{"id":"8a772114-4911-4181-817d-bc9fe0f0f483","arxiv_id":"2506.06961","paper_version":1,"verdict":"ACCEPT","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Langlands parameters and a Weil-Deligne group are defined for reductive groups over finite fields, reformulating Deligne-Lusztig theory and conjecturing a full correspondence.","lead":"The paper builds a Langlands parameter formalism for the symmetry groups attached to finite fields, using a newly defined Weil group and Weil-Deligne group. It recasts the existing Deligne-Lusztig classification in this language and states a conjecture about how representations should be packaged.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"No significant objection identified: the foundational torus bijection and the conjectural packet formula are internally consistent as far as the argument goes.","rationale":"The paper's central claim is the introduction of a Langlands-parameter formalism for finite fields: the new Weil and Weil-Deligne groups, the torus bijection in Proposition 3.8, the rigid-parameter translation of Deligne-Lusztig data in Proposition 3.11, and the conjectural Langlands correspondence in Conjecture 4.3. The reader's verdict of ACCEPT with moderate confidence is appropriate. The load-bearing arithmetic step is indeed the Frobenius fixed-point identification in equations (3.1)-(3.2), together with the equivalence relation on parameters; if either were wrong, the bijections would collapse. I checked this step carefully rather than taking it on faith. For a torus, a Weil L-parameter is a homomorphism from I_k to the dual torus together with a lift of the Frobenius; two parameters with different inertial parts cannot be equivalent because the first condition of Definition 3.3 would require them to be conjugate, and conjugation on an abelian dual torus is trivial. Parameters with the same inertial part but different Frobenius lifts are identified by the second condition, which is exactly what the character parametrization should do. The fixed-point computation then reduces to the norm-map surjectivity cited from Carter for all tori, and the explicit Corollary 3.7 gives the same conclusion directly. For general groups, the rigid-parameter equivalence in Definition 3.9 introduces the Weyl-group action, matching the quotient by the Weyl group of a rational maximal torus in Proposition 3.11(4). I also checked transitivity of the equivalence relation of Definition 3.3; it follows from the stabilizer behavior of the inertial part, so no hidden failure lurks there. The final Conjecture 4.3 is honestly labeled conjectural, and its component-group formula is consistent with the Lusztig-style packet-size expectation: Theorem 3.14 measures the size of the geometric-conjugacy packet, and the proposed A_phi is the natural canonical quotient from Lemma 4.2 and Lusztig's theory. The only concrete defect found is the dangling cross-reference to 'Theorem 5.9' in Definition 3.13, which is a typographical copyediting issue and does not alter the mathematical claims. The reader's weakest assumption matches my own assessment of the most sensitive point, so I agree with the reader's analysis and leave the verdict unchanged.","tokens_in":10117,"tokens_out":27804,"duration_ms":323682,"concrete_test":"Run an explicit check of Proposition 3.8 for a rank-one nonsplit torus, for example the norm-one torus over F_3 or F_5, by enumerating all Weil L-parameters modulo the equivalence of Definition 3.3 and comparing the resulting set with the character group of T(F_q). If the cardinalities differ, the foundational fixed-point computation or the equivalence relation fails; if they match, the main torus bijection is sound.","verdict_should_be":"UNCHANGED","load_bearing_attack":"I could not identify a load-bearing objection. The most delicate step is the identification in Proposition 3.8 of the Frobenius-fixed points of Hom(I_k, T∨(K)) with characters of T(F_q). This step depends on the norm-map surjectivity cited from Carter, and on the equivalence relation of Definition 3.3 being fine enough: for a torus the first condition in that equivalence separates parameters by their inertial homomorphism, while the second condition only quotients out the irrelevant torus factor in the sigma image, so the bijection is plausible and coheres with Corollary 3.7. The same structure propagates to the rigid-parameter bijection of Proposition 3.11(4), whose proof is admittedly terse but consistent with the explicit enumeration preceding it. Conjecture 4.3 is explicitly conjectural and matches the Lusztig-style packet-size expectation; no internal inconsistency was found. One clear editorial defect, not mathematical, is that Definition 3.13 cites a nonexistent 'Theorem 5.9'; the intended reference is evidently Theorem 3.12.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","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.","tokens_in":10285,"tokens_out":17091,"duration_ms":164367,"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.","major_comments":[],"minor_comments":[{"comment":"Definition 3.13 cites a nonexistent 'Theorem 5.9'; the intended reference is clearly Theorem 3.12.","section":"Definition 3.13"},{"comment":"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.","section":"Definition 3.3 and Definition 3.9"},{"comment":"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.","section":"Proposition 3.8 proof"},{"comment":"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.","section":"Proposition 3.11"},{"comment":"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).","section":"Definition 4.1"},{"comment":"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'.","section":"Introduction"}],"recommendation":"minor_revision","confidential_remarks":"I concur with the reader's assessment that there is no circularity and no unsupported data fitting. The manuscript is squarely within the journal's scope. The central claims are defensible, and the issues I found are local and easily fixed; I would be glad to see the revision accepted."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Dear colleague,\n\nYou 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.\n\nThe 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.\n\nSoft 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.\n\nWho 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.\n\nBest,\n[Your name]","headline":"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.","tokens_in":10858,"tokens_out":1842,"would_cite":true,"duration_ms":18377,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["20G40","11F70"],"pacs":[],"model":"deepseek-v4-flash","headline":"Finite fields get a Langlands correspondence: each L-packet is indexed by a component group's irreducible representations.","keywords":["Langlands correspondence","finite fields","reductive groups","Weil group","Weil-Deligne group","L-packets","Deligne-Lusztig theory","component groups"],"falsifier":"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.","tokens_in":9900,"feed_emoji":"🧮","tokens_out":10101,"duration_ms":97910,"temperature":0.7,"pith_summary":"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.","feed_headline":"Finite fields get Langlands parameters","feed_subtitle":"A Weil group for F_q organizes every representation into packets matching a component group's characters.","key_machinery":"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.","core_discovery":"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 $\\varphi(\\sigma_q)$.","pith_inferences":["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."],"forward_implications":["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."],"supporting_citations":[{"why":"supplies the model parametrization for $GL_n(\\mathbb{F}_q)$ and motivates defining the Weil group through inverse limits of norm maps.","marker":"[Mac80]"},{"why":"provides the Deligne\\textendash Lusztig virtual representations $R_T(\\theta)$, the geometric-conjugacy criterion, and the packet-size theorem that the paper reinterprets as a Langlands classification.","marker":"[DL76]"},{"why":"supplies the surjectivity of norm maps on tori quoted in Proposition 3.6(3), the arithmetic step behind the fixed-point computation (3.2).","marker":"[Car85]"},{"why":"defines Lusztig's canonical quotient used in forming $A_\\varphi$ and gives the finite-field character theory that Conjecture 4.3 refines.","marker":"[Lus84a]"},{"why":"supplies the structure theory for finite groups of Lie type, quoted as Proposition 4.4.1, that identifies Weyl groups of rational tori in Proposition 3.11(3).","marker":"[DM20]"},{"why":"states the shape of a Langlands classification with packets indexed by component groups and equivariant perverse sheaves, which the paper adapts to finite fields.","marker":"[ABV92]"},{"why":"introduces the Weil\\textendash Deligne modification of Langlands parameters that Definition 4.1 transplants to finite fields.","marker":"[Del73]"}],"fun_headline_variants":["Langlands parameters for finite reductive groups","A Weil group brings Langlands to finite fields","Finite fields get a Langlands correspondence","Langlands program extends to finite reductive groups","Reductive groups over finite fields have Langlands parameters"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"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.","fun_headline_variants_meta":{"raw":{"variants":["Langlands parameters for finite reductive groups","A Weil group brings Langlands to finite fields","Finite fields get a Langlands correspondence","Langlands program extends to finite reductive groups","Reductive groups over finite fields have Langlands parameters"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000167,"raw_usage":{"total_tokens":1255,"prompt_tokens":941,"completion_tokens":314,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":557,"completion_tokens_details":{"reasoning_tokens":244}},"tokens_in":557,"tokens_out":314,"duration_ms":3502,"temperature":1.0,"reasoning_tokens":244,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-07T05:45:33.536236+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"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.","supporting_citations":[],"review_version":1}