{"id":"d29c5538-95bb-4e05-8f14-1d6b95ec03ac","arxiv_id":"1908.05353","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"For symplectic type characters of wildly ramified quadratic extensions of p-adic fields, the epsilon factor sign is χ(-1)^d when the conductor is even and χ(-1)^l times a conjecturally ±1 Gauss sum when the conductor is odd.","lead":"This paper computes the sign of Tate's epsilon factor for a family of characters attached to wildly ramified quadratic extensions of p-adic fields, a case previously left open. The even-conductor case gives a clean answer, plus or minus one from the character evaluated at -1; the odd-conductor case is reduced to a Gauss sum the paper leaves unevaluated.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Equation (3.5) is not merely unproved: Theorem 1.2's quantification over all conductor-0 ψ trivial on F is false, since scaling ψ by units with ω_{K/F}(a)=-1 changes the epsilon factor by a factor ω_{K/F}(a)^{-1}.","rationale":"The reader's weakest assumption is exactly the right one. My stress-test confirms it and adds a decisive consequence: the unproved relation (3.5) is not a harmless gap but an unstated restriction on ψ. The family of additive characters in Theorem 1.2 is closed under scaling by units of F, and the epsilon factor transforms nontrivially under such scaling unless the symplectic character is trivial on those units, which it is not in the wild case. Hence the theorem as quantified cannot be true; a repair would require fixing a distinguished ψ (or replacing the constant χ(-1)^d by the full ψ-transformation law). I also verified the reader's auxiliary correction: the proof's claim that t is odd and t<2e fails for Q_2(√2), where t=2 and e=1. The algebra after (3.5) is fine given the relation, and Theorem 1.1's conductor dichotomy appears sound. Since the flaw is localized and a reformulation may rescue the main idea, I do not change the reader's CONDITIONAL verdict: the paper needs major revision, but rejection would be too harsh if the distinguished ψ and the transformation law are supplied.","tokens_in":9618,"tokens_out":18121,"duration_ms":186980,"concrete_test":"Take F=Q_2, K=Q_2(√2), fix an explicit symplectic type character χ of even conductor 2d (d>t+1), and construct ψ from Lemma 3.1 with n(ψ)=0. Pick a∈O_F^× with ω_{K/F}(a)=-1 (e.g., a=5; check that 5 is not a norm from K). Evaluate (3.5) for ψ and for ψ_a=ψ(a·) at a single representative x=π_K^d: compute χ(1+x), ψ(x/π_K^{2d}), and ψ(ax/π_K^{2d}). If the latter two differ, (3.5) cannot hold for both, so no single c'=π_K^{2d} can serve the whole stated family. Equivalently, compute ε(χ,ψ_a) from ε(χ,ψ) via the scaling law and compare with the theorem's formula; a mismatch confirms the missing unit condition.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The reader's flag on (3.5) is correct and can be sharpened to an internal inconsistency. The proof of Theorem 1.2 asserts that χ(1+x)=ψ(x/π_K^{2d}) for all v_K(x)≥d, but this relation determines the unit part of the Lamprecht–Tate parameter and is not implied by n(ψ)=0 and ψ|_F=1. Now let a∈O_F^× and put ψ_a(x)=ψ(ax). Since ψ=c·(ψ_F∘Tr) with Tr(c)=0, we have ψ_a=(ac)·(ψ_F∘Tr), Tr(ac)=0, and by Lemma 3.1 n(ψ_a)=n(ψ)+2v_F(a)=0; hence ψ_a lies in the same family. The standard local scaling law ε(χ,ψ_a)=χ(a)^{-1}ε(χ,ψ) (or its inverse, depending on convention) gives, because χ|_F=ω_{K/F}, a factor ω_{K/F}(a)^{±1}. But a(ω_{K/F})=t+1>1 in the wild case, so there are units a with ω(a)=-1. If Theorem 1.2 held for both ψ and ψ_a, its ψ-independent conclusion ε=χ(-1)^d would force ω(a)=1 for all units, contradiction. Thus the theorem as stated must be false; (3.5) can hold for at most one member of each scaling class, and that distinguished ψ is never specified. The proof's later assertion 't is odd and t<2e' is also false (e.g., K=Q_2(√2), t=2, e=1), but the scaling contradiction already invalidates the universal quantification.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper studies multiplicative characters χ of K^× for a wildly ramified quadratic extension K/F of a p-adic field of characteristic zero, with χ|_{F^×}=ω_{K/F} (symplectic type). Theorem 1.1 states that the conductor of such χ is either 2t+1 or an even integer strictly larger than 2(t+1), where t is the ramification break of K/F. Theorem 1.2 claims that for even conductor 2d (d>2) and any additive character ψ of K of conductor 0 that is trivial on F, the epsilon factor ε(χ,ψ) equals χ(-1)^d. Theorem 1.3 claims that for odd conductor 2t+1 and any ψ of conductor 2l+1 trivial on F, ε(χ,ψ)=χ(-1)^l G(Q), where G(Q) is a finite Gauss sum. The proofs use the Lamprecht-Tate formula, local duality, and class field theory, with the goal of completing Prasad's explicit sign computation in the wild case.","tokens_in":9835,"tokens_out":25316,"duration_ms":252834,"significance":"The topic is appropriate for a number theory journal: explicit epsilon factors for symplectic-type characters have applications to Tunnell-type results and to arithmetic invariants. The conductor restriction in Theorem 1.1 is useful and appears to follow from standard filtrations. The paper also correctly identifies the Lamprecht-Tate formula as the right tool. However, the main theorems as stated are too strong: the epsilon factor has a well-known dependence on the scaling of the additive character, and the proofs never control the unit part in the Lamprecht-Tate parameter. In consequence the central even-conductor claim is false over the stated family of additive characters, and the odd-conductor formula inherits the same defect. The positive core—the conductor bound and the possibility of choosing a distinguished additive character so that a formula of this shape holds—may well be salvageable, but it is not what the manuscript currently proves.","major_comments":[{"comment":"The proof of Theorem 1.2 asserts without demonstration that c'=π_K^{2d} satisfies χ(1+x)=(c'^{-1}ψ)(x) for all x with v_K(x)≥d. This is exactly the Lamprecht-Tate relation that fixes the unit part of c', and it is not a consequence of n(ψ)=0 and ψ|_F=1. The theorem's quantification over all such ψ is in fact false: for any unit a∈O_F^×, the character ψ_a(x)=ψ(ax) belongs to the same family (Tr(ac)=0 and n(ψ_a)=0), while the standard scaling law gives ε(χ,ψ_a)=ω_{K/F}(a)ε(χ,ψ). In the wild case there is a unit a with ω_{K/F}(a)=-1, so the claimed ψ-independent value χ(-1)^d cannot hold for both ψ and ψ_a. Thus either Theorem 1.2 must be restricted to a distinguished additive character satisfying (3.5), or the formula must be corrected by the factor χ(c') built from the actual Lamprecht-Tate parameter.","section":"Section 3, proof of Theorem 1.2, equation (3.5)"},{"comment":"The same defect is present in Theorem 1.3. The proof fixes c'=π_K^{2(t+l+1)} and uses ψ(c'^{-1})=1 because c'^{-1}∈F. But for a scaled ψ_a with a∈O_F^×, the admissible Lamprecht-Tate parameter is ac', and then χ(ac')=ω_{K/F}(a)χ(c'), while G(Q) is unchanged when c' is replaced by ac'. Hence the asserted identity ε(χ,ψ)=χ(-1)^lG(Q) is not valid for every ψ of conductor 2l+1 in the family; it can hold only after the unit part of c' is selected to satisfy the Lamprecht-Tate relation, and the statement needs to make that selection explicit.","section":"Section 3, Theorem 1.3"},{"comment":"Conjecture 1 is not an open problem given the paper's other claims. If Prasad's theorem, cited in the Introduction, says ε(χ,ψ)=±1 for symplectic-type characters and Theorem 1.3 is true, then G(Q)=χ(-1)^l ε(χ,ψ)∈{±1}. The paper therefore either has a theorem, not a conjecture, after Theorem 1.3, or the proof of Theorem 1.3 has not actually evaluated the sign; in the latter case the advertised completion of Prasad's explicit sign determination is incomplete. Remark 3.4, which identifies G(Q) with the 8th-root-valued gamma factor γ(Q), makes this gap concrete and needs to be reconciled with Conjecture 1.","section":"Section 3, Conjecture 1"}],"minor_comments":[{"comment":"The assertion that 't is odd and t<2e' is false; for K=Q_2(√2) one has t=2 and e=1. The conclusion d>2 needed in the proof follows already from d>t+1 and t≥1.","section":"Proof of Theorem 1.2, paragraph after Theorem 1.1"},{"comment":"The displayed computation ε(χ,ψ)=χ(c')ψ^{-1}(c'^{-1})·G(Q) contains an inverse on ψ; Corollary 2.3(2) gives ψ(c'^{-1}), not ψ^{-1}(c'^{-1}).","section":"Proof of Theorem 1.3"},{"comment":"The application in equation (3.9) does not specify in which ramification context the displayed formula is asserted; please clarify whether it is intended for the wild case or for the tame/unramified discussion that precedes it.","section":"Remark 3.5"}],"recommendation":"reject","confidential_remarks":"The manuscript contains a salvageable core: Theorem 1.1 and the idea of choosing a distinguished additive character satisfying the Lamprecht-Tate relation. But the central theorems are false as stated, so I recommend rejection. A resubmission would need to restate Theorems 1.2 and 1.3 for a distinguished additive character, prove the existence and uniqueness of that character, and either compute G(Q) or explicitly acknowledge the remaining Gauss-sum evaluation."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The paper attacks a genuine gap in Prasad's explicit sign program for symplectic type characters: the wildly ramified quadratic case. Theorem 1.1's conductor dichotomy (odd conductor is exactly 2t+1, even conductor is > 2(t+1)) is new, and the argument using class field theory and norm filtration is mostly sound. The strategy of plugging into the Lamprecht–Tate formula is also sensible, and the paper is clearly written.\n\nThe trouble is Theorem 1.2 as stated. The proof picks c' = pi_K^{2d} and asserts (3.5) with 'Here we see that'—no verification. That relation is load-bearing, and it cannot hold for every psi in the stated family. If a is a unit of F with omega_{K/F}(a) = -1 (which exists in the wild case), then psi_a(x) = psi(ax) still has conductor 0 and is trivial on F, so it belongs to the same family. But the standard scaling law gives epsilon(chi, psi_a) = omega(a)^{-1} epsilon(chi, psi) = -epsilon(chi, psi). If Theorem 1.2 held for all such psi, it would force epsilon(chi, psi_a) = epsilon(chi, psi), a contradiction. So the theorem can only hold for a distinguished psi, and that psi is never specified. The proof's claim 't is odd and t < 2e' is also false; K = Q_2(sqrt(2)) has t = 2, e = 1.\n\nThe odd conductor case, Theorem 1.3, has a related issue: the same asserted relation picks a specific c', and the conclusion still contains an unevaluated Gauss sum G(Q). The abstract says 'explicit signs,' but this is only a reduction to a Gauss sum that is conjectured to be ±1. In fact Conjecture 1 follows immediately from Prasad's theorem that epsilon is ±1, so it is not an independent open problem.\n\nIf the author revises to fix the psi quantification, supply the missing verification of (3.5) for a distinguished psi, and correct the auxiliary claims, the even-conductor theorem could become a real contribution. As it stands, the main theorem is not correct in the stated generality. Still, the underlying idea and the conductor result merit referee attention—this is not a desk-reject. I would encourage a revision, with a careful statement of the chosen psi and a proof of the unit-part compatibility.","headline":"Genuinely open wild-case computation with a sound conductor dichotomy, but Theorem 1.2 as stated is false: unit scaling of psi breaks the claimed psi-independence.","tokens_in":10549,"tokens_out":2305,"would_cite":false,"duration_ms":23287,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["11S37","22E50"],"pacs":[],"model":"deepseek-v4-flash","headline":"Wildly ramified quadratic extensions now have explicit signs for the epsilon factors of all symplectic type characters, with conductor $2t+1$ or even $>2(t+1)$.","keywords":["local fields","epsilon factors","symplectic type characters","wildly ramified quadratic extensions","Lamprecht-Tate formula","Gauss sums","character conductors","p-adic fields"],"falsifier":"Take $F=\\mathbb{Q}_2$ and $K=F(\\sqrt{\\pi_F})$, choose a symplectic type character $\\chi$ of even conductor $8$ and a conductor-zero additive character $\\psi$ trivial on $F$, and compute $\\epsilon(\\chi,\\psi)$ from formula (2.1), comparing it with $\\chi(-1)^4$. Also check whether $\\chi(1+x)=\\psi(\\pi_K^{-8}x)$ holds for all $x$ with $v_K(x)\\ge 4$; a mismatch in the first comparison would refute Theorem 1.2, while a failure of the identity would show the proof's central reduction does not hold.","tokens_in":9177,"feed_emoji":"➕","tokens_out":24639,"duration_ms":228663,"temperature":0.7,"pith_summary":"This paper determines the explicit sign of the local epsilon factor for every character of $K^{\\times}$ of symplectic type, where $K/F$ is a wildly ramified quadratic extension of a $p$-adic field. A symplectic type character is one whose restriction to $F^{\\times}$ is the quadratic character attached to $K$ by class field theory. The paper proves that such a character has conductor either $2t+1$ or an even integer larger than $2(t+1)$, with $t$ the ramification break of $K/F$ (Theorem 1.1). When the conductor is an even number $2d$, the epsilon factor is $\\chi(-1)^d$ for a conductor-zero additive character trivial on $F$ (Theorem 1.2); when the conductor is $2t+1$, it is $\\chi(-1)^l$ times a normalized Gauss sum $G(Q)$ (Theorem 1.3). Together these fill the wild-case gap left in earlier explicit computations of these signs for unramified and tamely ramified quadratic extensions.","feed_headline":"Wild quadratic cases now have explicit epsilon factor signs","feed_subtitle":"Symplectic-type characters now get explicit signs: $\\chi(-1)^d$, or $\\chi(-1)^l$ times a Gauss sum.","key_machinery":"The load-bearing mechanism is the Lamprecht-Tate formula (Theorem 2.2 and Corollary 2.3), which replaces an epsilon factor by $\\chi(c)$ times a normalized sum over unit classes once a suitable $c$ satisfies $\\chi(1+y)=\\psi(c^{-1}y)$ on a deep congruence subgroup. The proof combines this with two structural inputs: the conductor analysis of Theorem 1.1, which uses the norm map's behavior on unit groups in wildly ramified extensions, and the explicit choices $c'=\\pi_K^{2d}$ for even conductor and $c'=\\pi_K^{2(t+l+1)}$ for odd conductor. These choices force $\\psi(c'^{-1})=1$ and $\\chi(c')=\\chi(-1)^d$ in the even case; in the odd case the surviving finite sum is the Gauss sum $G(Q)$.","core_discovery":"The central discovery is that the conductor and epsilon factor of a symplectic type character in a wildly ramified quadratic extension are controlled by the single ramification break $t$. The conductor is either $2t+1$ or an even integer strictly larger than $2(t+1)$. In the even case, the Lamprecht-Tate formula with $c'=\\pi_K^{2d}$ reduces the epsilon factor to $\\chi(\\pi_K^{2d})\\psi(\\pi_K^{-2d})$; because $\\psi$ is trivial on $F$, $\\pi_F=N_{K/F}(\\pi_K)=-\\pi_K^2$, and $\\pi_F$ lies in the norm group, this becomes $\\chi(-1)^d$. In the odd conductor case $2t+1$, the same reduction leaves a normalized finite sum $G(Q)$ over $P_K^t/P_K^{t+1}$, giving $\\epsilon(\\chi,\\psi)=\\chi(-1)^l G(Q)$, where $l$ is fixed by $n(\\psi)=2l+1$. The paper conjectures that $G(Q)=\\pm 1$, which would make every wild symplectic epsilon factor equal to $\\pm 1$.","pith_inferences":["The proof's reliance on equation (3.5) suggests that Theorem 1.2 is established for those additive characters for which that identity holds; extending it to all conductor-zero $\\psi$ trivial on $F$ would require an additional independence argument.","Remark 3.4 identifies $G(Q)$ as a quadratic Gauss sum over the residue field and an 8th root of unity, so the conjecture $G(Q)=\\pm 1$ could be settled by a finite residue-field computation of the associated quadratic character's discriminant.","Explicit local signs of this form are natural inputs for global arguments that need root numbers of symplectic representations, so the wild-case formulas may feed into multiplicity and central-value questions beyond the local statement."],"forward_implications":["In the even wild case the epsilon factor is read off directly from the character: $\\epsilon(\\chi,\\psi)=\\chi(-1)^d$, so no integral needs to be evaluated in applications.","In the odd wild case the remaining object is a single normalized finite sum $G(Q)$; if the paper's conjecture $G(Q)=\\pm 1$ is correct, every wild symplectic epsilon factor is $\\pm 1$.","The conductor classification $a(\\chi)=2t+1$ or even $>2(t+1)$ gives a necessary condition on symplectic type characters in the wild case.","Together with the unramified and tame formulas, the wild formulas complete the explicit description of symplectic epsilon factor signs for all quadratic extensions of $p$-adic fields."],"supporting_citations":[{"why":"Supplies the Lamprecht-Tate formula connecting each epsilon factor to a normalized finite sum over unit classes.","marker":"[9]"},{"why":"Source of the even- and odd-conductor reduction formulas quoted as Corollary 2.3, which the main theorems apply.","marker":"[4]"},{"why":"Establishes that symplectic type epsilon factors are $\\pm 1$ and gives explicit signs in unramified and tame cases, the gap this paper fills.","marker":"[7]"},{"why":"Provides the norm-map facts on unit groups used to prove the conductor classification in Theorem 1.1.","marker":"[8]"},{"why":"Source of the conductor formula for additive characters of the form $c\\cdot(\\psi_F\\circ\\mathrm{Tr}_{K/F})$ used in Lemma 3.1.","marker":"[13]"}],"fun_headline_variants":["Wild symplectic epsilon factors pinned down explicitly","One break t decides epsilon factor sign in wild cases","Epsilon factors for wild symplectic characters: explicit signs","Conductor and epsilon factor from a single ramification break"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is an identity stated without proof in the proof of Theorem 1.2: the chosen additive character must satisfy $\\chi(1+x)=\\psi(\\pi_K^{-2d}x)$ for all $x$ with valuation at least $d$ (equation (3.5)); if that identity fails, the reduction $\\epsilon(\\chi,\\psi)=\\chi(-1)^d$ collapses.","fun_headline_variants_meta":{"raw":{"variants":["Wild symplectic epsilon factors pinned down explicitly","One break t decides epsilon factor sign in wild cases","Epsilon factors for wild symplectic characters: explicit signs","Conductor and epsilon factor from a single ramification break"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000185,"raw_usage":{"total_tokens":1268,"prompt_tokens":838,"completion_tokens":430,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":454,"completion_tokens_details":{"reasoning_tokens":377}},"tokens_in":454,"tokens_out":430,"duration_ms":4037,"temperature":1.0,"reasoning_tokens":377,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T13:19:56.335912+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Take $F=\\mathbb{Q}_2$ and $K=F(\\sqrt{\\pi_F})$, choose a symplectic type character $\\chi$ of even conductor $8$ and a conductor-zero additive character $\\psi$ trivial on $F$, and compute $\\epsilon(\\chi,\\psi)$ from formula (2.1), comparing it with $\\chi(-1)^4$. Also check whether $\\chi(1+x)=\\psi(\\pi_K^{-8}x)$ holds for all $x$ with $v_K(x)\\ge 4$; a mismatch in the first comparison would refute Theorem 1.2, while a failure of the identity would show the proof's central reduction does not hold.","supporting_citations":[{"cited_title":"Tate, Local Constants, Algebraic Number Fields (L-functions and Galois properties), Proceedings of Symposium, Edited by A","cited_arxiv_id":null,"evidence_quote":"Supplies the Lamprecht-Tate formula connecting each epsilon factor to a normalized finite sum over unit classes."},{"cited_title":"Langlands, On the functional equation of the Artin L-function, https://publications.ias.edu/sites/ default/files/a-ps.pdf","cited_arxiv_id":null,"evidence_quote":"Source of the even- and odd-conductor reduction formulas quoted as Corollary 2.3, which the main theorems apply."},{"cited_title":"Prasad, On an extension of a theorem of Tunnell, Compositio Math","cited_arxiv_id":null,"evidence_quote":"Establishes that symplectic type epsilon factors are $\\pm 1$ and gives explicit signs in unramified and tame cases, the gap this paper fills."},{"cited_title":"Serre, Local fields, Graduate texts in mathematics; 67, 1979, Springer-Verlag New York Inc","cited_arxiv_id":null,"evidence_quote":"Provides the norm-map facts on unit groups used to prove the conductor classification in Theorem 1.1."},{"cited_title":"Weil, Basic Number Theory, Third edition, Springer-Verlag, 1974","cited_arxiv_id":null,"evidence_quote":"Source of the conductor formula for additive characters of the form $c\\cdot(\\psi_F\\circ\\mathrm{Tr}_{K/F})$ used in Lemma 3.1."}],"review_version":1}