{"id":"3efe8978-c925-4f61-be17-df562024ca04","arxiv_id":"1908.04833","paper_version":2,"verdict":"ACCEPT","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"low","formal_verification":"none","parameter_count":0,"one_line_summary":"The average twelfth power of Dirichlet L-function central values for odd prime power moduli q=p^n is at most p^A q^{2+ε}, the expected order up to q^ε.","lead":"This paper proves a sharp upper bound on a difficult average associated with prime-power building-block objects in number theory. The result is the prime-power version of a classical moment bound for the Riemann zeta function and complements a recent square-free case.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"No significant objection identified.","rationale":"The reader identified Lemma 2 as the weakest assumption. I agree that the proof cannot proceed without it, but the lemma is a standard quoted result (Postnikov, [11, Lemma 13]) and is used only in the regime n >= 3, where its statement is correct. I therefore do not regard it as a load-bearing concern. The most technically dense part is Proposition 2, so I examined the displayed derivative and the stationary-phase congruence. The gcd and delta_q factors are consistent for Q > 1, and the count of O(1) solutions modulo rt*(Q) follows from the unit square-root identities and Hensel's lemma as claimed. The remaining ingredients: Lemma 1 as a dyadic approximate functional equation, Lemma 5 as the standard completion device, the reduction of the j, h double sum in Section 3.2, and the final fourth-moment/Weyl bookkeeping in Section 6 are all standard and mutually consistent. The small-modulus cases q = p, p^2, and p^3 can be absorbed into the p^A factor using the quoted Weyl bound, so the statement for all n is not exposed by the n = 1 issue in Lemma 2. No targeted numerical or analytical check in my reading exposed a gap, so the reader's ACCEPT verdict should stand.","tokens_in":15737,"tokens_out":43092,"duration_ms":437756,"concrete_test":"For p = 3, q = 3^5, and tilde q = 9, choose two primitive characters chi and chi' modulo q whose associated units satisfy A - A' ≡ 0 (mod 27) but not (mod 81), so that Q = q/tilde q is large; compute the complete sum in (30) for every unit v modulo Q by directly evaluating the K± factors from Lemma 6, and check that |sum| ≤ C Q^{1/2} for a small absolute constant C. A violation would indicate a hidden failure in Proposition 2's stationary-phase count and would undermine Proposition 3.","verdict_should_be":"UNCHANGED","load_bearing_attack":"I read the paper as a proof of the q-aspect twelfth moment for prime powers by transplanting Heath-Brown's and Nunes' architecture into the p-adic stationary phase setting. The logical chain is Theorem 2 -> summation by parts -> Theorem 1, with Proposition 3 as the aggregate short-second-moment engine and Lemma 6 plus Proposition 2 as the arithmetic core. I checked the main dependencies and found no load-bearing flaw. Lemma 2 is quoted from [11, Lemma 13] and is standard for primitive characters modulo p^n with n >= 2; every application here has tilde q > p^2, hence n >= 3, so the harmless n = 1 edge case never enters. The compressed parts of Proposition 2 are consistent: the derivative identity d/du A g±(u/A; tilde q) = 1/s±(u/A; tilde q) checks out, and for Q > 1 the quantity (tilde q, delta_q(chi, chi')) equals delta_q(chi, chi'), so the displayed stationary-phase congruence has the correct p-adic valuation. The reduction to O(1) stationary points modulo rt*(Q) is therefore coherent. I do not have a significant objection to the central claim.","agreement_with_reader":"disagree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proves a q-aspect analogue of Heath-Brown's twelfth moment estimate for Dirichlet L-functions, with the modulus restricted to odd prime powers. The main theorem states that for every odd prime p and every q=p^n, the sum over characters modulo q of |L(1/2,chi)|^{12} is bounded by O_epsilon(p^A q^{2+epsilon}). The proof follows the Heath-Brown/Nunes architecture: it establishes a short second moment S_2(chi), evaluates the relevant complete character sums K_chi by p-adic stationary phase (Lemma 6), proves square-root cancellation for sums of products of these sums with additive twists (Proposition 2), and then aggregates the short second moments over sets of characters (Proposition 3). This yields the large-value estimate Theorem 2, from which Theorem 1 follows by summation by parts. The paper is largely self-contained apart from Lemma 2, which is quoted from Milićević's earlier work and expresses primitive characters on principal units through the p-adic logarithm.","tokens_in":15898,"tokens_out":23239,"duration_ms":224450,"significance":"If the proof is correct, the paper establishes the Lindelöf-consistent upper bound for the twelfth moment in the depth aspect, complementing Nunes' result for smooth square-free moduli. The main innovations are the explicit evaluation of the complete character sum K_chi via p-adic stationary phase and the square-root cancellation estimate for sums of products in Proposition 2; these are substantial technical contributions that should be of independent use. The paper also organizes the Heath-Brown large-sieve argument into a clean general proposition, and the polynomial dependence on p is explicit. The result is exactly the expected q-aspect analogue of the Heath-Brown twelfth moment bound, with the same large-value consequence.","major_comments":[],"minor_comments":[{"comment":"The proof of the key estimate (30) compresses the verification that the phase sigma satisfies the hypotheses of Lemma 3 into the phrases \"Expanding the difference of roots...\" and \"a moment's reflection.\" I recommend expanding this verification, including explicit p-adic valuations of sigma^{(k)}/k! and a clear derivation that the stationary-phase congruence has only O(1) solutions modulo rt*(Q) via the four congruences and Hensel's lemma. The steps appear correct, but this is the central arithmetic input and the present level of detail makes independent verification unnecessarily hard.","section":"4.2, Proposition 2"},{"comment":"Theorem 1 is stated for all characters chi modulo q, while the proof in Section 6 counts only primitive characters. Add an explicit reduction showing that imprimitive characters contribute no more than a p^A q^epsilon factor; for example, use the local factor at p and sum the primitive contribution over conductors d|q. Without this comment the reader must reconstruct the reduction.","section":"1 (Theorem 1) and 6 (proof of Theorem 2)"},{"comment":"The displayed evaluation uses the notation p^{n/2} when n may be odd, which is a real power rather than an integer. Clarify how this is interpreted together with the factors Delta_f, or rewrite the exponent in terms of floor(n/2) and ceil(n/2). This would remove an avoidable source of confusion in a foundational lemma.","section":"2.3, Lemma 4"},{"comment":"Add a remark that when chi = chi' one has delta_q(chi,chi') = q/p and hence Q = 1 for every proper divisor tilde q of q. The \"vanishes unless |v|_p = 1\" assertion therefore implicitly concerns distinct characters; without this remark the diagonal case appears to contradict the statement.","section":"4.2, statement (2) of Proposition 2"},{"comment":"The passage from the double sums over h' and j' to the single sum over m = h'j' with coefficients A(m;p^eta) should state explicitly that each m arises from divisor pairs in the stated ranges, and that the divisor bound gives A(m;p^eta) << m^epsilon. This bookkeeping is currently too terse.","section":"3.1, equations (23)-(24)"}],"recommendation":"minor_revision","confidential_remarks":"I concur with the positive assessment of the paper's correctness. The main recommendation is to expand the verification in Proposition 2 and to clarify the reduction from all characters to primitive characters. These are local and presentation-level; I do not see a load-bearing mathematical error."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Punchline: this is a genuine q-aspect twelfth moment for prime-power moduli, a family where none existed before. It is not a trivial rewrite of Nunes or Heath-Brown; the p-adic stationary-phase machinery is doing real work, and the paper deserves a serious referee.\n\nThe new content is concentrated in Lemma 6 and Proposition 2. Lemma 6 gives a complete evaluation of the relevant character-exponential sum K_χ and splits it into two analytic pieces; Proposition 2 gives square-root cancellation in complete sums of products with additive twists. I checked the main dependency chain: Theorem 2 -> summation by parts -> Theorem 1, with Proposition 3 as the aggregate engine. The stress-test note's verification of the derivative identity and the stationary-phase congruence in Proposition 2 is consistent with the text. I did not find a circular step. Lemma 2 is quoted from the author's earlier paper, but that is a published lemma with a parameter-free proof, and all uses here have tilde{q} > p^2, so the edge cases never enter.\n\nSoft spots are real but not disabling. The proof of Proposition 2 is compressed; the verification of the hypotheses of Lemma 3 is summarized rather than fully displayed, and a referee should ask for the expanded version. The same goes for a couple of \"easily\" steps in section 6. The introduction's \"large sieve\" language is informal; the actual argument is Cauchy-Schwarz, which is fine, but readers should not expect a formal sieve. The bound has an unspecified p^A and the paper does not try to optimize it; that is consistent with the claim. The paper also explicitly notes that the sharper range result analogous to Nunes's Theorem 1.2 is not obtained; that is an honest limitation, not a defect.\n\nThe citation pattern looks appropriate: Heath-Brown, Nunes, Milićević's sub-Weyl paper, Postnikov, Blomer–Milićević. No data fitting, no invented entities, no hidden assumption of the target bound.\n\nWho is this for: analytic number theorists working on moments and subconvexity, especially anyone interested in p-adic analogues of archimedean methods. I would bring it to a reading group. It is a within-subfield advance that fills a clear gap. Recommendation: send it to peer review; it deserves careful referee time. My own confidence in the main theorem is moderate rather than high only because the p-adic arguments are dense and not machine-checked, but I saw no red flag.","headline":"This paper proves the first sharp q-aspect twelfth moment for Dirichlet L-functions to prime power moduli, and the p-adic stationary-phase arguments at its core deserve careful refereeing.","tokens_in":16473,"tokens_out":2290,"would_cite":true,"duration_ms":22022,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["11M06","11L07","11L40","26E30"],"pacs":[],"model":"deepseek-v4-flash","headline":"This paper proves that the twelfth power moment of Dirichlet L-functions summed over all characters modulo an odd prime power q=p^n is bounded by p^A q^{2+ε}, the conjecturally sharp size.","keywords":["Dirichlet L-functions","twelfth moment","prime power moduli","p-adic stationary phase","exponential sums","large values","short second moment","moments"],"falsifier":"Compute the complete exponential sums Kχ(m;p^n) explicitly for a small odd prime such as p=3 and conductors n=3,4, for all primitive characters and all m, and compare the results with the evaluation in Lemma 6; any mismatch there would trace back to the p-adic logarithmic representation of characters and would directly refute the main estimate. Equivalently, one can test the identity χ(1+kp)=e(A log_p(1+kp)/p^n) numerically for all k modulo p^n for every primitive character of a small conductor.","tokens_in":15513,"feed_emoji":"🧮","tokens_out":10507,"duration_ms":102291,"temperature":0.7,"pith_summary":"This paper proves that the twelfth power moment of Dirichlet L-functions at the central point, summed over all characters modulo an odd prime power q=p^n, is bounded by p^A $q^{{2+ε}}$ for some absolute constant A>0. This is the conjecturally sharp size for this family, up to the usual q^ε loss, and it is the q-aspect analogue of the classical twelfth moment bound for the Riemann zeta function. The proof passes through an equivalent large-value statement: for any V, the number of primitive characters with |L(1/2,χ)|>V is at most p^A $q^{{2+ε}}$ $V^{{-12}}$. The reader should care because a moment bound of this strength is the sharpest uniform control available on how often central values of these L-functions are large, and because the proof introduces a p-adic large-sieve mechanism that may work for other families.","feed_headline":"For prime powers, twelfth moment hits conjectured bound","feed_subtitle":"The sum of |L(1/2,χ)|^12 over characters of q=p^n is at most p^A q^{2+ε} for every odd prime p.","key_machinery":"The load-bearing mechanism is the p-adic method of stationary phase, packaged as Lemmas 3 and 4: a complete exponential sum with a p-adically analytic phase reduces to the contribution of points where the derivative vanishes, and each such contribution is evaluated explicitly through quadratic Gauss sums. The bridge into this framework is Lemma 2, which writes any primitive character modulo p^n as χ(1+kp)=e(A log_p(1+kp)/p^n) for a p-adic unit A. This turns character sums into the p-adically analytic exponential sums Kχ(m;Q1), whose square-root cancellation is shown in Lemma 6 and whose orthogonality over pairs of characters is shown in Proposition 2, a bound of size $Q^{{1/2}}$ for complete sums of products with an additive twist. These two estimates are what make the large-sieve step work.","core_discovery":"The central claim is Theorem 1: for every odd prime p and every q=p^n, ∑_{χ mod q} |L(1/2,χ)|^{12} ≪_ε p^A $q^{{2+ε}}$. Since summation by parts makes the moment and large-value estimates equivalent, the paper proves Theorem 2, which bounds the number of primitive characters with central value exceeding V by p^A $q^{{2+ε}}$ $V^{{-12}}$. The new work is in the short second moment: after an approximate functional equation and a trace-function decomposition, the argument reduces the bound to square-root cancellation in complete sums of products of the exponential sums Kχ(m;Q1), established by p-adic stationary phase. A Cauchy-Schwarz and large-sieve step then converts this orthogonality into the displayed moment bound. If correct, the theorem confirms the sharp predicted moment in the prime power aspect for every odd prime.","pith_inferences":["The orthogonality mechanism in Proposition 2 is the seed of a genuinely p-adic large sieve; it should extend to products of more than two Kχ factors and therefore to moments of order higher than 12, with new stationary-phase bookkeeping at degenerate critical points.","A natural next target is a hybrid or multi-prime-power modulus with finitely many well-separated prime power factors, where the same machinery could in principle unify the prime-power and square-free results.","A concrete check of the method would be to compute the matrices (Kχ(m;Q1)) for small p and n and test approximate row orthogonality numerically, providing independent evidence for the large-sieve step before any theoretical extension is attempted.","Because the paper's overview isolates the weighted trace sum as the only obstacle, a further application of stationary phase to more complicated phases should yield the sharper large-value statement in the intermediate range without changing the moment bound itself."],"forward_implications":["The moment bound is sharp up to q^ε, so the twelfth moment of these L-functions is as small as the conjectural optimal bound predicts for every odd prime power modulus.","The equivalent large-value bound shows that very few characters can have very large central values: the count above V is at most p^A q^{2+ε} V^{-12}.","A single character with |L(1/2,χ)| > q^{1/6+ε/2} would force the total twelfth moment above the theorem's bound, so the theorem recovers the standard subconvexity bound L(1/2,χ) ≪ q^{1/6+ε} for primitive characters modulo prime powers.","Together with the smooth square-free modulus case, the result covers both families where the sharp twelfth moment is now known, pointing to the remaining difficulty in moduli with several prime factors.","Because the moment and large-value statements are equivalent via integration by parts, the theorem also constrains the distribution of the large set R(V;q) at every scale of V."],"supporting_citations":[{"why":"Supplies the t-aspect twelfth moment bound and the strategy of deducing large-value estimates from the moment.","marker":"[7]"},{"why":"Supplies the p-adic logarithmic description of primitive characters and the p-adic stationary phase lemmas used throughout.","marker":"[11]"},{"why":"Establishes the smooth square-free modulus analogue that the present result complements.","marker":"[13]"},{"why":"Gives the q^{1/6} subconvexity bound for prime power moduli that delimits the range of V in Theorem 2.","marker":"[15]"},{"why":"Provides the fourth moment bound used for small values of V in Theorem 2.","marker":"[6]"},{"why":"Improves the fourth moment estimate and is used in the same small-V range.","marker":"[16]"},{"why":"Provides the approximate functional equation used to open the short second moment.","marker":"[10]"},{"why":"Supplies the p-adic analytic twist and stationary phase formalism underlying Lemmas 3 and 4.","marker":"[1]"}],"fun_headline_variants":["Prime power moduli: twelfth moment matches prediction","Dirichlet L-functions: sharp twelfth moment for prime powers","Twelfth moment bound proven for prime power moduli","Conjectured moment bound achieved for prime powers","For prime powers, L-function moments hit predicted size"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The proof's load-bearing premise is the quoted representation of primitive characters modulo p^n as exponentials of p-adic logarithms with a p-adic unit multiplier (Lemma 2); if that representation fails at some prime or precision, the complete exponential sums and the square-root cancellation estimates no longer follow, and with them the moment bound.","fun_headline_variants_meta":{"raw":{"variants":["Prime power moduli: twelfth moment matches prediction","Dirichlet L-functions: sharp twelfth moment for prime powers","Twelfth moment bound proven for prime power moduli","Conjectured moment bound achieved for prime powers","For prime powers, L-function moments hit predicted size"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000215,"raw_usage":{"total_tokens":1349,"prompt_tokens":783,"completion_tokens":566,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":399,"completion_tokens_details":{"reasoning_tokens":490}},"tokens_in":399,"tokens_out":566,"duration_ms":5850,"temperature":1.0,"reasoning_tokens":490,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T13:32:32.805844+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute the complete exponential sums Kχ(m;p^n) explicitly for a small odd prime such as p=3 and conductors n=3,4, for all primitive characters and all m, and compare the results with the evaluation in Lemma 6; any mismatch there would trace back to the p-adic logarithmic representation of characters and would directly refute the main estimate. Equivalently, one can test the identity χ(1+kp)=e(A log_p(1+kp)/p^n) numerically for all k modulo p^n for every primitive character of a small conductor.","supporting_citations":[{"cited_title":"Sub-Weyl subconvexity for DirichletL-functions to prime power moduli","cited_arxiv_id":null,"evidence_quote":"Supplies the p-adic logarithmic description of primitive characters and the p-adic stationary phase lemmas used throughout."},{"cited_title":"On the sum of characters with respect to a modulus equal to a power of a prime number","cited_arxiv_id":null,"evidence_quote":"Gives the q^{1/6} subconvexity bound for prime power moduli that delimits the range of V in Theorem 2."},{"cited_title":"The fourth moment of Dirichlet L-functions","cited_arxiv_id":null,"evidence_quote":"Improves the fourth moment estimate and is used in the same small-V range."},{"cited_title":"p-adic analytic twists and strong subconvexity","cited_arxiv_id":null,"evidence_quote":"Supplies the p-adic analytic twist and stationary phase formalism underlying Lemmas 3 and 4."}],"review_version":1}