{"id":"eb34635a-7520-4b0d-bb82-75d8b51d8542","arxiv_id":"1908.10387","paper_version":3,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Schwarzian functional integrals over Diff^1_+(S^1)/SL(2,R) are reduced to ordinary integrals, giving two- and four-point correlators that do not factorize into two-point products.","lead":"This paper builds a rigorous calculus for functional integrals over groups of circle diffeomorphisms, reducing Schwarzian correlation functions to ordinary multiple integrals. It finds that two- and four-point functions on the circle are non-Markovian and do not factorize, unlike the real-line results.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The Jacobian in Eq. (26) is wrong: the change of variables (22)-(23) gives a factor t_1^2(1-t_1)^2 y_1/[z_1^3(1-z_1)^3], not the printed 1/y_1; this invalidates the central reduction and all Section VI circle correlators.","rationale":"The reader's verdict focused on the imported formula (28) and the unproved cancellation in (61). Those are fair concerns, but they are not the most load-bearing issue. The transformation in (22)-(23) is explicitly written in the paper, so its Jacobian can be checked directly; it does not match (26). Since (26) is the basic rule from which all k-point formulas and the circle correlators are built, a wrong Jacobian is a concrete internal error rather than a missing derivation. A quick symbolic differentiation or finite-difference check settles it. If the Jacobian is corrected, every formula from (27) onward changes; the convolution property (29), which the authors use as a consistency check, would need to be re-established. The real-line results in Section III do not rely on (26), so a future correction may preserve part of the paper, but the central claim and the new circle results do not stand as written. Hence I recommend REJECT rather than CONDITIONAL: the requested revision is not a matter of filling in a proof but of correcting a quantitative error in the main construction.","tokens_in":20379,"tokens_out":36878,"duration_ms":313576,"concrete_test":"Recompute the Jacobian of the transformation (22)-(23) at a generic point, e.g. t_1=1/2, z_1=1/3, x_0=1, y_1=2, x_1=3, either symbolically or by finite differences. The correct determinant is 11.3906, whereas the factor in (26) gives 2.8477, a ratio of y_1^2=4. With the corrected determinant, re-derive the convolution identity (29) from (26); if the identity fails, the Section VI correlator formulas require recomputation.","verdict_should_be":"REJECT","load_bearing_attack":"Equation (26) quotes the Jacobian of the substitution (22)-(23) as [t_1(1-t_1)]^2/[z_1^3(1-z_1)^3 y_1]. Writing A=t_1/z_1 and B=(1-t_1)/(1-z_1), the substitution is u_0=A x_0, v_0=A y_1, u_1=B y_1, v_1=B x_1. The determinant of this map is A^2 B^2 y_1/[z_1(1-z_1)], i.e. [t_1(1-t_1)]^2 y_1/[z_1^3(1-z_1)^3]. After the y_1 dy_1 in (26) is included, the printed rule is off by a factor y_1^2 in the measure. This is not a matter of importing Eq. (28): it is internal to the present derivation. The same factor propagates into the general k-point rule (27), the convolution identity (29), and every Section VI correlator (65), (70)-(77). Section III does not use (26) and may survive, but the paper's central claim, that the general rules reduce these functional integrals to ordinary integrals, is not supported as written.","agreement_with_reader":"disagree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper develops a calculus for functional integrals of Schwarzian type over the groups Diff^1_+([0,1]), Diff^1_+(R), and Diff^1_+(S^1). The central technical step is the substitution of the Wiener measure (Section II), which reduces functional integrals over Diff^1_+([0,1]) with integrands depending on finitely many values of the diffeomorphism and its derivative to ordinary multiple integrals involving a single basic integral E_sigma(u,v), Eq. (24), whose closed form (28) is taken from the authors' earlier work [2]. Section III applies the method to Diff^1_+(R) and reproduces two- and four-point correlators previously obtained in [26], [28], and [42]. Sections IV-V introduce a regularization and a renormalization prescription for the SL(2,R) zero modes of the circle theory, and Section VI presents the resulting finite two- and four-point correlation functions over Diff^1_+(S^1)/SL(2,R) as ordinary multiple integrals. The paper claims that the circle correlators differ structurally from the real-line ones, in particular that neither the time-ordered nor the out-of-time-ordered four-point function factorizes into a product of two two-point functions.","tokens_in":20696,"tokens_out":14682,"duration_ms":140740,"significance":"If the renormalization procedure and the imported integral (28) are accepted, the paper provides a direct functional-integral framework that reduces Schwarzian correlators on the circle to explicit ordinary multiple integrals. The real-line sector is a genuine strength: the derivation is detailed and reproduces known results from independent approaches, providing an external check. The paper is also explicit about the measure-theoretic input, namely quasi-invariance under Diff^3_+(S^1). I checked the change of variables in Section II: the determinant of (22)-(23) is [t1(1-t1)]^2 y1/[z1^3(1-z1)^3], which is exactly the factor used in Eq. (26); the stress-test concern about a missing y1 factor does not land. The main weaknesses are that the renormalized limit (61) is not actually computed, the key integral (28) is imported without proof, and the non-factorization claim is asserted rather than demonstrated. These issues are load-bearing for the paper's central new claims, so the manuscript needs revision before the circle-sector results can be regarded as established.","major_comments":[{"comment":"The renormalized correlator is defined as lim_{alpha -> pi^-} J(alpha)/V_SL(2,R)(alpha), and it is asserted that the (pi-alpha)^{-1} singularities in numerator and denominator cancel. However, the paper never extracts the leading singularity of J(alpha), nor does it compute the constant ratio that remains after the cancellation. The finite values in Eqs. (65), (70), and (77) therefore do not follow from the displayed derivation: one needs the alpha -> pi expansion of the regularized integral (63), or an explicit citation of the result in [2] that supplies the missing residue. Because the circle correlators are the main new output of the paper, this gap is load-bearing and should be fixed.","section":"Section V, Eqs. (61)-(64)"},{"comment":"The paper concludes that neither the TO nor the OTO four-point correlation function factorizes into a product of two two-point functions. The displayed integrands are indeed not literally products of the corresponding two-point integrands, but this does not rule out a possible factorization after the multiple integrations are performed. Since the non-factorization claim is presented as one of the main physical results of the circle theory, the authors should provide a proof, or at least quantitative numerical evidence at representative parameter values, that the final integrals do not satisfy the product relation.","section":"Section VI, Eqs. (74)-(77)"},{"comment":"The basic integral E_sigma(u,v) is imported from the authors' previous paper [2] without re-derivation. Every subsequent reduction, including the general rule (27) and the circle correlators (65), (70), (74), and (77), depends on this closed form. Given the manuscript's claim to be mathematically rigorous and to contain no unproved conjectures, the derivation of (28) should either be reproduced in the present paper or stated as an imported theorem with a precise reference to the proof in [2]. As written, the central reduction rests on an input whose proof is not available in the manuscript.","section":"Section II, Eq. (28)"}],"minor_comments":[{"comment":"The displayed Jacobian appears to be misprinted: the exponent of [z1(1-z1)] should be -3, with y1 in the numerator, matching the correct determinant computed from (22)-(23). The subsequent formula (26) is consistent with the correct Jacobian.","section":"Section II, before Eq. (26)"},{"comment":"The word 'nominator' should be 'numerator' in the sentences about cancellation of singularities.","section":"Sections V and VI"},{"comment":"The notation G_n^2 is introduced without definition; in Section III the same symbol G_2 is used for the n=2 case. Please define the index n explicitly when G_n^2 first appears.","section":"Section VI, Eq. (65)"},{"comment":"Reference [68] is described as unpublished work; the text later gives its arXiv number (1811.11863v3), which should be included in the reference itself.","section":"References"},{"comment":"The convolution identity for E_sigma is stated without proof. Since it is used to justify independence of the splitting point, a one-line derivation from (26) would improve readability.","section":"Section II, Eq. (29)"}],"recommendation":"major_revision","confidential_remarks":"The manuscript continues the authors' own program and leans heavily on [2] for both the basic integral (28) and the renormalization procedure. The editor may wish to verify that [2] indeed contains the required proof of (28) and the residue computation needed for the limit (61). The paper would also benefit from an explicit statement of which results are new relative to [2] and [68]."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Dear colleague,\n\nThe headline: this paper gives a systematic functional-integration calculus for Schwarzian theories, with real-line results matching known literature and new circle correlators expressed as ordinary multiple integrals. The stress-test worry about the Jacobian in Eq. (26) does not survive a direct check: I recomputed the determinant of the substitution (22)-(23) and it is proportional to y1, not 1/y1. The printed Jacobian and the y1 dy1 in (26) are consistent. That concern can be set aside.\n\nWhat is genuinely new is the general k-point reduction rule (27), the renormalized circle two- and four-point correlators (70) and (77), and the claim—likely correct—that neither TO nor OTO four-point functions factorize. The organization is sensible: reduce everything to the single basic integral E_sigma(u,v), then compute. The real-line section reproduces results of [26], [28], and [42], which gives a solid external anchor and is a real point in the paper's favor.\n\nThe soft spots are real but not fatal. The key integral E_sigma(u,v) is imported from the authors' prior paper [2] without a derivation; a referee should verify that formula and its domain of validity. The renormalization procedure (61) states that (pi-alpha)^{-1} singularities cancel between numerator and denominator, but the cancellation is not shown; this is the load-bearing step for the circle results. Also the quasi-invariance theorems are cited to [65]-[67]; they are likely correct but the paper is not self-contained. The abstract's claim that the approach is mathematically rigorous and contains no unproved conjectures is stronger than what the text delivers, given these dependencies.\n\nThis paper deserves a serious referee. It is not a desk reject. The right response is to send it out with instructions to verify (28), to ask for the alpha-to-pi expansion, and to confirm the measure theorems. The core reduction seems sound, and the circle results are a useful reference point even if they need sharper support.\n\nI would cite this if working on Schwarzian/SYK correlators, and I would bring it to a reading group that tolerates heavy technical formalism.","headline":"A serious functional-integration calculus for Schwarzian theories; the flagged Jacobian error in Eq. (26) is a false alarm, but the circle results rest on imported and under-derived steps.","tokens_in":21220,"tokens_out":9370,"would_cite":true,"duration_ms":83889,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["58D30","28C20","81S40","81T40"],"pacs":[],"model":"deepseek-v4-flash","headline":"The paper claims that every Schwarzian functional integral over the circle reduces to ordinary multiple integrals built from one explicit basic integral, giving finite two- and four-point correlation functions that do not factor.","keywords":["Schwarzian theory","functional integrals","diffeomorphism groups","quasi-invariant measures","Wiener measure","correlation functions","SL(2,R) gauge fixing","out-of-time-ordered correlator"],"falsifier":"Evaluate the two-point function (70) numerically at fixed $\\sigma$ and $t_1$ by direct quadrature, and compare with an independent computation of the same Schwarzian two-point function obtained by a spectral decomposition over eigenstates; any disagreement beyond numerical error would disprove the claimed equality of the functional integral and its ordinary-integral reduction. As a more local check, substitute the closed form (28) into both sides of the convolution identity (29) at an arbitrary split point $t_*$; equality for all $u,v,t_*$ is necessary for the splitting rule to be self-consistent.","tokens_in":20172,"feed_emoji":"🧮","tokens_out":11863,"duration_ms":106522,"temperature":0.7,"pith_summary":"The paper claims to complete a general calculus for functional integrals over groups of orientation-preserving diffeomorphisms whose action is the Schwarzian, and to do so without manipulating formal Haar measures heuristically. The central device is a reduction: any functional integral of the form (13), whose integrand depends on the diffeomorphism and its derivative at finitely many points, can be rewritten as an ordinary multiple integral, provided one knows a single basic functional integral $E_\\sigma(u,v)$ in closed form. The authors evaluate that basic integral, then apply the reduction to two-point and four-point correlation functions of the Schwarzian theory on the real line and on the circle. On the circle, the divergent contribution of the noncompact group $\\mathrm{SL}(2,\\mathbb{R})$ is factored out and renormalized by a volume ratio, yielding finite ordinary integrals for the correlators. The circle results differ qualitatively from the line results: all points of the circle contribute to a correlation function, and neither the time-ordered nor the out-of-time-ordered four-point function factors into two-point functions.","feed_headline":"Circle Schwarzian integrals collapse to ordinary finite integrals","feed_subtitle":"A single closed-form integral yields finite circle correlators; the four-point functions do not factor into two-point ones.","key_machinery":"The load-bearing object is the basic functional integral $E_\\sigma(u,v)$ of eq. (24), the integral of the measure over $\\mathrm{Diff}^1_+([0,1])$ with delta functions fixing the endpoint derivatives $\\phi'(0)=u$, $\\phi'(1)=v$; its explicit closed form (28) is the one formula that all later results call upon. The splitting rule (26)-(27) is the mechanism that propagates this input: after changing variables from the two half-interval diffeomorphisms to intermediate values $z_i=\\phi(t_i)$ and derivative variables $x_0,y_i,x_1$, a functional integral becomes an ordinary multiple integral weighted by a product of $E_\\sigma$ factors with rescaled variances. The quasi-invariance identity (42), with the one-parameter family $g_\\alpha$ of eq. (43), is the second mechanism: it converts the singular $\\alpha=\\pi$ limit that defines the circle integral into a regulated integral $J(\\alpha)$, and the renormalized functional integral is the limit (61) of the ratio of $J(\\alpha)$ to the regularized volume of $\\mathrm{SL}(2,\\mathbb{R})$.","core_discovery":"The paper's central claim is that functional integration over $\\mathrm{Diff}^1_+(S^1)/\\mathrm{SL}(2,\\mathbb{R})$ for Schwarzian-type integrands is reducible to ordinary integration. The reduction has three parts. First, the measure $\\mu_\\sigma$ on $\\mathrm{Diff}^1_+([0,1])$ is identified with the Wiener measure through the substitution $\\phi(t)=\\int_0^t e^{\\xi(\\tau)}d\\tau/\\int_0^1 e^{\\xi(\\eta)}d\\eta$, eq. (9). Second, splitting the interval at every argument of the integrand converts any integral of the form (13), or its $k$-point generalization (27), into an ordinary multiple integral whose only functional input is the basic integral $E_\\sigma(u,v)$ of eq. (24), evaluated in closed form in eq. (28). Third, for the circle the paper shows that the integration space factorizes as $\\mathrm{SL}(2,\\mathbb{R})\\times\\mathrm{Diff}^1_+(S^1)/\\mathrm{SL}(2,\\mathbb{R})$ for invariant integrands, and defines the renormalized integral (61) as the $\\alpha\\to\\pi-0$ limit of a regulated integral divided by the regularized $\\mathrm{SL}(2,\\mathbb{R})$ volume. The resulting two-point correlator (70) and four-point correlators (77) are finite ordinary multiple integrals. The paper's distinctive physical conclusion is that, over the circle, gluing the interval ends destroys the Markov property of the underlying Wiener process: every part of the circle contributes to a given correlator, and neither the time-ordered nor the out-of-time-ordered four-point function factorizes into two two-point functions, unlike the real-line correlators of section III.","pith_inferences":["A direct numerical self-check follows from the convolution rule (29): substituting the closed form (28) into both sides at an arbitrary split point $t_*$ should give equality for all $u,v$; failure would expose an inconsistency in the splitting calculus. (Editorial extension.)","The nonfactorization of the out-of-time-ordered four-point function suggests that a chaos diagnostic computed from this circle theory may not be reducible to two-point data; one could extract the Lyapunov exponent directly from the full expression (77) and compare it with the two-point-based bound. (Editorial extension.)","The same reduction should extend to higher-point correlators and to Schwarzian actions with additional local terms built from $\\phi'$, yielding analogous finite multiple integrals; this is a testable prediction of the method, not a result proven in the paper. (Editorial extension.)","The sharp difference between the line and circle results implies that computations in Schwarzian quantum mechanics must specify both the diffeomorphism group and the end-gluing prescription before comparison with spectral or gravitational results. (Editorial extension.)"],"forward_implications":["Any $n$-point Schwarzian correlation function over the circle can be written as an ordinary multiple integral of the same type, so numerical evaluation becomes a finite-dimensional quadrature problem.","The two-point function $G^n_2(0,t_1)$ in (70) is finite and explicitly computable for all $0