{"id":"5b7f39c0-b91b-4814-b135-bdf8d344db82","arxiv_id":"2512.11054","paper_version":3,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":3,"one_line_summary":"Strong level repulsion produces damped crystalline oscillations of the spectral form factor, with a Debye-Waller suppression, a new plateau time scale t* ≈ t_H sqrt(β/4), and predictable derivative singularities.","lead":"This paper shows that quantum systems with very strong eigenvalue repulsion can have 'crystalline' spectral form factors—oscillations at multiples of the Heisenberg time that are damped by a Debye-Waller factor. It derives this damping in three models, bridging random-matrix and permutation-circuit statistics.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Gaussian truncation of the log-gas Hamiltonian is not validated in the deep tail; existing numerics cover only e^{-2W}>0.01, leaving the plateau-time prediction untested.","rationale":"The reader's weakest assumption correctly identifies the Gaussian truncation as the load-bearing step. However, the paper's own Supplemental estimate indicates that anharmonic corrections are suppressed by powers of 1/√β and only become noticeable for τ ~ β, which is far beyond the τ ~ √β times of interest for large β. Thus the concern is not that nonlinearities are known to contribute, but that the numerical validation does not extend to the deep tail where e^{-2W} < 0.01. Other potential issues—the undetermined constant C ≈ 3.6 and the non-self-averaging of the SFF—are secondary because they do not affect the qualitative new regime or the scaling t* ~ t_H√β. The proposed test (larger d, β = 100, τ up to 10) directly probes the untested region and would settle whether the Gaussian truncation holds where the SFF approaches the plateau. Since the reader already issued a CONDITIONAL verdict, this concern does not change that verdict; it sharpens the condition under which the central claim can be accepted.","tokens_in":16507,"tokens_out":28612,"duration_ms":261973,"concrete_test":"Compute the SFF of the circular β-ensemble via the CMV representation for d = 4096 and β = 100 at integer τ = 1, ..., 10, and compare with the Gaussian prediction Eq. (9). At τ = 5 (the claimed t*), e^{-2W} ≈ log(d)/d ≈ 0.002, below the paper's validation threshold of 0.01; at τ = 10, e^{-2W} is far smaller. If the relative error between the exact CMV SFF and Eq. (9) exceeds a few percent for τ ≥ 5, the Gaussian truncation fails before the predicted plateau onset; if it remains small, the central prediction is supported.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central Debye-Waller factor and the plateau onset t* rest on expanding the log-gas Hamiltonian (2) to quadratic order around the equidistant saddle point (Eq. (8)) and neglecting cubic and higher nonlinearities. The paper itself (Supplemental Sec. II and footnote [49]) states that the Gaussian approximation fails for τ ≫ β and that for rational β/2 = p/q the exact SFF has no singularities for τ > p. Yet the quantitative predictions—Eq. (11) and t* ≈ t_H√(β/4)—are evaluated in the crossover and tail regimes where the truncation has not been rigorously controlled beyond one-loop order. The numerical support is limited to β > 10 and e^{-2W} > 0.01; this covers the early peaks but not the deep tail where the SFF actually approaches the plateau. At t* itself, e^{-2W} ~ log(d)/d, which for d = 512 is ≈ 0.012, barely inside the tested region; for larger d it falls below 0.01, so the crossover is not tested. If anharmonic terms contribute at the level of a few percent at times τ ~ √β, the exponent e^{-2W} and the plateau time t* would shift, undermining the central claim.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies the spectral form factor (SFF) of unitary systems with very strong eigenvalue repulsion. It models the eigenvalues of the circular β-ensemble as a low-temperature Coulomb gas, expands the log-gas Hamiltonian around the equidistant 'crystal' configuration, and derives a Debye-Waller damping of the crystalline SFF: K(t_H τ) ≈ d + e^{-2W} d², with e^{-2W} given by the three asymptotic regimes in Eq. (11). It predicts that the plateau is reached only at t* ≈ t_H√(β/4), estimates the order of SFF singularities at integer multiples of the Heisenberg time via Eq. (12), and reproduces the crystalline oscillation in two further models: the perturbed permutation circuit U = e^{-igH}S and the random Lax-matrix ensemble. Numerical checks using d = 512 and β > 10 are reported for the circular β-ensemble, with stated agreement to a few percent in the regime e^{-2W} > 0.01.","tokens_in":16790,"tokens_out":10731,"duration_ms":110516,"significance":"If correct, the predicted long-lived crystalline regime with periodically peaked SFF up to t* ≫ t_H is a substantial addition to spectral statistics beyond the standard Wigner-Dyson classes, interpolating between β = ∞ permutation circuits and finite-β random matrix behavior. The derivation is internally consistent and not circular: the Debye-Waller factor follows from the Coulomb-gas Hamiltonian, and the Lax result is taken from the independent work of Bogomolny et al. The paper also provides reproducible numerical checks based on the Killip-Nenciu representation of the circular β-ensemble. The main weaknesses are the unexplained numerical constant C in Eq. (9)/(11) and the fact that the numerical verification does not extend into the deep-tail/plateau-onset regime; both need to be addressed before the central time-scale claim is fully supported.","major_comments":[{"comment":"The constant C ≈ 3.6 is simply stated as a 'numerical coefficient'. Because C enters the exponent of the tail prediction (β/τ²d)(Cπ)^{-4τ²/β} in Eq. (11), an unexplained fitted C would reduce the claimed derivation to a two-parameter fit. The elementary evaluation of the momentum sum in Supplemental Eq. (8) gives C = 2e^γ ≈ 3.56; please show that derivation explicitly (or state precisely which regularization of the divergent sum is used).","section":"SFF of the Coulomb gas, Eq. (9) and Supplemental Eq. (8)"},{"comment":"The numerical support is limited to β > 10 and e^{-2W} > 0.01. However, the central plateau-time prediction t* ≈ t_H√(β/4) concerns the regime in which the Debye-Waller damping has already reduced the peaks to a small fraction of d²; for d = 512 the crossover is at the edge of the tested window, and for larger d it lies outside. Please either provide numerical data extending to e^{-2W} ≲ 0.001 (including larger d), or give an analytic bound showing that anharmonic terms beyond the one-loop estimate in Supplemental Sec. II do not shift e^{-2W} by a nonperturbative amount up to τ ≈ √β/2. This is load-bearing because t* is one of the headline results.","section":"Eq. (11), Fig. 2, and t* claim"}],"minor_comments":[{"comment":"The first line appears to be a typographical error: S|n⟩|−⟩ = S|n⟩|+⟩ should presumably read S|n⟩|−⟩ = |n⟩|+⟩, i.e., the particle flips but does not move. Please correct.","section":"Eq. (4)"},{"comment":"The three asymptotic expressions do not match at the formal boundary 4τ²/β = 1: the first regime gives d^{-1}, the second gives log(d)/d. State the matching convention used, or define exactly which quantity is plotted in Fig. 2 so the reader can see how the crossover is obtained.","section":"Eq. (11)"},{"comment":"The model predicts α = d/g² for the pinning potential, but the inset fits α to the SFF instead of comparing with this prediction. Please report the fitted values alongside d/g² so the Debye-Waller factor in the perturbed permutation circuit is tested without a fitted parameter.","section":"Fig. 3 inset and Eq. (14)"},{"comment":"There is a typo 'Supplemental Meterial' in the Conclusion; also the sample counts in figure captions (e.g., '106 samples') should be typeset as 10^6.","section":"General"}],"recommendation":"major_revision","confidential_remarks":"The manuscript is readable and the Gaussian mechanism is plausible, and both major points are addressable within the current scope: deriving C = 2e^γ and adding deep-tail numerical data or a more rigorous anharmonic bound. I therefore recommend major revision rather than rejection. The unexplained C and untested plateau-onset regime should not be left as-is for publication."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Dear colleague,\n\nThis paper is worth a serious referee. Its central new result is the large-β behavior of the spectral form factor in the circular β-ensemble: the peaks at multiples of the Heisenberg time are suppressed by a Debye-Waller factor, and the plateau is only reached at t* ≈ t_H √(β/4). That time scale is new, and the singularity order formula (12) recovering β = 1, 2, 4 is a nice benchmark. The perturbed permutation circuit U = e^{−igH}S is a simple, workable interpolation between CUE and a perfect crystal. The Lax ensemble part is mostly quoted from Bogomolny et al., and the authors are clear about that.\n\nWhat the paper does well: the Gaussian (harmonic) expansion of the log-gas is internally consistent, the numerics are extensive (β from 1 to 20000, d = 512), and the error stays at the few-percent level in the stated regime. The authors also explicitly flag the τ ≪ β validity bound and the rational-β caveat. The Supplemental Material even estimates the leading cubic correction and argues it only becomes relevant at τ ≫ β, which is far beyond t*. So the stress-test worry about anharmonic terms is partially answered, though not with a rigorous all-orders proof.\n\nThe soft spots are real but fixable. First, the constant C ≈ 3.6 in the DW exponent is just presented as a numerical factor; it presumably comes from converting the discrete sum to a log integral, but that step should be shown or at least justified as a cutoff-dependent constant of order one. Second, the crossover region around t* is barely tested numerically—for d = 512 the DW factor at crossover is about 0.012, right at the edge of the claimed 0.01 accuracy, and it would be good to see larger d or a direct test of the t* scaling. Third, the α in Fig. 3 is obtained by fitting, which is minor and acceptable for a toy model.\n\nI would send this to peer review. The core physics is novel, the derivations are mostly transparent, and the limitations are honestly stated. The report should ask for a derivation or at least a careful discussion of C, and for a stronger numerical check of the deep tail. After that, I expect it to be accepted.","headline":"Defines a genuinely new large-β SFF regime with a Debye-Waller damped crystal and a new plateau time scale; the claims mostly hold up and the paper deserves refereeing, with requests to clarify C≈3.6 and test the deep tail.","tokens_in":17385,"tokens_out":2732,"would_cite":true,"duration_ms":31776,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"The paper claims that for unitary systems with extremely strong eigenvalue repulsion, the spectral form factor keeps crystal-like periodic peaks damped by a Debye-Waller factor until a new plateau time t* ≈ t_H√(β/4), much later than the He","keywords":["spectral form factor","circular beta-ensemble","Coulomb gas","Debye-Waller factor","level repulsion","Heisenberg time","permutation circuits","random matrix theory"],"falsifier":"Measure the SFF of the circular β-ensemble numerically for a fixed large β (e.g., β ≈ 500, d ≈ 512) and check the heights of the peaks at τ = 2 and 3: the ratio K(t_H τ)/d^2 should follow e^{-2W} ≈ d^{-4τ²/β}. If the peaks decay faster than this, or if the SFF reaches its plateau before t_H√(β/4), the quadratic approximation is not valid and the central prediction fails.","tokens_in":16337,"feed_emoji":"💎","tokens_out":8900,"duration_ms":83408,"temperature":0.7,"pith_summary":"This paper argues that the spectral form factor (SFF) of quantum systems with extremely strong eigenvalue repulsion does not settle to its plateau at the Heisenberg time t_H. Instead, at integer multiples of t_H the SFF keeps periodic 'Bragg' peaks, damped by a Debye-Waller factor e^{-2W} that the authors derive within a quadratic approximation for the circular β-ensemble Coulomb gas. The predicted time scale t* ≈ t_H√(β/4) marks when the peaks vanish and the plateau is reached, which is much later than t_H when β is large. The same crystalline oscillation appears in perturbed permutation circuits and in a random Lax-matrix ensemble, suggesting a broader universality class of intermediate level statistics. If correct, strongly level-repelling quantum systems have a new long-time dynamical regime.","feed_headline":"Strongly repelled spectra keep crystal order far past Heisenberg time","feed_subtitle":"New theory derives Debye-Waller damping of SFF peaks and predicts plateau at t_H √(β/4).","key_machinery":"The central object is the Debye-Waller factor e^{-2W}, borrowed from Bragg diffraction, which suppresses the periodic SFF peaks as thermal fluctuations move eigenvalues away from exactly evenly spaced positions. It is computed from a quadratic expansion of the log-gas (Coulomb gas) Hamiltonian around the equidistant saddle point, treating eigenvalue displacements as independent momentum modes; the same damping arises in the other two models from the perturbation parameter g (permutation circuits) and the rod length g (Lax ensemble).","core_discovery":"The paper's central claim is that the SFF of the circular β-ensemble at inverse temperature β ≫ 1 behaves near multiples of the Heisenberg time as K(t_H τ) ≈ d + e^{-2W} d^2, with Debye-Waller exponent e^{-2W} ≈ d^{-4τ²/β} for 4τ²/β ≪ 1, log(d)/d at the crossover 4τ²/β = 1, and (β/τ²d)(Cπ)^{-4τ²/β} for 4τ²/β ≫ 1. The plateau is therefore reached only at t* ≈ t_H√(β/4). The paper further estimates that the SFF's singularities at integer multiples of the Heisenberg time have order γ = 4τ²/β − 1, recovering the known β = 1, 2, 4 results at τ = 1 (and τ = 2 for β = 4), and it reproduces the same damped crystalline oscillations in a perturbed permutation circuit and in a random Lax-matrix ensembl","pith_inferences":["The same Debye-Waller mechanism should appear in other systems with large-β level statistics, such as the Gaussian β-ensemble or weakly disordered Anderson chains; a direct check would be to compute the SFF at τ = 1, 2 and fit the peak envelope to Eq. (11).","The paper's Gaussian truncation is expected to fail for non-rational β/2 at times τ ≫ β; a higher-order expansion, including the cubic vertex computed in the supplement, could predict how the SFF peak heights deviate from the pure Debye-Waller form at intermediate τ, giving a sharper falsifier.","Because the SFF is the Fourier transform of the two-level correlation function, the prediction of periodic peaks up to t* implies that the eigenvalue density autocorrelation retains sharp Bragg-like features out to distances ~√β times the mean spacing, which could be measured in level-spacing histograms.","The hard-rod gas underlying the Lax ensemble suggests an alternative experimental route: instead of tuning β, one tunes the rod length g; observing the late-time period d/g with amplitude ~d/((1−g)t) would distinguish this universality class from the circular β-ensemble."],"forward_implications":["The SFF of strongly level-repelling unitary systems does not plateau at the Heisenberg time; it keeps periodic crystal-like peaks up to t* ≈ t_H√(β/4).","The singularities of the SFF at integer multiples of the Heisenberg time have order γ = 4τ²/β − 1; for β = 1, 2, 4 this reproduces known features at τ = 1 (and τ = 2 for β = 4), showing those standard cases are remnants of Bragg peaks.","The perturbed permutation circuit U = e^{−igH}S and the random Lax-matrix ensemble provide explicit models that interpolate between β = 2 (CUE) and β = ∞ (picket fence), with SFF peak damping controlled by g.","The delay of the plateau to t* sets a new, longer time window over which eigenbasis dephasing is incomplete, affecting correlation functions and entanglement dynamics in these systems.","The Debye-Waller formula is testable on quantum processors, since the models can be implemented with shallow circuits."],"fun_headline_variants":["Crystal-like SFF plateaus at t_H√(β/4)","Strong repulsion freezes SFF oscillations to t_H√(β/4)","Debye-Waller formula tames spectral form factor peaks","Coulomb gas yields crystalline SFF fingerprints","SFF crystal order persists to Heisenberg time scaled"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The derivation assumes the fluctuations of eigenvalues around their evenly spaced positions stay small enough that only the quadratic part of the repulsion energy matters up to the predicted plateau time; if cubic and higher-order terms become significant earlier, the Debye-Waller exponent, the singularity orders, and the new time scale t* would all change.","fun_headline_variants_meta":{"raw":{"variants":["Crystal-like SFF plateaus at t_H√(β/4)","Strong repulsion freezes SFF oscillations to t_H√(β/4)","Debye-Waller formula tames spectral form factor peaks","Coulomb gas yields crystalline SFF fingerprints","SFF crystal order persists to Heisenberg time scaled"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000581,"raw_usage":{"total_tokens":2553,"prompt_tokens":703,"completion_tokens":1850,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":447,"completion_tokens_details":{"reasoning_tokens":1764}},"tokens_in":447,"tokens_out":1850,"duration_ms":15860,"temperature":1.0,"reasoning_tokens":1764,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-03T16:57:33.340887+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure the SFF of the circular β-ensemble numerically for a fixed large β (e.g., β ≈ 500, d ≈ 512) and check the heights of the peaks at τ = 2 and 3: the ratio K(t_H τ)/d^2 should follow e^{-2W} ≈ d^{-4τ²/β}. If the peaks decay faster than this, or if the SFF reaches its plateau before t_H√(β/4), the quadratic approximation is not valid and the central prediction fails.","supporting_citations":[],"review_version":1}