{"id":"676557b6-5ebe-45b2-8e96-bbaff5c4784b","arxiv_id":"2502.06154","paper_version":2,"verdict":"ACCEPT","confidence":"MODERATE","novelty_score":8.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"The pro-unipotent automorphism group of the associated graded Goldman-Turaev Lie bialgebra on a closed surface is explicitly described in terms of a divergence map and the kernel of a reduced coproduct.","lead":"This paper determines the pro-unipotent automorphism group of the associated graded Goldman-Turaev Lie bialgebra for a closed surface, completing the uniqueness part of the formality problem in that case. It does so by reformulating Kashiwara-Vergne groups in terms of non-commutative connections and by constructing a basis of the relevant trace space via rewriting rules.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Theorem 6.2's proof only characterizes Lie bialgebra automorphisms induced by Der^+(L̂(H)_ω); it never proves every pro-unipotent automorphism of the associated graded Goldman–Turaev Lie bialgebra is induced by such a Hopf algebra automorphism.","rationale":"The reader's weakest assumption concerned Lemma 6.4 and its reliance on CBEG07 and the new rewriting basis. While I agree those are intricate and worth scrutiny, the more load-bearing issue is structural: the proof of the main theorem does not bridge from arbitrary Lie bialgebra automorphisms to the Hopf-algebra automorphisms in which krv(g,0) lives. The associated graded Goldman–Turaev Lie bialgebra is a Lie bialgebra on the trace space |T̂(H)_ω/K1|; its automorphism group is a priori a group of linear automorphisms of that vector space, not of L̂(H)_ω. The proof only analyzes the subgroup of automorphisms that are induced by Aut^+(L̂(H)_ω). Since the stated equality is an equality of the full pro-unipotent automorphism group with exp(krv(g,0)), this missing containment is essential. It may be a standard lifting result from AKKN23 or the author's prior work, but it is not stated or proved in this paper. I recommend conditional acceptance: the theorem's statement and proof should be completed by adding (or explicitly citing) a lemma that every pro-unipotent automorphism of the associated graded Goldman–Turaev Lie bialgebra on a closed surface is induced by an automorphism of T̂(H)_ω preserving the Hopf structure. The proposed genus-1 test is a fast, concrete way to detect whether the missing step is actually false in a simple case.","tokens_in":37943,"tokens_out":49424,"duration_ms":416453,"concrete_test":"For g = 1, Der^+(L̂(H)_ω) = 0, so the theorem predicts the pro-unipotent automorphism group of the associated graded Goldman–Turaev Lie bialgebra on |Ŝ(H)_{>0}| is trivial. Compute the Lie algebra of derivations of this Lie bialgebra in degrees 1 through 4 (using the explicit Goldman bracket and Turaev cobracket for the torus). If any nonzero derivation exists, Theorem 6.2 is false as stated. If none exists, run the same check for g = 2 in low degrees: determine whether every derivation of the Lie bialgebra that is homogeneous of degree ≥ 1 is induced by a derivation of L̂(H)_ω, which would support adding the missing lifting lemma.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"The central theorem states that the pro-unipotent automorphism group of (|T̂(H)_ω/K1|, [·,·]_gr, δ_gr) equals exp(krv(g,0)), with krv(g,0) ⊂ Der^+(L̂(H)_ω). The proof of Theorem 6.2 begins: 'Suppose that G ∈ Aut^+(L̂(H)_ω) induces an automorphism...' and then shows that, among such G, preserving the Lie bialgebra is equivalent to the divergence condition. This establishes only the inclusion exp(krv(g,0)) ⊆ Aut(g, [·,·]_gr, δ_gr). The converse — that every pro-unipotent automorphism of the Lie bialgebra on the trace space is induced by an automorphism of the completed Hopf algebra T̂(H)_ω, or equivalently by an element of Aut^+(L̂(H)_ω) — is never proved or cited. The surrounding text says the proof 'imitates [AKKN23]', but the closed-surface case lacks the boundary conditions that in the boundary case force a tangential automorphism. Without a lifting lemma, the theorem overstates the result: it characterizes the intersection of the automorphism group with Aut^+(L̂(H)_ω), not the automorphism group itself. This is not a problem internal to Lemma 6.4; even if Lemma 6.4 and Theorem 5.22 are correct, the stated equality does not follow from the written proof.","agreement_with_reader":"disagree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies the formality problem for the Goldman–Turaev Lie bialgebra on a closed surface. It first reformulates the Kashiwara–Vergne groups and associators for surfaces with boundary in terms of non-commutative connections and divergence maps, building on the author's earlier work. It then constructs a new basis of the trace space |T(H)_ω| via a rewriting system, proves several structural lemmas, and applies them to identify the pro-unipotent automorphism group of the associated graded Goldman–Turaev Lie bialgebra with exp(krv(g,0)), where krv(g,0) consists of derivations of L̂(H)_ω whose divergence lies in the kernel of the reduced coproduct. The last section computes that kernel in low degrees and states a conjecture for degree 5.","tokens_in":38243,"tokens_out":21746,"duration_ms":200499,"significance":"If the main theorem is correct, it solves the uniqueness side of the formality problem for closed surfaces in higher genus, a question left open in [AKKN23]. The reformulation in terms of connections (Theorem 4.21) is a genuine conceptual step, and the rewriting basis of Theorem 5.22 as well as the low-degree computations of Section 7 are explicit and checkable. The paper is carefully written and supplies detailed proofs of the combinatorial ingredients. However, the proof of the main theorem has a load-bearing gap in the direction from arbitrary Lie bialgebra automorphisms to automorphisms of the Hopf algebra, so the stated equality is not fully established.","major_comments":[{"comment":"The proof of Theorem 6.2 begins with 'Suppose that G in Aut+(L̂(H)_ω) induces an automorphism ...' and then characterizes such G. This establishes only that exp(krv(g,0)) is contained in the pro-unipotent automorphism group of (|T̂(H)_ω/K1|, [·,·]_gr, δ_gr). The reverse inclusion is not proved: no argument shows that an arbitrary pro-unipotent automorphism of this Lie bialgebra is induced by an automorphism of the completed Hopf algebra T̂(H)_ω, or equivalently by an element of Aut+(L̂(H)_ω). In the boundary case of [AKKN23] the tangential boundary conditions provide such a lifting; for a closed surface no boundary condition is available and the paper neither proves nor cites a lifting lemma. Since Theorem 6.2 asserts an equality of automorphism groups, the missing surjectivity (and the faithfulness of the action on trace space) is a load-bearing gap.","section":"6, Theorem 6.2"}],"minor_comments":[{"comment":"In Definition 2.2, the concatenation α ∗_p β is defined in Definition 2.1 as a free loop, but σ(α)(β) is required to be an element of Kπ (based loops). Please clarify how α ∗_p β is based at the base point in the setting of the Kawazumi–Kuno action.","section":"2, Definition 2.2"},{"comment":"In Lemma 5.18, the collection move and the re-distribution move are only described via Figure 3; giving explicit algebraic formulas for m^col_w and m^red_w would make the proof easier to verify, especially because the lemma is used to identify all loop holonomies with multiples of hol(m^s_t ... m^s_1).","section":"5, Lemma 5.18"},{"comment":"In Lemma 6.4, the reduction from the completed space to the non-completed T(H)_ω by gradedness and the step 'if deg(y) ≥ 2 ... we obtain [ỹ,x̃]_gr = 0' are terse; spelling out the degree and invariance argument would increase confidence, as this lemma is used in the proof of Theorem 6.2.","section":"6, Lemma 6.4"}],"recommendation":"major_revision","confidential_remarks":"The main gap is a missing lifting/surjectivity statement for Theorem 6.2. The paper also relies heavily on the author's earlier work ([Tan24a], [Tan24b], [Tan25]), including the factorization of the Turaev cobracket and the family of algebraic operations; the editor may wish to confirm these are available and that the cited statements cover the completed/closed-surface settings. If the missing lifting lemma cannot be supplied, the theorem should be restated as a characterization of the automorphisms induced by Aut^+(L̂(H)_ω) rather than as the full automorphism group."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Quick take: the paper has real results inside it, but the main theorem is likely stronger than the proof demonstrates. Theorem 6.2 claims the pro-unipotent automorphism group of the associated graded Goldman–Turaev Lie bialgebra on a closed surface equals exp(krv). The proof begins by assuming an automorphism G of the completed free Lie algebra L̂(H)_ω, then shows that among those, preserving the cobracket is equivalent to the divergence condition. It never proves that every pro-unipotent automorphism of the trace Lie bialgebra (|T̂(H)_ω/K1|, [·,·]_gr, δ_gr) is induced by such a Hopf algebra automorphism. The boundary case gets a lifting mechanism from tangential automorphisms; closed surfaces lack that, and I could not find a lifting lemma or a reference supplying one. So the equality in Theorem 6.2 is not established. The written proof gives exp(krv) ⊆ Aut and the characterization of Aut^+(L̂(H)_ω) ∩ Aut, but not the converse inclusion. That is load-bearing: if trace-space automorphisms that do not lift exist, the theorem as stated is false. The gap is not internal to Lemma 6.4 or the rewriting system; it is in the framing of the theorem.\n\nWhat is genuinely good: the connection-theoretic reformulation of the KV groups in Theorem 4.21 is elegant and useful—expressing divergences as traces of adjoint connections is a clean framework. The rewriting basis of |T(H)_ω| in Section 5 is substantial and original; the skew-weighted graph argument (Lemma 5.24) is terse but seems sound. The low-degree kernel computations in Section 7 are honest and flagged as partial. The paper relies on the author's own to-appear work for the divergence factorization, which is not a flaw by itself but makes independent verification harder. Lemma 6.4's use of CBEG07 with a gradedness extension to the completed setting looks plausible.\n\nWho should read this: specialists in Goldman–Turaev theory and deformation quantization. The tools in Sections 4 and 5 have value even if the main theorem needs adjustment. It deserves a serious referee, but the referee should push hard on the lifting question. If a lifting lemma can be added, the paper is strong; if not, Theorem 6.2 should be restated as a characterization of the image of Aut^+(L̂(H)_ω).","headline":"Strong technical content, but Theorem 6.2 as written only characterizes automorphisms that lift to the Hopf algebra—the closed-surface case needs a lifting lemma before the stated equality is justified.","tokens_in":788,"tokens_out":852,"would_cite":true,"duration_ms":105037,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["16D20","17A61","53D30","57K20","58B34"],"pacs":[],"model":"deepseek-v4-flash","headline":"A single divergence condition determines the pro-unipotent automorphism group of the associated graded Goldman–Turaev Lie bialgebra on any closed surface.","keywords":["Goldman–Turaev Lie bialgebra","formality problem","Kashiwara–Vergne groups","higher genus associators","non-commutative connections","divergence maps","closed surfaces","necklace Lie bracket"],"falsifier":"For a genus-2 closed surface, take the element of Ker(|Δ̄_ω|) in degree 4 that is not in |HL(H)_ω^(3)|, whose existence is asserted in Remark 7.2, express it as the divergence of a derivation in Der⁺(L̂(H)_ω), and test directly whether that derivation preserves the graded Turaev cobracket on a finite truncation. A single failure of the equivalence in either direction — a cobracket-preserving derivation whose divergence lies outside Ker(|Δ̄_ω|), or a divergence in the kernel that does not preserve the cobracket — would falsify Theorem 6.2.","tokens_in":37688,"feed_emoji":"🌀","tokens_out":8231,"duration_ms":70051,"temperature":0.7,"pith_summary":"This paper determines the pro-unipotent automorphism group of the associated graded Goldman–Turaev Lie bialgebra for a closed surface: the symmetries are exactly the exponentials of derivations whose non-commutative divergence lands in the kernel of the reduced coproduct. This settles the non-uniqueness side of the formality problem for closed surfaces, complementing the previously known existence of formality isomorphisms. The author also reformulates the higher-genus Kashiwara–Vergne groups and associators in terms of connections on modules over Hopf groupoids, giving a uniform framework that covers both boundary and closed cases. The proof rests on a new explicit basis of the trace space of the quotient tensor algebra together with a quoted theorem about the centre of the preprojective algebra.","feed_headline":"Closed-surface automorphism group pinned down by divergence","feed_subtitle":"The pro-unipotent symmetries are exactly exponentials of derivations with divergence in the reduced-coproduct kernel.","key_machinery":"The machinery is threefold. First, the non-commutative connection ∇'_{•,H} and its associated divergence map, taken from the author's previous work, factor the Turaev cobracket as a composition Div_{∇•,H} ∘ σ_gr of a divergence and the graded Kawazumi–Kuno action; this factorization is what lets a derivation condition stand in for a cobracket condition. Second, a new rewriting basis of |T(H)_ω| (Theorem 5.22), built from rewriting rules ρ and ρ₂ and analysed by skew-weighted graphs and holonomy loops, proves the crucial Lemmas 6.3 and 6.4; Lemma 6.4 uses the Crawley–Boevey–Etingof–Ginzburg centre theorem applied to the necklace Lie bracket to show (g⊗g)^g = 0. Third, the reduced coproduct Δ̄_ω and its kernel provide the target in which the divergence must land, and the paper computes that kernel in degrees up to 4, giving explicit generators such as ∧³H/|Hω| in degree 3.","core_discovery":"The central claim is Theorem 6.2: for a closed surface of genus g, the pro-unipotent automorphism group of the associated graded Goldman–Turaev Lie bialgebra (|T̂(H)_ω/K1|, [·,·]_gr, δ_gr) is exp(krv_(g,0)), where krv_(g,0) = {g ∈ Der⁺(L̂(H)_ω) : div_{∇'_{•,H}}(g) ∈ Ker(|Δ̄_ω|)}. Here L̂(H)_ω is the completed free Lie algebra on the first homology H modulo the symplectic element ω, ∇'_{•,H} is a flat homological connection on a resolution of the trivial module, and Δ̄_ω is the reduced coproduct. The proof reduces preservation of the graded Turaev cobracket to the vanishing of the invariant space (g⊗g)^g, which is shown to be zero using the identification of the centre of the trace space of the preprojective algebra with the constants; this reduction uses a new explicit rewriting basis of the trace space |T(H)_ω|.","pith_inferences":["The divergence condition is cohomological in nature, so the automorphism group might be expressible in terms of Hochschild or cyclic cohomology of the preprojective algebra; the paper does not pursue this identification.","The rewriting basis of |T(H)_ω| constructed in Section 5 is likely reusable for computations in necklace Lie algebra cohomology and for the Johnson homomorphism problem, independently of the formality question.","Conjecture 7.6, if true, would imply that for (g,d) ≠ (2,4) every kernel element in the quotient algebra lifts from the free algebra, so the exceptional genus-2 degree-4 behaviour may be an isolated pathology in the automorphism group computation.","A finite linear-algebra truncation at degree 4 or 5 for a small genus could be used to test the theorem computationally, since the paper's kernel computations make both sides of the condition explicit."],"forward_implications":["The automorphism group of the associated graded Goldman–Turaev Lie bialgebra on a closed surface is now explicitly described, so the non-uniqueness part of the formality problem is settled for every genus.","Every automorphism in the group preserves the graded Hamiltonian flow σ_gr automatically, so the graded Lie bracket and the Kawazumi–Kuno action are fixed before the cobracket condition is imposed.","The divergence condition can be checked degree by degree: in low degrees the kernel of the reduced coproduct is zero in degree 2, equals ∧³H/|Hω| in degree 3, and equals |HL(H)_ω^(3)| in degree 4 for g ≠ 2, giving explicit constraints on the lowest potentially non-trivial automorphisms.","Because krv_(g,0) is pro-nilpotent, the group KRV_(g,0) is itself determined by the same divergence condition, and the set of formality isomorphisms remains a torsor over this group."],"supporting_citations":[{"why":"Introduces the Goldman–Turaev Lie bialgebra and the Kashiwara–Vergne problem in higher genera; provides the existence of formality isomorphisms and the torsor framework used for the closed-surface case.","marker":"[AKKN23]"},{"why":"Supplies Theorem 8.6.1(ii), identifying the centre of the trace space of the preprojective algebra with |K1|, which is used in Lemma 6.4 to prove (g⊗g)^g = 0.","marker":"[CBEG07]"},{"why":"The author's previous paper defines non-commutative divergence and the factorization of the Turaev cobracket via connections, providing the core tool that the closed-surface argument builds on.","marker":"[Tan24b]"},{"why":"Establishes the Hopf-groupoid and tangential-derivation framework, including the connections ∇'_{C,fr} and ∇'_{H,fr} on modules over the fundamental groupoid.","marker":"[Tan24a]"},{"why":"Supplies the identification of the necklace Lie bracket on |T(H)_ω| with the Goldman bracket, enabling the use of quiver-variety results in Lemma 6.4.","marker":"[KK16]"},{"why":"A companion paper whose Corollary A.5 equates d(Div) with Tr(ad), used in the proof of Theorem 6.2 to pass from divergence preservation to cobracket preservation.","marker":"[Tan25]"}],"fun_headline_variants":["Formality of Goldman–Turaev on closed surfaces resolved","Closed-surface Turaev symmetries: divergence determines all","Pro-unipotent automorphisms of closed-surface Turaev bracket","Non-commutative connections crack closed-surface Goldman–Turaev"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The proof relies on the quoted theorem that the centre of the trace space of the preprojective algebra is exactly the constants, extended from the uncompleted to the completed algebra, together with the new rewriting basis for |T(H)_ω|; if that extension fails, the equivalence between preserving the cobracket and the divergence condition breaks down.","fun_headline_variants_meta":{"raw":{"variants":["Formality of Goldman–Turaev on closed surfaces resolved","Closed-surface Turaev symmetries: divergence determines all","Pro-unipotent automorphisms of closed-surface Turaev bracket","Non-commutative connections crack closed-surface Goldman–Turaev"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001199,"raw_usage":{"total_tokens":4892,"prompt_tokens":844,"completion_tokens":4048,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":460,"completion_tokens_details":{"reasoning_tokens":3975}},"tokens_in":460,"tokens_out":4048,"duration_ms":26751,"temperature":1.0,"reasoning_tokens":3975,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-08T16:32:04.815264+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"For a genus-2 closed surface, take the element of Ker(|Δ̄_ω|) in degree 4 that is not in |HL(H)_ω^(3)|, whose existence is asserted in Remark 7.2, express it as the divergence of a derivation in Der⁺(L̂(H)_ω), and test directly whether that derivation preserves the graded Turaev cobracket on a finite truncation. A single failure of the equivalence in either direction — a cobracket-preserving derivation whose divergence lies outside Ker(|Δ̄_ω|), or a divergence in the kernel that does not preserve the cobracket — would falsify Theorem 6.2.","supporting_citations":[],"review_version":1}