{"id":"f6453b25-4da4-4917-b782-972f48f283f6","arxiv_id":"2507.20151","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"The paper establishes ancestor and descendant Virasoro constraints for topological recursion on meromorphic spectral curves, with explicit operators and non-perturbative claims.","lead":"This paper proves that many generating functions built from topological recursion obey Virasoro constraints, which are explicit differential equations encoding hidden symmetries. Specialists care because Virasoro constraints are a powerful organizing principle in enumerative geometry, integrable systems, and matrix models.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Proposition 3.1's bridge from ancestor residue (2) to descendant residue (27) rests entirely on the imported CohFT/TR identification (10); if that identification fails, Theorem 1's proof collapses.","rationale":"The reader's weakest_assumption identifies exactly the same load-bearing dependency: equation (10) imported from [11,16,27] is used in Proposition 3.1 to pass from the ancestor residue identity (2) to the descendant identity (27). My read of the proof confirms this is the single most consequential step: without (10), the replacement of \\bar z by z in the critical-point residues has no justification, and Theorems 1 and 3.2 do not follow. The paper does not re-derive (10) or state precisely which hypotheses on the spectral curve are needed for it, so the descendant theorem inherits an unverified structural assumption. I agree that this warrants a conditional verdict rather than full acceptance. I do not see a reason to move to REJECT: the cited identification is a well-established theorem by Dunin-Barkowski–Orantin–Shadrin–Spitz and related works, so the concern is a dependency to be checked, not an observed contradiction. The omitted proof of Theorem 4.2 is a further gap, but the reader already noted it and I treat it as secondary because the parallel argument is plausibly mechanical once the perturbative case is secured. Therefore the appropriate verdict remains CONDITIONAL, and my stress-test does not change the reader's verdict.","tokens_in":26099,"tokens_out":10095,"duration_ms":104939,"concrete_test":"Independently re-derive Proposition 3.1 for the smallest non-trivial cases (g,n)=(1,1) and (0,3) directly from the recursion (1), without invoking the CohFT/TR identification (10). For these cases, ω_{1,1}, ω_{0,3}, ω_{0,2} and the Bergman kernel can be computed explicitly from the recursion, and the local involution \\bar z is determined by x. One then checks by direct residue computation whether the right-hand side of (2) equals the right-hand side of (27). If the equality fails for either case, the proof's bridge is broken and Theorem 1 is unsupported as written. If it holds, run the same check for (g,n)=(1,2) and for a non-trivial curve such as the deformed r-Bessel curve with r=3, ε≠0, to test whether the mechanism generalizes without importing (10).","verdict_should_be":"UNCHANGED","load_bearing_attack":"The crucial step in the paper is Proposition 3.1, which converts the ancestor Virasoro identity (2) into the descendant identity (27) by moving residues from the critical points of x to the boundary points. For the case 2g-2+n>0, the proof of Proposition 3.1 explicitly uses equation (10) to replace the product ω_{g-1,n+2}(z, \\bar z, z[n]) + Σ' ω_{g_1,|I|+1}(z,z_I) ω_{g_2,|J|+1}(\\bar z,z_J) by its 'diagonal' version with both arguments equal to z, up to a sign. This replacement is not a formal consequence of the topological recursion alone: the residues at z=z_γ of x^{m+1}/dx times these products depend on the full polar structure of ω_{g,n} at the critical points, and equation (10) is exactly the statement that supplies that structure. The local expansion (17) and the involution \\bar z(η)=-η are not sufficient without the global ancestor expansion, because the cross terms in the products have a delicate diagonal behavior that only (10) pins down. The paper cites [11,16,27] for (10), but Theorem 1 is stated for arbitrary meromorphic x,y with the pole condition, and it is not verified in the paper that the hypotheses of those citations cover exactly this class of spectral curves. The same identification is reused in the proof of Theorem 3.2 when expression (28) is identified with equation (27). Thus the proof of the paper's central descendant theorem inherits a load-bearing, un-re-derived dependency. The secondary gap in Theorem 4.2 (proof omitted as 'parallel') would also need to be closed, but the dependency on (10) is the more substantive mathematical risk.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper derives Virasoro constraints for topological recursion. It gives a direct residue-based proof of the ancestor Virasoro constraints for the ancestor generating series A(s;ℏ) of an arbitrary spectral curve, using the definition of the topological recursion and a CohFT/TR dictionary. It then establishes descendant Virasoro constraints for the TR descendant generating series Z(p;ℏ) under the condition that x,y are meromorphic and x^{m+1}y has poles only at the boundary points, with an explicit operator L_m. The same operator is claimed to annihilate the non-perturbative descendant generating series Z^{NP}_{μ,ν}(p;ℏ;w). Several examples are presented: the Airy curve, the deformed r-Bessel curve, the extended Grothendieck dessins curve, and the Weierstrass curve, comparing TR Virasoro constraints with geometric descendant Virasoro constraints.","tokens_in":26442,"tokens_out":5753,"duration_ms":62404,"significance":"If the main theorems hold, this paper gives a substantive unification: topological recursion partition functions in a broad class are Virasoro vacua, with explicit operators, and the same constraints control perturbative and non-perturbative levels. The ancestor part is largely self-contained and includes the delicate (g,n)=(1,0) residue computation. The explicit comparison examples, especially the Weierstrass-curve example with its flat-basis computations and the observation that non-perturbative contributions do not alter the TR Virasoro operator, are valuable. However, the descendant theorem and the non-perturbative theorem depend on the authors' companion papers and on an unproved conjecture from [21], so the significance is conditional until those dependencies are either verified or supplied.","major_comments":[{"comment":"The proof of Proposition 3.1 for 2g-2+n>0 is load-bearing and rests entirely on the imported identification (10) from [11,16,27], applied to replace products of ω_{g,n} at (z, z̄) by their diagonal versions at (z,z). The manuscript does not verify that the hypotheses under which (10) is proved in those references cover exactly the class of spectral curves admitted in Theorem 1, where x is allowed to be a possibly multi-valued function of z on Σ\\{boundaries} with meromorphic differential and y is only required to be holomorphic near the critical points. This is not a formal consequence of the topological recursion itself: the residues at z=z_γ of x(z)^{m+1}/dx(z) times the products in (2) depend on the full polar structure of ω_{g,n} at the critical points, and equation (10) is precisely the statement supplying that structure. The same dependency is inherited by Theorem 3.2 when expression (28) is identified with equation (27). The authors should either prove (10) under the stated hypotheses or cite a theorem whose hypotheses are visibly satisfied and state the verification.","section":"§3, Proposition 3.1, Eq. (27)"},{"comment":"The proof of the non-perturbative Virasoro constraints is omitted with the sentence 'the proof of this equation is parallel to the perturbative version... We omit the details here.' Theorem 4.2 is one of the two central advertised results (Theorem 2 in the introduction), so an omission of this size is a material gap. The reduction from the non-perturbative series Z^{NP}_{μ,ν,τ'}(p;ℏ;w) to the perturbative proof requires checking how the B-cycle periods, the theta function, the operators e^{(1/2)ω^{(2)}_{0,0}∇^2_w}, and the shift by φ·∇_w interact with the Virasoro operator L_m. A sentence saying the proof is parallel does not establish this. The authors should include the full argument or a precise statement showing that Theorem 4.2 is formally implied by Theorem 3.2 and the formula for Z^{Λ,NP}_{μ,ν,τ'}.","section":"§4.2, Theorem 4.2"},{"comment":"The claim that D(t(p);ℏ)=Z(p;ℏ) for the Weierstrass curve, and hence that the geometric and TR Virasoro operators coincide, explicitly relies on [21, Conjecture 0.4], with only the remark that it 'can be proved by following a similar argument as the one used for the spectral curve of the Hurwitz space M_{1,1}.' Since this is an unproved conjecture and not a theorem, the example should either supply the proof, state the result as conditional on [21, Conjecture 0.4], or replace the invocation with a proved statement. As written, the final comparison in the Weierstrass example is not fully established.","section":"§5.4, Weierstrass curve"},{"comment":"The hypotheses on x are inconsistent as stated: Definition 0.1 allows x to be 'possibly multi-valued' on Σ\\{boundaries}, while Theorem 1 and Proposition 3.1 assume 'x,y are meromorphic functions on Σ.' For a Riemann surface of genus g>0, a function with meromorphic differential need not be a single-valued meromorphic function, and the objects x^{m+1}y, the residue condition at boundary points, and the choice of local coordinate λ_i with x=λ_i^{r_i} require a single-valued x. The authors should either clarify that 'meromorphic' means single-valued meromorphic and explain why the multi-valued language in Definition 0.1 is not needed, or state and prove the theorems under a precise hypothesis on the multivaluedness of x.","section":"Definition 0.1 and Theorem 1"}],"minor_comments":[{"comment":"The symbol m is used both for the number of boundary points in the spectral curve data and for the Virasoro index m≥−1; this is occasionally confusing in formulas such as Theorem 1, though the meaning is usually clear from context.","section":"Throughout"},{"comment":"The statement that equation (2) is equivalent to the ancestor Virasoro constraint might be made more reader-friendly by stating explicitly that the equivalence is at the level of the genus expansion of L_m A(s;ℏ), as is done in the proof of Proposition 2.3.","section":"§0.2, Proposition 0.5"},{"comment":"In the definition of L_m, the term δ_{m,-1} includes a sum over a=0,...,r_i of \tilde p_a \tilde p_{r_i-a}, but the variables p_i^0 have been set to zero; the formula would be clearer if the convention for p_i^0 and for \tilde p_i^0 were stated in the theorem itself rather than only in the proof.","section":"§3.2, Theorem 3.2"},{"comment":"The sentence comparing Z_{r,ϵ}(p;ℏ) with D_{r,ϵ}(t(p);ℏ) contains a footnote distinguishing the definition of Z_{r,ϵ} here from that in [21]; it would help to state in the main text whether the resulting Virasoro operators are unchanged by this difference.","section":"§5.2"},{"comment":"In the Weierstrass example, the formulas for the flat basis vectors φ_i and the matrix of E would benefit from a short derivation or a reference to the equation numbers in [21, §7.1], since several identities involving Eisenstein series are used without comment.","section":"§5.4"}],"recommendation":"major_revision","confidential_remarks":"The paper relies heavily on the authors' own companion works [20] and [21], including an unproved Conjecture 0.4 in [21]. I would ask the editor to require that the authors clearly delineate which statements in [21] are theorems and which are conjectural, and to supply the omitted proof of Theorem 4.2. The central idea is appealing and the explicit residue computations are valuable, but the descendant and non-perturbative theorems are currently not verifiable without those external inputs."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Dear colleague,\n\nThe thing worth knowing: this paper gives the first general descendant Virasoro constraints for topological recursion, not just curve-by-curve examples. The route from the ancestor residue identity to the descendant identity by moving poles to the boundary is a real technical step, and the proof of Theorem 3.2 is mostly explicit residue bookkeeping, including the delicate (g,n)=(1,0) case. If the theorem holds, TR partition functions in this class are Virasoro vacua, with the same operator governing perturbative and non-perturbative levels. That is a meaningful advance.\n\nWhat I like: the ancestor constraints are reproved from scratch in Lemma 2.1 without leaning on the full CohFT machine, and the examples (Airy, deformed r-Bessel, dessins) are worked out carefully enough to compare TR operators with the geometric ones. The paper is readable for anyone already inside this series.\n\nSoft spots, in order of concern. First, Theorem 4.2 is stated as a theorem and then its proof is dismissed as \"parallel\" with details omitted. The abstract sells it as a result. That gap should be closed, or the statement downgraded to a conjecture. Second, the Weierstrass example invokes [21, Conjecture 0.4] as if it were valid, and the claimed equality D(t(p))=Z(p) depends on it. The example is therefore conditional. Third, the bridge in Proposition 3.1 rests on equation (10), the CohFT/TR identification imported from [11,16,27]. The stress-test note is right that the paper does not verify that those citations cover exactly the class of meromorphic x,y satisfying the pole condition. I do not see this as fatal circularity—(10) is an established identification for broad spectral curve classes—but the authors should state the precise theorem they are importing and check its hypotheses, because the descendant theorem inherits whatever extra assumptions sit inside that identification.\n\nOverall, the central perturbative claim looks plausible and is substantially proven. The gaps are in secondary claims, not in the core residue computation. I would send this to a serious referee, with instructions to ask for a complete proof, or at least an explicit conjecture, for Theorem 4.2, and for a resolution of the Weierstrass conjecture dependency.","headline":"A genuinely new general descendant Virasoro theorem for topological recursion with a mostly rigorous residue proof; the non-perturbative extension and one example lean on unproved imports, so referee should demand details but not desk-reject.","tokens_in":26997,"tokens_out":2661,"would_cite":true,"duration_ms":28870,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["14N35","53D45"],"pacs":[],"model":"deepseek-v4-flash","headline":"Topological recursion partitions are Virasoro vacua under a boundary-pole condition.","keywords":["Virasoro constraints","topological recursion","spectral curve","cohomological field theory","descendant invariants","nonperturbative generating series","Bergman kernel","moduli spaces of curves"],"falsifier":"For the elliptic curve of Section 5.4, compute the genus-one part of $L_0Z(p;\\hbar)$ directly from the recursion's $\\omega_{1,1}$: the constant $\\delta_{m,0}/16$ term must cancel against contributions from the $\\lambda^1$ and $\\lambda^5$ coefficients of $y\\,dx$. Equivalently, verify the $(g,n)=(1,0)$, $m=0$ instance of identity (27) by evaluating both residues on that curve; a mismatch would falsify Theorem 1 and the transfer from critical-point to boundary residues.","tokens_in":25849,"feed_emoji":"","tokens_out":13659,"duration_ms":120494,"temperature":0.7,"pith_summary":"This paper proves that the generating series built from topological recursion data are Virasoro vacua. For a spectral curve $(\\Sigma,x,y)$ with meromorphic $x,y$, whenever $x^{m+1}y$ has poles only at the boundary points, an explicit second-order differential operator $L_m$ annihilates the descendant generating series $Z(p;\\hbar)$ for every $m\\ge -1$; the same operator annihilates the nonperturbative series $Z^{\\mathrm{NP}}_{\\mu,\\nu}(p;\\hbar;w)$ on higher-genus curves. On the ancestor side, the paper derives the Virasoro constraints directly from the recursion identity, without assuming the pole condition. The result matters because it shows that the many invariants produced by the recursion assemble into a single object with the Virasoro algebra familiar from intersection theory on moduli spaces of curves, and the symmetry survives the nonperturbative completion.","feed_headline":"Recursion series obey Virasoro constraints under a pole condition","feed_subtitle":"The same operators kill the descendant and nonperturbative series for boundary-only pole curves.","key_machinery":"The load-bearing objects are the residue identities (2) and (27), together with the ancestor expansion (10). Identity (2) equates the residue of $x^{m+1}y\\,\\omega_{g,n+1}$ at the critical points with half the residue of the recursion input $R_{g,n}$ against $x^{m+1}/dx$, and is proved directly from the recursion. Identity (27) is the descendant analogue: under the pole condition the same residue can be evaluated at the boundary points, with the diagonal Bergman-kernel subtraction $\\tilde\\omega_{0,2}=\\omega_{0,2}-dx_1dx_2/(x_1-x_2)^2$ handling the base case $(g,n)=(0,2)$. Equation (10), imported from earlier work, identifies the recursion differentials with ancestor correlators of the semisimple cohomological field theory attached to the spectral curve, and this identification is what turns the residue identities into operator equations on the generating series.","core_discovery":"The central claim is Theorem 1: for a spectral curve $(\\Sigma,x,y)$ with meromorphic $x,y$ and with $x^{m+1}y$ having poles only at boundary points, the TR descendant generating series $Z(p;\\hbar)$ satisfies $L_mZ(p;\\hbar)=0$ for $m\\ge -1$, where $L_m$ is the explicit differential operator displayed in Section 3.2. Theorem 2 extends the same statement to the nonperturbative descendant series $Z^{\\mathrm{NP}}_{\\mu,\\nu}(p;\\hbar;w)$ with the same operator. The proof runs through residue identities: the ancestor identity (2) holds for every spectral curve, and the descendant identity (27) follows when the pole condition lets one move residues from critical points to boundary points; expanding the resulting boundary residues in coordinates $x=\\lambda_i^{r_i}$ converts the identity into the operator statement. The paper also derives the ancestor Virasoro constraints directly from the recursion, without invoking the classification of semisimple CohFTs as an input.","pith_inferences":["Because the pole condition is used only to move residues from critical points to boundary points, one could classify the spectral curves satisfying it; the paper does not offer such a classification, and it may cut out a finite-dimensional family.","The fact that the nonperturbative series satisfies the identical operators suggests that the theta-function completion acts as a Virasoro intertwiner rather than a deformation; the paper proves the constraint but does not phrase it this way.","The examples with homogeneous underlying CohFTs suggest a broader coincidence between TR descendant constraints and geometric descendant Virasoro constraints; proving that coincidence for all homogeneous curves would be a natural next step beyond the paper's examples."],"forward_implications":["Every spectral curve in the allowed class has its TR descendant potential annihilated by the full family of operators $L_m$, $m\\ge -1$.","The same operators annihilate the nonperturbative series, so including B-cycle periods and theta functions does not alter the Virasoro structure.","The ancestor constraints hold without the boundary-pole condition, because equation (2) is proved directly from the recursion.","In the basic genus-zero example, the deformed $r$-Bessel curve, the extended dessins d'enfants example, and the elliptic curve of Section 5.4, the TR descendant operators coincide with the geometric descendant Virasoro operators after explicit coordinate changes.","The Virasoro equations give recursions among the TR descendant invariants, usable as checks or alternative computations of the generating series."],"supporting_citations":[{"why":"Defines the topological recursion recursion kernel and the multi-differentials $\\omega_{g,n}$ on which the paper's identities act.","marker":"[13]"},{"why":"Establishes the identification of recursion differentials with ancestor correlators of a semisimple CohFT, which is equation (10) and the bridge to Virasoro operators.","marker":"[11]"},{"why":"Provides the Airy-coordinate and R-matrix construction of the CohFT from the spectral curve, supporting equation (10).","marker":"[16]"},{"why":"Gives ancestor Virasoro constraints for local recursion in singularity theory, which Proposition 0.5 globalizes to arbitrary spectral curves.","marker":"[27]"},{"why":"First paper in the series, defining descendant generating series and Virasoro conjectures for calibrated CohFTs with vacuum, whose operator formalism is used here.","marker":"[20]"},{"why":"Defines TR descendant invariants, the nonperturbative generating series, and their relation to geometric descendants and KP tau-functions.","marker":"[21]"}],"fun_headline_variants":["Boundary-only poles yield Virasoro constraints for recursion","Recursion series pass Virasoro constraints under pole condition","Virasoro constraints proven for descendant and nonperturbative recursion","Topological recursion satisfies Virasoro constraints with boundary poles"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The proof uses, without re-deriving, the imported identification that the differentials produced by the recursion coincide with the ancestor expansion of the associated semisimple cohomological field theory; if that identification fails for some curve satisfying the pole condition, the bridge from residue identity to descendant Virasoro constraint breaks.","fun_headline_variants_meta":{"raw":{"variants":["Boundary-only poles yield Virasoro constraints for recursion","Recursion series pass Virasoro constraints under pole condition","Virasoro constraints proven for descendant and nonperturbative recursion","Topological recursion satisfies Virasoro constraints with boundary poles"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000236,"raw_usage":{"total_tokens":1454,"prompt_tokens":848,"completion_tokens":606,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":464,"completion_tokens_details":{"reasoning_tokens":537}},"tokens_in":464,"tokens_out":606,"duration_ms":5390,"temperature":1.0,"reasoning_tokens":537,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-15T17:49:11.109010+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"For the elliptic curve of Section 5.4, compute the genus-one part of $L_0Z(p;\\hbar)$ directly from the recursion's $\\omega_{1,1}$: the constant $\\delta_{m,0}/16$ term must cancel against contributions from the $\\lambda^1$ and $\\lambda^5$ coefficients of $y\\,dx$. Equivalently, verify the $(g,n)=(1,0)$, $m=0$ instance of identity (27) by evaluating both residues on that curve; a mismatch would falsify Theorem 1 and the transfer from critical-point to boundary residues.","supporting_citations":[{"cited_title":"Eynard, N","cited_arxiv_id":null,"evidence_quote":"Defines the topological recursion recursion kernel and the multi-differentials $\\omega_{g,n}$ on which the paper's identities act."},{"cited_title":"Dunin-Barkowski, N","cited_arxiv_id":null,"evidence_quote":"Establishes the identification of recursion differentials with ancestor correlators of a semisimple CohFT, which is equation (10) and the bridge to Virasoro operators."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the Airy-coordinate and R-matrix construction of the CohFT from the spectral curve, supporting equation (10)."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Gives ancestor Virasoro constraints for local recursion in singularity theory, which Proposition 0.5 globalizes to arbitrary spectral curves."},{"cited_title":"Cohomological Field Theory with vacuum and its Virasoro constraints","cited_arxiv_id":"2502.18895","evidence_quote":"First paper in the series, defining descendant generating series and Virasoro conjectures for calibrated CohFTs with vacuum, whose operator formalism is used here."}],"review_version":1}