{"id":"5cb04c09-3582-49ac-84dd-7fcd2a554e7c","arxiv_id":"2501.02332","paper_version":1,"verdict":"REJECT","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"high","formal_verification":"none","parameter_count":0,"one_line_summary":"The paper claims the Macdonald correspondence is uniquely characterized by central character and pair epsilon-factors, but the proof of this characterization is incomplete.","lead":"This paper reviews a classical 1980 bijection between representations of finite general linear groups and local Galois representations, then claims a new uniqueness theorem for it using epsilon-factors. The proof of the main theorem has a gap that leaves the central characterization unsupported.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Theorem 2 proof applies Corollary 5.0.3 to a mixed pair without establishing the required mixed-pair epsilon identity; the assumptions only equate same-correspondence pairs, so the characterization is unsupported.","rationale":"The strongest claim is the uniqueness characterization in Theorem 2. The proof's only forward mechanism is Corollary 5.0.3, a converse theorem in the style of Henniart and Nien. To apply it one must know a mixed epsilon identity involving τρ and πρ′. The paper proves Theorem 1, which identifies ε(πρ1 × πρ2) with ε(ρ1 ⊗ ρ2) when both cuspidal-side representations come from the Macdonald correspondence, and assumes property (2), which identifies ε(τρ1 × τρ2) with ε(ρ1 ⊗ ρ2) when both come from the candidate T. Since T is not known to equal M, knowing the two same-side pairs are equal does not permit replacing τρ′ by πρ′ in one factor. The step \"Equations (5.0.4) and (5.0.5) allow us to apply Corollary 5.0.3\" is therefore invalid. This is a genuine logical gap, not a disagreement with consensus. The review portions and Theorem 1 may be useful, but the abstract's advertised characterization is not established by the written proof. Minor typographical issues in the corollary do not repair the gap.","tokens_in":12450,"tokens_out":7955,"duration_ms":74859,"concrete_test":"Reconstruct the proof of Theorem 2 in a formal or semi-formal derivation: starting from property (2), Theorem 1, equations (3.0.5) and Definition 3.0.6, attempt to prove ε(τρ × πρ′, ψ) = ε(ρ ⊗ ρ′, ψ) for all ρ′. If the derivation requires an assumption that T agrees with M on ρ′ (or on the pair), the proof is circular. As a secondary check, for a small field such as F_2 or F_3 and n = 2, compare the mixed-pair epsilon value ε(τρ × πρ′) with ε(ρ ⊗ ρ′) for a candidate T not yet known to equal M; a mismatch would demonstrate the missing premise is substantive.","verdict_should_be":"REJECT","load_bearing_attack":"Corollary 5.0.3 requires, for the chosen ρ and every m ≤ n/2 and every irreducible m-dimensional ρ′, the equality ε(π × πρ′, ψ) = ε(ρ ⊗ ρ′, ψ). In the proof one takes π = τρ and wants to conclude τρ = πρ. The established identities are: (i) property (2): ε(τρ × τρ′, ψ) = ε(ρ ⊗ ρ′, ψ); (ii) Theorem 1: ε(πρ × πρ′, ψ) = ε(ρ ⊗ ρ′, ψ); and hence (5.0.5) ε(πρ × πρ′) = ε(τρ × τρ′). None of these controls the mixed pair ε(τρ × πρ′, ψ), because T is not yet known to coincide with M. The sentence \"Equations (5.0.4) and (5.0.5) allow us to apply Corollary 5.0.3\" conflates the same-side identities with the mixed-pair identity demanded by the corollary. Since the converse theorem is the only mechanism forcing τρ = πρ, the advertised uniqueness of the Macdonald correspondence does not follow from the written argument.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper revisits Macdonald's 1980 bijection between irreducible representations of GL_n(k) and inertial equivalence classes of tame n-dimensional Weil-Deligne representations of W_F. It constructs the correspondence via Lusztig's classification of finite groups of Lie type, recalls its compatibility with the depth-zero local Langlands correspondence and Shintani descent, and defines epsilon factors for pairs of cuspidal representations using Soudry-Zelingher's Shahidi gamma factors. Theorem 1 states that these pair epsilon factors match Deligne epsilon factors under the Macdonald correspondence. The advertised new result is Theorem 2, which claims that any family of bijections T_n between the same two sets that preserves central characters and epsilon factors for pairs must equal the cuspidal Macdonald correspondence, with the proof resting on Nien's converse theorem through Corollary 5.0.3.","tokens_in":12670,"tokens_out":7325,"duration_ms":74188,"significance":"If valid, Theorem 2 would give a clean finite-field analogue of Henniart's characterization of the local Langlands correspondence, and the paper's construction of the Macdonald correspondence through Lusztig's classification is a useful perspective. Theorem 1 is a valuable matching statement for pair epsilon factors. However, the central characterization is not established by the written argument: the proof of Theorem 2 applies Corollary 5.0.3 to a mixed pair without proving the required mixed-pair epsilon identity. Since Theorem 2 is the main new claim of the paper, this is a load-bearing gap rather than a presentation issue.","major_comments":[{"comment":"Corollary 5.0.3 requires, for the chosen ρ and every m ≤ n/2 and every irreducible m-dimensional ρ′, the equality ε(π × πρ′, ψ) = ε(ρ ⊗ ρ′, ψ). In the proof one takes π = τρ and needs exactly this mixed equality. The text establishes Eq. (5.0.5), which equates ε(πρ1 × πρ2, ψ) with ε(τρ1 × τρ2, ψ), plus the equality of central characters. These are same-correspondence comparisons: Theorem 1 gives ε(πρ × πρ′) = ε(ρ ⊗ ρ′), and property (2) gives ε(τρ × τρ′) = ε(ρ ⊗ ρ′). Neither controls the mixed quantity ε(τρ × πρ′, ψ), because T is not yet known to coincide with M. The sentence \"Equations (5.0.4) and (5.0.5) allow us to apply Corollary 5.0.3\" therefore conflates same-side identities with the mixed-pair identity demanded by the corollary. This is not a minor omission: the conclusion τρ = πρ is precisely the statement needed to convert the same-side equalities into the mixed equality, so the reasoning is circular at the load-bearing step.","section":"§5, proof of Theorem 2 (Eq. (5.0.5))"},{"comment":"The statement of Corollary 5.0.3 contains an internal inconsistency: it quantifies over an integer m with 1 ≤ m ≤ n/2 but then requires equality for every irreducible n-dimensional representation ρ′ of W_F, leaving m unused; presumably ρ′ should be m-dimensional. More substantively, the proof says the corollary follows from Nien's Converse Theorem by using Definition 3.0.6 and (3.0.5), but Nien's theorem requires equality of gamma factors against every irreducible generic representation τ of GL_m(k), whereas the corollary only supplies equalities for cuspidal representations of the form πρ′. A reduction from arbitrary generic τ to cuspidal data, presumably via multiplicativity of gamma factors, is not supplied. Since Corollary 5.0.3 is the bridge from the hypotheses of Theorem 2 to Nien's converse theorem, this step needs to be written out.","section":"§5, Corollary 5.0.3"}],"minor_comments":[{"comment":"The formula reads ψ(a1 + a1 + ⋯ + an−1); the second occurrence of a1 should almost certainly be a2.","section":"§2.5, Eq. (2.5.1)"},{"comment":"There is a typo: \"In particuler\" should be \"In particular\".","section":"§2.3.2, paragraph before Eq. (2.3.16)"},{"comment":"The notation ε(σ1 × σ2, ψ) is introduced for pairs of cuspidal representations, but in Theorem 2 the same notation is used in property (2) for all ρ1, ρ2 without explicitly restating the cuspidal restriction; clarifying this at the start of §5 would help the reader.","section":"§3, Definition 3.0.6"}],"recommendation":"reject","confidential_remarks":"The main issue is not a missing detail but a structural gap: the hypotheses of Theorem 2 only constrain T on one side of the correspondence at a time, and the proof has no mechanism to relate the mixed pair (τρ, πρ′). Since the advertised characterization is the paper's central contribution, this appears to be a fundamental obstruction rather than a local repair. The author might be able to state a different theorem with a stronger hypothesis, but that would not be the result advertised here. The Lusztig-based construction and Theorem 1 may still be of interest, but the paper in its current form is not ready for publication."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nThe headline: the proof of Theorem 2 does not go through. The advertised characterization of the Macdonald correspondence is unsupported as written. That said, the paper is not hollow; Theorem 1 is correct and the survey portion is genuinely useful.\n\nWhat is actually new: the definition of epsilon-factors for pairs of cuspidal representations of finite general linear groups (Definition 3.0.6) using Shahidi gamma factors from Soudry–Zelingher, and Theorem 1 matching these with Deligne epsilon-factors under the Macdonald correspondence. That theorem is a clean consequence of known results (Ye–Zelingher, Zelingher) and appears right. The construction of the Macdonald correspondence via Lusztig's classification is clearly written and worth having as a reference.\n\nThe soft spot is Section 5. Corollary 5.0.3, the finite-field converse theorem, demands a mixed-pair identity: ε(π × πρ', ψ) = ε(ρ ⊗ ρ', ψ) for every relevant ρ'. In the proof of Theorem 2 one has property (2), which says ε(τρ × τρ') = ε(ρ ⊗ ρ'), and Theorem 1, which says ε(πρ × πρ') = ε(ρ ⊗ ρ'), and hence ε(πρ × πρ') = ε(τρ × τρ'). But nothing relates the mixed pair ε(τρ × πρ') to either. The sentence \"Equations (5.0.4) and (5.0.5) allow us to apply Corollary 5.0.3\" conflates same-side identities with the mixed-pair identity the corollary requires. That is a load-bearing gap, not a minor omission.\n\nMinor issues: unfinished cross-reference placeholders like \"(??)\" and a few typos, which suggest the manuscript is not polished.\n\nWho is this for? Readers interested in finite general linear groups and the p-adic local Langlands analogy will find the survey parts worth reading, and Theorem 1 is a tidy result. But if the paper is advertised as giving a characterization, that claim is not established. I would send it to a referee—the topic is important enough and the gap may well be repairable—but the referee should be told to focus on Section 5. I would not cite Theorem 2 in its current form.","headline":"The advertised characterization theorem is unproven as written — the converse theorem is applied to a mixed pair without the required mixed-pair epsilon identity — but the survey parts and Theorem 1 are solid.","tokens_in":13213,"tokens_out":2629,"would_cite":false,"duration_ms":24657,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["20C33","22E50","11S40"],"pacs":[],"model":"deepseek-v4-flash","headline":"Macdonald's correspondence is uniquely characterized by central characters and pair epsilon-factors.","keywords":["Macdonald correspondence","epsilon-factors for pairs","cuspidal representations","Weil-Deligne representations","finite general linear groups","local Langlands correspondence","gamma factors","characterization theorem"],"falsifier":"Compute the pair epsilon-factor with one representation taken from the Macdonald correspondence and the other from a candidate bijection, for a small rank where the two could disagree; any mismatch would show that the uniqueness proof as written does not go through.","tokens_in":12208,"feed_emoji":"🔢","tokens_out":11860,"duration_ms":106360,"temperature":0.7,"pith_summary":"The paper tries to establish that the Macdonald correspondence is the only bridge between inertial classes of tamely ramified Weil-Deligne representations and irreducible cuspidal representations of finite general linear groups that is compatible with the natural invariants. It reconstructs the correspondence through the classification of representations of finite groups of Lie type, defines epsilon-factors for pairs of cuspidal representations, and proves that these match the epsilon-constants on the Weil-Deligne side. The main theorem states that any collection of bijections preserving central characters and these paired epsilon-factors must be the Macdonald correspondence. If correct, this puts the finite-field correspondence on the same axiomatic footing as the p-adic characterization.","feed_headline":"Macdonald's map is the unique pair-epsilon preserving bijection","feed_subtitle":"The theorem rules out any alternative matching of cuspidal representations to Weil-Deligne classes that keeps these invariants.","key_machinery":"The central object is the epsilon-factor for pairs of irreducible cuspidal representations, built from a normalized gamma factor attached to parabolic induction. Three supporting mechanisms carry the argument: the parametrization of cuspidal representations by regular characters of elliptic tori, the correspondence between Galois orbits of such characters and inertial classes of tamely ramified Weil-Deligne representations, and a finite-field converse theorem that converts equality of all pair gamma factors into equality of representations. The pair epsilon-factor identity is what lets the author compare any candidate bijection with the Macdonald correspondence and force them to match.","core_discovery":"The central claim is Theorem 2: if a family of bijections from inertial classes of n-dimensional tamely ramified irreducible Weil-Deligne representations to irreducible cuspidal representations of finite general linear groups preserves central characters and the epsilon-factor for every pair of such classes, then the family coincides with the Macdonald correspondence. The proof shows that the inverse Macdonald map has both properties, then uses a converse theorem for finite general linear groups to show that any competitor must agree with it on every representation. The paper also records a construction of the correspondence and proves in Theorem 1 that its epsilon-factors for pairs agree with the matching epsilon-constants of Weil-Deligne representations.","pith_inferences":["If the mixed-pair equality used in the proof can be derived, the theorem would extend the p-adic uniqueness principle to finite fields: pair epsilon-factors alone determine the correspondence, without invoking the full local Langlands machinery.","The same strategy may characterize correspondences for other finite reductive groups whenever a compatible pair-factor theory and a converse theorem are available.","A direct computation of the mixed epsilon-factor for small q and small n would settle whether the only gap in the written proof is repairable.","The axiomatic formulation suggests viewing the Macdonald correspondence not as a list of matches but as the unique solution to a system of invariance equations."],"forward_implications":["If Theorem 2 is correct, the Macdonald correspondence is the unique matching of cuspidal representations to tamely ramified Weil-Deligne classes that preserves central characters and pair epsilon-factors.","The epsilon-factors for pairs defined from gamma factors agree with the epsilon-constants of the corresponding Weil-Deligne representations, giving a finite-field analogue of the local Langlands pair-factor compatibility.","The construction via the classification of finite groups of Lie type makes the Macdonald correspondence explicitly computable from regular characters and partitions.","The characterization implies that central characters together with all pair epsilon-factors form a complete set of invariants for this classification problem.","The compatibility with the depth-zero local Langlands correspondence locates the finite-field correspondence as the residue-field shadow of the p-adic correspondence."],"supporting_citations":[{"why":"Supplies the original correspondence and the zeta-function epsilon-factor identity that the paper reconstructs.","marker":"[Mac2]"},{"why":"Supplies the parametrization of irreducible cuspidal representations by regular characters, used to build the bijection.","marker":"[Gr]"},{"why":"Supplies the classification of representations of finite groups of Lie type used in the construction.","marker":"[Lu]"},{"why":"Defines the normalized gamma factor for finite general linear groups from which the pair epsilon-factor is built.","marker":"[SoZe]"},{"why":"Provides the finite-field converse theorem used to turn equality of pair gamma factors into equality of representations.","marker":"[Nie]"},{"why":"Establishes epsilon factors for pairs through the Macdonald correspondence, used for the unequal-rank case in Theorem 1.","marker":"[YZ]"},{"why":"Supplies the Bessel-function identity for equal-sized pairs, used for the equal-rank case in Theorem 1.","marker":"[Z]"},{"why":"Defines the epsilon-constants for Weil-Deligne representations that the Macdonald correspondence is required to match.","marker":"[Del]"}],"fun_headline_variants":["Macdonald's bijection is the unique pair-epsilon preserver","Only Macdonald's map keeps all pair epsilon-factors","Pair-epsilon uniqueness forces Macdonald's correspondence","Macdonald's map: the only pair-epsilon preserving bijection"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The proof assumes that epsilon-factors for pairs remain unchanged when one member is chosen by the candidate bijection and the other by Macdonald's map, and this mixed equality is not derived in the text.","fun_headline_variants_meta":{"raw":{"variants":["Macdonald's bijection is the unique pair-epsilon preserver","Only Macdonald's map keeps all pair epsilon-factors","Pair-epsilon uniqueness forces Macdonald's correspondence","Macdonald's map: the only pair-epsilon preserving bijection"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000558,"raw_usage":{"total_tokens":2609,"prompt_tokens":855,"completion_tokens":1754,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":471,"completion_tokens_details":{"reasoning_tokens":1696}},"tokens_in":471,"tokens_out":1754,"duration_ms":13144,"temperature":1.0,"reasoning_tokens":1696,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-10T22:14:11.358320+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute the pair epsilon-factor with one representation taken from the Macdonald correspondence and the other from a candidate bijection, for a small rank where the two could disagree; any mismatch would show that the uniqueness proof as written does not go through.","supporting_citations":[],"review_version":1}