{"id":"9973b282-22f2-492d-97a2-0da4a1f9aa9a","arxiv_id":"2607.08481","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.5,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Gram/Wishart/Stiefel reformulation of N=2 large-d BFSS/BMN endpoints absorbs -A into a shifted mass and recovers the universal continuum -2d DΛ-channel after non-polynomial transverse completion.","lead":"This paper rewrites the N=2 large-d BFSS/BMN endpoint theory in Gram/Wishart/Stiefel variables so holonomy invariants become radial Wishart eigenvalues plus relative Stiefel angles. It shows that absorbing the large-R linear -A term yields a shifted mass and recovers the universal continuum -2d DΛ-channel after a non-polynomial transverse completion, while the anisotropic βΛ-channel stays outside pure B-type models.","discovery_kind":"extension","skeptic_critique":{"model":"grok-4.5","headline":"The balanced +A split is an ad-hoc numerical fix that may not preserve the physical continuum mass shift.","rationale":"The pure -A Cartesian benchmark and the Bessel/HCIZ prefactor that cancels one Wishart block are solid and independently checkable; they correctly produce the shifted mass and the -2d law in the pure -A theory. The load-bearing step for the full transverse claim is the subsequent bookkeeping that keeps this same unshifted M after the non-polynomial completion is introduced. The reader correctly isolates the balanced split as the weakest link: it is fixed only numerically for one pair (R_*, A_d), is not a convex decomposition, and is required to cancel an otherwise O(1) residual slope that would spoil the continuum coefficient. Because the paper already shows that without the split one obtains M_flat \neq M, the concern is internal and decisive for the strongest claim. No stronger objection (e.g., failure of the pure -A kernel itself) appears. The verdict therefore remains CONDITIONAL, pending an analytic or continuum-controlled construction of the split (or an alternative regularization that preserves M without it).","tokens_in":57332,"tokens_out":701,"duration_ms":8234,"concrete_test":"Solve the two flat-point equations (8.47) analytically or by high-precision root-finding for a continuous range of R_* around 1.545 (and for A_d = (d-1)/(2(d-2)) at large d). If no real solution with finite w exists for R_* in a neighborhood of the constrained saddle, or if the resulting continuum free-energy coefficient deviates from -2(d-2), the balanced-split construction fails and the universal -2d claim does not hold.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"The strongest claim requires that after absorbing -A into the Gaussian sector (producing M = α_Λ - 1/2) and replacing the residual transverse potential by its summed completion, a balanced compensating split +A = (1-w)A + wA with w ≈ 3.75 cancels the residual linear slope at a new flat point, so that the Wishart free energy retains the unshifted M and yields log(Z_W(x)/Z_W(0)) = -2(d-2)x^{2} + O(x^{4}) (Eq. (8.60)). Section 8.2.1 fixes w by the simultaneous numerical conditions log cosh eB_w = w eA_w and K_tot_w = 0 for the single values R_* ≈ 1.545, A_d ≈ 1/2, producing w ≈ 3.75 (a large add-subtract, not a convex partition). No existence/uniqueness proof is given for general R_* or continuum scaling of R_*, and the paper itself notes that without the split one obtains M_flat = M + (1/2)K_flat, which would replace the universal -2d coefficient by a different continuum expansion. Thus the recovery of the physical shifted mass (and therefore the claimed D_Λ-channel law) rests on an unproven numerical cancellation rather than a derived identity.","agreement_with_reader":"agree"},"referee_report":{"model":"grok-4.5","summary":"The paper reformulates the N=2, large-d planar endpoint theory of lattice BFSS/BMN matrix quantum mechanics in Gram/Wishart/Stiefel variables: rank-two Wishart eigenvalues and relative Stiefel angles. The large-R aligned holonomy asymptotics contribute a universal linear term −A, which is absorbed into the Gaussian sector to produce the shifted mass (α_Λ)_eff = α_Λ − 1/2. The exact O(2) integral yields a rank-two Bessel kernel; the pure −A theory, solved exactly in Cartesian variables, fixes the leading Bessel/HCIZ structure (aligned exponential plus a prefactor that cancels the spurious doubled Wishart entropy). Applied to the transverse B-type expansion and a non-polynomial toy completion V_toy = −log cosh B, finite polynomial truncations produce an apparent large-d perturbativity bound incompatible with the continuum; after a summed local completion and a balanced compensating split of the +A term, the Wishart saddle is recovered with the physical shifted mass and the continuum free-energy ratio log(Z_W(x)/Z_W(0)) = −2(d−2)x² + O(x⁴), reproducing the universal D_Λ-channel −2d contribution. The anisotropic β_Λ-channel is stated to lie outside pure transverse B-type descriptions.","tokens_in":57663,"tokens_out":1594,"duration_ms":20911,"significance":"If the construction holds, it supplies a systematic radial/angular dictionary for the planar endpoint theory and isolates how the universal D_Λ-channel −2d is already encoded in the shifted Wishart measure after Bessel resummation of −A. The pure −A Cartesian identity (6.32)–(6.33) is an independent, parameter-free benchmark that cleanly fixes the leading kernel prefactor and cancels doubled Wishart entropy; that step is a genuine technical contribution. The continuum coefficient is not fitted but follows from the shifted mass M = α_Λ − 1/2. The work is part of a multi-paper program and is most valuable as a structural reformulation rather than a new physical prediction; its significance for matrix quantum mechanics and large-d holonomy dynamics is real but incremental, and depends on controlling the residual approximations (weak-minor truncation, toy completion, and the balanced +A split).","major_comments":[{"comment":"§8.2.1, Eqs. (8.47)–(8.50): the recovery of the physical mass M = α_Λ − 1/2 (and hence the claimed continuum law (8.60)) rests on a balanced compensating split +A = (1−w)A + wA with w fixed numerically so that the residual linear slope vanishes at a new flat point. For R_* ≃ 1.545 and A_d ≃ 1/2 the paper reports w ≃ 3.75 (a large add–subtract, not a convex partition). No existence/uniqueness argument is given for general R_* or under continuum scaling of R_*, and the text itself notes that without the split one obtains M_flat = M + (1/2)K_flat, which would alter the continuum expansion. This is load-bearing for the abstract’s claim that the Wishart saddle is recovered with the physical shifted mass. Either a general existence proof (or continuum-controlled construction) of the split, or a clear statement that the −2d law is conditional on this bookkeeping device and fails if the residual","section":null},{"comment":"§§4.4–5.2 and 7.1: the continuum analysis is carried out under the weak-minor (or improved weak-minor) truncation δ → 0, with the residual κ-dependent determinant splitting suppressed while ρ_± still retain two-branch memory. The paper does not quantify the error of this truncation on the continuum free-energy coefficient, nor show that the omitted minor sector cannot renormalize M or generate O(d) corrections that compete with −2(d−2)x². Since the pure −A sector is treated nonperturbatively but the transverse sector is not, a controlled estimate (or an explicit argument that the minor contribution is subleading in the continuum scaling of R) is needed before the D_Λ-channel claim can be regarded as established within the Gram formulation.","section":null},{"comment":"§6.3.3, Eq. (6.45) and §7.1, Eq. (7.8): the rank-two orthogonal Bessel/HCIZ kernel is introduced as a structural ansatz whose exponential is fixed by large-argument alignment and whose prefactor is fixed by the Cartesian identity (6.33). The residual factor P(v,w;ℓ) is asserted to be subleading for continuum scaling, but no bound or asymptotic control on P is given. Because the cancellation of one full Wishart entropy block is essential to obtaining a single soft Gaussian block (and thus −2d rather than ∼−4d), the continuum coefficient inherits any uncontrolled contribution from P. A sharper statement of what is proven versus assumed for K_d would strengthen the central pure −A result.","section":null}],"minor_comments":[{"comment":"Figure 2 is referenced as contrasting the toy and quartic B-potentials but is not described in enough detail in the caption for a reader who cannot see the plot; a short quantitative caption (e.g., location of artificial minima of the quartic) would help.","section":null},{"comment":"Notation for the shifted mass is not fully uniform: (α_Λ)_eff, M, and meff appear in different sections for closely related quantities; a single convention table early in §1.3 would reduce friction.","section":null},{"comment":"The manuscript relies extensively on the author’s prior endpoint and large-d papers [76–81]. For a standalone reading, a short self-contained recap of the constrained saddle R_* ≃ 1.545 and of the exact continuum split C_ex_2 → d(d−1) would help non-specialists.","section":null},{"comment":"§3.1: the Bernoulli asymptotics of c_{2n} are given, but the radius of convergence of the toy series in the physical B-range is not stated; one sentence would clarify that the series is only a local diagnostic.","section":null},{"comment":"Typos and style: “eVhol” / “eVtoy” notation for completed potentials is easy to misread in plain text; consider a more distinctive accent or a roman “comp” subscript consistently. Occasional long sentences in §§1.3 and 8.1 could be split for readability.","section":null}],"recommendation":"major_revision","confidential_remarks":"The pure −A Cartesian identity and the associated kernel prefactor are the strongest, cleanest parts of the manuscript and would stand even if the balanced-split continuum recovery were demoted to a consistency check. The main risk for the journal is that §8 currently presents a numerical bookkeeping fix as if it restores a universal law; after revision that distinction should be transparent. Fit to hep-th is clear; novelty is incremental within the author’s series rather than a standalone breakthrough. I would not reject on the basis of the self-citation pattern alone, but the editor may wish to ensure the paper can be read without the full prior sequence."},"author_rebuttal":null,"desk_editor":{"model":"grok-4.5","letter":"The new piece here is the Gram/Wishart/Stiefel rewrite of the N=2 planar endpoint theory, plus the exact O(2) integral that produces a rank-two Bessel kernel. The pure -A model, which is solvable in Cartesian variables, cleanly fixes the leading exponential (alignment) and the prefactor that cancels the spurious doubled Wishart entropy. That identity is the strongest part of the paper; everything else is built around it.\n\nOnce that kernel is in place, the large-R linear -A is absorbed into the Gaussian sector, giving the shifted mass (α_Λ)_eff = α_Λ - 1/2. Finite polynomial truncations of the transverse B-potential then produce an apparent large-d perturbativity bound that is incompatible with the continuum. The paper shows this is a truncation artifact: after the non-polynomial completion (toy model plus compensating +A) and a balanced split of the +A term, the Wishart saddle reappears with the physical shifted mass and the free-energy ratio becomes log(Z_W(x)/Z_W(0)) = -2(d-2)x^{2} + O(x^{4}). That is exactly the universal D_Λ-channel coefficient they were after. The β_Λ-channel is correctly declared out of scope for a pure B-theory.\n\nThe soft spot is real but limited. The balanced split that cancels the residual linear slope is fixed numerically (w ≈ 3.75 for R_* ≈ 1.545, A_d ≈ 1/2). It is a large add-subtract, not a convex partition, and there is no general existence proof. Without it one gets a shifted M_flat and loses the clean -2d law. Weak-minor and factorized-kernel approximations are also present, but they are controlled and stated as such. The paper sits inside the author’s own series, so novelty is incremental rather than revolutionary; the pure -A Cartesian check keeps it from being circular.\n\nThis is for people already working on large-d BFSS/BMN endpoint formulations or Molien–Weyl holonomy dynamics. It is not a first paper for outsiders. Math and citation pattern look solid for the subfield. I would send it to referees; the balanced-split construction needs firmer analytic control, but the core derivation is worth the referee time.","headline":"Solid technical step that recovers the continuum -2d DΛ-channel via Wishart/Stiefel variables and a pure -A benchmark; the balanced +A split is the real soft spot but does not sink the paper.","tokens_in":58275,"tokens_out":630,"would_cite":true,"duration_ms":9371,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"grok-4.5","headline":"Absorbing the large-R holonomy term −A into a shifted Gaussian recovers the universal −2d continuum free-energy law via Wishart saddles.","keywords":["BFSS matrix model","BMN matrix model","Wishart variables","Stiefel manifolds","holonomy dynamics","large-d expansion","Bessel kernel","planar endpoint theory"],"falsifier":"Evaluate the continuum free-energy ratio for the completed transverse potential with the compensating +A term left unsplit (or with weight w = 1) and check whether it still yields −2(d−2)x^{2} or instead produces a renormalized mass that spoils the universal coefficient.","tokens_in":58139,"feed_emoji":"⚛️","tokens_out":1036,"duration_ms":26047,"temperature":0.7,"pith_summary":"The paper rewrites the N=2 large-d planar endpoint theory of BFSS/BMN matrix quantum mechanics on the lattice in Gram/Wishart/Stiefel variables: endpoint vectors become rank-two Wishart eigenvalues plus relative Stiefel angles. The dominant large-R aligned piece of the holonomy potential is the linear term −A; absorbing it into the Gaussian sector produces the shifted mass (α_Λ)_eff = α_Λ − 1/2, which the exact O(2) integral encodes as a rank-two Bessel kernel. Finite polynomial truncations of the remaining transverse B-potential create an artificial large-d perturbativity bound incompatible with the continuum, but summing the local completion and balancing the compensating +A term restores the physical Wishart saddle. The continuum free-energy ratio then behaves as log(Z_W(x)/Z_W(0)) = −2(d−2)x^{2} + O(x^{4}), reproducing the universal D_Λ-channel contribution. A sympathetic reader cares because this isolates how holonomy dynamics and continuum scaling already sit inside the radial Wishart measure and angular kernel of matrix quantum mechanics, without needing the full anisotropic β_Λ channel.","feed_headline":"Wishart saddles recover the universal −2d free-energy law","feed_subtitle":"Mass shift from large-R holonomy plus non-polynomial completion erases fake large-d bounds and isolates the D-channel.","key_machinery":"The rank-two orthogonal Bessel/HCIZ angular kernel obtained from the exact O(2) integral over relative Stiefel and planar angles; its exponential selects the aligned configuration while its prefactor cancels the spurious doubled Wishart entropy, fixing the pure −A theory and enabling the shifted Wishart analysis of the transverse sector.","core_discovery":"After the large-R linear contribution −A is absorbed into the Gaussian sector (producing the shifted mass (α_Λ)_eff = α_Λ − 1/2) and the residual transverse potential is replaced by its summed non-polynomial completion with a balanced compensating +A split, the Wishart saddle is recovered with the physical shifted mass; the continuum free-energy ratio then satisfies log(Z_W(x)/Z_W(0)) = −2(d−2)x^{2} + O(x^{4}), reproducing the universal −2d D_Λ-channel of the exact endpoint theory.","pith_inferences":["The same Bessel/HCIZ factorization may allow reduced Wishart Monte-Carlo sampling of continuum free energies in higher-N or higher-d regimes without sampling the full matrix coordinates.","A variational principle that systematically cancels residual linear slopes after completion could replace the numerical balanced split in related endpoint formulations.","Hybrid A-B formulations that keep the longitudinal mode dynamical are required if one wishes to recover the full d(d+1) anisotropic channel alongside the universal −2d piece."],"forward_implications":["Pure transverse B-type models capture only the universal D_Λ-channel −2d contribution; the anisotropic β_Λ-channel requires the longitudinal invariant A to remain dynamical.","Apparent large-d perturbativity bounds arising from finite polynomial truncations are truncation artifacts and disappear once the full non-polynomial completion is restored.","The endpoint Wishart measure already encodes the −2d coefficient via the shifted Gaussian normalization, before any residual holonomy interaction is added.","Continuum fine-tuning of the shifted mass is controlled by large-R holonomy asymptotics rather than by the bare Gaussian data alone."],"fun_headline_variants":["Wishart saddle recovers universal −2d free-energy after mass shift","Shifted mass plus non-polynomial completion restores Wishart −2d law","Large-R −A absorption yields physical Wishart continuum free energy","Gram-Wishart data fix free-energy ratio to −2(d−2)x^{2}","B-type completion erases fake bounds and isolates DΛ channel"],"cache_read_input_tokens":49280,"weakest_assumption_plain":"The compensating +A term must be split by a large non-convex weight so that residual linear slopes cancel at a flat point; without a consistent such split the continuum mass is corrupted and the universal −2d law fails.","fun_headline_variants_meta":{"raw":{"variants":["Wishart saddle recovers universal −2d free-energy after mass shift","Shifted mass plus non-polynomial completion restores Wishart −2d law","Large-R −A absorption yields physical Wishart continuum free energy","Gram-Wishart data fix free-energy ratio to −2(d−2)x^{2}","B-type completion erases fake bounds and isolates DΛ channel"]},"model":"grok-4.5","effort":"low","cost_usd":0.005786,"raw_usage":{"total_tokens":1687,"prompt_tokens":988,"num_sources_used":0,"completion_tokens":86,"cost_in_usd_ticks":57860000,"prompt_tokens_details":{"text_tokens":988,"audio_tokens":0,"image_tokens":0,"cached_tokens":256},"completion_tokens_details":{"audio_tokens":0,"reasoning_tokens":613,"accepted_prediction_tokens":0,"rejected_prediction_tokens":0}},"tokens_in":988,"tokens_out":86,"duration_ms":5530,"temperature":1.0,"reasoning_tokens":613,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-07-10T06:46:19.435761+00:00","model_set":{"reader":"grok-4.5"},"falsifier":"Evaluate the continuum free-energy ratio for the completed transverse potential with the compensating +A term left unsplit (or with weight w = 1) and check whether it still yields −2(d−2)x^{2} or instead produces a renormalized mass that spoils the universal coefficient.","supporting_citations":[],"review_version":1}