{"id":"cd3473bf-31fa-4389-9942-8afc5cf522b8","arxiv_id":"2509.02300","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":8.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"The Isaacs-Navarro Galois conjecture is proved: for every finite group G and prime l, there is an H0-equivariant bijection between the l'-degree characters of G and those of the normalizer of a Sylow l-subgroup.","lead":"This paper proves the Isaacs-Navarro Galois conjecture, a 2002 strengthening of the McKay conjecture in the representation theory of finite groups. The result gives a Galois-equivariant bijection between two character sets for every finite group and every prime.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Lemma 6.1's D4(q) case depends on an unverified MAGMA assertion; if the stated 3-structure is wrong, the local extension maps and non-regular inductive step for D4(q) are not established.","rationale":"The reader's weakest_assumption is exactly Lemma 6.1, and I agree that this is the most load-bearing technical step in the manuscript. The proof of Theorem A rests on the H0-equivariant extension maps and transversals of Sections 6–7, which in turn rely on the structural decomposition of N-hat = N_{GE(GF)}(L) asserted in Lemma 6.1. That lemma is a sweeping statement over all Lie types, and its proof mixes references to prior work with one unverified MAGMA computation. The D4(q) case is not a finite exceptional case: it concerns infinitely many groups and primes with d = 3 or 6. If the MAGMA assertion is wrong, the conclusion of Lemma 6.1 fails for those groups, and the proof of the inductive Isaacs–Navarro condition for D4(q) collapses. Since the manuscript explicitly states the computation without providing an artifact, the appropriate verdict remains CONDITIONAL: the authors should supply the code/certificate or replace the computation with a proof. I do not see a more fundamental internal inconsistency; the deductive structure from Lemma 6.1 to Theorem A is coherent, and the reliance on [RSST25] and [CS25] is clearly flagged. Therefore my read does not change the reader's conditional verdict.","tokens_in":23266,"tokens_out":37916,"duration_ms":371529,"concrete_test":"Run an independent computation (GAP/MAGMA) reproducing the D4 extended Weyl group and the group V0Γ used in Lemma 6.1; compute a Sylow 3-subgroup V̂3, verify V̂3 ≅ C3×C3, exhibit an element v ∈ V̂3∩V0 with v³ ∈ H0 that is a Sylow 3-twist, and check that v is γ-stable for the graph automorphism γ of order 2. Publish the script and output. If the computation fails, Lemma 6.1 (and hence Corollaries 6.2/6.4, Lemma 6.7, Lemma 6.11, and Proposition 7.2) is unsupported for the D4(q) family.","verdict_should_be":"UNCHANGED","load_bearing_attack":"In the proof of Lemma 6.1 (Section 6.1, page 9), for G = D4(q) with d ∈ {3,6}, the authors write: “A calculation in MAGMA shows that V̂3 ≅ C3×C3.” No code, input, or certificate is supplied. This is not a peripheral remark: the asserted structure is used to select a Sylow 3-twist v and build V̂0, Ê0, from which Corollary 6.2 derives the Sylow ℓ-subgroup decomposition D = Z(L)ℓ ⋊ V0 and the equality N_Ĝ(D) = N_G(D)Ê. Those consequences are then used in Corollary 6.4, Lemma 6.7, Lemma 6.11, and Proposition 7.2 — i.e., in the extension maps and transversals needed for the non-regular case of the inductive Isaacs–Navarro condition. If the MAGMA assertion is false or the construction of v is invalid, the proof does not cover the infinite family D4(q) with ℓ | q²+q+1 or ℓ | q²−q+1 (d = 3 or 6). The surrounding Lie-type arguments are dense and cite prior work, but this particular case is the only spot where a computational claim is made without a reproducible artifact; the correctness of Lemma 6.1 is load-bearing for the rest of the proof.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proves the Isaacs–Navarro Galois conjecture (Theorem A): for every finite group G and every prime ℓ, with D ∈ Syl_ℓ(G), there is an H_0-equivariant bijection Irr_{ℓ′}(G) → Irr_{ℓ′}(N_G(D)), where H_0 is the Galois group acting trivially on ℓ′-roots of unity. The proof follows the standard reduction of Navarro–Späth–Vallejo to a condition on simple groups, handles the non-Lie-type and defining-characteristic cases via prior results, and then treats groups of Lie type in non-defining characteristic. For Lie-type groups the authors use the criteria of Ruhstorfer–Schaeffer Fry–Späth–Taylor and construct the required extension maps and transversals, with a new extension lemma (Lemma 6.1) and an induction argument for the non-regular case (Theorem 7.4). Corollaries B and C are derived, giving the exponent-of-Sylow-abelianization statement and Hung’s p-rationality-level conjectures.","tokens_in":23572,"tokens_out":5067,"duration_ms":58045,"significance":"If the proof is correct, this settles a central conjecture from 2002 and provides substantial corollaries on Galois actions on character values, extending the completed McKay conjecture. The paper is well-organized, builds on the recent proof of the McKay conjecture by Cabanes–Späth and on the reduction work of Navarro–Späth–Vallejo and Ruhstorfer–Schaeffer Fry–Späth–Taylor, and it isolates the new Lie-type extension problems clearly. The internal logic is consistent: the induction on the order of G/Z(G) is legitimate, and the paper does not assume the target conjecture. The main weakness is that one load-bearing step, the D_4(q) case of Lemma 6.1, relies on an unreproduced MAGMA computation. Because that step supports several later results used for the non-regular inductive step, the manuscript needs to make this computation verifiable before the proof can be considered complete.","major_comments":[{"comment":"The proof of Lemma 6.1 for G = D4(q) with d ∈ {3,6} contains the sentence: “A calculation in MAGMA shows that V̂3 ≅ C3×C3.” No MAGMA code, input file, or output is supplied. This is not a peripheral remark: the asserted structure is used to choose a Sylow 3-twist v, then to construct V̂0 and Ê0. Corollary 6.2 derives from this the Sylow ℓ-decomposition D = Z(L)_ℓ ⋊ V0 and the equality N_Ĝ(D) = N_G(D)Ê; those consequences are subsequently used in Corollary 6.4, Lemma 6.7, Lemma 6.11, and Proposition 7.2. If the computational assertion is false, the extension maps and transversals for the infinite family D_4(q) with ℓ | q²±q+1 are not established. Please provide the MAGMA code and transcript, or a hand proof of the claim that a Sylow 3-subgroup of V0Γ is C3×C3.","section":"Section 6.1, Lemma 6.1, D4(q) case (page 9)"},{"comment":"The line “By our assumption and applying [NSV20, Thm. A], the Isaacs–Navarro Galois conjecture holds for the group W~λ” is compressed. The induction hypothesis covers simple groups involved in groups of order smaller than |G/Z(G)|. For the argument to be valid, the relative Weyl group W~λ = N_G~(L, λ~)/L~ must have all its composition factors within that class. Since W~λ is a section of N_G~(L)/L~, which is a subgroup of the Weyl group (up to the regular-embedding center), its order is bounded uniformly in q and is indeed smaller for all but possibly trivial small q. Please spell out this order comparison explicitly, as the induction step is the mechanism that completes the proof for all non-regular d.","section":"Section 7.2, proof of Theorem 7.4"}],"minor_comments":[{"comment":"The Galois group and the finite group are both denoted G: “the Galois group G := Gal(Q^ab/Q) acts naturally on Irr(G).” This is confusing; use a different letter, e.g. 𝒢, for the Galois group.","section":"Introduction, Section 1"},{"comment":"Typo: “Isaccas–Navarro” should be “Isaacs–Navarro”.","section":"Section 8"},{"comment":"The notation V̂3 is introduced as “a Sylow 3-subgroup of V0Γ”, but the sentence “A calculation in MAGMA shows that V̂3 ≅ C3×C3” would benefit from stating explicitly which group (V0Γ or W0Γ) the computation concerns.","section":"Section 6.1, Lemma 6.1"},{"comment":"Some inline formulas are poorly broken across lines (e.g. “N G(D) = NG(D)Ê” and “L NG(D)”). Please improve typesetting for readability.","section":"Throughout"}],"recommendation":"major_revision","confidential_remarks":"The paper is a strong and important contribution, and the overall strategy is credible. The only reason I recommend major revision rather than acceptance is the unreproduced MAGMA computation in Lemma 6.1, which is load-bearing. If the authors provide the code/certificate and add the small clarification in Theorem 7.4, I would be happy to support acceptance."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Bottom line: this is the real thing. The paper proves the Isaacs–Navarro Galois conjecture for every finite group and every prime, which is a major theorem in finite group representation theory. The structure is exactly what you'd hope: NSV20 reduction, then the heavyweight Lie-type work in Sections 5–7. The genuinely new content is the H0-equivariant extension maps and transversals for non-type-A groups, especially Lemmas 6.6, 6.7, 6.10, and Proposition 7.2, and the inductive argument in Theorem 7.4 that handles the non-regular case. The type-A case and the regular case fall out cleanly once that machinery is up. The paper is also honest about what it takes from [RSST25] and [CS25], and the internal logic is coherent as far as I can check.\n\nThe soft spots, in proportion. The proof leans on two to-appear or unrefereed companions, [RSST25] and [CS25]. That is common in this area and not a mathematical flaw, but a referee needs stable versions and a precise statement of which hypotheses of Theorem 5.1 come from where. The more concrete issue is Lemma 6.1. For G = D4(q) with d = 3 or 6, the authors write that a MAGMA calculation shows V̂3 is C3×C3, and no code or certificate is provided. The stress-test note is right that this is load-bearing: Corollary 6.2’s Sylow decomposition and Proposition 7.2’s extension maps for that D4(q) family depend on it. If the 3-structure were wrong, the proof would not cover that infinite family. I think the assertion is very likely true and checkable in an afternoon, but “a calculation in MAGMA” is not a proof. This is a minor-to-moderate gap, not a fatal one. The central argument holds up.\n\nOne point in favor worth naming: the paper does not oversell. It proves the H0 version, not the full Hℓ McKay–Navarro conjecture, and says so clearly. The corollaries are genuine consequences, and the discussion of what does not follow is a model of restraint.\n\nMy verdict is close to the reader’s conditional. I would send this to a serious referee. Before publication I would ask for either the MAGMA input or a short proof for the D4(q) claim, and for stable links to the companion papers. Desk rejection would be wrong.","headline":"Full proof of the Isaacs–Navarro Galois conjecture for all primes; solid Lie-type machinery, but the D4(q) MAGMA assertion and unrefereed companions need referee scrutiny.","tokens_in":24098,"tokens_out":2617,"would_cite":true,"duration_ms":26639,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["20C15","20C33","20D20"],"pacs":[],"model":"deepseek-v4-flash","headline":"This paper proves the Isaacs–Navarro Galois conjecture: for every finite group and prime ℓ, an H0-equivariant bijection exists between the ℓ′-degree irreducible characters and the characters of the normalizer of a Sylow ℓ-subgroup.","keywords":["Isaacs–Navarro Galois conjecture","McKay–Navarro conjecture","Galois action on characters","irreducible characters","groups of Lie type","ℓ-rationality","local–global conjectures","Sylow subgroups"],"falsifier":"Exhibit a prime ℓ and a finite reductive group G^F, e.g. a D4(q) case or an exceptional family not covered by the direct argument, for which no subgroup V-hat of N-hat = N_{GE(G^F)}(L) satisfies N-hat = L V-hat with V-hat ∩ L a 2-group contained in Z(L) and an ℓ′-subgroup E-hat ≤ V-hat with N-hat = N E-hat. A computer search using the Steinberg presentation of the group would settle the existence. If such a group exists, Corollary 6.4's extension maps fail, so the inductive Isaacs–Navarro condition for that simple group would not follow.","tokens_in":23118,"feed_emoji":"","tokens_out":9410,"duration_ms":101220,"temperature":0.7,"pith_summary":"This paper proves the Isaacs–Navarro Galois conjecture, the 2002 refinement of the McKay conjecture that tracks Galois fields of values alongside character degrees. The main theorem asserts that for any finite group G and prime ℓ, there is a bijection between the irreducible characters of ℓ′-degree in G and those in the normalizer of a Sylow ℓ-subgroup D, and this bijection commutes with the Galois automorphisms in H0. Since H0 controls which ℓ′-roots of unity appear in character values, the bijection preserves more than degrees: it matches fields of values up to the ℓ-part of their conductors. From this the paper derives consequences that had remained open for odd primes, including an equivalence between the exponent of D/D′ and the fixed points of certain Galois automorphisms, and a conjecture about the ℓ-rationality levels of ℓ′-degree characters. The reason to care is that these are local–global statements: data visible in the ordinary character table of G now determine structural facts about Sylow subgroups.","feed_headline":"Isaacs–Navarro Galois conjecture proven for all finite groups","feed_subtitle":"A Galois-symmetric character bijection now exists for every prime ℓ, settling the odd-ℓ cases that had stayed open.","key_machinery":"The machinery is the H0-equivariant extension map assembled via a structural splitting of the normalizer of a Sylow d-torus centralizer. Lemma 6.1 asserts that for N-hat = N_{GE(G^F)}(L), there is a subgroup V-hat with N-hat = L V-hat and V-hat ∩ L a 2-group contained in Z(L), plus an ℓ′-subgroup E-hat inside V-hat with N-hat = N E-hat. This splitting lets Lemma 3.2 extend characters from L to their inertia groups while preserving their stabilizers under H0 and the relevant automorphisms. Around that core, the proof organizes the Lie-type case through the reduction criteria of [RSST25] and [NSV20], reducing the inductive condition to three manageable requirements: an equivariant extension ma","core_discovery":"The central claim, Theorem A, is that the Isaacs–Navarro Galois conjecture holds for all finite groups and all primes ℓ. Explicitly, if D is a Sylow ℓ-subgroup of G, there is an H0-equivariant bijection Irr_{ℓ′}(G) → Irr_{ℓ′}(N_G(D)), where H0 is the Galois group of automorphisms of Q^ab that act trivially on ℓ′-roots of unity and have ℓ-power order. The proof uses the reduction of the Galois–McKay family to simple groups and verifies the resulting inductive Isaacs–Navarro condition. The bulk of the verification concerns finite groups of Lie type in non-defining characteristic, where the authors construct the required extension maps and stable transversals, treating type A and regular d-valu","pith_inferences":["The proof does not by itself advance the stronger H_ℓ version of the McKay–Navarro conjecture; H0 cannot see the difference between ℓ-rationality levels 0 and 1, so the ℓ-rationality gap phenomenon remains outside its reach.","A direct application one could test now is algorithmic: in concrete small groups, use the corollary to recover exp(D/D′) from the character table, which would make the classical character-table detection question computable rather than existential.","The splitting lemma used here may transfer to other local–global problems: any future equivariant bijection for Lie-type groups at odd primes will likely need a similar decomposition of N-hat into L times a 2-group with an ℓ′-complement, so the lemma is a reusable structural tool.","One open direction the paper leaves implicit is whether the H0-equivariant bijections constructed here can be upgraded to H_ℓ-equivariant bijections family-by-family; the missing ingredient is control of rationality levels 0 versus 1, not the underlying extension maps."],"forward_implications":["If Theorem A is right, then for every finite group G and prime ℓ, the exponent of D/D′ is bounded by ℓ^e exactly when every ℓ′-degree character is fixed by the Galois automorphism σ_e, and this is equivalent to the same fixed-point statement for the principal ℓ-block (Corollary B).","If Theorem A is right, then ℓ′-degree characters with ℓ-rationality level e ≥ 2 force characters at every intermediate level, and the existence of such characters is invariant under passage to subgroups of ℓ′-index (Corollary C).","If Theorem A is right, the original Isaacs–Navarro Galois conjecture is settled uniformly for odd primes, not just the prime 2 case that was known via separate arguments.","If Theorem A is right, the inductive Isaacs–Navarro condition holds for every nonabelian simple group, so the reduction theorem of [NSV20] can be applied in full scope."],"supporting_citations":[{"why":"introduced the conjecture (their Conj. C) and showed it implies the first equivalence in Corollary B.","marker":"[IN02]"},{"why":"reduction theorem reducing the conjecture to the inductive condition on simple groups; used as Theorem 2.1 with H0.","marker":"[NSV20]"},{"why":"provides the refined Lie-type criteria and the equivariant bijection eΩ that Theorem 5.1 and Section 7 build on.","marker":"[RSST25]"},{"why":"proof of the McKay conjecture, used for the base result and for transversal/extension results such as Thm. 2.18 and Prop. 4.8.","marker":"[CS25]"},{"why":"results on height-zero characters and generic Sylow normalizers, used to restrict to groups with N_G(D) ≤ N.","marker":"[Ma07]"},{"why":"extension maps for exceptional groups and the Sylow d-twist setup used in Lemma 6.8 and Section 6.","marker":"[Sp09]"},{"why":"proves the McKay–Navarro conjecture for ℓ=2, used in Lemma 4.1(5).","marker":"[RSF25]"},{"why":"defining-characteristic case of the Navarro refinement, used in Lemma 4.1(1).","marker":"[Ruh21]"}],"fun_headline_variants":["Galois conjecture proven for all finite groups","Isaacs–Navarro settled for every finite group","All finite groups satisfy Isaacs–Navarro Galois","Galois-symmetric character bijections proven universal","Isaacs–Navarro conjecture holds for all groups"],"cache_read_input_tokens":2688,"weakest_assumption_plain":"The Lie-type part of the proof depends on Lemma 6.1's structural splitting: in every case, the normalizer of the Sylow d-torus centralizer must factor as L times a small group whose intersection with L is a 2-group inside the center, with a complementary ℓ′-piece. If one group family fails that splitting, the extension maps that carry the H0-equivariant bijection cannot be constructed by this proof.","fun_headline_variants_meta":{"raw":{"variants":["Galois conjecture proven for all finite groups","Isaacs–Navarro settled for every finite group","All finite groups satisfy Isaacs–Navarro Galois","Galois-symmetric character bijections proven universal","Isaacs–Navarro conjecture holds for all groups"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000184,"raw_usage":{"total_tokens":1072,"prompt_tokens":579,"completion_tokens":493,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":323,"completion_tokens_details":{"reasoning_tokens":418}},"tokens_in":323,"tokens_out":493,"duration_ms":5262,"temperature":1.0,"reasoning_tokens":418,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-05T11:42:02.223712+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Exhibit a prime ℓ and a finite reductive group G^F, e.g. a D4(q) case or an exceptional family not covered by the direct argument, for which no subgroup V-hat of N-hat = N_{GE(G^F)}(L) satisfies N-hat = L V-hat with V-hat ∩ L a 2-group contained in Z(L) and an ℓ′-subgroup E-hat ≤ V-hat with N-hat = N E-hat. A computer search using the Steinberg presentation of the group would settle the existence. If such a group exists, Corollary 6.4's extension maps fail, so the inductive Isaacs–Navarro condition for that simple group would not follow.","supporting_citations":[],"review_version":1}