{"id":"39fb1058-33ad-49af-952f-0db1d5820028","arxiv_id":"2608.11550","paper_version":1,"verdict":"ACCEPT","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"low","formal_verification":"none","parameter_count":0,"one_line_summary":"The paper computes the Schwarzian anomaly term for boundary deformations of Liouville amplitudes, derives local Ward identity recursions, and uses them to recover factorization, the Shapovalov form, chiral vertex operators, smoothness, and BPZ equations.","lead":"This paper computes the scalar anomaly term that appears when the boundary parametrization of a Liouville conformal field theory surface is smoothly changed, and shows its holomorphic part is a Schwarzian boundary integral. It then derives local Ward identities on disks, annuli, and pairs of pants and uses them to re-derive central bootstrap structures, including the Shapovalov form and BPZ equations.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Ward identities and factorization depend on an unproved analytic-continuation hypothesis from [GKRV21, Section 11.2]; if the continuation domain is narrower or the weighted estimates fail for descendant states, the recursions in Propositions 4.3 and 4.7 do not follow from Theorem 1.1.","rationale":"The reader's weakest_assumption is exactly the analytic continuation of generalized Liouville amplitudes to complex boundary weights and descendant states. My reading of the proof confirms this is the main external dependency: Theorem 1.1 itself is proved from the anomaly functional by a direct variation, and I could not find a defect in that computation. The subsequent Ward identities, however, are not self-contained: they are proven probabilistically in the Seiberg region and extended by quoting [GKRV21, Section 11.2] and [BGKR24]. The extension is load-bearing because the paper's main applications operate on states Ψ_{Q+iP,ν} lying on the boundary Re(α)=Q of the probabilistic region, and the recursions require uniform estimates for arbitrarily high descendant levels. The paper states this limitation implicitly by saying 'wherever the corresponding generalized amplitudes are defined' and by citing rather than proving the continuation. This is not a fatal internal inconsistency, but it makes the central applications conditional on an external result that is not independently checked here. A conditional accept is therefore the appropriate verdict: the Schwarzian anomaly theorem appears sound, and the Ward-identity machinery is correct provided the cited analytic-continuation estimates hold as stated. No ad hominem is intended; the concern is entirely about the logical dependency chain of the proof.","tokens_in":27203,"tokens_out":44149,"duration_ms":500297,"concrete_test":"Verify the continuation hypothesis at the exact corner used in Proposition 4.3: take the round-disk amplitude A_{D^c_x}(Ψ_{Q+iP,ν}) with two bulk insertions and a fixed finite partition ν, and check that P ↦ A(Ψ_{Q+iP,ν}) is holomorphic in a complex neighborhood of P=P_0>0 with estimates uniform in |ν|, by re-deriving the statement from the spectral decomposition of H and the weighted bounds in [BGKR24, Theorem 1.2]. A complementary numerical check: for Q=1.5, P=1, and a non-real boundary weight α=Q+i, compare the level-one Ward identity A(L_{-1}Ψ_α)=D_1 A(Ψ_α) against the derivative of the DOZZ three-point function; agreement to numerical precision would support the continuation, while any disagreement would invalidate the recursion.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"The central anomaly computation in Theorem 1.1 is explicit, local, and self-contained, and I found no sign or factor error in the Schwarzian variation formula. The load-bearing weakness is elsewhere: the paper's Ward identities and the factorization theorem are applied to generalized descendant states Ψ_{α,ν} with α = Q+iP and Re(α) = Q, i.e. on the boundary of the probabilistic Seiberg region. The proofs of Proposition 4.3 and Theorem 4.7 first work in the probabilistic region and then assert: 'The identities extend to the analytically continued generalized states wherever the corresponding generalized amplitudes are defined' (Proposition 4.3), and 'the general case follows from analytic continuation' (Theorem 4.7). The continuation domain and the weighted estimates are not proved here; they are quoted from [GKRV21, Section 11.2] and [BGKR24]. This matters concretely because the recursion for, say, L_{-n}Ψ_{α,ν} inserts the same state on the opposite boundary and must be iterated for arbitrary finite |ν|; this requires uniform control of the generalized amplitude in a neighborhood of Re(α)=Q and, in Theorem 4.7, in a domain with Re(α_j)≤Q and Σ Re(α_j)>2Q. If the continuation exists only for Re(α)<Q, or if the estimates grow badly with |ν|, then the finite recursions and the factorization (1.2) would fail, even though the Schwarzian anomaly formula itself would remain correct. The paper explicitly flags this by citing, not reproving, the continuation; the reader's weakest-assumption analysis identifies the same point.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper develops a geometric framework for local Ward identities in probabilistic Liouville conformal field theory on surfaces with analytic boundary. Its central computation is the holomorphic variation of the Liouville anomaly under boundary deformations: for a genus-zero admissible surface with b incoming boundaries and a meromorphic vector field v, Theorem 1.1, proved as Theorem 3.1, identifies the combination (1/2)(∂_t S_L^0(Σ_t^v)-i∂_t S_L^0(Σ_t^{iv}))|_0 - (1/24)Σ_k [v_k]_0 with the Schwarzian boundary integral (i/24π)Σ_k ∮ v(z) S_{ζ_k^{-1}}(z) dz. The authors combine this formula with the differentiability theorem of Baverez, Guillarmou, Kupiainen, and Rhodes to derive local Ward identities for disks, annuli, and pairs of pants (Lemma 4.1, Propositions 4.3 and 4.4). These yield finite Virasoro recursions, from which the paper recovers the polynomial factorization of normalized pair-of-pants descendant coefficients (Theorem 4.7), the Shapovalov limit of annulus coefficients (Corollary 4.6), and the formal chiral vertex-operator coefficients (Proposition 4.10 and Corollary 4.11). Section 5 gives a geometric proof of smoothness of bulk correlation functions (Theorem 5.2) and derives genus-zero BPZ equations for degenerate insertions (Proposition 5.4 and Corollary 5.5).","tokens_in":27539,"tokens_out":23670,"duration_ms":258070,"significance":"The anomaly formula is the clearest new contribution; the proof of Theorem 3.1 is explicit, local, and effectively self-contained, and I found no sign or factor error in the Schwarzian contour term. If the result stands, it gives a concrete geometric origin for the Virasoro central term as a boundary Schwarzian integral. The paper also adds value by showing that the local Ward identities provide a uniform derivation of several known bootstrap facts, namely polynomial factorization, the Shapovalov limit, and chiral vertex-operator coefficients, and by offering a geometric proof of smoothness of correlations. The authors are appropriately explicit about what is quoted from GKRV21 and BGKR24. The main reservation is that the analytic continuation of generalized amplitudes to complex boundary weights, and the weighted estimates needed for descendant states, are not proved in the manuscript; the cited papers may cover these points, but the precise statements are not located. Overall the contribution is solid, yet the advertised domain of the Ward-identity applications needs to be tightened.","major_comments":[{"comment":"The extension of the Ward identities and of the factorization theorem to the physical boundary Re(alpha)=Q of the Seiberg region is asserted by analytic continuation: Proposition 4.3 says that the identities extend to analytically continued generalized states wherever the corresponding generalized amplitudes are defined, and Theorem 4.7 says that the general case follows from analytic continuation. Neither the continuation domain nor the required weighted estimates are stated or proved in this paper. The recursion for L_{-n}Psi_{alpha,nu} is finite, but it must be iterated through intermediate descendant states on the opposite boundaries, so it requires the continuation to be defined in a fixed neighborhood of Re(alpha)=Q for every descendant that appears in the recursion; Theorem 4.7 additionally requires the domain Re(alpha_j)≤Q with sum Re(alpha_j)>2Q. The cited [GKRV21, Section 11.2] and [BGKR24] may well supply this, but the manuscript does not identify the exact theorem or verify that it covers descendant insertions on this domain. Since the factorization (1.2), the Shapovalov limit (4.18), and the chiral vertex identification (4.32) all depend on this step, the proof is incomplete as written. The authors should add a precise analytic-continuation lemma and check that it applies to all descendant states appearing in the recursions.","section":"Sec. 4.1-4.2 (Prop. 4.3, Thm. 4.7)"},{"comment":"Lemma 5.1 is the key analytic input for the new geometric proof of smoothness of correlation functions. Its proof consists almost entirely of citations to [GKRV21, Theorem 4.4, Propositions 6.4 and 11.13] and [BGKR24, Lemma 4.10], and the differentiability assertion in equation (5.1) is asserted rather than demonstrated. Because Theorem 5.2 is advertised as an application of the Ward-identity method, the authors should either provide the weighted-estimate argument in detail or state explicitly that Theorem 5.2 is conditional on the collected estimates from the cited papers.","section":"Sec. 5.1 (Lemma 5.1, Thm. 5.2)"}],"minor_comments":[{"comment":"The introduction refers to a boundary-local version as Proposition 3.1, but Section 3 contains only Theorem 3.1; the cross-reference should be corrected.","section":"Introduction and Sec. 3"},{"comment":"The footnote says Proposition 4.1 was stated in [GKR24, Proposition 6.6], but the statement proved in the text is Lemma 4.1; the labels should be harmonized.","section":"Footnote 1, Sec. 4"},{"comment":"The phrase \"tripe Young diagrams\" is a typo for \"triple Young diagrams,\" and the role of the points x_j=psi_j(0) should be stated explicitly, since the statement of the theorem is otherwise ambiguous.","section":"Theorem 4.7"},{"comment":"The word \"implicityly\" should read \"implicitly.\"","section":"Lemma 5.1"},{"comment":"The notation Q_{N≥0} for the formal completion of V_{alpha_3} is confusing; it is presumably the direct product over levels, and should be typeset as a product symbol.","section":"Prop. 4.10"},{"comment":"In the definition of E(O), the summation index M is not specified; it should be a finite sum over the finitely many generators U_a in V_fin^0.","section":"Thm. 5.2 proof"}],"recommendation":"major_revision","confidential_remarks":"For the editor: the core anomaly computation in Theorem 3.1 is sound, explicit, and worth publishing; the main weakness is the unproved analytic continuation from the probabilistic Seiberg region to the boundary Re(alpha)=Q, which is load-bearing for the factorization theorem and its corollaries. This gap is likely repairable by a precise citation of the relevant statements in GKRV21 and BGKR24, or by a short self-contained lemma. The paper also needs a careful language and notation pass. I recommend major revision rather than rejection."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The genuinely new result is Theorem 1.1: the holomorphic variation of the Liouville anomaly under boundary deformations is a Schwarzian boundary integral, with a -1/24 zero-mode shift. That computation is explicit, self-contained, and I checked the sign and factor conventions in Theorem 3.1; they work. The paper then uses this to derive local Ward identities on disks, annuli, and pants, and from them finite recursions for descendant matrix coefficients. Most of the later results are re-derivations—factorization, Shapovalov form, chiral vertex coefficients—but the route is genuinely geometric and the paper is honest about what is new.\n\nThe applications are also good: the smoothness proof via zero-weight insertion and the BPZ derivation from the level-two null vector are clean. The anomaly calculation is direct enough to serve as a proof; the paper ships no code or machine-checked proofs, but the core computation is sound.\n\nThe soft spot is the analytic continuation. The recursions are first proved on the probabilistic Seiberg region and then extended to generalized descendant states with Re(α)=Q, on the boundary of that region. The extension is quoted from [GKRV21, Section 11.2] and [BGKR24], not reproved. If that continuation only holds for Re(α)<Q, or if the weighted estimates grow badly with descendant level, then the recursions and the factorization theorem would not follow, even though Theorem 1.1 itself would survive. The paper flags this by citing, but the reader is being asked to take a substantial load-bearing input on faith. The same applies to the weighted extension in Lemma 5.1.\n\nI think this is a real but bounded weakness. It does not sink the paper; it makes the ancestry of the main results conditional. A referee should ask the authors to state the continuation theorem as a separate lemma, with the exact domain and estimates, and either prove it or mark it as an assumption. The compressed iteration in Theorem 4.7 also deserves expansion.\n\nWho this is for: anyone working in the Liouville bootstrap program, especially on the geometric side. The anomaly formula itself is a useful toolkit piece.\n\nRecommendation: send it to peer review. The paper deserves serious refereeing; the central computation is sound, and the cited dependencies are likely correct, but the reliance on unproved analytic continuation needs to be explicit before publication.","headline":"The Schwarzian anomaly formula is new and clean; the descendant recursions rest on an unproved analytic continuation from [GKRV21], so a referee should demand that dependency be made explicit.","tokens_in":28076,"tokens_out":2025,"would_cite":true,"duration_ms":22279,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["81T40","17B68","81R10"],"pacs":[],"model":"deepseek-v4-flash","headline":"A Schwarzian boundary integral computes the holomorphic variation of the Liouville anomaly in genus zero, making the Virasoro central term a geometric quantity.","keywords":["Liouville conformal field theory","Ward identities","Virasoro algebra","Schwarzian derivative","Liouville anomaly","conformal bootstrap","BPZ equations","Shapovalov form"],"falsifier":"Take a three-holed sphere with one boundary parametrized by $w+\\varepsilon w^2$ and the other two round; evaluate the left-hand side of Theorem 1.1 by differentiating the Liouville anomaly along the flow of $v$ numerically, and compare with the Schwarzian contour integral on the right. Any mismatch at order $\\varepsilon^2$ would falsify the theorem.","tokens_in":26987,"feed_emoji":"🌀","tokens_out":8996,"duration_ms":104894,"temperature":0.7,"pith_summary":"This paper aims to identify the scalar anomaly that appears when one infinitesimatically changes the analytic boundary parametrizations of a Liouville conformal field theory amplitude. The paper claims that in genus zero this scalar term is a boundary integral of the Schwarzian derivative of the inverse boundary coordinate, giving a geometric origin for the Virasoro central term. From that formula it derives local Ward identities on disks, annuli, and pairs of pants, and uses them to obtain finite recursions for descendant matrix coefficients. The recursions recover known structural facts—polynomial factorization of pair-of-pants coefficients, the Shapovalov form as an annular zero-weight limit, and the formal chiral vertex-operator coefficients—and yield a geometric proof of smoothness of bulk correlations plus genus-zero BPZ equations for degenerate insertions. A sympathetic reader should care because the paper converts a numerical anomaly into contour data and then into a bootstrapping mechanism for the whole genus-zero theory.","feed_headline":"Schwarzian integrals explain the Virasoro central term","feed_subtitle":"A geometric derivation turns boundary deformations into finite recursions for Liouville descendant coefficients.","key_machinery":"The load-bearing mechanism is the pairing between infinitesimal boundary deformations and Virasoro modes through meromorphic vector fields. Given a surface with boundary parametrized by $\\zeta_k$, a vector field $v$ with a pole at one boundary component pulls back to a Laurent series $-\\sum_n [v_k]_n w^{n+1}\\partial_w$, so deforming along $v$ inserts $\\sum_n([v_k]_n L_n+\\overline{[v_k]_n}\\tilde L_n)$ into the amplitude; the holomorphic combination isolates $L_{-n}$. The anomaly term in the same deformation is expressed, by Theorem 1.1, as a contour integral of $v$ against the Schwarzian derivative of the inverse boundary coordinate. The local Ward identity balances these Virasoro insertions against the correction term that moves interior marked points, and the recursion closes because $L_k$ acting on a level-$|\\nu|$ descendant vanishes for $k>|\\nu|$. This yields finite recursions, with polynomial coefficients in conformal weights, that reduce descendant coefficients to primary coefficients.","core_discovery":"The central claim is Theorem 1.1: for a genus-zero admissible surface with $b\\ge 1$ incoming boundary components and filling $(\\hat\\Sigma,\\hat g)\\simeq(\\hat{\\mathbb C},\\hat g)$, the holomorphic variation of the Liouville anomaly under a meromorphic vector field $v=v(z)\\partial_z$ is $$\\frac12\\left(\\partial_t $S_L^{0}$(\\Sigma_t^v)-i\\partial_t $S_L^{0}$(\\$Sigma_t^{{iv}}$)\\right)\\Big|_{t=0}-\\frac1{24}\\sum_{k=1}^b [v_k]_0 = \\frac{i}{24\\pi}\\sum_{k=1}^b \\oint_{\\partial_k\\Sigma} v(z)\\mathcal{S}_{\\$zeta_k^{{-1}}$}(z)\\,dz,$$ where $\\mathcal{S}_{\\zeta_k^{-1}}$ is the Schwarzian derivative of the inverse boundary parametrization and the contours follow the incoming orientation. In plain terms, the holomorphic half of the anomaly is completely recorded by Schwarzian boundary integrals; no interior information is needed in genus zero. The paper then combines this formula with the variational formula for Liouville amplitudes, Weyl covariance, and diffeomorphism covariance to obtain a local Ward identity. Choosing vector fields with a pole at one boundary and holomorphy elsewhere turns the identity into recursions that lower the Virasoro level on one boundary at a time, while the Schwarzian term accounts exactly for the central charge. These recursions are the paper's main working tool, applied to disks, annuli, and pairs of pants.","pith_inferences":["Because the boundary-local Proposition 3.1 isolates the anomaly as a contour term plus an interior $\\bar\\partial V$ correction, the same computation should extend to higher genus; the choice of projective connection would then enter as an obstruction, making the central-charge term computable on arbitrary compact surfaces.","The finite recursions are driven entirely by local data such as pole orders and Schwarzian coefficients, so iterating them over a trinion decomposition may yield a deterministic bootstrap algorithm for higher-point and higher-genus correlation numbers, bypassing modular bootstrap constraints.","The zero-weight annular limit identifies a probabilistic amplitude with the Shapovalov form; testing this identification at low levels with explicit Gram matrices would provide a numerical check of both the recursions and the module structure.","The smoothness argument uses only a zero-weight insertion and the weighted Ward identity; if the same weighted estimates hold for boundary amplitudes, it should prove smoothness of bulk insertions in boundary Liouville CFT as well."],"forward_implications":["In the round-boundary case, every normalized descendant coefficient on a disk, annulus, or pair of pants is computed by a finite number of applications of the differential operators $D_{n,z,\\alpha}$, so no integration over moduli is needed.","The normalized pair-of-pants coefficient factors as $\\omega_{\\mathcal P_\\psi,\\nu}(\\Delta_\\alpha)\\omega_{\\mathcal P_\\psi,\\tilde\\nu}(\\Delta_\\alpha)$, with the holomorphic and anti-holomorphic sectors separated; this factorization follows from the recursion itself rather than from a separate bootstrap argument.","The zero-weight limit of the normalized annulus coefficient is the Shapovalov form of the Virasoro Verma module, connecting probabilistic amplitudes to the representation-theoretic inner product.","Bulk correlation functions of Liouville CFT are smooth in the insertion points on any closed Riemann surface, with a proof that reduces each derivative to finitely many descendant amplitudes.","For a degenerate bulk insertion of weight $\\alpha_{r,s}$, the null vector in the Liouville module becomes a differential equation of order at most $rs$ with principal part $\\partial_x^{rs}$ in flat coordinates; the level-two case is the standard BPZ equation."],"supporting_citations":[{"why":"Supplies the starting-point variational formula: analytic boundary-parametrization changes act differentiably on Liouville amplitudes, with derivative given by Virasoro operators plus the scalar anomaly that Theorem 1.1 computes.","marker":"[BGKR24]"},{"why":"Provides the probabilistic construction of Liouville amplitudes, their sewing laws, and the analytic continuation in boundary weights used when the recursions are extended off the real Seiberg region.","marker":"[GKRV21]"},{"why":"Establishes the analytic family of generalized primary and descendant states and the Virasoro algebra action on the Liouville Hilbert space, letting the Ward identities be evaluated on descendants.","marker":"[BGK+24]"},{"why":"Recorded the local Ward identity as Proposition 6.6; the present paper supplies a proof and turns it into a geometric recursion.","marker":"[GKR24]"},{"why":"Gives the irreducible Virasoro module structure, the Kac-type classification, and the singular vectors used to turn null-vector relations into BPZ differential equations.","marker":"[BW26]"},{"why":"Defines the formal chiral vertex-operator matrix coefficients and their commutation relations, the objects matched by the two-incoming-boundary pair-of-pants coefficients.","marker":"[Tes01]"},{"why":"Constructs the Poisson-scattering family of generalized primary states used as the base of the descendant recursion.","marker":"[GKRV20]"}],"fun_headline_variants":["Schwarzian boundary integrals compute Virasoro central term","Ward identities from Schwarzian geometry yield finite recursions","Geometric derivation of Ward identities via Schwarzian anomaly","Boundary deformations encode Virasoro anomaly in Schwarzian terms","From Schwarzian integrals to descendant recursions in genus zero"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The argument relies on the analytically continued generalized Liouville amplitudes: it assumes the extension to complex boundary weights and to descendant states exists with enough regularity for the recursions, a continuation taken from prior work rather than reproved here.","fun_headline_variants_meta":{"raw":{"variants":["Schwarzian boundary integrals compute Virasoro central term","Ward identities from Schwarzian geometry yield finite recursions","Geometric derivation of Ward identities via Schwarzian anomaly","Boundary deformations encode Virasoro anomaly in Schwarzian terms","From Schwarzian integrals to descendant recursions in genus zero"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000321,"raw_usage":{"total_tokens":1848,"prompt_tokens":1028,"completion_tokens":820,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":644,"completion_tokens_details":{"reasoning_tokens":738}},"tokens_in":644,"tokens_out":820,"duration_ms":9074,"temperature":1.0,"reasoning_tokens":738,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-16T00:35:49.551501+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Take a three-holed sphere with one boundary parametrized by $w+\\varepsilon w^2$ and the other two round; evaluate the left-hand side of Theorem 1.1 by differentiating the Liouville anomaly along the flow of $v$ numerically, and compare with the Schwarzian contour integral on the right. Any mismatch at order $\\varepsilon^2$ would falsify the theorem.","supporting_citations":[],"review_version":1}