{"id":"6c5eb82d-80b2-4c32-817e-7b23ce451e96","arxiv_id":"2412.08964","paper_version":2,"verdict":"ACCEPT","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"low","formal_verification":"none","parameter_count":0,"one_line_summary":"For hierarchical integer-valued Gaussian and Coulomb gas fields, the covariance and fractional-charge exponents are computed sharp asymptotically up to and slightly beyond the BKT critical point, including explicit iterated-logarithmic corrections at criticality.","lead":"This paper proves sharp formulas for how correlations decay in hierarchical two-dimensional integer-valued fields (discrete Gaussian and sine-Gordon models) below, at, and just above the Berezinskii-Kosterlitz-Thouless critical temperature. It finds logarithmic correlations in all three regimes, with an extra log-log correction exactly at criticality, and gives a rare rigorous handle on critical exponents in a tractable random-surface family.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"No significant objection identified: the central theorems are internally consistent, and the main fragility (Assumption 1.2) is a disclosed scope restriction rather than a flaw.","rationale":"The reader's verdict is ACCEPT with moderate confidence, and my stress-test pass supports that. The strongest claim—sharp asymptotic (1.14) and (1.18) under Assumptions 1.1–1.2—is established by a long but structured RG argument. I looked specifically at the places where a hidden assumption could break the proof: (i) the ratio iteration for Fourier coefficients (Lemmas 3.7–3.8) requires strict positivity and the ratio bound, which is Assumption 1.2; this is disclosed and excludes only models outside the stated scope; (ii) the supercritical fixed-point contraction (Lemma 6.12) contains explicit numerical conditions, and the constants chosen in Lemma 6.13 satisfy them with comfortable margins for b≥4, the admissible hierarchical branching (b=L^2 with integer L≥2); (iii) the critical covariance computation correctly yields a negative iterated-log correction, consistent with the physical intuition that periodic modulation reduces variance; (iv) the fractional-charge exponent κ(α,β) and the critical correction τ(α) are mathematically consistent with the near-critical expansion (1.20); (v) the dependence on the sequence {d_k} is handled under the stated extra summability hypotheses, and the bounds in (4.55) and (5.63) are correct. The manuscript also includes explicit remarks about limitations (Section 1.4), which strengthens trust. The only respect in which I depart from the reader's framing is that I do not treat Assumption 1.2 as an objection: it is a hypothesis of the theorem, satisfied by the principal examples, and the authors are transparent about its restrictiveness. Therefore the verdict should remain ACCEPT/UNCHANGED.","tokens_in":67578,"tokens_out":24217,"duration_ms":229044,"concrete_test":"Independently verify the supercritical contraction conditions by running the truncated RG iteration (3.19) numerically for the smallest admissible hierarchical branching b=4 (so L=2), choosing β with bθ−1=10^{-3} and sine-Gordon initial data with κ=1. Check that λ_k(1) converges to the value predicted by Lemma 6.7, that λ_k(2)/(bθ−1) approaches 1/((b−1)(b+1)^3), and that the distance to the fixed point decays at rate e^{−c k (bθ−1)} for some c>0. This directly tests the contraction constants in Lemma 6.12 and the conclusions of Theorem 3.6 for a parameter regime covered by the theorem.","verdict_should_be":"UNCHANGED","load_bearing_attack":"After a close reading of the argument, I do not find a load-bearing internal gap in the proof of Theorems 1.3 and 1.4. The central claims are conditional on Assumptions 1.1–1.2, and the most fragile premise is indeed Assumption 1.2 (strict positivity of all Fourier coefficients plus the ratio bound sup_q a(q+1)/a(q)<∞), used at Eq. (3.21), Lemma 3.7, and throughout the supercritical contraction argument. This assumption excludes the GFF and the hard-core Coulomb gas (1.10), but the authors explicitly state this limitation in Section 1.4 and do not claim those cases. For the models covered (DG, sine-Gordon), Assumption 1.2 holds. The supercritical proof in Section 6 is the most delicate part: the contraction in Lemma 6.12 involves explicit constants, and the paper shows that the required inequalities (6.60)–(6.61) hold for t=7/5, A=(10/7)√b. For the admissible hierarchical branching b=L^2≥4 these margins are not tight, and the algebra leading to the claimed contraction is coherent. The critical and near-critical coefficients in (1.15) and (1.20) match the same fixed-point constants, providing an internal consistency check. I also checked the sign of the iterated-log correction in (1.14): although the extracted text of (4.54) shows a plus sign, the derivation through (4.5) and (4.25) yields a minus sign, consistent with the stated theorem. Overall, the paper is a careful, conditional proof with no identified circularity or omitted essential step.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies hierarchical random fields on a square box in two dimensions, with law proportional to exp((β/2)(φ,Δ_n φ)) times a product of a 1-periodic single-spin measure ν, covering the hierarchical DG model and sine-Gordon models among others. Under Assumptions 1.1 and 1.2 on the hierarchical Laplacian and the Fourier coefficients of ν, the authors prove sharp asymptotic formulas for the covariance ⟨φ_x φ_y⟩ and the fractional charge ⟨e^{2π i α(φ_x−φ_y)}⟩ in the subcritical (β<β_c), critical (β=β_c), and slightly supercritical (β>β_c) regimes, with explicit constants including an iterated-log correction at criticality. The proof combines a renormalization-group analysis of the Fourier coefficients, a tree-indexed Markov chain representation of the field, and a new fixed-point contraction argument for the supercritical flow. The subcritical and critical flow estimates are adapted from the authors' earlier work [15], while the supercritical analysis is proved from scratch in Section 6.","tokens_in":67858,"tokens_out":4040,"duration_ms":44759,"significance":"If correct, the paper establishes a sharp, model-independent description of the BKT-type transition in hierarchical Z-modulated fields, including the nontrivial supercritical fixed point and explicit critical corrections. The main strengths are the uniformity in the single-spin measure, the explicit constants in the covariance and fractional-charge asymptotics, and the detailed proof structure with explicit error terms. The paper also provides internal consistency checks between the critical and near-critical coefficients. The restriction to Assumption 1.2, which excludes the GFF and the hard-core Coulomb gas, is disclosed in Section 1.4 and is a scope condition rather than a gap for the models covered.","major_comments":[],"minor_comments":[{"comment":"The displayed formula (4.54) has a plus sign in front of the log(n/k) term, whereas both the derivation through (4.25) and the statement of Theorem 1.3 require a minus sign; this appears to be a typographical sign error and should be corrected.","section":"Section 4.3, Eq. (4.54)"},{"comment":"The sentence introducing v_k contains the duplicated word 'with with' and should read 'each v_k is a C^8 function with v'_k and v''_k uniformly bounded'.","section":"Theorem 3.6"},{"comment":"The word 'trivally' appears in the last paragraph of the proof and should be corrected to 'trivially'.","section":"Lemma 5.10"},{"comment":"The notation 'op(1)' is used in asymptotic statements where the asymptotic variable is not always explicit; for clarity, the authors should specify that the error terms tend to 0 uniformly in the indicated parameters, e.g., uniformly in z in (3.29).","section":"Eqs. (3.29) and (5.28)"}],"recommendation":"minor_revision","confidential_remarks":"The manuscript is technically very detailed and the central conditional claims appear sound. The main fragility is the disclosed Assumption 1.2, which is a genuine scope restriction rather than an internal inconsistency. The reliance on the authors' earlier work [15] is explicit and does not appear circular. The only issue requiring correction before publication is the sign typo in Eq. (4.54), together with a few minor wording errors."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Let me get straight to it: this is a strong, careful paper that delivers sharp asymptotics for covariance and fractional-charge correlations in hierarchical integer-valued Gaussian and sine-Gordon models below, at, and slightly above beta_c. The main theorems (1.3, 1.4) are new — earlier work had bounds or fixed-point existence, not these explicit formulas with log corrections and near-critical expansions. The proof strategy, tracking the exponential of renormalized potentials rather than linearizing, is the right call and gives uniformity across models.\n\nThe rigor looks solid. Proofs are lemma-based, assumptions explicit, error terms tracked. The supercritical section is done from scratch via contraction on a carefully chosen set, with explicit constants for b=L^2>=4. The internal consistency between the critical log correction and the near-critical slope of sigma^2 is good evidence the constants are right. Reliance on the authors' earlier paper [15] for subcritical and critical flow is openly stated, and [15] proves different things, so no circularity.\n\nThe soft spots are real but disclosed. Assumption 1.2 (strictly positive Fourier coefficients with ratio bound) excludes the hard-core Coulomb gas and the GFF; the authors say in Section 1.4 they expect the results to extend but don't prove it. The supercritical regime is only slightly above beta_c, which they call a technicality. The proof is long and not machine-checked; I would not bet the farm on every constant without a second pass, but I found no gap.\n\nWho is this for: anyone working on hierarchical models, BKT transition, or Z-valued interface models in 2D. It will be a useful reference for the explicit critical behavior. It deserves a serious referee — I'd send it to review without hesitation.","headline":"A careful, sharp RG analysis of hierarchical integer-valued Gaussian/Coulomb gas models across the BKT transition; the main limitation is disclosed and does not undermine the theorems.","tokens_in":68418,"tokens_out":2059,"would_cite":true,"duration_ms":21997,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["60K35","60G60","82B27","82B28"],"pacs":[],"model":"deepseek-v4-flash","headline":"This paper proves sharp asymptotic formulas for the covariance and fractional-charge correlations of hierarchical integer-valued Gaussian and Coulomb gas fields, uniformly across a whole family of single-spin measures, with an explicit…","keywords":["hierarchical random fields","integer-valued Gaussian model","Coulomb gas","sine-Gordon model","BKT transition","renormalization group","fractional charge","logarithmic correlations"],"falsifier":"For b = 2 and a single-spin measure satisfying the paper's positivity condition, numerically iterate the coefficient recursion (3.19) at beta = beta_c and at beta just above beta_c. The claimed asymptotics predict a_k(1)/a_k(0) ~ A/sqrt(k) at criticality and convergence to lambda_*(1) at a rate proportional to sqrt(beta - beta_c) above criticality; any deviation larger than the stated error bounds would falsify the core renormalization-group flow theorem.","tokens_in":67344,"feed_emoji":"📐","tokens_out":9535,"duration_ms":97899,"temperature":0.7,"pith_summary":"This paper proves that a large class of two-dimensional hierarchical random fields with periodically modulated values, including the integer-valued Gaussian model and the sine-Gordon model, undergoes a sharp phase transition at inverse temperature beta_c = 2*$pi^{2}$/log b, where b is the branching number. Below beta_c the covariance is that of a Gaussian free field at scale 1/$\\beta$; above beta_c it remains logarithmic but with a strictly smaller amplitude; exactly at beta_c an explicit iterated-logarithm correction appears. The fractional-charge correlation decays as a power of distance with an exponent that is smaller above beta_c than below, signalling partial screening, and receives a power-of-log correction at criticality. The formulas are uniform in the system size and in all single-spin measures satisfying the stated positivity condition, which is the sense in which the result is universal.","feed_headline":"Hierarchical Coulomb gas covariance solved at all temperatures","feed_subtitle":"Below, at, and above the transition each has its own law, with an explicit iterated-log term at criticality.","key_machinery":"The proof rests on the hierarchical Laplacian's finite-range decomposition, which turns the Gibbs measure into the law of a tree-indexed Markov chain. The renormalization-group flow is tracked through the Fourier coefficients a_k(q) of $e^{{-v_k}}$, which iterate by a convolution recursion with Gaussian damping; the key ratio inequality for a_k(q+1)/a_k(q) is what lets all three regimes be controlled. At criticality the ratio a_k(1)/a_k(0) decays as A/$\\sqrt$(k), producing the iterated-log corrections, while just above beta_c the flow is drawn to a nontrivial fixed point lambda_* characterized by an explicit fixed-point equation, with contraction measured in a weighted metric. The covariance and fractional-charge asymptotics are then read off from the tree-indexed Markov chain via martingale identities and a coupling bound that compares the field's fractional parts to the stationary law at the fixed point.","core_discovery":"The central assertion is that, for every admissible single-spin measure and every b >= 2, the covariance <phi_x phi_y> equals $sigma^{2}$($\\beta$) log_{$b^{{1/2}}$}(diam/(1+d)) + O(1) for $\\beta$ != beta_c, and equals (1/beta_c) log_{$b^{{1/2}}$}(diam/(1+d)) - c_bar log(log diam / log(2+d)) + O(1) at $\\beta$ = beta_c, with all error terms uniform in the box size and in x,y. The fractional charge <$e^{{2*pi*i*alpha*(phi_x-phi_y)}}$> decays as (C_n + o(1)) $d^{{-kappa(alpha,beta)}}$ for $\\beta$ != beta_c and as (C_n + o(1)) $d^{{-kappa(alpha,beta)}}$ (log d)^{tau($\\alpha$)} at criticality. The exponent kappa($\\alpha$,$\\beta$) is defined implicitly through a fixed-point equation, equals 4*beta_c*$alpha^{2}$/$\\beta$ below beta_c, is strictly smaller above beta_c, and satisfies kappa($\\alpha$,$\\beta$) = 4*beta_c*$sigma^{2}$($\\beta$)*$alpha^{2}$ + o($alpha^{2}$) for small $\\alpha$. The paper's way of saying this is that the renormalization-group flow converges to the trivial Gaussian fixed point below and at criticality, slowly at criticality, and to a nontrivial fixed point above beta_c.","pith_inferences":["The paper's main theorems exclude the hard-core Coulomb gas, whose Fourier coefficients vanish for |q| >= 2 and so violate the ratio condition; if the same formulas hold there, as the authors expect, then Assumption 1.2 is sufficient but not necessary, and the real threshold is that the renormalization-group flow enters the contractive region.","At beta = beta_c the iterated-log correction should change the second-order term in the law of the field's maximum away from the Gaussian free field value; the authors identify this as an open problem, and the tree-indexed Markov chain representation offers a concrete route to test it.","The exponent kappa(alpha,beta) is defined implicitly and solved numerically, so one could test the near-critical slope in (1.20) by direct simulation of the hierarchical model for small alpha and small beta - beta_c.","Because the covariance remains logarithmic at all beta, the extremal process should follow the general log-correlated pattern, but the critical correction may alter the second-order extremal statistics; the paper does not address this, and it is a natural next question."],"forward_implications":["Below beta_c every admissible model has the same logarithmic covariance amplitude 1/beta as the Gaussian free field, uniformly in the single-spin measure.","Above beta_c the covariance amplitude sigma^2(beta) is strictly smaller than 1/beta and has a square-root cusp at beta_c, with explicit expansion sigma^2(beta) = 1/beta - const*(beta - beta_c) + O((beta - beta_c)^{3/2}).","Exactly at beta_c the covariance contains the explicit iterated-log term -c_bar log(log diam / log(2+d)) and the fractional charge contains the factor (log d)^{tau(alpha)}, with constants given in closed form.","The fractional-charge exponent obeys kappa(alpha,beta) = 4*beta_c*sigma^2(beta)*alpha^2 + o(alpha^2), so phi_x - phi_y is asymptotically Gaussian after normalization, while higher-order corrections are expected to be non-Gaussian once beta >= beta_c.","Above beta_c the energy cost of inserting opposite fractional charges still grows logarithmically with distance, but with a reduced coefficient, meaning that screening is only partial in these long-range hierarchical models."],"supporting_citations":[{"why":"Supplies the tree-indexed Markov chain representation and the subcritical and critical renormalization-group estimates that Theorems 3.4 and 3.5 adapt to variable coefficients.","marker":"[15]"},{"why":"Provides the earlier hierarchical Coulomb gas renormalization-group analysis, including the subcritical fractional-charge upper bound and the nontrivial fixed point, which this paper extends and makes uniform.","marker":"[56]"},{"why":"Establishes the massive sine-Gordon renormalization-group scheme whose coefficient assumptions motivate Assumption 1.1.","marker":"[10]"},{"why":"Proves existence of a nontrivial supercritical fixed point and identifies the eigenvector identity used in Remark 4.3 for the covariance calculation.","marker":"[11]"},{"why":"Treats the hierarchical Coulomb gas renormalization-group flow and represents the limitations, excluding the DG model, that this paper removes.","marker":"[44]"},{"why":"Supplies the correlation inequalities used to prove monotonicity of sigma^2(beta) in the supercritical regime.","marker":"[36]"},{"why":"Determines the fractional-charge asymptotic for the lattice sine-Gordon model at criticality, providing the comparison for the iterated-log correction.","marker":"[34]"},{"why":"Computes critical exponents for the two-dimensional Coulomb gas at the BKT transition, giving the corresponding lattice result that the hierarchical formulas are compared against.","marker":"[35]"}],"fun_headline_variants":["Covariance and charge exponents solved in hierarchical Coulomb gas","Criticality adds iterated-log term to hierarchical Gaussian gas","Sharp asymptotics for integer Gaussian fields at, below, above beta_c","Beta_c splits scaling: iterated logs at criticality, distinct exponents"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The argument needs the single-spin measure's Fourier coefficients to be strictly positive with a uniformly bounded ratio a(q+1)/a(q); without that, the key ratio iteration and the fixed-point contraction lose their footing, which is why the main theorems do not cover the hard-core Coulomb gas or the Gaussian free field.","fun_headline_variants_meta":{"raw":{"variants":["Covariance and charge exponents solved in hierarchical Coulomb gas","Criticality adds iterated-log term to hierarchical Gaussian gas","Sharp asymptotics for integer Gaussian fields at, below, above beta_c","Beta_c splits scaling: iterated logs at criticality, distinct exponents"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000923,"raw_usage":{"total_tokens":4048,"prompt_tokens":1128,"completion_tokens":2920,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":744,"completion_tokens_details":{"reasoning_tokens":2847}},"tokens_in":744,"tokens_out":2920,"duration_ms":20873,"temperature":1.0,"reasoning_tokens":2847,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-11T17:21:50.103058+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"For b = 2 and a single-spin measure satisfying the paper's positivity condition, numerically iterate the coefficient recursion (3.19) at beta = beta_c and at beta just above beta_c. The claimed asymptotics predict a_k(1)/a_k(0) ~ A/sqrt(k) at criticality and convergence to lambda_*(1) at a rate proportional to sqrt(beta - beta_c) above criticality; any deviation larger than the stated error bounds would falsify the core renormalization-group flow theorem.","supporting_citations":[{"cited_title":"Marchetti and J.F","cited_arxiv_id":null,"evidence_quote":"Provides the earlier hierarchical Coulomb gas renormalization-group analysis, including the subcritical fractional-charge upper bound and the nontrivial fixed point, which this paper extends and makes uniform."},{"cited_title":"Benfatto, G","cited_arxiv_id":null,"evidence_quote":"Establishes the massive sine-Gordon renormalization-group scheme whose coefficient assumptions motivate Assumption 1.1."},{"cited_title":"Benfatto and J","cited_arxiv_id":null,"evidence_quote":"Proves existence of a nontrivial supercritical fixed point and identifies the eigenvector identity used in Remark 4.3 for the covariance calculation."},{"cited_title":"Guidi and D.H.U","cited_arxiv_id":null,"evidence_quote":"Treats the hierarchical Coulomb gas renormalization-group flow and represents the limitations, excluding the DG model, that this paper removes."},{"cited_title":"Fr ¨ohlich and Y.M","cited_arxiv_id":null,"evidence_quote":"Supplies the correlation inequalities used to prove monotonicity of sigma^2(beta) in the supercritical regime."},{"cited_title":"Falco (2012)","cited_arxiv_id":null,"evidence_quote":"Determines the fractional-charge asymptotic for the lattice sine-Gordon model at criticality, providing the comparison for the iterated-log correction."}],"review_version":1}