{"id":"f0a94e5f-c94c-469c-aced-520d0a50b263","arxiv_id":"2507.20492","paper_version":1,"verdict":"ACCEPT","confidence":"HIGH","novelty_score":4.0,"correctness_risk":"low","formal_verification":"none","parameter_count":0,"one_line_summary":"The vertex-graded 1-cocycles of the ribbon graph complex RGC1 are one-dimensional and spanned by the graph G1 corresponding to the Enomoto-Satoh trace.","lead":"This paper proves that the ribbon graph complex has exactly one 1-cocycle, the one that produces the Enomoto-Satoh trace used to study the Johnson cokernel. This negative result says that searching for new 1-cocycles of this kind cannot reveal more of the Johnson cokernel.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The proof transfers the cohomology computation from RGC0 to RGC1 via an asserted identity-map isomorphism that is not verified; if the identity map is not a chain map, the main theorem is unsupported.","rationale":"The proof of Theorem 4.1 is short and mostly sound: the degree computation for one-vertex ribbon graphs correctly reduces the candidate factors to the (g,n) terms with cohomological degree 4g+2n-5, and the vanishing arguments (Harer vcd, CFP/MSS) eliminate all but G1. The step I find least secure is the transfer from RGC0 to RGC1. The paper asserts an isomorphism given by the identity map, but the total degree shift |G|_1 - |G|_0 = -2g(G) depends on the genus, so the complexes are not graded-isomorphic by the identity; the claim must be that the underlying (ungraded) differentials coincide. In deformation complexes of properads, signs in the differential typically depend on the degree parameter d, so this requires verification. The paper cites MW15, but the reader cannot check the citation from the text, and the proof itself does not use any property of RGC1 other than the asserted identity. If the identity is not a chain map, the RGC0 computation says nothing about RGC1 cocycles, and Theorem 4.1 is unproven. This is a genuine gap in the exposition, not a disagreement with the external vanishing theorem. I therefore recommend a conditional verdict: the result is plausible and likely correct, but the manuscript should either prove the identity-map chain property or give a precise reference with the sign conventions. The proposed test (comparing δ of one-vertex graphs in RGC0 and RGC1) would settle the matter directly.","tokens_in":8236,"tokens_out":22675,"duration_ms":223745,"concrete_test":"Using the definitions in Merkulov–Willwacher (arXiv:1511.07808, Section 4.3), compute the differential δ of the one-vertex, two-loop ribbon graph (genus 1, one boundary component) in RGC0 and in RGC1, taking the identity map between the two complexes. Compare the two resulting linear combinations of ribbon graphs: if δ0 and δ1 differ, the identity is not a chain map and the transfer step in the proof of Theorem 4.1 is invalid. If they agree, repeat for a one-vertex, three-loop graph (or an arbitrary one-vertex graph) to confirm the sign conventions are d-independent; if the differentials agree on all one-vertex generators, the proof's transfer is justified.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"The central step in the proof of Theorem 4.1 is the identification H•(RGC0) ∩ (#V=1) ≅ H•(RGC1) ∩ (#V=1), justified by the statement in Section 3 that 'the complexes RGCd are isomorphic to each other up to degree shift, and the isomorphism is given by the identity map of the underlying vector space.' This is load-bearing: the entire computation of the vertex-degree-1 cocycles is performed in RGC0 using the Kontsevich isomorphism (1), and the conclusion about 1-cocycles in RGC1 depends on this transfer. However, the identity map is not obviously a chain map. The total degrees in RGC0 and RGC1 differ by |G|_1 - |G|_0 = -2g(G), which is not a constant shift, so the identity cannot be a graded isomorphism; whether it is a chain map on the ungraded vector space depends on the signs in the vertex expansion differential, which in the deformation complex construction generally depend on d. The paper cites MW15 for this isomorphism but gives no verification that δ0 and δ1 coincide under the identity. If the differentials differ, the kernel of δ on the vertex-degree-1 subspace could change, and the theorem's conclusion would not follow from the RGC0 computation. This is the most concrete point where the proof could fail; the external vanishing theorems (CFP/MSS) are well-established and not a genuine risk.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies the ribbon graph complex RGC1 introduced by Merkulov and Willwacher, equipped with the vertex grading. The main theorem (Theorem 4.1) states that the space of 1-cocycles with respect to the vertex grading is one-dimensional, spanned by the unique one-vertex-one-edge graph G1 that corresponds to the graded Turaev cobracket and, via the natural map to the Chevalley-Eilenberg complex of symplectic derivations, to the Enomoto-Satoh trace. The proof computes the relevant piece of the cohomology of RGC0 using the Kontsevich isomorphism (1), shows that all moduli-space summands vanish by Harer's virtual cohomological dimension bound and the Church-Farb-Putman/Morita-Sakasai-Suzuki vanishing theorem, and then transfers the result to RGC1 using the asserted isomorphism of the complexes RGCd given by the identity map.","tokens_in":8580,"tokens_out":34237,"duration_ms":331230,"significance":"If the result is correct, it is a clean and useful negative result: no ribbon-graph 1-cocycle other than the universal ES trace can exist, so the detection of the Johnson cokernel cannot be improved within this particular complex. The argument is short and transparent, and it depends on strong external results (Kontsevich's theorem, Harer's vcd, and CFP/MSS vanishing) rather than on new heavy machinery. The degree computation leading to 4g+2n-5 is carefully done, and the vanishing checks are correct. The paper is appropriately cautious in framing the result as a statement about the ribbon graph complex, leaving open the Lie graph analogue in Question 4.4.","major_comments":[],"minor_comments":[{"comment":"The statement 'The complexes RGCd are isomorphic to each other up to degree shift, and the isomorphism is given by the identity map of the underlying vector space' is load-bearing for the proof of Theorem 4.1, but no precise reference to a proposition in [MW15] is given, and the phrase 'up to degree shift' is ambiguous because the degree difference |G|_1 - |G|_0 = -2g(G) is not a constant shift. Please specify the exact statement in [MW15] and explicitly confirm that the differentials δ0 and δ1 coincide on the underlying vector space, or provide the correct chain isomorphism if the identity is not a chain map.","section":"Section 3, paragraph on RGCd isomorphisms"},{"comment":"The expression 'H•(RGC0,δ)∩(#V(G)=1)' is informal; it should be defined as the subspace of cohomology classes admitting a representative that is a cocycle with exactly one vertex. Also, the displayed equality with the product of moduli cohomology groups is asserted 'by the isomorphism (1)' without explaining why the vertex-count condition corresponds exactly to those summands; a containment statement would suffice for the vanishing argument, and equality would then follow from G1 itself being such a cocycle.","section":"Proof of Theorem 4.1"},{"comment":"The transfer 'since the complexes RGCd are isomorphic via the identity map' should be spelled out in the context of the vertex grading: one should explain that a 1-cocycle with one vertex is the unique representative of its cohomology class, so the identification of one-vertex cocycles follows from the identification of the underlying vector spaces. This would also address the reader's concern about the non-constant degree shift.","section":"Proof of Theorem 4.1, final paragraph"},{"comment":"There are a few typos: 'fundamntal' should be 'fundamental' in Section 2, and 'qudratic' should be 'quadratic' in the title of reference [MP98]. In the introduction, the name 'Kassabov' is spelled 'Kasabov' in the phrase 'Conant, Kassabov and Vogtmann'; please make the spelling consistent.","section":"Section 2 and references"},{"comment":"The phrase 'the moduli space M_1^1 is the modular curve, which is a cusped disk' is imprecise: M_{1,1} is an orbifold whose coarse moduli space is the affine line (a disk-like space), and its rational cohomology is indeed trivial. Rephrasing this sentence would avoid confusion.","section":"Proof of Theorem 4.1, (g,n)=(1,1) case"}],"recommendation":"minor_revision","confidential_remarks":"The only point that gives me pause is the transfer from RGC0 to RGC1 via the asserted identity-map isomorphism. I believe the cited fact from [MW15] is standard and the proof is sound, but the authors should make the citation precise and state explicitly that the differentials coincide. This is a local clarity issue rather than a correctness problem, so minor revision seems appropriate."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"This is a short, honest paper that proves a modest negative result: in the ribbon graph complex, the only 1-cocycle in vertex degree one is the graph G1 corresponding to the Enomoto–Satoh trace. That rules out this complex as a source of new detectors for the Johnson cokernel. The result itself, Theorem 4.1, is new; I don't think it appears in the cited literature. The proof is a clean application of two known black boxes: the Kontsevich isomorphism identifying RGC0 cohomology with moduli-space cohomology, and the Church–Farb–Putman/Morita–Sakasai–Suzuki vanishing of the relevant top cohomology. The degree computation is correct: for one-vertex graphs the moduli degree lands at 4g+2n−5, and Harer's vcd plus the vanishing theorem kill every factor except G1. No free parameters, no post-hoc fitting.\n\nThe one soft spot is the transfer from RGC0 to RGC1. The paper says the complexes are isomorphic via the identity map, up to degree shift, and cites Merkulov–Willwacher. This is less transparent than the rest of the argument. The degree formulas differ by a genus-dependent shift, so \"degree shift\" cannot mean a constant shift; and whether the identity is actually a chain map depends on signs in the vertex-expansion differential. If those signs depend on d, the kernel on the vertex-degree-one subspace could in principle change. I suspect this is a standard identification that the author knows, but a referee should ask for the precise chain map and a sentence explaining why the #V=1 cocycle condition is preserved. It is the only load-bearing step I would want verified in writing.\n\nThe citation pattern is fine. The single self-citation appears in Example 4.3 and is not used in the main proof. The external vanishing theorems are solid. This is not a breakthrough; it closes a small door. The right audience is people working on the Johnson homomorphism, the ES trace, or ribbon graph cohomology. I would send it to a serious referee rather than desk-reject: it is short, clear, and correct-looking, and the requested clarification is well within the scope of a normal revision.","headline":"A short, correct-looking negative result: the ES trace graph is the only vertex-graded 1-cocycle in the ribbon graph complex; the only step worth probing is the RGC0/RGC1 identification via the identity map.","tokens_in":9054,"tokens_out":15866,"would_cite":true,"duration_ms":174649,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["18M70","20F34","32G15"],"pacs":[],"model":"deepseek-v4-flash","headline":"In the ribbon graph complex, the vertex-graded 1-cocycles form a one-dimensional space spanned by the one-vertex, one-edge graph G1.","keywords":["Johnson homomorphism","ribbon graph complex","Enomoto–Satoh trace","moduli space of Riemann surfaces","Johnson cokernel","Turaev cobracket","1-cocycles","vertex grading"],"falsifier":"Find a nonzero class in $H^{4g-3}(M_g^1;\\mathbb{Q})$ for some $g\\ge 2$, or directly compute $H^1$ of the vertex-graded ribbon graph complex in low genus and find a class not proportional to $G_1$; either would contradict Theorem 4.1.","tokens_in":1872,"feed_emoji":"🔗","tokens_out":2075,"duration_ms":100941,"temperature":0.7,"pith_summary":"The paper establishes that the space of vertex-graded 1-cocycles in the ribbon graph complex is one-dimensional, spanned by the unique graph with one vertex and one edge, denoted G1. This graph is the universal version of the Enomoto–Satoh trace, the known invariant that annihilates the image of the Johnson homomorphism, so the result says the ribbon graph complex offers no other independent 1-cocycle for detecting the Johnson cokernel. The significance is that a natural route toward new Johnson-cokernel invariants closes at cocycle degree one, while higher-degree cocycles remain an open possibility. The proof achieves this by transporting the question, through the Kontsevich isomorphism, into known vanishing results for the cohomology of moduli spaces of Riemann surfaces.","feed_headline":"Only one ribbon-graph 1-cocycle exists","feed_subtitle":"A moduli-space vanishing theorem rules out every other 1-cocycle, so the Johnson cokernel gains no new detectors from this complex.","key_machinery":"The central object is the ribbon graph complex $\\mathrm{RGC}_d$, a chain complex spanned by labelled ribbon graphs (fat graphs) whose differential is vertex expansion; it arises from the properad of ribbon graphs, a multi-input/multi-output algebraic structure encoding Lie bialgebras. The paper works in the vertex grading, i.e. grading by the number of vertices, and isolates the one-vertex, one-edge graph $G_1$, which represents the graded Turaev cobracket and hence the Enomoto–Satoh trace. The load-bearing identity is the Kontsevich isomorphism, which identifies the cohomology of $\\mathrm{RGC}_0$ with a product of cohomology groups of moduli spaces of marked Riemann surfaces; filtering this isomorphism by vertex count turns the one-dimensionality claim into a vanishing computation in moduli-space cohomology.","core_discovery":"The main result, Theorem 4.1, is that the space of 1-cocycles in $\\mathrm{RGC}_1$ with respect to the vertex grading has dimension one and is spanned by $G_1$, the one-vertex, one-edge ribbon graph. Under the canonical representation of the ribbon graph properad, $G_1$ maps to the graded Turaev cobracket, which in turn gives the Enomoto–Satoh trace on symplectic derivations; the theorem therefore makes the ES trace the unique 1-cocycle of this kind. The proof uses the Kontsevich isomorphism to identify the vertex-count-one part of the cohomology of $\\mathrm{RGC}_0$ with a direct sum of cohomology groups $H^{4g+2n-5}(M_g^n;K) \\otimes \\mathrm{sgn}_n$ of moduli spaces of $n$-marked genus-$g$ curves. That sum vanishes: the $n \\ge 2$ summands exceed Harer's virtual cohomological dimension, the $n=1$, $g\\ge 2$ summand $H^{4g-3}(M_g^1;\\mathbb{Q})$ vanishes by the external top-dimensional vanishing theorem for one-marked moduli spaces, and the $(g,n)=(1,1)$ term is zero because the modular curve has trivial $H^1$. Since the differential in $\\mathrm{RGC}_1$ raises vertex count by one and there are no zero-vertex graphs, cohomology classes with one vertex are exactly the 1-cocycles, and the identity isomorphism between $\\mathrm{RGC}_0$ and $\\mathrm{RGC}_1$ completes the identification.","pith_inferences":["A natural extension is to run the same moduli-space comparison in other vertex degrees and for other values of $d$; a plausible pattern is that low-vertex cohomology classes in $\\mathrm{RGC}_d$ are governed entirely by virtual-cohomological-dimension vanishing, so new detectors, if any, would have to come from high vertex number or from other complexes such as Lie graph complexes.","A direct computational check in low genus at vertex number one would test the theorem independently of the moduli-space vanishing: explicit graph-homology computation of $H^1(\\mathrm{RGC}_1)$ should return exactly the class of $G_1$.","If the external vanishing theorem were ever strengthened or refuted in a range, the one-dimensionality statement would change correspondingly, so the paper's conclusion is conditional on that theorem rather than on the graph complex alone.","The paper's open question about Lie graph complexes suggests an analogous uniqueness statement may or may not hold there; if it does, it would provide a Lie-world version of the ES trace and potentially new invariants of the Johnson cokernel."],"forward_implications":["Any vertex-graded 1-cocycle in the ribbon graph complex is a scalar multiple of $G_1$, so the associated Lie algebra 1-cocycle on symplectic derivations is always the Enomoto–Satoh trace up to scale.","The ribbon graph complex in vertex degree one contains no new Johnson-cokernel detector: any 1-cocycle there annihilates the Johnson image only through the already known mechanism.","The uniqueness is sharp at degree one: higher-degree cocycles in $\\mathrm{RGC}_1$ remain abundant, and the paper poses the question whether any $k$-cocycle for $k>1$ can detect parts of the Johnson cokernel beyond the ES trace.","The vanishing argument transfers a computation over moduli spaces into a statement about combinatorial graph cocycles, so the same comparison can be used to constrain cocycles in other vertex degrees."],"supporting_citations":[{"why":"Constructs the ribbon graph properad and the ribbon graph complex RGC_d, and gives the canonical representation sending G1 to the graded Turaev cobracket.","marker":"[MW15]"},{"why":"One of the references for the Kontsevich isomorphism (1) identifying RGC0 cohomology with moduli-space cohomology.","marker":"[MP98]"},{"why":"Computes the virtual cohomological dimension of mapping class groups, used to eliminate all n >= 2 summands in the vertex-count-one part.","marker":"[Har86]"},{"why":"Supplies the top-dimensional vanishing of rational mapping-class-group cohomology needed for the n=1, g>=2 summands.","marker":"[CFP11]"},{"why":"Provides the companion vanishing result H^{4g-3}(M_g^1;Q)=0 for g>=2, used in the same summand.","marker":"[MSS13]"},{"why":"Establishes the equivalence between the ES trace and the graded Turaev cobracket, making G1 the universal ES trace.","marker":"[AKKN23]"}],"fun_headline_variants":["Ribbon-graph 1-cocycles: only one exists","No other 1-cocycles for Johnson cokernel","Unique 1-cocycle spans ribbon graph cohomology","One vertex, one edge: the sole 1-cocycle","Moduli space vanishing: no extra 1-cocycles"],"cache_read_input_tokens":11264,"weakest_assumption_plain":"The load-bearing premise is the external vanishing theorem that the top-dimensional rational cohomology $H^{4g-3}(M_g^1;\\mathbb{Q})$ is zero for every $g\\ge 2$; if that theorem were false, additional 1-cocycles could survive in the vertex-count-one part of the ribbon graph complex.","fun_headline_variants_meta":{"raw":{"variants":["Ribbon-graph 1-cocycles: only one exists","No other 1-cocycles for Johnson cokernel","Unique 1-cocycle spans ribbon graph cohomology","One vertex, one edge: the sole 1-cocycle","Moduli space vanishing: no extra 1-cocycles"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000543,"raw_usage":{"total_tokens":2679,"prompt_tokens":1105,"completion_tokens":1574,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":721,"completion_tokens_details":{"reasoning_tokens":1485}},"tokens_in":721,"tokens_out":1574,"duration_ms":12838,"temperature":1.0,"reasoning_tokens":1485,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-15T17:43:16.267284+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Find a nonzero class in $H^{4g-3}(M_g^1;\\mathbb{Q})$ for some $g\\ge 2$, or directly compute $H^1$ of the vertex-graded ribbon graph complex in low genus and find a class not proportional to $G_1$; either would contradict Theorem 4.1.","supporting_citations":[],"review_version":2}