{"id":"251d1ed2-c378-478d-bb8f-22221db64c18","arxiv_id":"2502.07262","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"For tame Kazhdan-Patterson and Savin covers of GL_r, the Gelfand-Graev representation decomposes as explicit simple-type Hecke algebra modules, yielding the Whittaker dimension of discrete series as |X(λ)/S_k|.","lead":"For Kazhdan-Patterson and Savin's covers of the general linear group over a p-adic field, the author identifies the Gelfand-Graev representation as a Hecke algebra module and computes its explicit structure. The main corollary gives the Whittaker dimension of every discrete series representation, a quantity that is no longer always one for covering groups.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The main theorem rests on Assumption 4.5 (H ≅ A⊗_C H0), quoted from the author's unpublished [34, Conj. 7.16] via [34, Cor. 7.12] without proof; if this isomorphism fails for any KP/Savin simple type, the orbit decomposition and the dimension formula (10.2) do not follow.","rationale":"The reader's conditional verdict is appropriate. I read the argument through Sections 6–10; the reduction to the cuspidal block, the Jacquet isomorphism, the finite-group computation in §8, and the combinatorial completion in §9.3–9.4 are internally coherent, and I found no concrete algebraic error in the steps carried out in the paper. The decisive input is the Hecke algebra isomorphism of Assumption 4.5. The paper states that this is a conjecture in [34] and that a corollary covers the two classes, but the proof is not included here. Since the central theorem and its dimension corollary are literally conditional on that external result, the correct verdict remains conditional acceptance pending verification of [34, Corollary 7.12]. I therefore do not move the reader's verdict.","tokens_in":32497,"tokens_out":18609,"duration_ms":170695,"concrete_test":"Obtain arXiv:2308.16143 ([34]) and verify the proof of Corollary 7.12 in full, checking that it proves the untwisted affine-Hecke isomorphism H(G,λ) ≅ A⊗_C H0 for every simple type of KP and Savin covers and that it does not invoke Theorem 5.1 of the present paper. As a numerical probe, compute the smallest non-depth-zero case (e.g., r=2, k=2, an n-fold KP or Savin cover with n=2 or 3) directly from the cocycle σ: write H(G,λ) by intertwiners and confirm the finite generators satisfy (T_s+1)(T_s−q0)=0 and the A-H0 cross-relation (4.5) with no extra terms; if extra terms appear, Assumption 4.5 fails and the orbit decomposition and dimension formula are not established.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The load-bearing point is Assumption 4.5, not any internal step I could falsify. The paper needs H = H(G,λ) ≅ A⊗_C H0 with the Bernstein presentation (4.5). This is used twice: in §9.2–9.4 to conclude V^λ = ⊕ (A⊗H0)∗Φ_t with summands A⊗(H0⊗_{H_O}ε_O), and in §10 through the adjunction Hom_H(A⊗M, π^λ) ≅ Hom_{H0}(M, ε0). For KP and Savin covers the paper cites [34, Corollary 7.12], but [34] is an unpublished preprint and the proof of that corollary is not reproduced; Assumption 4.5 is described in the text as [34, Conjecture 7.16]. If the cross-relations in (4.5) acquired cocycle corrections, or if the support of the intertwiners were larger than stated, the H0-submodule H0∗Φ_t need not be isomorphic to H0⊗_{H_O}ε_O, and the parametrization by X(λ)/S_k, hence Proposition 10.1 and formula (10.2), would be unsupported. This is a dependency concern, not an internal inconsistency; the paper is transparent about it.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies the Gelfand–Graev representation V = ind_G^U(ψ) of a tame finite central cover of GL_r(F), and determines its image V^λ under the Hecke algebra functor associated to a simple type (J,λ). The main theorem (Theorem 5.1) asserts, for Kazhdan–Patterson and Savin covers, a decomposition V^λ ≅ ⊕_{O∈X(λ)/S_k} A ⊗_C (H0 ⊗_{H_O} ε_O) as modules over the affine Hecke algebra H = A⊗_C H0. The proof proceeds in three steps: a cuspidal case via Jacquet modules (§7), an H0-module structure via finite field computations (§8), and an H-module structure using the Bernstein presentation of the affine Hecke algebra (§9). The main application is Proposition 10.1, which gives dim_C Hom_G(ind_G^U(ψ),π) = |X(λ)/S_k| for every discrete series π with type (J,λ), and the explicit formulas (10.2). The paper is clearly written and transparent about its reliance on the author's prior preprint [34] for the simple type theory and for Assumption 4.5.","tokens_in":32769,"tokens_out":8189,"duration_ms":66627,"significance":"If the cited results from [34] are correct, this is a substantial contribution. It provides the first explicit calculation of Whittaker dimensions for all discrete series representations of tame Kazhdan–Patterson and Savin covers, with a structural decomposition of the Gelfand–Graev module that generalizes earlier work of Gao–Gurevich–Karasiewicz and Chan–Savin. The dimension formula (10.2) is a simple, testable prediction. The paper also demonstrates a general three-step strategy that may apply to other covers once the relevant type theory is developed. The main weakness is the heavy dependence on the author's unpublished work [34]; the manuscript does not reproduce the proof of the central Hecke algebra isomorphism (Assumption 4.5) or of the properties of π^λ used in §10.","major_comments":[{"comment":"The paper's central object, the Hecke algebra H, is identified with the affine Hecke algebra A⊗_C H0 only under Assumption 4.5, which is stated as [34, Conjecture 7.16] and for Kazhdan–Patterson and Savin covers is cited to [34, Corollary 7.12] without proof. This isomorphism is load-bearing: it is used in §9.2 (strategy (9.1)), in Lemma 9.3 to obtain the multiplication law (4.5), and in §10 (Eq. (10.1)) for the adjunction between H and H0. If the isomorphism fails or the Bernstein presentation has cocycle corrections, the orbit decomposition of V^λ and the dimension formula (10.2) do not follow. Because [34] is an unpublished preprint, the main theorem is not currently verifiable from the manuscript alone. The author should either prove Assumption 4.5 for the two families of covers in this paper or provide a complete statement and proof in an appendix.","section":"4.3, Assumption 4.5"},{"comment":"The proof that π^λ is one-dimensional and restricts to the sign character of H0 is cited from [34, Proposition 9.10], another unpublished result. This property is essential for the dimension count: without it, the computation dim Hom_H(V^λ,π^λ) = |X(λ)/S_k| is unsupported. The author should state and prove this property, or provide a precise published reference, before the main corollary can be accepted.","section":"10, Proposition 10.1"}],"minor_comments":[{"comment":"The title contains 'GELF AND–GRAEV' with a missing 'D'; it should read 'GELFAND–GRAEV'.","section":"Title"},{"comment":"Several LaTeX artifacts appear in the displayed text, such as '/bracehtipupleft' and '/bracehtipdownright'; these should be cleaned up before publication.","section":"§4.2, §4.3"},{"comment":"The phrase 'we assume the readers to have a basic knowledge on the simple type theory' should be 'of the simple type theory'.","section":"§1.2"},{"comment":"The definition of X(λ) would benefit from an explicit description of the S_k-action in coordinates, for example σ·(s_1,...,s_k) = (s_{σ^{-1}(1)},..., s_{σ^{-1}(k)}).","section":"§4.2"},{"comment":"Corollary 8.12 is stated as a 'bonus' but is not used later; consider moving it to a remark or clarifying its role in the paper.","section":"§8"}],"recommendation":"major_revision","confidential_remarks":"The paper depends heavily on the author's own unpublished preprint [34] for the simple type theory and, crucially, for Assumption 4.5 and Proposition 9.10. The editor should ensure that [34] has been rigorously refereed and is available before making a final decision. If [34] is not yet accepted, this manuscript should not be accepted until the dependency is resolved, either by publication of [34] or by inclusion of the needed proofs in this paper."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nThe thing to know: this is a real extension of the known Iwahori-component result to all simple types, for Kazhdan-Patterson and Savin covers of GL_r(F). The main theorem (Thm 5.1) decomposes the Gelfand-Graev module as a direct sum of A ⊗ (H0 ⊗_{H_O} ε_O), and the corollary (Prop 10.1) gives the Whittaker dimension of every discrete series as |X(λ)/S_k|. For n=1 it recovers [9], and for depth-zero Iwahori types it recovers [13]. That is a genuine step forward.\n\nThe proof is careful and the three-step strategy is new. The author proves the required support lemmas in detail, and the paper is transparent about what is being assumed. The main soft spot is exactly Assumption 4.5: the isomorphism H ≅ A ⊗ H0, quoted from the author's own preprint [34, Conj. 7.16] via [34, Cor. 7.12], with the proof not reproduced. The entire orbit decomposition and the dimension formula sit on that. If [34] is correct, the argument here goes through; if not, it doesn't. That makes the paper conditional, but not in a way that signals an error. It is a dependency problem, and the paper names it honestly.\n\nThe other cited inputs (the Jacquet-module isomorphism, Paskunas-Stevens, uniqueness of Whittaker models) are standard or published elsewhere. The only other grumble is that the claim about 'theoretically' getting all irreducibles via Zelevinsky is a bit loose, since the KL coefficients are not explicit; but that is clearly stated as a theoretical consequence.\n\nWho should read it: people working on metaplectic covers, simple types, and Whittaker dimensions. It is a within-field paper, and the referee will need access to [34] (which is to appear in J. Inst. Math. Jussieu). I would send it to a serious referee, conditional on the availability and correctness of [34]. The paper deserves refereeing, not desk rejection.","headline":"A solid, honest generalization of the Gelfand-Graev decomposition to arbitrary simple types for KP and Savin covers, conditional on the author's unpublished [34] for the key Hecke algebra isomorphism.","tokens_in":33318,"tokens_out":2404,"would_cite":true,"duration_ms":21096,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["11F70","22E50","19C09"],"pacs":[],"model":"deepseek-v4-flash","headline":"The Gelfand-Graev representation of a tame Kazhdan-Patterson or Savin cover of GL_r(F) decomposes over the Hecke algebra of any simple type into summands indexed by S_k-orbits on a finite abelian group X(λ), yielding an exact Whittaker…","keywords":["Gelfand-Graev representation","Whittaker dimension","simple types","metaplectic covers","Hecke algebras","p-adic groups","Kazhdan-Patterson cover","Savin's cover"],"falsifier":"Compute the Gelfand-Graev dimension for a discrete series representation of a small Kazhdan-Patterson or Savin cover (for instance, n=2, r=2, depth-zero Iwahori type) and compare with |X(λ)/S_k|; any discrepancy would disprove Theorem 5.1. Alternatively, exhibit a simple type of one of these covers for which the Hecke algebra H(G,λ) is not isomorphic to A ⊗_C H0, contradicting Assumption 4.5.","tokens_in":32260,"feed_emoji":"🔢","tokens_out":6460,"duration_ms":52131,"temperature":0.7,"pith_summary":"For an n-fold Kazhdan-Patterson cover or Savin's cover of GL_r(F) over a non-archimedean local field F, the paper realizes the Gelfand-Graev representation V as a module over the Hecke algebra of any simple type (J,λ). It proves that V^λ decomposes into summands indexed by the orbits of the symmetric group S_k on a finite abelian group X(λ) built from λ, each summand being a tensor product of a lattice algebra A and an induced sign representation of a finite Hecke algebra. The main corollary is an exact count of the Whittaker dimension of every discrete series representation: it equals the number of these orbits, |X(λ)/S_k|, with explicit binomial formulas for the two families of covers. This matters because for metaplectic covers Whittaker models are no longer unique, so a complete dimension formula is a structural result that has been missing.","feed_headline":"Orbit count gives Whittaker dimension for metaplectic GL(r)","feed_subtitle":"Decomposing the Gelfand-Graev representation over simple types yields exact formulas for every discrete series.","key_machinery":"The load-bearing objects are simple types (J,λ) for the metaplectic cover, constructed from simple strata and cuspidal representations of finite general linear groups, together with the associated Hecke algebra H(G,λ). The proof uses the Whittaker datum constructed in [23] to position the Gelfand-Graev representation in a good relative position with λ, so that the normalized Jacquet functor identifies the Jacquet module of V with the Gelfand-Graev representation of the Levi component M. The finite abelian group X(λ) = T(b)/T(b,ρ) carries an S_k-action, and the argument shows that the H0-module generated by each support-normalized vector in V^λ is the induced sign representation H0 ⊗_{H_O} ε_O; assembling these over the S_k-orbits gives the decomposition.","core_discovery":"Theorem 5.1 states that for a tame Kazhdan-Patterson cover or Savin's cover of GL_r(F), the Gelfand-Graev representation V = ind_G^U(ψ), viewed through the simple-type functor M_λ, decomposes as an H(G,λ)-module as ⊕_{O∈X(λ)/S_k} A ⊗_C (H0 ⊗_{H_O} ε_O). Here H(G,λ) is isomorphic to A ⊗_C H0, an affine Hecke algebra of type A, where A is the group algebra of the lattice T(b,ρ), H0 is the finite Hecke algebra of S_k, and H_O is the Hecke algebra of the stabilizer of the orbit O. The paper then proves Proposition 10.1: for any discrete series representation π whose inertial class is typed by (J,λ), the Gelfand-Graev dimension dim_C Hom_G(ind_G^U(ψ), π) equals |X(λ)/S_k|. For Kazhdan-Patterson covers this reads binom(k+n0-1, k) · d0/n0 and for Savin's cover binom(k+n0-1, k). Since every discrete series carries a simple type, the Whittaker dimension of all discrete series is thereby determined explicitly.","pith_inferences":["The method suggests that for any Brylinski-Deligne cover for which a simple type theory and the isomorphism H ≅ A ⊗_C H0 are available, the same decomposition should hold, with the same dependence on the finite group X(λ) and its S_k-orbits; the two families treated here are the cases where those hypotheses are currently verified.","The explicit binomial formulas imply that Whittaker dimensions grow polynomially in the parameters k and n0, so they can be used to test conjectures about L-packets of covering groups, where multiplicities are often controlled by such orbit counts.","A testable extension: compute the H0-structure of V^λ for a tame Brylinski-Deligne cover with c,d outside the KP/Savin range in a small example (say n=2, r=2) to see whether failure of Assumption 4.5 changes the decomposition or only the parameters."],"forward_implications":["Every discrete series representation of a tame Kazhdan-Patterson or Savin cover has an explicitly computed Gelfand-Graev dimension, |X(λ)/S_k|, so the Whittaker dimension question for these representations is closed.","Combined with Zelevinsky's classification and metaplectic tensor products, the dimension of any irreducible genuine representation can in principle be written as an integer combination of these orbit counts, with coefficients given by Kazhdan-Lusztig polynomials.","The decomposition of V^λ provides a template for the doubling method: verifying multiplicity one for degenerate Whittaker models of generalized Speh representations reduces to checking which S_k-orbits contribute.","For the two-fold Kazhdan-Patterson cover, the formula gives numerical evidence for the conjectured relation between Whittaker dimensions and multiplicities of tempered representations in the space of functions on symmetric matrices."],"supporting_citations":[{"why":"Provides the simple type theory for metaplectic covers, the intertwining description of λ, the affine Hecke algebra structure H ≅ A ⊗_C H0 (Assumption 4.5), and the type-theoretic classification of discrete series.","marker":"[34]"},{"why":"Constructs the Whittaker datum (U,ψ) in good relative position to λ and provides the intertwining and Jacquet-module facts used throughout Sections 6-8.","marker":"[23]"},{"why":"Supplies the normalized Jacquet functor isomorphism r_N(V) ≅ V_M (Proposition 6.2) used to pass from G to the Levi component M.","marker":"[6]"},{"why":"Gives the uniqueness of Gelfand-Graev models over finite fields and identifies the sign character of the finite Hecke algebra, used in Corollary 8.9.","marker":"[28]"},{"why":"The pro-p Iwahori Hecke algebra decomposition that Theorem 5.1 generalizes, and which is recovered in the depth-zero Iwahori case.","marker":"[13]"},{"why":"The Bernstein-Zelevinsky derivative approach and the three-step strategy (free A-basis, H0-modules with controlled support) that the paper adapts to general simple types.","marker":"[9]"}],"fun_headline_variants":["Orbit count gives Whittaker dimension for metaplectic GL(r)","Simple-type orbit count fixes Whittaker dimension","Gelfand-Graev module computes Whittaker dimension","Tame covers: Whittaker dimension from orbit counts","Discrete series Whittaker dimension via orbits"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The argument depends on Assumption 4.5, taken from the author's prior simple-type theory: the Hecke algebra H(G,λ) is isomorphic to the affine Hecke algebra A ⊗_C H0; the paper cites its verification for Kazhdan-Patterson and Savin covers but does not reproduce the proof.","fun_headline_variants_meta":{"raw":{"variants":["Orbit count gives Whittaker dimension for metaplectic GL(r)","Simple-type orbit count fixes Whittaker dimension","Gelfand-Graev module computes Whittaker dimension","Tame covers: Whittaker dimension from orbit counts","Discrete series Whittaker dimension via orbits"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000685,"raw_usage":{"total_tokens":3098,"prompt_tokens":930,"completion_tokens":2168,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":546,"completion_tokens_details":{"reasoning_tokens":2092}},"tokens_in":546,"tokens_out":2168,"duration_ms":15887,"temperature":1.0,"reasoning_tokens":2092,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-08T13:17:57.319477+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute the Gelfand-Graev dimension for a discrete series representation of a small Kazhdan-Patterson or Savin cover (for instance, n=2, r=2, depth-zero Iwahori type) and compare with |X(λ)/S_k|; any discrepancy would disprove Theorem 5.1. Alternatively, exhibit a simple type of one of these covers for which the Hecke algebra H(G,λ) is not isomorphic to A ⊗_C H0, contradicting Assumption 4.5.","supporting_citations":[{"cited_title":"Paskunas and S","cited_arxiv_id":null,"evidence_quote":"Constructs the Whittaker datum (U,ψ) in good relative position to λ and provides the intertwining and Jacquet-module facts used throughout Sections 6-8."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the normalized Jacquet functor isomorphism r_N(V) ≅ V_M (Proposition 6.2) used to pass from G to the Levi component M."},{"cited_title":"Silberger and E.-W","cited_arxiv_id":null,"evidence_quote":"Gives the uniqueness of Gelfand-Graev models over finite fields and identifies the sign character of the finite Hecke algebra, used in Corollary 8.9."},{"cited_title":"↑2, 4, 13","cited_arxiv_id":null,"evidence_quote":"The pro-p Iwahori Hecke algebra decomposition that Theorem 5.1 generalizes, and which is recovered in the depth-zero Iwahori case."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"The Bernstein-Zelevinsky derivative approach and the three-step strategy (free A-basis, H0-modules with controlled support) that the paper adapts to general simple types."}],"review_version":1}