{"id":"52a07bc6-9ecd-4c37-8139-c438f2e70d6b","arxiv_id":"2607.19037","paper_version":1,"verdict":"REJECT","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"high","formal_verification":"none","parameter_count":0,"one_line_summary":"The zero-area limit of the SU(N) Yang-Mills measure on compact surfaces is realized at the level of random distributional connections and identified with the Atiyah-Bott-Goldman measure.","lead":"The authors build new function spaces for random connections and show that, as the area of a surface goes to zero, the two-dimensional Yang-Mills random connection converges to a flat random object. The limit is identified with the Atiyah-Bott-Goldman symplectic measure, a connection-level version of earlier holonomy-level results.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Central theorem depends on an unpublished construction: Appendix A, the only in-manuscript support, leaves the holonomy used for conditioning undefined for curves crossing unstable curves.","rationale":"I read the paper as attempting to prove, at the level of random distributional connections, that the small-area Yang–Mills measure converges to the Atiyah–Bott–Goldman symplectic volume in the Morse gauge. For that claim to hold, three ingredients are needed: (i) a Borel probability measure A_{Σ,σ} on the anisotropic spaces B^{α,p,s,ℓ}_{YM}, constructed by conditioning the free-boundary measure; (ii) tightness of the family t↦A_{Σ,tσ}; (iii) identification of every subsequential limit with the ABG measure via the extraction maps. The reader's verdict identifies the first ingredient as the weak point, and I agree. The free-boundary measure and its holonomy functional are the sole input to the conditioning procedure, and the manuscript's own Appendix A is explicitly a sketch. Moreover, the appendix contains an admission that the second-order rough path term is not defined exactly in the situation needed for the conditioning level curve. Formula (A.11) is a concatenation recipe, not a proof. Since the unpublished companion [6] is 'in preparation', the preprint as written does not demonstrate existence of the object whose limit is the theorem. I also noted two secondary internal issues: the parameter ranges in Theorem 2.4 are inconsistent with those in Theorem 6.3 (ℓ<−1 vs ℓ<0, and s<0 in both but with swapped roles relative to Theorem 2.1), and the tightness argument in Theorem 6.1 is missing the strictly stronger regularity space required for a compact injection to yield tightness. These do not change the verdict but reinforce that the manuscript is not self-contained. A single concrete analytical check would settle the primary concern: complete the Appendix A construction for curves that intersect unstable curves, including Chen's identity and moment bounds. Until that is done, the central claim is not supported by the manuscript alone.","tokens_in":54518,"tokens_out":16635,"duration_ms":168052,"concrete_test":"Complete the missing step in Appendix A for type-II curves: give a rigorous definition of the second-order process W_{s,t} for level curves whose endpoints lie on unstable curves, prove the Chen relation for the concatenation (A.11), and verify the Kolmogorov-type bound E sup_{s≠t} |W_{s,t}|^q / |t−s|^{2αq} < ∞ for some α>1/3 and q large enough. If this cannot be done without invoking unpublished results from [6], then the conditioning in Definition 5.3 is unsupported and the central convergence theorem fails as a self-contained statement.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The claim that matters is Theorem 2.4/6.3: (A_{Σ,tσ}) converges in B^{α,p,s,ℓ}_{YM} to the Morse-gauge representative of the ABG measure. The object A_{Σ,σ} is defined in §5.2–5.3 by conditioning the free-boundary measure A^Free_{S,σ} on Hol_{maxf−ε}(A^Free)=1. This conditioning requires a measurable, integrable holonomy functional along a level curve that intersects the unstable curves. Appendix A is the only in-manuscript justification and is explicitly labelled a sketch. In §A.3, footnote 13 concedes that when a curve intersects an unstable curve, it is no longer clear how to make sense of the Wick square defining the second-order rough path. Formula (A.11) then concatenates elementary pieces with jumps, but no proof of Chen's identity across the jumps, nor of the required moment bounds, is supplied. The construction is delegated to the companion paper [6], which is 'in preparation'. If this holonomy is not measurable on B^{α,p,s,ℓ}_{YM}, Definition 5.3 is not well defined, so Theorem 2.1 has no object and Theorem 2.4 has no limit to identify. This is the load-bearing condition for the paper's central claim, and the manuscript does not establish it. Secondary but reinforcing gaps: the parameter ranges of Theorem 2.4 (ℓ<−1, s<0) are inconsistent with Theorem 6.3 (ℓ<0), and the tightness proof of Theorem 6.1 invokes compact injections while only proving boundedness in the same space, which does not imply tightness without a strictly better regularity index.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper develops a Morse-theoretic gauge fixing to construct Yang--Mills random connections as probability measures on anisotropic Besov-type spaces B^{α,p,s,ℓ}_{YM} of distributional 1-forms on compact surfaces. The geometric part (Section 3) represents flat connections by singular currents supported on unstable curves of a Morse gradient flow and gives a formula for the Atiyah--Bott--Goldman symplectic form in these 'Morse gauge' coordinates. The probabilistic part (Sections 5--6) defines the Yang--Mills measure by conditioning a free-boundary measure on trivial boundary holonomy, and claims that in the zero-area limit the family (A_{Σ,tσ}) converges in B^{α,p,s,ℓ}_{YM} to a measure supported on the finite-dimensional space F, identified with the normalized ABG symplectic volume. This is the content of Theorems 2.4 and 6.3. The proof of the existence and measurability of the free-boundary measure and its holonomy functional is delegated to Appendix A and to the unpublished companion paper [6], with Appendix A explicitly labeled a sketch.","tokens_in":54876,"tokens_out":9030,"duration_ms":81613,"significance":"If the main theorem is correct, it would be a substantial advance: it would give the first distribution-level semiclassical limit of two-dimensional Yang--Mills fields, going beyond the holonomy-level results of Witten, Forman, Liu, and Sengupta, and would answer a question of Lévy. The paper also contains genuinely interesting geometric material: the Morse-gauge description of the character variety in Section 3 is concrete and elegant, and the anisotropic spaces of Section 4 are designed in a principled way, with the decomposition and compactness arguments worked out in unusual detail. However, the central probabilistic claim rests on an unfinished construction: the companion paper [6] is 'in preparation', Appendix A is a sketch, and at the exact point where the conditioning holonomy crosses unstable curves the manuscript concedes that the required second-order rough-path term is not defined. In addition, the tightness proof in Section 6.1 contains a logical gap. For these reasons the paper cannot be accepted in its present form.","major_comments":[{"comment":"Definition 5.3 defines the Yang--Mills pre-measure by conditioning on Hol_{maxf−ε}(A^Free_{S,σ})=1, so the holonomy functional must be a well-defined measurable function on the anisotropic space. The only in-manuscript support for this is Appendix A, which is explicitly a sketch. The gap is not cosmetic: in §A.3, footnote 13 states that 'when γ intersects unstable curves, it is no longer clear how to make sense of the Wick square', and equation (A.11) simply concatenates elementary parallel transports with group-element jumps, without proving Chen's identity across the jumps or the moment bounds needed to control the concatenated rough path. Since the level curve {f=maxf−ε} intersects ∪_a W^u(a), this is exactly the case needed for the conditioning. Thus Definition 5.3 is not fully justified, and Theorems 2.1 and 2.4 do not currently have a well-defined object.","section":"Appendix A, Definition 5.3"},{"comment":"The proof of tightness states: 'Using the compact injections of the spaces B^{α,p,s,ℓ}_{YM}, it is enough to show sup_t E[∥A_{Σ,tσ}∥^p_{B^{α,p,s,ℓ}_{YM}}] < ∞.' This inference is invalid. Boundedness in the same Banach space B^{α,p,s,ℓ}_{YM} does not imply tightness there, even if compact injections into that space exist; one would need a uniform bound in a strictly finer space that compactly injects into B^{α,p,s,ℓ}_{YM}. Propositions 4.11 and 4.12 provide such finer spaces, but no uniform bounds in those finer spaces are proved. Consequently Theorem 6.1, and with it the convergence statements in Theorems 6.3 and 2.4, are not established by the arguments given.","section":"Section 6.1, proof of Theorem 6.1"},{"comment":"The parameter ranges of the main statements are mutually inconsistent. Theorem 2.4 requires α∈(1/3,1/2), p≥2, s<0, ℓ<−1, while Theorem 6.3 concludes the same convergence for α<1/2, p≥1, ℓ<0, and Theorem 5.4 states the existence of the measure for ℓ<1, s<0. The bounds proved in Proposition 5.7 and Proposition 5.8 involve conditions such as ℓ−1/p<0 and 2pℓ−2+dimG/q′−p<0, not ℓ<−1. Moreover the introduction indicates that the expected support condition is s<−1, while Theorem 2.4 assumes s<0. The proof therefore does not justify the range claimed in the paper's headline theorem, and the statements in Sections 5 and 6 do not agree with it.","section":"Theorems 2.4, 5.4, 6.3 and Propositions 5.7–5.8"}],"minor_comments":[{"comment":"The displayed second difference in the flow-box norm contains W(x1,y2) twice; presumably one of the terms should be W(x1,y1). As written the expression reduces to a first-order difference and would define a different space.","section":"Definition 4.1"},{"comment":"The phrase 'From Paragraph 20' does not correspond to any numbered paragraph; it should refer to Section 5.3 or to a specific displayed estimate.","section":"Section 6.2, proof of Proposition 6.2"},{"comment":"The Cauchy estimate in L^p(Ω,W^{α−ε,α−1−ε,p}) is followed by the conclusion of convergence in W^{α,α−1,p;s,ℓ}, but the loss of ε in the local regularity is never addressed.","section":"Appendix A.1"},{"comment":"Theorem 6.3 does not specify the allowed range of the weight s, although the spaces B^{α,p,s,ℓ}_{YM} depend on s; this makes the statement of the theorem incomplete.","section":"Theorem 6.3"}],"recommendation":"reject","confidential_remarks":"The core issue is that the main theorem depends on the unpublished companion paper [6] and on an appendix the authors themselves label a sketch; at the key technical point the manuscript concedes (footnote 13 in §A.3) that the required second-order rough path is not defined when the conditioning curve meets the unstable curves. This is not a routine missing detail but an unresolved question about the existence of the object under study. If the companion paper is completed and the construction is made fully rigorous, the manuscript may be worth reconsidering, but in its current form the central claim is not proved."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"First thing to know: this is not a finished proof of the zero-area limit. It is a research announcement with substantial scaffolding. The new object — the anisotropic spaces B and the noise-flat decomposition, Theorem 2.1 — is genuinely novel and the right kind of tool for the problem. The Morse-gauge description of the character variety and the ABG symplectic form formula are also nice, though much of Section 3 is recollection. If the construction works, Theorem 2.4 is a real advance over the holonomy-level results of Witten, Forman, Liu, Sengupta, and Levy, and it answers Levy's question in the intended sense.\n\nWhere it falls short: the measure A_{Σ,σ} is defined in Section 5 by conditioning the free-boundary measure on Hol=1, but the holonomy functional needed for that conditioning is not established for curves that cross unstable curves. Appendix A is the only in-manuscript support, it is explicitly a sketch, and footnote 13 concedes that the Wick square, and hence the second-order rough path, is unclear at exactly those crossings. Formula (A.11) concatenates pieces with jumps, but no Chen identity or moment bounds across the jumps are supplied. That is not a cosmetic gap; it is the definition of the object whose limit is the theorem. The companion paper [6] is 'in preparation', so the referee cannot verify the input. Second, the parameter statements do not match: Theorem 2.4 demands ℓ<−1, s<0 while Theorem 6.3 only needs ℓ<0, and the tightness argument in 6.1 claims compact injections make a uniform bound in B suffice, but the bound is in the same space, not a strictly better one. Without a better regularity index the compactness step does not give tightness. These are fixable but real.\n\nThe geometric and algebraic parts are mostly solid and honestly attributed. The citation pattern is fine; self-citation is not a problem here because the cited results are the actual foundation.\n\nWho this is for: people working on functional-analytic 2D Yang–Mills and low-regularity gauge theory. It deserves a serious referee — the idea and the partial proofs are worth the referee time — but the decision should be to require the companion paper and a repaired holonomy construction before acceptance. I would not cite the main theorem in its current form.","headline":"A promising but under-built preprint: the zero-area limit is the right target and the anisotropic spaces are a real idea, but the central theorem is not proven in the manuscript — it leans on an unpublished companion and an appendix that concedes the key obstacle.","tokens_in":55395,"tokens_out":2923,"would_cite":false,"duration_ms":26677,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["58D27","60B10","81T13","57R70"],"pacs":[],"model":"deepseek-v4-flash","headline":"On compact surfaces, the zero-area Yang–Mills limit is the Atiyah–Bott–Goldman symplectic measure, realized on singular currents from a Morse flow.","keywords":["Yang-Mills measure","semiclassical limit","flat connections","Atiyah-Bott-Goldman symplectic form","anisotropic Banach spaces","Morse gauge","distributional connections","compact surfaces"],"falsifier":"On the torus with $G=\\mathrm{SU}(2)$, the proof predicts that the noise moment $E\\|\\pi_{\\mathrm{noise}} A_{\\Sigma,t\\sigma}\\|^{2p}_{B^{\\alpha,2p,s,\\ell}}$ decays like a positive power of $t$ determined by $p$ and $\\dim G$; if a direct computation shows the decay is slower than any positive power, tightness fails and the limit retains noise. Alternatively, if the holonomy functional $\\mathrm{Hol}_{\\max f-\\epsilon}(A^{\\mathrm{Free}}_{S,\\sigma})$ cannot be realized as a measurable function of the $B$-space element, the conditioning that defines the closed-surface measure is undefined.","tokens_in":54342,"feed_emoji":"","tokens_out":9228,"duration_ms":80926,"temperature":0.7,"pith_summary":"The paper aims to prove that, for compact surfaces and gauge group $\\mathrm{SU}(N)$, the two-dimensional Yang–Mills random connection converges as the area goes to zero to a measure carried by flat distributional connections, and that the limiting measure is exactly the normalized Atiyah--Bott--Goldman symplectic volume written in a Morse gauge. The convergence takes place in new anisotropic Banach spaces of distributional one-forms tailored to the Morse flow. If correct, this upgrades earlier results, which identified the limit only through holonomies, to a statement about random gauge potentials themselves, and gives a probabilistic interpretation of the symplectic volume of the moduli space of flat connections. It also answers the question of whether the semiclassical limit can be seen at the level of connections rather than only loop observables. The proof depends on a companion construction of the free-boundary Yang-Mills measure, which is only sketched in an appendix.","feed_headline":"Small-area Yang-Mills limit lands on flat-connection symplectic volume","feed_subtitle":"Zero-area random gauge potentials on any compact surface converge to the Atiyah-Bott-Goldman measure in Morse gauge.","key_machinery":"The carrying object is the decomposition $B^{\\alpha,p,s,\\ell}_{\\mathrm{YM}}=W^{\\alpha,p,s,\\ell}\\oplus(\\mathfrak{g}\\otimes\\mathrm{span}\\{U_a\\})$, where $W^{\\alpha,p,s,\\ell}$ is an anisotropic Banach space of distributional one-forms with local Gagliardo regularity $\\alpha$ along the flow and $\\alpha-1$ transversely, weighted by parameters $s$ and $\\ell$ near saddle points and extrema, and the $U_a$ are integration currents over unstable curves. The Morse gauge itself is the gauge transformation generated by the twisted transport equation $\\partial_t g_t + L_V g_t + A(V)g_t=0$, whose large-time solution turns any smooth flat connection into the combinatorial object $\\sum_a\\left(\\int_{W^s(a)}A\\right)U_a$. This geometry supplies a formula for the Atiyah--Bott--Goldman symplectic form as a sum over intersection numbers of unstable curves, and the extraction maps $\\pi_{\\mathrm{noise}}$ and $\\mathrm{Ext}_a$ split any random connection into its vanishing noise part and its surviving flat part.","core_discovery":"The central claim (Theorem 2.4) is that, for $\\alpha\\in(1/3,1/2)$, $p\\ge 2$, $s<0$, $\\ell<-1$, the family of Yang--Mills measures $(A_{\\Sigma,t\\sigma})_{t>0}$ converges weakly on $B^{\\alpha,p,s,\\ell}_{\\mathrm{YM}}$ to a measure supported on the finite-dimensional space $F$ spanned by the currents $U_a$ of integration over unstable curves of the Morse flow. The limit coincides with the push-forward of the normalized Atiyah--Bott--Goldman measure under the map $(g_a)\\mapsto\\sum_a \\log(g_a)U_a$ restricted to the locus $\\mathrm{Hol}_{\\partial S}((g_a))=1_G$. The proof proceeds by uniform moment bounds that make the family tight, then shows the Gaussian noise component vanishes as $t\\to 0$ while the finite-dimensional flat component survives and is identified with the symplectic volume through the heat-kernel regularization of $\\mathrm{Hol}_{\\partial S}$. This yields a semiclassical limit at the level of random distributional connections.","pith_inferences":["If the identification is right, the Atiyah--Bott--Goldman measure can be sampled by drawing the finite-dimensional flat component from the conditioned Haar law and discarding the noise; this suggests numerical estimators for symplectic volumes from the distributional limit.","The Morse-gauge description of flat connections is effectively a non-abelian analogue of the Morse--Witten complex, so the same device might compute intersection-pairing or symplectic invariants of character varieties from purely dynamical data on the surface.","The same probabilistic convergence is expected for general compact Lie groups, but the geometric identification with the Atiyah--Bott--Goldman measure would need to handle reducible representations and non-trivial centers separately.","A testable extension would be to prove that the conditioned holonomy process converges to the combinatorial holonomy in a rough-path topology, giving a quantitative rate for the zero-area limit rather than only weak convergence."],"forward_implications":["Bounded continuous functions on the anisotropic space $B^{\\alpha,p,s,\\ell}_{\\mathrm{YM}}$ become admissible observables for both the Yang--Mills measure and the limiting Atiyah--Bott--Goldman measure.","In the zero-area limit the Gaussian noise component of the connection vanishes, so the limiting random connection is almost surely a finite linear combination of integration currents over unstable curves of the Morse flow.","The normalized Atiyah--Bott--Goldman symplectic volume acquires an explicit representative as the push-forward of a conditioned Haar measure on $G^{2g}$ under $(g_a)\\mapsto\\sum_a \\log(g_a)U_a$.","For genus at least two, singular points of the moduli space carry zero symplectic volume and can be ignored in the limit, while genus one requires and receives a separate treatment.","Because the Yang--Mills action depends on the metric only through the area form, the small-area limit is the same as the strong-coupling, small-temperature, and semiclassical limits."],"supporting_citations":[{"why":"Constructs the free-boundary Yang--Mills measure on surfaces via Morse theory; supplies the measure whose conditioning, moment bounds, and holonomy measurability the present proof assumes.","marker":"[6]"},{"why":"Builds the Yang--Mills measure as a scaling limit of lattice models in anisotropic spaces; provides the companion construction of the random field and its noise component.","marker":"[15]"},{"why":"Proves convergence of the pulled-back connection $\\varphi^{-t*}A$ to Morse-theoretic currents under the gradient flow, the dynamical step behind the Morse-gauge description.","marker":"[16]"},{"why":"Establishes the small-volume limit of two-dimensional Yang--Mills and identifies the limiting volume with Reidemeister torsion or symplectic volume on regular points.","marker":"[24]"},{"why":"Supplies the heat-kernel and moduli-space arguments identifying the normalized limit with the Atiyah--Bott--Goldman symplectic volume.","marker":"[39]"},{"why":"Gives the semiclassical limit for gauge theory on the torus, the genus-one input for the identification.","marker":"[57]"},{"why":"Provides the computation linking the Yang--Mills partition function in the small-area limit to the symplectic volume of the moduli space of flat connections.","marker":"[62]"}],"fun_headline_variants":["Zero-area Yang-Mills limit equals Atiyah-Bott-Goldman law","Morse gauge semiclassical Yang-Mills: flat-connection volume","Yang-Mills zero-area limit lands on Atiyah-Bott-Goldman","Semiclassical Yang-Mills on compact surfaces: flat-connection measure"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The conclusion depends on the free-boundary Yang--Mills measure existing as a random element of these anisotropic spaces with measurable holonomy and the moment bounds used in Propositions 5.6--5.8; that construction is deferred to a companion paper and only sketched in Appendix A.","fun_headline_variants_meta":{"raw":{"variants":["Zero-area Yang-Mills limit equals Atiyah-Bott-Goldman law","Morse gauge semiclassical Yang-Mills: flat-connection volume","Yang-Mills zero-area limit lands on Atiyah-Bott-Goldman","Semiclassical Yang-Mills on compact surfaces: flat-connection measure"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001471,"raw_usage":{"total_tokens":5930,"prompt_tokens":975,"completion_tokens":4955,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":591,"completion_tokens_details":{"reasoning_tokens":4873}},"tokens_in":591,"tokens_out":4955,"duration_ms":29512,"temperature":1.0,"reasoning_tokens":4873,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-15T15:32:03.338108+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"On the torus with $G=\\mathrm{SU}(2)$, the proof predicts that the noise moment $E\\|\\pi_{\\mathrm{noise}} A_{\\Sigma,t\\sigma}\\|^{2p}_{B^{\\alpha,2p,s,\\ell}}$ decays like a positive power of $t$ determined by $p$ and $\\dim G$; if a direct computation shows the decay is slower than any positive power, tightness fails and the limit retains noise. Alternatively, if the holonomy functional $\\mathrm{Hol}_{\\max f-\\epsilon}(A^{\\mathrm{Free}}_{S,\\sigma})$ cannot be realized as a measurable function of the $B$-space element, the conditioning that defines the closed-surface measure is undefined.","supporting_citations":[{"cited_title":"Bonthonneau, R","cited_arxiv_id":null,"evidence_quote":"Constructs the free-boundary Yang--Mills measure on surfaces via Morse theory; supplies the measure whose conditioning, moment bounds, and holonomy measurability the present proof assumes."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Proves convergence of the pulled-back connection $\\varphi^{-t*}A$ to Morse-theoretic currents under the gradient flow, the dynamical step behind the Morse-gauge description."},{"cited_title":"Forman, Small volume limits of 2-d Yang–Mills,Communications in Mathematical Physics 151(1993), 39–52","cited_arxiv_id":null,"evidence_quote":"Establishes the small-volume limit of two-dimensional Yang--Mills and identifies the limiting volume with Reidemeister torsion or symplectic volume on regular points."},{"cited_title":"Liu, Heat kernel and moduli space,Mathematical Research Letters3(1996), 743–762","cited_arxiv_id":null,"evidence_quote":"Supplies the heat-kernel and moduli-space arguments identifying the normalized limit with the Atiyah--Bott--Goldman symplectic volume."},{"cited_title":"Sengupta, The Semiclassical Limit for SU(2) and SO(3) Gauge Theory on the Torus, Commun","cited_arxiv_id":null,"evidence_quote":"Gives the semiclassical limit for gauge theory on the torus, the genus-one input for the identification."},{"cited_title":"Witten, On quantum gauge theories in two dimensions,Communications in Mathematical Physics141(1991), 153–209","cited_arxiv_id":null,"evidence_quote":"Provides the computation linking the Yang--Mills partition function in the small-area limit to the symplectic volume of the moduli space of flat connections."}],"review_version":2}