{"id":"551e76c9-6a59-4a73-8f88-69bf02f53aaf","arxiv_id":"2601.00052","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"For Galois normic bundles N_{K/k}(z)=P(x), the vertical unramified Brauer group is isomorphic to an explicit quotient of character groups determined by the Galois group G and the multiplicities of the irreducible factors of P.","lead":"An arithmetic-geometry preprint derives explicit formulas for the vertical unramified Brauer group of varieties defined by norm equations N_{K/k}(z)=P(x) when K/k is Galois. The result converts a cohomological description into a combinatorial quotient built from the Galois group of K/k and the irreducible factors of P(x).","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Unproven reduction to n|deg P in §3 is the load-bearing gap: it is needed for the U1/U2 gluing and to preserve Br_vert, but no fiber-preserving birational map can change the degree modulo n.","rationale":"The reader's weakest assumption correctly identifies the unproved reduction to n|deg P (and e_i<n). My analysis shows this is indeed the most load-bearing step: the proof of Theorem 4.1 relies on Wei22's partial compactification, whose construction uses the gluing z'=z/x^{m/n}, which is not algebraic when n∤m. The paper merely asserts a birational reduction without specifying the map or proving it preserves the vertical Brauer group. In fact, for the vertical part the reduction must be fiber-preserving, but a fiber-preserving birational map over P^1 preserves the set of degenerate fibers. For a conic bundle over P^1 arising from a normic bundle with n=2, the degenerate fibers are precisely the roots of P; the fiber at infinity is smooth. Therefore the parity of deg P (the number of roots over \\bar{k}) is a fiber-preserving birational invariant. A cubic (odd degree) cannot be fiberwise birational to any even-degree polynomial. A non-fiber-preserving map would change the base field k(x), and hence Br_vert = Br ∩ Br(k(x)) would not be preserved. Thus the reduction to n|deg P is either false or changes the object being computed. This means Theorem 4.1 does not, as written, establish the claimed computation for all Galois normic bundles; it is at best proven for the case n|deg P. The e_i<n reduction is less problematic, since it can be achieved by absorbing a factor P_i(x) into one of the norm variables, but it is also stated without proof. Since the reader already gave a conditional verdict, my read does not change the verdict: still CONDITIONAL, pending a proof of the reduction or a restriction of the theorem's scope. If the reduction turns out to be false, the theorem would need to be stated for n|deg P only.","tokens_in":7024,"tokens_out":43402,"duration_ms":419221,"concrete_test":"Take k=Q, K=Q(i) (n=2), P(x)=x^3-x. (1) Independently compute Br_vert of the conic bundle X: y^2+z^2=x^3-x using the standard conic-bundle residue sequence (cf. [CTHS04]); the expected group is Z/2. (2) Enumerate all even-degree Q(x) that could arise from the claimed reduction (e.g., x^4-x^2, x^4-1, x^2(x^2-1), (x^2-1)^2, and their Möbius transforms) and compute the right-hand side of Theorem 4.1 for each. If no Q yields a group isomorphic to Z/2, the reduction cannot preserve Br_vert, settling the concern. Alternatively, verify the parity argument that the number of degenerate fibers is invariant under fiber-preserving birational maps, which already rules out such a reduction for n=2.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The paper's Section 3 asserts 'We will always assume that n|m, otherwise U1 is birational to an affine variety of this form where the degree of the polynomial on the right side is divisible by n' and 'we can always assume that ei<n', without proof. The subsequent glueing of U2 via z'=z/x^{m/n} is algebraic only if m/n is an integer, so the entire construction of the partial compactification X (and hence Theorem 3.1 and Theorem 4.1) depends on n|m. For the vertical part, the reduction must preserve the fibration over the x-line. A fiber-preserving birational map induces an automorphism of P^1 and therefore preserves the set of degenerate fibers of the conic bundle N_{K/k}(z)=P(x). For n=2, the degenerate fibers are exactly the roots of P; the fiber at infinity is smooth because the homogenized equation at infinity is N(z)=c x^m with c≠0. Hence a cubic P has three degenerate fibers over \\bar{k}, while any even-degree Q has an even number. Thus no fiber-preserving birational map can relate deg P odd to deg P even. A non-fiber-preserving map would change the subfield k(x), so Br_vert(X)=Br(X)∩Br(k(x)) is not birational invariant. Therefore Theorem 4.1, as stated for 'Galois normic bundles' generally, is not justified for n∤deg P; at best it applies to the case n|deg P.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies the vertical unramified Brauer group of a Galois normic bundle, i.e. the variety N_{K/k}(z)=P(x) with K/k a finite Galois extension of degree n. It recalls Wei's construction of a partial compactification X and Wei's exact sequence for Br(X)/Br(k), then proves Theorem 4.1, which gives a purely combinatorial description of Br_vert(X)/Br(k) as a quotient of a subgroup of ⊕ H^1(G_i,Q/Z) by the diagonal restrictions of characters of G=Gal(K/k). The description involves the multiplicities e_i of the irreducible factors of P, the fields L_i=k[x]/(P_i), and the intersections L_i∩K. The proof is a cohomological computation using Shapiro's lemma, Mackey's formula, and corestriction/restriction maps.","tokens_in":7408,"tokens_out":19686,"duration_ms":199247,"significance":"If Theorem 4.1 is correct, it is a useful and genuinely general result: it reduces the computation of the vertical unramified Brauer group of Galois normic bundles to finite-group cohomology and the factorization data of P. This goes beyond previous work, which treated special cases. The formula is internally consistent (the denominator is contained in the numerator precisely because n|m forces m·χ=0) and there are no fitted parameters or circular dependencies: the argument starts from Wei's external exact sequence. Because the paper is short and telegraphic, however, several load-bearing identifications and the birational reduction are asserted rather than proved; those gaps need to be filled before the theorem can be accepted as stated.","major_comments":[{"comment":"The paper states 'We will always assume that n|m ... Similarly, we can always assume that e_i<n' without proof. This assumption is load-bearing: the gluing of U_2 uses z'=z/x^{m/n}, which is algebraic only if m/n is an integer, and the compatibility of the numerator and denominator in Theorem 4.1 uses m·χ=0, i.e. n|m. If the reduction is meant to be by a birational transformation that changes the base coordinate (e.g. inversion), the invariance of Br_vert(X)=Br(X)∩Br(k(x)) is not automatic and must be proved. If the reduction is not valid, the theorem applies only when n|deg P and e_i<n. The author should either supply a precise argument for the reduction or state Theorem 4.1 with these hypotheses explicitly.","section":"§3, p. 3 (reduction assumptions)"},{"comment":"The proof begins by asserting 'By Lemma 2.1, one can easily check that H^1(k,Z_P)=H^1(k,D)=0' and then identifies the kernel of H^2(k,Z_P)→H^2(k,Z_P⊗Z[K/k]) with ⊕ \\hat G_i. This is a central step and is not demonstrated. In particular, the double-coset decomposition of Res_{\\Gamma_{L_i}} Ind_{\\Gamma_K}^{\\Gamma_k} Z has [L_i:L_i∩K] summands, and the map from H^1(L_i,Q/Z) to each summand is a restriction map. The conclusion that the kernel is exactly ⊕ H^1(Gal(KL_i/L_i),Q/Z) should be written out; as it stands, the passage from 'the map is just the restriction map' to the stated kernel is too fast for a claim on which the main theorem depends.","section":"§4, proof of Theorem 4.1 (H^1(k, \\hat T') identification)"},{"comment":"The diagram used to prove that the map from H^2(k,Z_P) to H^2(k,Z) becomes l_i·Cor_i on \\hat G_i is garbled in the text and the equality '\\hat G_i = H^1(L_i,Q/Z)' is not literally true: \\hat G_i is naturally a subgroup of H^1(L_i,Q/Z) via inflation, not the whole group. The relation Cor_{L_i/k}∘Inf = l_i·Cor_i needs a precise statement and proof. Without this, the condition ∑ l_i·Cor_i(χ_i)=0 in the numerator of Theorem 4.1 is not rigorously established.","section":"§4, proof of Theorem 4.1 (corestriction condition)"},{"comment":"The proof equates the kernel of the map to H^2(k,S) with the subgroup \\hat G'_i of characters vanishing on elements of order dividing e'_i. The argument depends on properties of the action of G_i on the set Ω_i of e'_i-element subsets of G. The assertions 'the exponent of G_{i,M} is dividing e'_i' and 'for any g of order dividing e'_i there exists M with g∈G_{i,M}' are only stated as 'one can check'. They are true for the left-regular action, but the proof should indicate that the action is free outside the identity, so that invariant subsets are unions of cycles of length ord(g). This is a necessary step and should be made explicit.","section":"§4, proof of Theorem 4.1 (definition of \\hat G'_i)"}],"minor_comments":[{"comment":"The notation \\tilde P(x') is introduced as x'^m P(1/x'), but the letter eP later appears as 'eP(x ′)'. The text should be consistent.","section":"§3, p. 3"},{"comment":"The theorem statement does not explicitly list the assumptions n|deg P and e_i<n; these appear only in the earlier 'We will always assume...' sentence. The theorem should be rephrased so that its hypotheses are self-contained.","section":"§4, Theorem 4.1"},{"comment":"There are numerous typos and formatting errors: 'Leemma', 'arguement', 'corestristion', 'bG' versus '\\widehat G', and the commutative diagram in the middle of the proof is badly typeset. These should be corrected.","section":"Throughout"},{"comment":"The Mackey formula is stated for a discrete N-module M, but the special case 'if M is a discrete G-module with trivial G-action' is used later; the notation in this special case is a little loose and could be clarified.","section":"§2, Lemma 2.3"}],"recommendation":"major_revision","confidential_remarks":"The core computation is plausible and likely correct, but the manuscript is too compressed for the claims it makes. The main theorem rests on a birational reduction that is asserted without proof, and the key cohomological identifications are sketched rather than demonstrated. I would encourage the editor to ask for a revised version with these points filled in; if the author can supply the missing arguments, the paper would be a solid contribution. The paper's novelty is incremental (building on Wei's framework), but the resulting explicit formula is useful enough for publication in an appropriate journal."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Short version: the main formula is a real extension, and the cohomology part hangs together; but the opening reduction in §3 is not justified and, as stated, likely false. I would send it to a referee, with a request that the reduction be proved or the theorem restricted.\n\nWhat's new: Theorem 4.1 turns Wei's abstract quotient H^1(k,\\hat T')/j^*H^1(k,\\hat T) into a finite abelian group written in terms of G, the irreducible factors of P, and the intersections L_i∩K. That is genuinely useful: if you need Br_vert for a Galois normic bundle, you can compute characters instead of chasing cohomology. The proof is a standard Shapiro/Mackey computation and it is coherent: the restriction/corestriction relations and the G'_i condition line up, and the n|m hypothesis makes the denominator sit inside the numerator. I found no circularity, and the self-citation [LL25] is only contextual, not load-bearing. The paper is honestly built on Wei22's exact sequence, and if that sequence is correct, the computation is likely right in the reduced case.\n\nThe soft spot is the one the stress-test flags. Section 3 says \"We will always assume that n|m ... and e_i<n\" without proof, and the gluing z' = z/x^{m/n} is algebraic only when m/n is an integer. The natural coordinate change—multiplying z by x^r—multiplies the right-hand side by x^{rn}, so it preserves m mod n. A Möbius change of the base changes the relevant degree to the order of P at the point sent to infinity, which is not always 0 mod n; for n=2 and a cubic P, no point of the divisor has even order, so no fiber-preserving birational model gives even degree. I think the stress-test's degenerate-fiber argument points the same way, even if its exact counting at infinity is a bit quick. So the theorem as stated for all Galois normic bundles is not supported. Either provide the birational map (I doubt there is one in general) or state the theorem for n|deg P and e_i<n.\n\nThere are also several \"one can easily check\" steps in §4—the identification of the map to H^2(k,Z) as corestriction, the stabilizer exponent argument—that a referee will want expanded. They looked routine, not fatal.\n\nWho it is for: people working on Brauer–Manin obstructions for normic bundles. It is a real step beyond Wei22's abstract sequence. Once the reduction is fixed or the statement narrowed, I would cite it. Recommendation: engage, send to peer review, and make the verdict conditional on §3.","headline":"A genuine extension of Wei's vertical Brauer computation, but the §3 reduction to n | deg P is asserted without proof and probably false in the stated generality; the theorem should be read as conditional on that reduction.","tokens_in":784,"tokens_out":920,"would_cite":true,"duration_ms":189555,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["14F22"],"pacs":[],"model":"deepseek-v4-flash","headline":"The vertical unramified Brauer group of a Galois normic bundle is isomorphic to a quotient of Galois character groups determined by the factor multiplicities of P, and is therefore explicitly computable from the equation.","keywords":["normic bundle","vertical unramified Brauer group","Brauer group","Galois extension","character groups","norm form","Brauer–Manin obstruction","cohomology of tori"],"falsifier":"For a cyclic extension K/k of prime degree p and a polynomial P with deg P not divisible by p, compute Br_vert(X)/Br(k) from the smooth model available for cyclic extensions, then redo the calculation after applying the paper's reduction to n | deg P; if the two groups differ, the reduction changes the unramified Brauer group and Theorem 4.1 applies only to the reduced variety.","tokens_in":6906,"feed_emoji":"🧮","tokens_out":7812,"duration_ms":71096,"temperature":0.7,"pith_summary":"This paper computes the vertical part of the unramified Brauer group of a Galois normic bundle, a variety given by an equation of the form Norm_{K/k}(z)=P(x) with K/k a Galois extension and P a one-variable polynomial. The main theorem gives a closed formula: the group is isomorphic to a quotient of a direct sum of character groups attached to the Galois groups of the fields generated by the irreducible factors of P, with a single corestriction condition in the numerator and the images of restrictions in the denominator. This is the first systematic computation of such groups for an arbitrary Galois extension, turning a cohomological construction into data read directly off the equation. A sympathetic reader should care because the unramified Brauer group is the standard obstruction for counting rational points on these rationally connected varieties, so an explicit formula is a concrete step toward Brauer–Manin computations.","feed_headline":"Brauer groups of Galois normic bundles reduced to character data","feed_subtitle":"A quotient of Galois character groups determines the vertical unramified Brauer group, making it computable from the polynomial's factors.","key_machinery":"The load-bearing object is the exact sequence reviewed in Section 3, which identifies Br_vert(X)/Br(k) with H^1(k, bT')/j*H^1(k, bT), where T is the norm torus R_{K/k}G_m and bT' is a character module modified by the multiplicities and degrees of the factors of P. The computation proceeds by Shapiro's lemma (writing H^1 of induced modules as cohomology of the fields L_i and KL_i), Mackey's formula (decomposing the restriction of the induced module along double cosets), and the observation that the three kernel conditions defining bT' collapse to the single equation Σ l_i Cor_i(χ_i)=0. The denominator is then identified with the image of the restriction maps multiplied by the exponents e_i.","core_discovery":"The paper's central claim is Theorem 4.1: after a reduction to the case where [K:k] divides deg P and every multiplicity e_i of an irreducible factor satisfies e_i < [K:k], the vertical unramified Brauer group modulo constants is isomorphic to the quotient { (χ_i) ∈ ⊕ Ĝ'_i : Σ l_i·Cor_i(χ_i)=0 } / { (e_i·Res_i(χ))_i : χ ∈ Ĝ }, where G=Gal(K/k), G_i=Gal(KL_i/L_i), L_i is the field of a root of the i-th factor, l_i=[L_i : L_i∩K], Ĝ=H^1(G,Q/Z), and Ĝ'_i holds characters of G_i vanishing on elements of order dividing gcd(e_i,n). In words, the group is determined by the Galois group of K/k and the factorization data of P, with no further geometric input.","pith_inferences":["If the reduction n | deg P is truly harmless, the formula implicitly extends to all P by first replacing P with a birationally equivalent polynomial; checking this for cyclic extensions of prime degree, where independent smooth compactifications exist, would turn the paper's normalization step into a proven statement.","A natural next step is to compute the remaining X^2ω term in the exact sequence using the same character-group data, which would complete the full unramified Brauer group rather than just its vertical part.","For number fields, the formula suggests that Br_vert(X) depends only on the Galois group and the fields L_i, not on other invariants of k; testing this against explicit Brauer–Manin computations on small normic bundles would clarify how much arithmetic of k survives in the quotient."],"forward_implications":["The vertical unramified Brauer group of any Galois normic bundle is finite and computable from the Galois group of K/k, the fields L_i, the multiplicities e_i, and the intersections L_i∩K.","The formula specializes to the previously known cases (P irreducible with K/k abelian, and deg P=2 with [K:k]=4) as instances of one uniform statement.","The quotient structure shows how the group varies with the multiplicities: changing an exponent e_i changes only the character subgroup Ĝ'_i and the denominator factor, so the group can be read off for families of polynomials.","Because the denominator is exactly the image of restrictions from Ĝ, the formula isolates which numerator classes are constant (from the base field), exposing the genuinely new part of the obstruction."],"fun_headline_variants":["Normic bundle Brauer groups reduced to Galois and factor data","Vertical unramified Brauer group from Galois group and polynomial factors","Brauer group of Galois normic bundles computable from character data","Theorem: Brauer group determined by Galois structure and factor multiplicities","Galois normic bundles: vertical Brauer group via combinatorial formula"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The paper assumes, with a one-line justification, that one can always reduce to the case where the degree of P is divisible by [K:k] and every multiplicity e_i is less than [K:k] without changing the unramified Brauer group; the birational equivalence proving this is not supplied, and the theorem is stated only under this reduction.","fun_headline_variants_meta":{"raw":{"variants":["Normic bundle Brauer groups reduced to Galois and factor data","Vertical unramified Brauer group from Galois group and polynomial factors","Brauer group of Galois normic bundles computable from character data","Theorem: Brauer group determined by Galois structure and factor multiplicities","Galois normic bundles: vertical Brauer group via combinatorial formula"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000187,"raw_usage":{"total_tokens":1113,"prompt_tokens":638,"completion_tokens":475,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":382,"completion_tokens_details":{"reasoning_tokens":381}},"tokens_in":382,"tokens_out":475,"duration_ms":4374,"temperature":1.0,"reasoning_tokens":381,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-03T13:12:10.736077+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"For a cyclic extension K/k of prime degree p and a polynomial P with deg P not divisible by p, compute Br_vert(X)/Br(k) from the smooth model available for cyclic extensions, then redo the calculation after applying the paper's reduction to n | deg P; if the two groups differ, the reduction changes the unramified Brauer group and Theorem 4.1 applies only to the reduced variety.","supporting_citations":[],"review_version":1}