{"id":"f8cee1b8-1735-4cc2-a193-eac58676f585","arxiv_id":"2607.24170","paper_version":1,"verdict":"ACCEPT","confidence":"HIGH","novelty_score":7.0,"correctness_risk":"low","formal_verification":"none","parameter_count":1,"one_line_summary":"Explicit finite-energy Wigner and Gaussian-regularised Bernoulli ensembles violate the conjectured dimensional monotonicity of quadratically penalised logarithmic energy for mean ESDs.","lead":"The paper builds explicit random-matrix counterexamples showing that a conjectured decrease of penalised logarithmic energy with dimension can fail. This redirects work on free-entropy-style monotonicity toward sharper hypotheses or large-dimension thresholds.","discovery_kind":"extension","skeptic_critique":{"model":"moonshotai/kimi-k3","headline":"Theorem 1.2 reduces to the sign of L3 − C2 computed in huge rationals, but independent decimal recomputation confirms the margin (~+4.6e-3); the flagged arithmetic appears correct.","rationale":"I read the paper in good faith as a pure-math counterexample construction and asked what must be true for the central claim: (a) the semicircle gap formula (Prop. 2.2), (b) the explicit 2×2 event in Theorem 1.1, (c) the limsup/liminf collision-mass framework (Lemmas 2.3–2.4, 4.1–4.2, 4.5–4.6), and (d) the explicit rational comparison L3 > C2. Items (a)–(c) survive careful reading; the proofs are modular and I found no hidden assumptions (weak convergence of β_{n,ε}, uniform second moments, disk disjointness, P(Gε)→1, and the one-sided nature of both bounds — tightness is not needed — all check out). Item (d) is the genuinely load-bearing spot, exactly as the reader identified: the gap is small (~4.6e-3) and the fractions are enormous. However, rather than only pointing at the risk, I re-derived the key quantities independently in floating point — all sixteen characteristic polynomials, the atom masses (including the non-obvious 3p²q² at 30/√62), the collision-mass formula, the expansion (27) (validated at p=0 and p=1), C2 ≈ 0.3927048, a0's factor 2219 = 2·31²+9·31+18, and L3 ≈ 0.3972757 — and every number I could check reproduces the paper's printed fractions. So the reader's flagged concern, while correctly located, does not appear to land; the residual risk is a transcription slip that my checks already partially exclude and that a CAS pass would fully exclude. I therefore agree with the reader's weakest_assumption and leave the ACCEPT verdict unchanged, recommending the exact-arithmetic verification as a worthwhile confirmatory step rather than a condition.","tokens_in":15562,"tokens_out":15948,"duration_ms":427016,"concrete_test":"Run exact-rational CAS verification (e.g. SymPy/Mathematica): (i) symbolically compute the characteristic polynomials of all 16 matrices [[a,b],[c,d]]/√2 with entries in {−1/√31, √31}, cluster the eigenvalues into distinct atoms with masses, and sum squared masses to obtain C2 exactly — compare with (27) and 431783475867/1099511627776; (ii) recompute a0, a1 from (32)–(33) and evaluate sign(L3 − C2) in exact arithmetic at p=1/32. If the sign is not positive or any value differs from the printed fractions, Theorem 1.2's separation fails. As a cheap independent cross-check, Monte Carlo: sample 2×2 Bernoulli matrices, average the empirical spectral law over ~10^7 draws, and estimate Col(ν0_2) via binning near the 14 atoms; agreement with 0.3927048 within Monte Carlo error confirms the enumeration.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The reader's weakest_assumption is the right one: the entire i.i.d. counterexample funnels into L3 > C2 (eq. 37), where C2 = Col(ν0_2) comes from a 14-atom enumeration (Lemma 4.3, Tables 1–2) and L3 = a0² + a1² from selected zero/outlier branches (Lemma 4.4). The structural machinery is sound on reading: Lemma 2.3's limsup bound (the 1/4 constant in (14) is (1/2π)·(π/2) ✓; (15) follows from angular averaging ✓), the traceless-Ginibre decoupling in Lemma 4.1, the Rouché/Riesz stability in Lemma 4.5, the branch counts in Lemma 4.4 (s2 = ce(3δ−1), ce(3δ−2), nonzero since ce = −32/31 ✓; 9+18 configuration counts ✓), and Lemma 2.4 applied to two disjoint shrinking disks in Prop. 4.6. So the only load-bearing fragility is arithmetic. I independently recomputed in floating point: the 16 characteristic polynomials of Table 1 all check (e.g. one-v-diagonal gives z²−30z−32; two-v-diagonal (z−30)(z−32) with product 960 = 62·480/31 ✓); the mass of atom 30/√62 is 3p²q² (four row/column configs plus diagonal and anti-diagonal, each weight 1/2) ✓; Col formula (1/4)(q⁴+4p²q²+p⁴)² + (1/4)(q⁸+p⁸) + 4p²q⁶ + 4p⁶q² + (19/2)p⁴q⁴ ✓; expansion (27) evaluates to 1/2 at both p=0 and p=1 ✓ and gives C2 ≈ 0.3927048 at p=1/32, matching 431783475867/2⁴⁰; a0 = 31⁷·2219/(3·32⁹) ≈ 0.5783875 (2·31²+9·31+18 = 2219 ✓), a1 ≈ 0.2504863, L3 ≈ 0.3972757, gap ≈ +0.0046 > 0, consistent with (37). The margin is thin (~1.2% of C2), so a single slip could in principle flip it — but my spot checks all reproduce the paper's values. On the Wigner side, Theorem 1.1 verifies cleanly: the 2×2 eigenvalue formula, L+ > 3/√2 on A, vsc(3/√2) = 3/4 − log 2 > 0 (e^{3/4} ≈ 2.117 > 2), and m2(ρn) = 1. The concern is real but does not appear to land.","agreement_with_reader":"agree"},"referee_report":{"model":"moonshotai/kimi-k3","summary":"The manuscript constructs two counterexamples to the dimensional monotonicity questions for penalised logarithmic energies of mean empirical spectral distributions posed by Chafaï, Dadoun, and Youssef [5]. Theorem 1.1 takes a real symmetric Wigner sequence whose entries have the semicircle law itself, so ρ1 = σ; on the positive-probability event that all three entries of M2 lie in (3/2,2), the top eigenvalue exceeds 3/√2 > 2, and a gap identity (Proposition 2.2, FR(µ) − 3/4 = ∫vsc dµ + a nonnegative Fourier term) then yields FR(ρ2) ≥ FR(ρ1) + (p0³/2)(3/4 − log 2) > FR(ρ1). Theorem 1.2 regularises a Bernoulli law (rare atom probability p = 1/32) with a small circular Gaussian, and shows FC(ν3,ε) > FC(ν2,ε) for all small ε by separating limsup and liminf of FC/log(1/ε): an upper bound by the collision mass C2 of the exactly enumerated 14-atom measure ν0_2 (Lemmas 2.3, 4.2–4.3), and a lower bound L3 = a0² + a1² from stable zero/outlier eigenvalue branches of selected 3×3 configurations (Lemmas 2.4, 4.4–4.5, Prop. 4.6), with L3 − C2 given as an explicit positive rational.","tokens_in":16132,"tokens_out":11062,"duration_ms":301225,"significance":"The paper settles, in the negative, Questions 1.1 and 1.2 of Chafaï–Dadoun–Youssef (RMTA 2024) for unrestricted entry laws — a conjecture backed in [5] by exact Gaussian computations and extensive numerics. The strengths are concreteness and verifiability: Theorem 1.1 is essentially parameter-free (the only quantity is p0 = σ((3/2,2))), rests on an exact semicircle gap identity (Prop. 2.2) of independent interest, and the strictness reduces to the elementary log 2 < 3/4. Theorem 1.2 is fully explicit: p = 1/32, exact collision mass of ν0_2 from a 14-atom enumeration, and a final comparison in exact rational arithmetic with no fitted constants. I spot-verified the load-bearing arithmetic (characteristic polynomials of Table 1, Col formula, C2 ≈ 0.392705, a0 ≈ 0.578388, L3 ≈ 0.397276, gap ≈ +0.0046) and it reproduces. The work leaves open, and clearly frames, the interesting question of monotonicity under restrictions or above a dimension threshold.","major_comments":[],"minor_comments":[{"comment":"Lemma 4.1, proof: the covariance display 'E[G0(Gii−Tn)] = 1/√n − 1/√n = 0' only reads correctly with a complex conjugate on the second factor, i.e. E[G0 \\overline{(Gii−Tn)}]. For proper complex Gaussians one needs vanishing of both the covariance and the pseudo-covariance; both do vanish here and the conclusion (independence of G0 and Rn) is correct and standard, but as printed the two 1/√n terms arise only in the conjugated product, so the display should be corrected and a half-sentence of justification added.","section":"§4.1, Lemma 4.1"},{"comment":"Proposition 4.6, proof: the displayed value '|c|√3 = √3/31' should be |c|√3 = √3/√31 = √(3/31) ≈ 0.311. Only positivity is used, so the argument is unaffected.","section":"§4.3, Proposition 4.6"},{"comment":"Table 1: the column header '√62 spec(A)' is confusing on first reading; the entries listed are the eigenvalues of √62A (equivalently, the atoms of ν0_2 scaled by √62). A one-line caption stating this, and noting that each empirical eigenvalue carries weight 1/2 (which explains the masses in Table 2), would make the enumeration easier to audit.","section":"§4.2, Table 1"},{"comment":"Proposition 2.2: the gap identity (8) is the equilibrium-condition form of the classical Fourier representation of the logarithmic kernel; a brief remark situating it relative to known relative-energy identities (e.g. in the free-entropy/large-deviation literature [1, 8]) would help readers, even if the self-contained proof is retained.","section":"§2.2, Proposition 2.2"},{"comment":"The margin in (37) is L3−C2 ≈ 4.6×10⁻³, about 1.2% of C2; since the entire Theorem 1.2 funnels through this sign, a short appendix or ancillary file documenting the rational arithmetic (the expansion of (27), the 14-atom mass table, and the final subtraction) would materially assist referees and readers.","section":"§4.4, Eq. (37)"},{"comment":"There are scattered typographical artifacts, presumably from source formatting: 'generalβ-ensembles' and 'For ann×nmatrixA' in §1, and run-together text around equations (e.g. 'Sincerε→0', 'log(1/sε)+1/4' line-break). A careful proofread of spacing around inline mathematics is advised.","section":"General"}],"recommendation":"accept","confidential_remarks":"This is a direct answer (negative) to an open question posed by Chafaï, Dadoun, and Youssef, and fits the journal's scope well. The manuscript includes an unusually prominent AI-assistance disclosure; that is the author's prerogative and does not affect the mathematics, but the editor may wish to confirm it complies with journal policy. I independently verified the load-bearing rational arithmetic in (37) to floating-point precision; it is correct as printed. The only caveat I would flag to the editor is that the paper's conclusions are entirely rigorous but the Theorem 1.2 comparison is numerically delicate; publication of a short verification script would be a valuable addition, though I do not require it."},"author_rebuttal":null,"desk_editor":{"model":"grok-4.5","letter":"The paper settles the unrestricted versions of the 2024 monotonicity questions in the negative with two concrete constructions. That is the news.\n\nWhat is new is the pair of finite-energy counterexamples inside the exact classes asked about: a real Wigner sequence whose entries are themselves semicircle (so m2=1 and energy finite for every n) with FR(ρ2)>FR(ρ1)=3/4, and a one-parameter Gaussian-smoothed Bernoulli family for which FC(ν3,ε)>FC(ν2,ε) for all small ε. The Wigner argument is short once the semicircle-gap identity is in hand: positive mass past 2 on a simple event, vsc(3/√2)=3/4-log2>0. The i.i.d. argument is longer but modular—collision-mass limsup/liminf via Gaussian convolution, exact 2\times2 atom enumeration, selected 3\times3 zero/outlier clusters controlled by Rouché/Riesz, then a rational comparison L3>C2.\n\nThe soft spot is real but narrow. Theorem 1.2 funnels into the sign of one explicit difference of huge rationals (eq. 37). The structural lemmas check out (angular averaging, traceless Ginibre decoupling, configuration counts, disk separation). Independent floating-point recomputation of the characteristic polynomials, masses, and the final fractions reproduces the paper’s values and a positive gap of about 4.6e-3. So the arithmetic appears correct; it is just the only place a transcription slip could flip the claim. Everything else is standard and cleanly written. The author also flags the natural remaining questions (monotonicity under stronger assumptions or after a dimension threshold).\n\nThis is for people who work on free entropy, logarithmic energies, or large deviations for ESDs. It does not invent new machinery, but it cleanly redirects the program. I would send it to a serious referee without hesitation; the constructions are explicit and the residual risk is ordinary arithmetic verification, not conceptual fragility.","headline":"Clean, explicit counterexamples that kill the unrestricted monotonicity conjectures of Chafaï–Dadoun–Youssef; the Wigner case is short and solid, the i.i.d. case rests on a thin but independently checkable rational inequality.","tokens_in":17046,"tokens_out":568,"would_cite":true,"duration_ms":10919,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["60B20","15B52","31A15"],"pacs":[],"model":"grok-4.5","headline":"Dimensional monotonicity of penalised logarithmic energy fails for general Wigner and i.i.d. random matrices.","keywords":["logarithmic energy","Wigner matrices","mean empirical spectral distribution","dimensional monotonicity","circular law","collision mass","Gaussian regularisation"],"falsifier":"Recompute the fourteen atoms and masses of the unregularised two-by-two Bernoulli spectral measure and the selected zero and outlier masses in dimension three; if the resulting L3−C2 is not strictly positive, the small-ε separation FC(ν3,ε)>FC(ν2,ε) collapses.","tokens_in":16542,"feed_emoji":"📉","tokens_out":861,"duration_ms":17880,"temperature":0.7,"pith_summary":"Random-matrix limit laws minimise logarithmic energies with quadratic penalties, and it was conjectured that the same energies decrease with dimension along the mean empirical spectral distributions of Wigner matrices and of matrices with i.i.d. entries. This paper shows the unrestricted conjecture is false. It builds a real symmetric Wigner sequence whose entries are themselves semicircle-distributed, so that the mean spectral measure already equals the semicircle law in dimension one, yet the penalised energy strictly rises from dimension one to two while keeping second moment one and finite energy. It also builds a one-parameter family of Gaussian-smoothed two-point entry laws for which the circular-law energy is larger in dimension three than in dimension two once the smoothing is small enough. The constructions leave open whether monotonicity could still hold under extra assumptions or only after a large enough dimension.","feed_headline":"Log-energy monotonicity fails for general random matrices","feed_subtitle":"Explicit Wigner and Bernoulli-entry counterexamples break the conjectured decrease with dimension","key_machinery":"The semicircle gap identity expressing FR(μ)−3/4 as the integral of the positive potential vsc plus a nonnegative Fourier discrepancy, together with collision-mass bounds that control how logarithmic energy of Gaussian-smoothed atomic spectral measures scales with log(1/ε).","core_discovery":"The conjectural decrease of the quadratically penalised logarithmic energies FR and FC along mean empirical spectral distributions fails in general: there is a finite-energy real Wigner sequence with m2(ρn)=1 for which FR(ρ2)>FR(ρ1)=FR(σ)=3/4, and a concrete centred unit-variance Gaussian-regularised Bernoulli family for which FC(ν3,ε)>FC(ν2,ε) for all sufficiently small ε>0.","pith_inferences":["The same collision-mass comparison may extend to other sparse or lattice-supported entry laws, giving a systematic source of low-dimensional counterexamples.","Whether monotonicity recovers for all entry laws after some N0, or only inside classical ensembles, is now the natural next quantitative question.","The semicircle-gap identity may be reusable to certify energy increase whenever a positive mass leaks outside [−2,2]."],"forward_implications":["Unrestricted dimensional monotonicity of FR and FC for mean ESDs does not hold for general Wigner or i.i.d. matrices.","Any true monotonicity statement must impose further restrictions on the entry law or a large enough dimension threshold.","Finite-energy counterexamples exist even when every mean ESD has a bounded density and all moments.","Gaussian regularisation of discrete entry laws is enough to produce rigorous energy comparisons via collision masses."],"fun_headline_variants":["Wigner counterexample breaks log-energy monotonicity conjecture","Finite-energy Wigner sequence violates dimensional decrease of FR","Bernoulli-entry family counters FC monotonicity in low dimensions","Log-energy fails to decrease with dimension for random matrices","Explicit counterexamples disprove conjectured FR and FC decrease"],"cache_read_input_tokens":128,"weakest_assumption_plain":"The i.i.d. counterexample stands or falls on a strict numerical inequality between an exact two-dimensional collision mass and a partial three-dimensional collision mass obtained by enumerating atoms and selected eigenvalue branches.","fun_headline_variants_meta":{"raw":{"variants":["Wigner counterexample breaks log-energy monotonicity conjecture","Finite-energy Wigner sequence violates dimensional decrease of FR","Bernoulli-entry family counters FC monotonicity in low dimensions","Log-energy fails to decrease with dimension for random matrices","Explicit counterexamples disprove conjectured FR and FC decrease"]},"model":"grok-4.5","effort":"low","cost_usd":0.004115,"raw_usage":{"total_tokens":1159,"prompt_tokens":601,"num_sources_used":0,"completion_tokens":83,"cost_in_usd_ticks":41148000,"prompt_tokens_details":{"text_tokens":601,"audio_tokens":0,"image_tokens":0,"cached_tokens":256},"completion_tokens_details":{"audio_tokens":0,"reasoning_tokens":475,"accepted_prediction_tokens":0,"rejected_prediction_tokens":0}},"tokens_in":601,"tokens_out":83,"duration_ms":7602,"temperature":1.0,"reasoning_tokens":475,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-07-31T22:00:40.183704+00:00","model_set":{"reader":"grok-4.5"},"falsifier":"Recompute the fourteen atoms and masses of the unregularised two-by-two Bernoulli spectral measure and the selected zero and outlier masses in dimension three; if the resulting L3−C2 is not strictly positive, the small-ε separation FC(ν3,ε)>FC(ν2,ε) collapses.","supporting_citations":[],"review_version":1}