusztig characters under automorphisms of $K$, a consequence the paper does not develop.
If Conjecture 4.3 holds, the finite-field correspondence could serve as a test bed for the local Langlands correspondence: taking a local parameter for a $p$-adic field and restricting its inertia image to a finite quotient should, at depth zero, produce finite-field packets whose component-group parametrization matches the one proposed here.
The special unipotent condition in Definition 4.1 suggests, by analogy with real groups, that special parameters correspond to representations that are temperate in a finite-group sense; identifying the finite-field analogue of temperedness and checking which packets of unipotent representations arise from parameters with trivial unipotent part would be a direct test of the analogy.

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

Dear colleague,

You should know this paper is a reformulation paper, but a genuinely useful one. Imai and Vogan define a Weil group for a finite field as an inverse limit of multiplicative groups under norm maps, semidirect product with the Frobenius, and a Weil–Deligne group over Z[1/p]. That is new, and it lets them write Deligne–Lusztig's classification as a Langlands correspondence: rigid Weil L-parameters match G(k)-conjugacy classes of pairs (T, θ), and the Deligne–Lusztig theorem becomes a statement about L-packets indexed by inertial parameters. The paper does not overclaim; the final correspondence (Conjecture 4.3) is flagged as conjectural, and the packet-size formula with component groups is stated as exactly that.

The proven parts look solid. The torus bijection (Proposition 3.8) rests on the norm-map surjectivity cited from Carter, and the fixed-point computation is sketched but plausible; the equivalence relation in Definition 3.3 is fine enough to make geometric conjugacy match inertial equivalence. Proposition 3.11's bijection is a translation of Deligne–Lusztig, but the translation is correct and useful for connecting to local and categorical Langlands. No circularity, no fitting.

Soft spots, in proportion. First, the proofs of Propositions 3.8 and 3.10 are terse; a referee should ask for a fuller explanation of the fixed-point identification in equation (3.2). Second, the main bijections are, as the authors acknowledge, restatements of existing classification results, so the paper's value is conceptual rather than new representation-theoretic output. Third, Definition 3.13 cites a nonexistent 'Theorem 5.9'; the intended reference is clearly Theorem 3.12. That is an editing defect, not a mathematical one. Finally, the paper depends on a fair amount of Deligne–Lusztig background, so the target audience is specialists, but the writing is unusually clear for this area.

Who is this for? Anyone working on finite groups of Lie type who wants to see them inside the Langlands parameter framework, and people working on local or categorical Langlands who want a finite-field analogue. It deserves a serious referee — the new definitions are worth checking carefully, and the conjectural packet formula will likely be cited. I would accept for review and suggest minor revision: fix the reference, expand the two sketched proofs, and make the dependence on Carter's norm surjectivity explicit. No mathematical red flags for me.

Best, [Your name]

Referee Report

0 major / 6 minor

Summary. The paper proposes a Langlands-parameter formalism for connected reductive groups over finite fields. It defines a Weil group W_k for a finite field k as an extension of the arithmetic Frobenius group by the profinite group I_k = lim← F_{q^m}^×, then defines Weil L-parameters and an equivalence relation (Definition 3.3), and rigid counterparts (Definition 3.9). The main proven results are Proposition 3.8, a natural bijection between equivalence classes of Weil L-parameters of a torus T and characters of T(k), and Proposition 3.11, which translates rigid L-parameters into G(k)-conjugacy classes of pairs (T, θ). The paper then packages Deligne–Lusztig theory into L-packets indexed by inertial parameters and states Conjecture 4.3, a Langlands correspondence for finite fields in which fibers are parametrized by irreducible representations of component groups.

Significance. If the results hold, the paper gives a natural formulation of a finite-field Langlands correspondence over Q_ℓ, avoiding the unnatural choices in earlier Deligne–Lusztig parametrizations and connecting the finite-field story to the local Langlands program. The proved bijections for tori and rigid parameters are explicit and are derived from standard root-datum and Deligne–Lusztig theory, with external results properly credited to Carter, Deligne–Lusztig, and Digne–Michel rather than reproved. The paper introduces no ad hoc free parameters and performs no post hoc data fitting. The most delicate step, Proposition 3.8, is internally consistent: for a torus, the homomorphism condition on the semidirect product imposes exactly the σ_q-fixed condition on Hom(I_k, T^∨(K)), so the claimed bijection with characters of T(F_q) is plausible and coherent. Conjecture 4.3 is explicitly marked as conjectural, and its packet-size prediction matches the Lusztig-style expectation.

minor comments (6)

[Definition 3.13] Definition 3.13 cites a nonexistent 'Theorem 5.9'; the intended reference is clearly Theorem 3.12.
[Definition 3.3 and Definition 3.9] The expressions L^G/Z_{G∨}(ϕ′_0) and L^G/T′∨ are used, but the subgroups Z_{G∨}(ϕ′_0) and T′∨ are not generally normal in L^G; the authors should state explicitly that these denote sets of left cosets and that equality of images means equality of left cosets.
[Proposition 3.8 proof] In the proof of Proposition 3.8, the arithmetic Frobenius is first called 'f' and then later 'σ_q'; please use σ_q throughout for the Galois automorphism and reserve F for the geometric Frobenius morphism.
[Proposition 3.11] There is a typographical issue in parts (4) and (5): the symbol 'bT' should almost certainly be 'T∨' in the display of rigid Weil L-parameters.
[Definition 4.1] In Definition 4.1, the action of W_k on G_a is written as 'where (σ_q^n, w) ∈ W_k acts', but elements of W_k are pairs of the form (w, σ_q^n); also the notation φ|Ga(K)(1) should be clarified, for example by spelling out that it denotes the image of the element 1 of G_a(K).
[Introduction] The first sentence of the Introduction contains a missing word: 'The goal of this paper is try to formulate' should read 'is to try to formulate'.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the torus bijection and packet conjecture derive from external arithmetic and representation theory, not from their own definitions.

full rationale

The paper's derivation chain is not circular. Weil L-parameters and their equivalence relation are introduced in Definitions 3.3 and 3.9 in terms of the dual group, and the bijections in Propositions 3.8 and 3.11 are then proved rather than stipulated. The key arithmetic step, the surjectivity of norm maps for tori in Proposition 3.6(3), is cited to the external source Carter [Car85, Proposition 3.2.2]; the fixed-point identification in equation (3.2) follows from Corollary 3.7, which is a direct consequence of that external norm-map result. The indexing of rational maximal tori by W-conjugacy classes in the Frobenius coset is cited to Deligne-Lusztig [DL76, Corollary 1.14], and the Weyl-group identification is cited to Digne-Michel [DM20, Proposition 4.4.1], both independent. The rigidity and equivalence of parameters are not defined by reference to the characters they are later matched with; the matching in Proposition 3.11(4) is a genuine bijection mediated by Proposition 3.8. Theorems 3.12 and 3.14 are quoted from Deligne-Lusztig, and the final Conjecture 4.3 is explicitly conjectural, asserting a new packet-size formula rather than deriving it from a fit or from the paper's own definitions. The only self-citation, [ABV92], appears as motivational context for the categorical form of the correspondence and is not load-bearing in any proof. One editorial defect exists: Definition 3.13 refers to a nonexistent 'Theorem 5.9', evidently meaning Theorem 3.12, but this is a typographical issue and does not affect circularity.

Assumptions & free parameters 0 free parameters · 4 assumptions · 0 invented entities

No free parameters are fitted; the paper introduces no numerical constants. The axioms are standard root-datum theory plus imported theorems from Deligne-Lusztig, Lusztig, and Carter, all cited. The Weil group and Weil-Deligne group are definitions, not empirical entities, so no physical 'graviton-like' postulates appear.

assumptions (4)

standard math Standard theory of based root data and existence and uniqueness of pinned reductive groups over separably closed fields (Theorem 2.1).
Invoked throughout Section 2 to define the dual group and L-group. This is standard background in the field.
domain assumption Surjectivity of norm maps for tori over finite fields (Proposition 3.6(3), citing Carter Proposition 3.2.2).
Used in Proposition 3.8 to identify direct limits of character groups and in Proposition 3.11(5) to compare geometric conjugacy with inertial parameter equivalence.
domain assumption Deligne-Lusztig virtual character theory: existence of R_T(theta), geometric conjugacy, orthogonality, and completeness (Theorem 3.12 and Theorem 3.14).
The paper explicitly imports these theorems from [DL76] and builds the packet structure on top of them.
domain assumption Lusztig's canonical quotient construction for component groups, referenced as [Lus84a, 13.1].
Used in Definition 4.1 to define the canonical quotient of A_{Z_{G^vee}(phi(I_k))^circ}(phi(G_a)). This is an external construction whose details are not reproduced.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Langlands parameters for reductive groups over finite fields." pith.science (2026). https://pith.science/paper/3EQYBDRY

@misc{pith2026250606961,
  author       = {Pith},
  title        = {Pith review of: Langlands parameters for reductive groups over finite fields},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/3EQYBDRY}},
  note         = {Machine review of arXiv:2506.06961}
}

read the original abstract

We define Langlands parameters for connected reductive groups over finite fields and formulate the Langlands correspondence for finite fields using these parameters.

Discussion (0). Continue with ORCID 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

Finite Langlands correspondence
math.NT 2025-08 conditional novelty 6.0 of 10

Every irreducible representation of a connected reductive group over a finite field is matched with a special Langlands parameter, with the fiber over each parameter given by irreducible representations of a finite co...

Reference graph

Works this paper leans on

10 extracted references · 6 canonical work pages · cited by 1 Pith paper

[1]

Adams, D

J. Adams, D. Barbasch and D. A. Vogan, Jr., The L anglands classification and irreducible characters for real reductive groups, vol. 104 of Progress in Mathematics, Birkh\" a user Boston, Inc., Boston, MA, 1992

work page 1992
[2]

R. W. Carter, Finite groups of L ie type, Pure and Applied Mathematics (New York), John Wiley & Sons, Inc., New York, 1985, conjugacy classes and complex characters, A Wiley-Interscience Publication

work page 1985
[3]

Deligne, Les constantes des \'equations fonctionnelles des fonctions L , in Modular functions of one variable, II ( P roc

P. Deligne, Les constantes des \'equations fonctionnelles des fonctions L , in Modular functions of one variable, II ( P roc. I nternat. S ummer S chool, U niv. A ntwerp, A ntwerp, 1972), pp. 501--597. Lecture Notes in Math., Vol. 349, Springer, Berlin, 1973

work page 1972
[4]

Deligne and G

P. Deligne and G. Lusztig, Representations of reductive groups over finite fields, Ann. of Math. (2) 103 (1976), no. 1, 103--161

work page 1976
[5]

Digne and J

F. Digne and J. Michel, Representations of finite groups of L ie type, vol. 95 of London Mathematical Society Student Texts, Cambridge University Press, Cambridge, second edn., 2020

2020
[6]

Lusztig, Characters of reductive groups over a finite field, vol

G. Lusztig, Characters of reductive groups over a finite field, vol. 107 of Annals of Mathematics Studies, Princeton University Press, Princeton, NJ, 1984

1984
[7]

Lusztig, Characters of reductive groups over finite fields, in Proceedings of the I nternational C ongress of M athematicians, V ol

G. Lusztig, Characters of reductive groups over finite fields, in Proceedings of the I nternational C ongress of M athematicians, V ol. 1, 2 ( W arsaw, 1983), PWN, Warsaw, 1984 pp. 877--880

work page 1983
[8]

Lusztig, On the representations of reductive groups with disconnected centre, 168, pp

G. Lusztig, On the representations of reductive groups with disconnected centre, 168, pp. 10, 157--166, 1988, orbites unipotentes et repr\' e sentations, I

1988

Show all 10 references

[9]

I. G. Macdonald, Zeta functions attached to finite general linear groups, Math. Ann. 249 (1980), no. 1, 1--15

1980
[10]

Weil, Sur la th\' e orie du corps de classes, J

A. Weil, Sur la th\' e orie du corps de classes, J. Math. Soc. Japan 3 (1951), 1--35

1951

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.

[1] [1]

Adams, D

J. Adams, D. Barbasch and D. A. Vogan, Jr., The L anglands classification and irreducible characters for real reductive groups, vol. 104 of Progress in Mathematics, Birkh\" a user Boston, Inc., Boston, MA, 1992

work page 1992

[2] [2]

R. W. Carter, Finite groups of L ie type, Pure and Applied Mathematics (New York), John Wiley & Sons, Inc., New York, 1985, conjugacy classes and complex characters, A Wiley-Interscience Publication

work page 1985

[3] [3]

Deligne, Les constantes des \'equations fonctionnelles des fonctions L , in Modular functions of one variable, II ( P roc

P. Deligne, Les constantes des \'equations fonctionnelles des fonctions L , in Modular functions of one variable, II ( P roc. I nternat. S ummer S chool, U niv. A ntwerp, A ntwerp, 1972), pp. 501--597. Lecture Notes in Math., Vol. 349, Springer, Berlin, 1973

work page 1972

[4] [4]

Deligne and G

P. Deligne and G. Lusztig, Representations of reductive groups over finite fields, Ann. of Math. (2) 103 (1976), no. 1, 103--161

work page 1976

[5] [5]

Digne and J

F. Digne and J. Michel, Representations of finite groups of L ie type, vol. 95 of London Mathematical Society Student Texts, Cambridge University Press, Cambridge, second edn., 2020

2020

[6] [6]

Lusztig, Characters of reductive groups over a finite field, vol

G. Lusztig, Characters of reductive groups over a finite field, vol. 107 of Annals of Mathematics Studies, Princeton University Press, Princeton, NJ, 1984

1984

[7] [7]

Lusztig, Characters of reductive groups over finite fields, in Proceedings of the I nternational C ongress of M athematicians, V ol

G. Lusztig, Characters of reductive groups over finite fields, in Proceedings of the I nternational C ongress of M athematicians, V ol. 1, 2 ( W arsaw, 1983), PWN, Warsaw, 1984 pp. 877--880

work page 1983

[8] [8]

Lusztig, On the representations of reductive groups with disconnected centre, 168, pp

G. Lusztig, On the representations of reductive groups with disconnected centre, 168, pp. 10, 157--166, 1988, orbites unipotentes et repr\' e sentations, I

1988