{"id":"25ebc2c0-3ea7-4058-8225-01673b90cc10","arxiv_id":"2610.40211","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"The paper defines point-splitting PS-cycles, their homology, and a canonical frame-dependent homology class for compact 3-manifolds, with a partition function satisfying the quantum master equation.","lead":"This paper constructs a mathematical framework, Point-Splitting Perturbative Chern-Simons, in which the frame of a 3-manifold is built into the configuration spaces of Feynman-diagram vertices. A generalist might read it because the framework aims to make the framing anomaly and skein relations tractable in perturbative Chern-Simons theory.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek/deepseek-v4-flash","headline":"Prop. 33's gluing of flat-approaching local chains is the load-bearing step for Thm. 23; the asserted 'faster than 1/n' overlap error is unproved and normal-coordinate expansion gives only O(1/n) on non-flat M, so the exact-cycle conclusion is unsupported.","rationale":"The reader's weakest-assumption analysis and this stress-test identify the same step: the gluing of local flat approximants in Proposition 33, on which Theorem 23 depends. The reader correctly notes that no estimates are given and that the paper acknowledges the non-flat mismatch. The present pass sharpens the concern: a direct normal-coordinate expansion predicts an O(1/n) overlap error after rescaling, not the faster-than-1/n decay asserted in the proof, because the flat cycle is only translation-invariant and the quadratic part of the exponential-map transition is generically nonzero. This makes the gap concrete and testable. I do not see a more load-bearing issue elsewhere: the QME computation in Section 4 is a genuine calculation, and the Section 6 reduced-PSH computation, while sketchy, is not the foundation of the canonical framing class. The concern does not by itself force a verdict change from the reader's CONDITIONAL: the gap is potentially fillable by a quantitative perturbation argument or by imposing enough Euclidean invariance on the flat cycles, but as written the central existence theorem is unsupported. If the proposed overlap computation confirms the O(1/n) rate, the appropriate verdict would move toward REJECT or UNVERDICTED for the current version; until then UNCHANGED matches the evidence.","tokens_in":44508,"tokens_out":12484,"duration_ms":90477,"concrete_test":"Take M = S^3 with the round metric and a global framing, and use the flat annulus component Z^{♭,⊚}_{G0} from Definition 7 (or the first nontrivial flat cycle from Proposition 24). Choose two centers z,z' with |z-z'| = c/n and compare the local approximants Z^{prel,z,n} = exp_z(Dil_{1/n}(D^♭_{ρ,0} ∩ Z^{♭,⊚}_{G0})) and Z^{prel,z',n} on the overlap D_{ρ/n,z} ∩ D_{ρ/n,z'}. After cancelling the translation v = n exp_{z'}^{-1}(z), compute E_n = || Dil_n(exp_{z'}^{-1} Z^{prel,z,n}) - (Z^{♭,⊚}_{G0} + v) ||_{C^k(D_{ρ,0})}. If E_n decays only like 1/n along a sequence of centers, Proposition 33's asserted faster-than-1/n gluing is false as stated and the exact-cycle step requires a new argument; if E_n = o(1/n), the stated mechanism survives this check.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Proposition 33 is the load-bearing step for Theorem 23: the canonical class [Z_fr] exists only if local flat approximants can be glued into an exact flat-approaching PS-cycle on a non-flat M. The proof asserts, after the paper's own sentence b541, that on overlaps the chains exp_z(Dil_{1/n}(D^♭_{ρ,0} ∩ Z^♭_{G,m,P})) agree up to errors decaying faster than 1/n in every C^k norm, and then invokes a 'standard transversality argument'. No estimate, perturbation lemma, or control of the obstruction to exactness is supplied. The asserted rate is the delicate point. For two centers z,z' with |z-z'| ∼ 1/n, the rescaled coordinate transition is F(u) = n exp_{z'}^{-1}(exp_z(u/n)) = v + A u + (1/n)Q_{z,z'}(u) + O(1/n^2), where v = n exp_{z'}^{-1}(z) is cancelled by the translation invariance of the flat cycle, but Q, built from curvature and second derivatives of the exponential map, is generically nonzero on non-flat M. The flat cycles are required to be invariant only under translations (Definition 7 and §5.1.6), not under this quadratic deformation. Hence the overlap mismatch is O(1/n), not o(1/n), in the rescaled C^k norm. If that is correct, the exact chain cannot be obtained by the stated argument, and Proposition 33, and therefore Theorem 23, is not established by the text. This is an analytic gap rather than a disagreement with consensus: corrected exact chains might exist, but the paper gives no mechanism controlling the correction terms.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek/deepseek-v4-flash","summary":"The paper proposes a point-splitting variant of perturbative Chern-Simons theory on an oriented compact 3-manifold. Instead of using the Axelrod–Singer compactified configuration spaces, the author works with singular chains on M^{H(G)} (with a forgetful-compatible lift to multi-loop spaces), organized by decorated graphs with markings. A 'point-splitting cycle' is defined through transversality, cycle, and forgetful compatibility conditions, with an 'annulus datum' encoding unstable graphs. Isotopy classes of 'nice' such cycles define PS-homology PSH(M|U), depending on an Euler structure. The paper defines a Chern-Simons partition function by integrating a propagator-based form over such cycles, and proves a quantum master equation (Proposition 22). The central structural result associates to every frame fr of M a canonical PSH-class [Z_fr] (Theorem 23), constructed as a 'flat-approaching' cycle obtained by transplanting, via the exponential map of the frame, a flat R^3 cycle built recursively in Section 5.1. The final section introduces reduced PSH and proves, by an inductive obstruction argument, that it is a rank-one free module over Q[[h,a]] generated by the coherent class (Proposition 35).","tokens_in":45145,"tokens_out":2801,"duration_ms":21195,"significance":"If the main claims are correct, the paper would provide a mathematically defined point-splitting perturbative Chern-Simons partition function whose framing anomaly arises naturally from the frame-dependence of the point-splitting data, together with a new chain-level homology theory (PSH) claimed in the introduction to be connected to skein relations and open Gromov-Witten theory. The QME computation in Section 4 is a genuine calculation using a propagator and the cycle conditions, and the flat-cycle recursion in Section 5.1 is explicit and detailed. These are substantial strengths. However, the significance is conditional: the existence of the canonical class [Z_fr] rests on a non-flat gluing step in Proposition 33 that the manuscript does not currently prove, and the reduced-PSH theorem rests on a recursively defined obstruction whose vanishing and contraction are only sketched. The paper is a theory-building paper with many new definitions and a large inductive machinery; its value will depend on whether the crucial analytic and transversality steps can be supplied for non-flat manifolds.","major_comments":[{"comment":"No estimates or perturbation lemma are supplied for the asserted 'faster than 1/n' overlap decay in Proposition 33. The argument is also invoked in Proposition 34 for isotopies, so the uniqueness statement inherits the same gap.","section":"§5.2.7, Proposition 33, sentences b526–b541","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The relation between the quotient by Z_null and the equivalence relation (154) is defined on sums over equivalence classes, but it is not shown that the obstruction used in the inductive construction is well-defined on these sums.","section":"§6.1, Proposition 35","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The chain model on L_S(M) in Appendix 2.4 is a central technical tool. The manuscript lists axioms (boundaries, pullbacks, orientations, Stokes formula (74)) but does not prove that the objects defined in §2.4 satisfy them; the proof of Proposition 22 relies directly on these properties. For a foundational paper, this is a load-bearing point that merits a rigorous treatment or a clear statement of which properties are taken as axioms.","section":"§2.2.1–§2.2.2 and Lemma 2","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The QME computation is the strongest computational section of the paper, and the cancellations of the δ_e and δ_v terms are plausible. However, the transition from the chain-level identity (74) to the final operator identity (78) for Δ is compressed: the identification of the RHS of (77) with -h Δ P(Zbar) depends on signs and orientations of the special boundary faces N_{0,v} introduced in §2.2.5, and these signs are not fully tracked. This is likely fixable, but as written it is not a complete verification.","section":"§4.2, Proposition 22","authors_acknowledge":null,"authors_acknowledge_p":null}],"minor_comments":[{"comment":"There are numerous typographical and notation errors: 'Perturbative' appears as 'Per turba tive' in the header, 'Pont-Splitting' in §1, 'conditino' in §5.1.7, 'Chen-Simons' in §4, and 'q standard transversality argument' in Lemma 5. A careful proofreading pass is needed.","section":"Throughout","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The definition of chains on L_S(M) uses 'up to triangulations' and several identifications that are not fully precise, e.g., the sentence 'wnot constant on each fiber' (b172) is grammatically unclear. The section would benefit from stating the equivalence relation on generators explicitly.","section":"§2.4","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The claimed equivalence Eul(M) ≅ Eul(M)^▲ is stated with the map V ↦ (graph(V),0) and the assertion that it is compatible with the H^1(M,Z)-action, but the proof is only a sketch. Since PSH(M|U) is defined through this equivalence, a fuller argument would help.","section":"§3.3.1","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The flat annulus data is defined by setting Z^flat_{G1} = Z^flat_{G2} = 0. It would be clarifying to state explicitly that this is the analog of the 'no annulus corrections' flat model, and to record how the Euler structure associated to this flat annulus data is defined.","section":"§5.1, Definition 7","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The notation PSH for the reduced space is reused without a new symbol (the text writes 'P SH = P SH / Z_null'). This is confusing and should be changed, e.g., to PSH^red.","section":"§6","authors_acknowledge":null,"authors_acknowledge_p":null}],"recommendation":"major_revision","confidential_remarks":"The paper is ambitious and contains ideas that could be important, but the central existence theorem (Theorem 23) currently rests on an unproved analytic gluing estimate in Proposition 33. The discrepancy between the asserted o(1/n) overlap error and the O(1/n) error that a normal-coordinate expansion suggests is serious enough that the result cannot be accepted in this form. The same is true, to a lesser degree, of the reduced-PSH computation in Proposition 35. I would not reject the paper: the framework and the Section 4 computation are substantial, and the gap is of a kind that might be repairable if the author supplies the missing perturbation lemma. But the revision needs to provide real estimates for the non-flat gluing, not just another invocation of 'standard transversality'. I also note that the paper cites two of its own papers as [7] and [8] with the former listed as 'Point-Splitting Perturbative Chern-Simons II' without an arXiv identifier; that is acceptable if the companion is available, but the dependence of the present paper on [7] for the skein connection should be flagged."},"author_rebuttal":null,"desk_editor":{"model":"deepseek/deepseek-v4-flash","letter":"Punchline: this is not a repackaging. Replacing Axelrod-Singer's compactified configuration spaces by ordinary chains on M^{H(G)}, with half-edges split by a transversal perturbation, is a genuinely different construction, and the cycle condition doing the work of the usual anomaly cancellations is a real idea. But the one step the whole canonical class rests on, Proposition 33, is not proved, and the stress-test's objection lands. I worked through the overlap estimate myself: for two centers with |z-z'| ~ 1/n, the rescaled transition is v + A u + (1/n)Q_{z,z'}(u) + O(1/n^2), and Q is generically nonzero on a non-flat manifold. The flat cycles are only translation-invariant, not invariant under this quadratic term, so the overlap mismatch is O(1/n), not faster than 1/n. The paper's own sentence b541 concedes the mismatch exists, then asserts the faster-than-1/n decay and invokes a standard transversality argument. That is a load-bearing analytical gap, not a cosmetic one: Theorem 23, the coherent class, and the framing-anomaly interpretation all depend on it.\n\nWhat the paper does well: Section 4 contains an honest computation. The QME comes out of Stokes plus the cycle condition, and the signs, the delta_e and delta_v cancellations, and the hbar factor are all shown rather than hand-waved. The framework definitions of PS-cycles, PS-homology, and the transfer maps are explicit enough to be checked. The reduced-PSH rank-one statement is new, and the quotient is a sensible way to isolate the universal part, though the proof is lighter than the claim deserves: the inductive adjustment of q and the n-independence are largely asserted.\n\nThe other soft spots are secondary. The extension lemmas in Sections 2 and 5 repeatedly use a standard transversality argument without control of the corrections, which matters most where gluing is actually nontrivial, i.e., Proposition 33. The announced applications in [7] and [8] are unpublished, so the paper asks the reader to take its motivation on faith. None of these are reasons to desk-reject; the framework is original and the QME part is checkable.\n\nBottom line: I would send this to a serious referee. The definitions deserve real scrutiny, and a good referee could either repair Proposition 33 or isolate precisely where the construction breaks. But I would not cite or build on the canonical frame class until the gluing claim has an actual estimate behind it.","headline":"A genuinely new point-splitting reformulation of perturbative Chern-Simons with a real QME computation, but the one analytic gluing claim on which the canonical frame class depends is asserted rather than proved and appears to be false as stated.","tokens_in":46244,"tokens_out":1017,"would_cite":false,"duration_ms":28803,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"pith_extraction":{"msc":["57R56","81T45","57K31"],"pacs":[],"model":"deepseek/deepseek-v4-flash","headline":"Choosing a framing of a compact oriented 3-manifold suffices to canonically define a point-splitting perturbative Chern-Simons partition function, with the framing anomaly arising from the dependence of the configuration spaces on the…","keywords":["Chern-Simons perturbation theory","framing anomaly","point-splitting","configuration spaces","quantum master equation","singular chains","3-manifold invariants","skein relations"],"falsifier":"Compute the lowest-order nontrivial graph contribution to the point-splitting partition function on $S^3$ with a non-flat frame and compare it with the standard perturbative Chern-Simons value corrected by the usual framing counterterm: any discrepancy would show the class $[Z_{fr}]$ depends on more than the frame. Alternatively, exhibit on some compact curved three-manifold a decorated graph whose overlap mismatch of the local flat approximants fails to decay faster than $1/n$, which would remove the exact chain on which Proposition 33 rests.","tokens_in":45988,"feed_emoji":"🪢","tokens_out":13236,"duration_ms":115633,"temperature":0.7,"pith_summary":"The paper aims to give a mathematically precise definition of perturbative Chern-Simons theory in which the framing of the three-manifold enters from the start, by splitting the points that standard graph-integral constructions force to coincide. The central object is Point-Splitting-Homology (PSH), made of singular chains on configuration spaces indexed by decorated graphs and constrained by transversality, cycle, and forgetful-compatibility conditions. The main result is that every framing carries one canonical PSH class, so a well-defined point-splitting partition function exists for each framing and moves naturally under isotopies of frames. A consequence of the cycle condition is that integrating the Chern-Simons differential forms produces no boundary terms, and the partition function satisfies the quantum master equation. If the construction is sound, the framing anomaly of Chern-Simons theory becomes a property of the configuration spaces themselves rather than a set of correction terms added afterwards.","feed_headline":"Frames define the Chern-Simons partition function","feed_subtitle":"One frame picks out one point-splitting class per 3-manifold, putting the framing anomaly inside configuration space.","key_machinery":"The load-bearing object is Point-Splitting-Homology (PSH), built from PS-cycles: collections of singular chains on the space $M^{H(G)}$ of points labelled by the half-edges of a decorated graph $G$, one chain per graph and marking. A PS-cycle must satisfy three conditions: transversality to the diagonals where split points re-coincide; a cycle condition expressing the boundary of each chain in terms of the graph operations $\\delta_e$ (contracting an edge) and $\\delta_v$ (removing a vertex); and forgetful compatibility, which forces chains to factor through an annulus datum when external edges are removed. Homology is taken from isotopy classes of 'nice' PS-cycles, sequences of PS-cycles whose supports shrink to the diagonal. From this machinery, the coherent class $[Z_{fr}]$ is characterized by flat-approaching PS-cycles, which converge after exponential-map dilation to the unique translation-invariant flat PS-cycle in $\\mathbb{R}^3$; integrating Chern-Simons propagator forms against these chains defines the partition function, and the cycle condition cancels the boundary terms that the quantum master equation would otherwise see.","core_discovery":"On the paper's own terms, the discovery is a canonical assignment: to each frame $fr$ of an oriented compact three-manifold $M$ there is a Point-Splitting-Homology class $[Z_{fr}] \\in PSH(M|fr)$, with the property that an isotopy of frames sends $[Z_{fr_0}]$ to $[Z_{fr_1}]$ (Theorem 23). The class is characterized through flat-approaching PS-cycles: chains that, after rescaling by $n$ in normal coordinates, converge to the unique translation-invariant flat PS-cycle on $\\mathbb{R}^3$. The same formalism yields the quantum master equation $d_\\lambda P + \\hbar \\Delta P = 0$ for the partition function (Proposition 22), and the reduced homology is a rank-one free module over $\\mathbb{Q}[[\\hbar, a]]$ generated by the coherent class (Proposition 35). The framing anomaly is therefore not a defect of the theory but the visible effect of the configuration space of points depending on the frame.","pith_inferences":["If the gluing step in Proposition 33 is supplied with the missing estimates, the point-splitting partition function should coincide order by order with the standard perturbative Chern-Simons value after the usual framing correction; comparing the two on a low-degree graph would test the entire construction.","The rank-one result for reduced PSH suggests reading the coherent class as a vacuum state: the commutative product on PSH would then organize all pure-Chern-Simons observables as a formal deformation of $\\mathbb{Q}[[\\hbar, a]]$ around that class.","The flat-approaching convergence condition is a hidden locality axiom: it presumes curvature near the diagonal dies out under $n$-dilation fast enough to be irrelevant. On manifolds with large curvature one might expect a genuine obstruction, and finding one would mark exactly where the construction's domain of validity ends."],"forward_implications":["Every framing of a compact oriented three-manifold carries a canonically defined point-splitting perturbative Chern-Simons partition function, and isotopies of frames transport the corresponding homology classes without ambiguity.","The partition function satisfies the quantum master equation $d_\\lambda P + \\hbar \\Delta P = 0$, so it is well defined up to quantum-master-equation isotopy and receives no boundary terms from the chain integration.","Because the cycle condition absorbs the boundary contributions, the framing anomaly appears as the dependence of the point-splitting configuration spaces on the frame rather than as a separate counterterm.","Reduced PSH being a rank-one free module over $\\mathbb{Q}[[\\hbar, a]]$ means the coherent class generates the entire reduced homology, with $\\hbar$ and $a$ parameterizing the loop order and the anomaly degree.","The announced isomorphism between a version of PSH and a version of the Skein gives a new ambient structure in which skein relations can be studied."],"supporting_citations":[{"why":"Defines the standard perturbative Chern-Simons construction on compactified configuration spaces that this paper modifies, supplying the graph-integral framework.","marker":"[1]"},{"why":"Completes the standard formulation whose diagonals and framing anomaly are replaced here by point-splitting chains on the half-edge configuration space.","marker":"[2]"},{"why":"Provides the Chern-Simons differential forms and master-equation technology that the propagator and the quantum master equation of this paper generalize.","marker":"[3]"},{"why":"Introduces the Multi-Curve-Homology that, in degree zero, reduces to the Point-Splitting-Homology defined here.","marker":"[6]"}],"fun_headline_variants":["Frames pick the point-splitting class","Framing anomaly lives in configuration space","Each frame defines its own Chern-Simons class","Flat cycles pin down the framed partition function","Frame isotopy moves the Chern-Simons class"],"cache_read_input_tokens":0,"weakest_assumption_plain":"Proposition 33 assumes that local flat approximants — chains built near each point of the manifold by dilating the flat cycle on $\\mathbb{R}^3$ — can be glued into one exact chain on any compact curved three-manifold, even though the approximants do not match on overlaps; the paper acknowledges the mismatch and asserts it decays faster than $1/n$ in every $C^k$ norm, then invokes a standard transversality argument to conclude an exact chain exists, without supplying estimates or a perturbation lemma, and Theorem 23 depends on this step.","fun_headline_variants_meta":{"raw":{"variants":["Frames pick the point-splitting class","Framing anomaly lives in configuration space","Each frame defines its own Chern-Simons class","Flat cycles pin down the framed partition function","Frame isotopy moves the Chern-Simons class"]},"model":"deepseek/deepseek-v4-flash","effort":"low","cost_usd":4.3e-05,"raw_usage":{"cost":4.28064e-05,"is_byok":false,"cost_details":{"upstream_inference_cost":4.28064e-05,"upstream_inference_prompt_cost":2.55164e-05,"upstream_inference_completions_cost":1.729e-05},"total_tokens":949,"prompt_tokens":854,"completion_tokens":95,"prompt_tokens_details":{"audio_tokens":0,"video_tokens":0,"cached_tokens":717,"cache_write_tokens":0},"completion_tokens_details":{"audio_tokens":0,"image_tokens":0,"reasoning_tokens":37}},"tokens_in":854,"tokens_out":95,"cost_source":"openrouter_usage","duration_ms":2772,"image_pages":[],"temperature":1.0,"billed_attempts":1,"reasoning_tokens":37,"transport_provider":"openrouter","cache_read_input_tokens":0,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-10-30T12:14:21.744974+00:00","model_set":{"reader":"deepseek/deepseek-v4-flash"},"falsifier":"Compute the lowest-order nontrivial graph contribution to the point-splitting partition function on $S^3$ with a non-flat frame and compare it with the standard perturbative Chern-Simons value corrected by the usual framing counterterm: any discrepancy would show the class $[Z_{fr}]$ depends on more than the frame. Alternatively, exhibit on some compact curved three-manifold a decorated graph whose overlap mismatch of the local flat approximants fails to decay faster than $1/n$, which would remove the exact chain on which Proposition 33 rests.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Defines the standard perturbative Chern-Simons construction on compactified configuration spaces that this paper modifies, supplying the graph-integral framework."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Completes the standard formulation whose diagonals and framing anomaly are replaced here by point-splitting chains on the half-edge configuration space."},{"cited_title":"Bullimore, T","cited_arxiv_id":null,"evidence_quote":"Provides the Chern-Simons differential forms and master-equation technology that the propagator and the quantum master equation of this paper generalize."},{"cited_title":"Tangles, Generalized Reidemeister Moves, and Three-Dimensional Mirror Symmetry","cited_arxiv_id":"1211.3730","evidence_quote":"Introduces the Multi-Curve-Homology that, in degree zero, reduces to the Point-Splitting-Homology defined here."}],"review_version":1}