{"id":"44bda210-7e53-4d1a-a8a3-e7084f7ff699","arxiv_id":"2508.20285","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"For positive symplectic rational surfaces of type D, the group of Hamiltonian diffeomorphisms is a connected component of the symplectomorphism group in the C0-topology.","lead":"The paper proves a rigidity theorem for four-dimensional rational surfaces: on most of them, Hamiltonian symmetries form an isolated connected component inside the larger group of all symplectic symmetries. This is the first nontrivial case where the Hamiltonian group is known to be C0-closed in the full symplectomorphism group.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The proof of Theorem 1.2 inherits its entire algebraic foundation from the unpublished, overlapping preprint [LLW22b]; if the mapping-class-group and fibration computations there fail, the diagram chase in Proposition 3.2 collapses, along with Theorem 3.1, Theorem 4.1, and the main rigidity result.","rationale":"The reader's CONDITIONAL verdict points to the LLW22b input as the weakest assumption, and my stress-test confirms that this is the single most load-bearing item. The diagram chase in Proposition 3.2 is the linchpin: all later C0 work presupposes that elements of Stab0(Σ) are Hamiltonian. The paper's new inflation argument in §3.2 extends the criterion to all pure type D forms, but it needs a base case, and that base case, together with the identification of the relevant mapping class groups, is taken from unpublished work. My internal check did not reveal an independent fatal flaw in the C0-control part: the omitted Torelli hypothesis in Theorem 4.1 is harmless for the intended application, since a sufficiently C0-small diffeomorphism is homotopic to the identity and hence Torelli; Lemma 4.7 is plausible and can be proved from the faithfulness of the action of π0(Diff^+(S2,n)) on arcs; and the codimension-2 genericity invoked in Construction 4.2 is standard, though it would benefit from a citation. The n=5 and non-pure sections are sketched, but they do not constitute a separate collapse once the pure type D core is granted. Thus the concern is precisely the reliance on LLW22b for Theorem 2.7, Proposition 2.8, and the fibration data (2.3)-(2.6). The proposed test pins down one concrete pure type D case and checks the three imported facts directly; if they hold there, the main structural concern is substantially mitigated, though not eliminated for all type D forms. I therefore keep the reader's verdict unchanged.","tokens_in":20486,"tokens_out":34055,"duration_ms":323923,"concrete_test":"Take the pure type D example X=CP2#6CP2 with [ω]=(1|7/10,3/20,...,3/20), which satisfies condition (2.4) for n=6 (λ=7/10 > 3/5). Independently write out the exact sequence (2.6) for this case: verify that π0(Symph)≅PB_5(S2)/(Z/2), that Symp_c(X\\Σ) is weakly contractible for this λ, and that the connecting map π1((S1)^6×Diff^+(S2,5))→π0(Stab0(Σ)) is surjective. If any one of these imported facts cannot be reproduced from [LLW22b] or from the published methods in [Eva11] and [LLW22a], then Proposition 2.8 and the base case of the inflation argument are unsupported.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The load-bearing step is not the C0-foliation argument but the smooth-level algebra that feeds it. Theorem 3.1 (Stab0(Σ)⊂Ham) is obtained from Proposition 3.2, whose proof uses the isomorphisms π0(Symph(X,ωλ))≅PB_{n-1}(S2)/(Z/2) and π0(Aut(S2,n-1))≅PB_{n-1}(S2)/(Z/2), together with the bihopfian property of this quotient, to force im(u)=im(u'). These isomorphisms are precisely Theorem 2.7, imported from [LLW22b], an unpublished arXiv preprint with overlapping authorship. Proposition 2.8, which supplies the base case v'∘u=0 under condition (2.4), depends on [LLW22b, Lemmas 5.5 and 5.6] asserting weak contractibility of Symp_c(X\\Σ) and on [LLW15, Lemma 2.9] for surjectivity of the connecting map in the exact sequence (2.6); only a sketch is given. If the weak-contractibility statement or the exact sequence is wrong, Proposition 2.8 fails, Lemma 3.3 has no base case, the inflation transfer in §3.2 cannot get started, and Theorem 3.1 is not established. Consequently Theorem 4.1 cannot correct a C0-small f into Stab0(Σ), and Theorem 1.2 has no proof. I found no circularity: LLW22b does not prove the C0 conclusion. The concern is therefore dependence on an unverified external algebraic input, not an internal inconsistency.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proves that for positive symplectic rational surfaces of type D, the Hamiltonian diffeomorphism group Ham(X, ω) is a connected component of Symp(X, ω) in the C0-topology, and in particular is C0-closed. The proof has two main stages. First, an algebraic stage (Theorem 3.1) shows that the pointwise stabilizer Stab0(Σ) of a filling divisor Σ is contained in the Hamiltonian group; this is achieved via a diagram chase in symplectic mapping class groups, using the pure braid group PB_{n−1}(S2)/(Z/2) and its bihopfian property, together with an inflation/Moser transfer argument that reduces the statement to a parameter range where the relevant compactly supported symplectomorphism group is weakly contractible. Second, an analytic stage (Theorem 4.1) shows that any symplectomorphism sufficiently C0-close to the identity can be composed with a Hamiltonian diffeomorphism so that the resulting map fixes a filling divisor pointwise; this is proved using C0-controlled J-holomorphic foliation data, a two-step pseudo-holomorphic isotopy, and a projection estimate on the divisor. Theorem 1.2 follows by taking C0-limits. The paper also discusses a Floer-theoretic alternative and proposes open questions.","tokens_in":20679,"tokens_out":12363,"duration_ms":118663,"significance":"If the proof is completed and the imported mapping-class-group inputs are valid, the result is a significant advance: it provides the first examples with non-trivial symplectic mapping class group for which Question 1.1 has a positive answer, showing that Ham(X, ω) is C0-closed in Symp(X, ω) for positive type D rational surfaces. The strategy of combining symplectic mapping class group computations with C0-controlled foliation and inflation techniques is original and is likely to be influential. The paper is also careful in indicating where it relies on external results, and I found no circularity in the argument: the cited works do not contain the C0-conclusion. The main risk is not internal inconsistency but dependence on the unpublished preprint [LLW22b] and on a few compressed technical assertions that need to be substantiated before the paper can be accepted.","major_comments":[{"comment":"The main theorem rests on an algebraic foundation that is almost entirely imported from the unpublished preprint [LLW22b], which has overlapping authorship with the present paper. Specifically, the diagram chase in Proposition 3.2 uses the isomorphism π0(Symph(X,ωλ)) ≅ PB_{n−1}(S2)/(Z/2) from Theorem 2.7 and the exactness of the rows in (3.1), while Proposition 2.8 uses the weak contractibility of Sympc(X\\Σ) from [LLW22b, Lemmas 5.5 and 5.6] and the surjectivity of the connecting map from [LLW15, Lemma 2.9]. These inputs are load-bearing: if any of them fails, Theorem 3.1, Theorem 4.1, and Theorem 1.2 collapse. The present manuscript gives only a sketch of Proposition 2.8 and does not reproduce the arguments behind Theorem 2.7. The authors should either include proofs of these statements, state the main theorem as conditional on the verification of [LLW22b], or provide a detailed and self-contained account of the exact sequences and isomorphisms used in Proposition 3.2.","section":"§2.2, Theorem 2.7 and Proposition 2.8"},{"comment":"The existence of the path (J_f^t)_{t∈[0,1]} in J^reg satisfying the three listed conditions is justified only by the sentence 'Such a path always exists since J\\J^reg is a union of submanifolds of codimension 2 or higher.' This codimension statement is not proved and no reference is supplied. The path is essential to Step A, since it produces the Hamiltonian isotopy that moves the curves f(D_i) back to D_i. A proof of the codimension statement for the specific configuration classes appearing in the type D filling divisor, or a precise reference establishing it, must be provided.","section":"§4.1.2, Construction 4.2"},{"comment":"The reduction from non-pure type D forms to pure type D forms is compressed into a single paragraph. It asserts that a C0-small symplectomorphism can first be adjusted to move the extra exceptional curves back, then descends to the blown-down pure type D surface, and that the conclusion follows from [LLW22b, Lemma 4.3]. Since Theorem 4.1 is stated for all type D forms, this step is load-bearing. The details of the isotopy of the extra exceptional curves, the C0-smallness control after blowing down, and the precise application of [LLW22b, Lemma 4.3] should be written out.","section":"§4.2.2, non-pure cases"},{"comment":"The proof of Theorem 4.1 invokes the statement 'since f is in the Torelli part' in order to identify f(D_i) with the unique f_*J_0-holomorphic representative of [D_i]. The theorem statement, however, only assumes d_C0(f,id) < ε. This is fixable: for ε smaller than the injectivity radius of the fixed Riemannian metric, C0-closeness to the identity implies that f is homotopic to the identity and hence acts trivially on homology. The authors should state this explicitly in the proof; as written, the proof appears to use an unstated hypothesis.","section":"§4.1, Step A"}],"minor_comments":[{"comment":"There is a typo in 'symplecitc surfaces' in the introduction; it should read 'symplectic surfaces.'","section":"§1, Abstract and Introduction"},{"comment":"The arrows in the iterated fibration diagram (2.3) are not all labeled, making it hard to tell which maps are fibrations and which are inclusions. A short legend or a more explicit chain of fibrations would greatly improve readability.","section":"§2.2, diagram (2.3)"},{"comment":"Lemma 4.7 is stated as an 'elementary fact' with the proof omitted. A one-sentence proof, or a reference, would help the reader verify the criterion for triviality of the mapping class of φ_H^1 ∘ f on D_1.","section":"§4.1.4, Lemma 4.7"},{"comment":"The notation J(X∗)^reg is introduced without an explicit definition; it should be defined analogously to J^reg_ω, and the regularity conditions for the classes B∗−∑ E_i∗ and E_1∗ should be stated.","section":"§4.2.1, n=5 case"},{"comment":"The phrase 'C0-small Hamiltonian isotopy' is ambiguous: it could mean that the diffeomorphisms ϕ_t are C0-close to the identity for all t, or that the path itself is C0-close to the constant path. Please clarify.","section":"§5.2, Question 5.2"}],"recommendation":"major_revision","confidential_remarks":"The manuscript's central theorem is conditional on the unpublished preprint [LLW22b], which has overlapping authorship. This is a verifiability concern that should be addressed editorially: the authors should be asked to make the relevant results from [LLW22b] available in a form that can be checked, or to include the needed statements and proofs in the present paper. The other main issue is the unproved codimension assertion in Construction 4.2, which is essential for the C0-small isotopy step. I do not see a circularity problem, and the overall strategy appears promising, so a major revision rather than rejection seems appropriate."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The main result is genuinely new: for positive rational surfaces of type D, Ham is a connected component of Symp in the C0-topology, the first case where C0-closedness of Ham inside Symp is proved when the symplectic mapping class group is nontrivial. The paper deserves to be read carefully by anyone working on C0 symplectic topology or symplectic mapping class groups.\n\nWhat the paper does well: the pure type D, n ≥ 6 case is argued in real detail. The inflation transfer in Lemma 3.4, the Moser argument in Lemma 3.5, and the projection estimate in Lemma 4.10 are substantial and look right. The diagram chase in Proposition 3.2 using the bihopfian property of PB_{n-1}(S2)/(Z/2) is clean. I found no circularity: none of the cited results contain the C0 conclusion, and the C0 part is separate from the smooth algebra that feeds it.\n\nThe soft spots are real but mostly not fatal. The biggest is that the algebraic foundation comes from the unpublished preprint [LLW22b] with overlapping authorship: the identification of the symplectic Torelli group with PB_k(S2)/(Z/2), the weak contractibility of Symp_c(X\\Sigma) under (2.4), and the surjectivity of certain boundary maps are all imported from there. If any of those inputs fail, Proposition 3.2 collapses and with it Theorem 3.1, Theorem 4.1, and the main result. This is an external dependency, not an internal inconsistency, but it is load-bearing.\n\nTwo smaller issues: Theorem 4.1 as stated omits the Torelli hypothesis that the proof uses (the text says \"since f is in the Torelli part\" without putting that in the statement); this is easily fixed because the application only needs Torelli maps. Lemma 4.7 is stated without proof, and the codimension-2 genericity asserted in Construction 4.2 is not justified by a proof or citation. The n = 5 and non-pure cases are sketched rather than written out in full.\n\nMy verdict: the central claim is defensible, the main case is argued in detail, and the issues are fixable. I would send this to a serious referee rather than desk-reject. I would also ask the authors to state the Torelli hypothesis explicitly, supply a proof or reference for Lemma 4.7, and make the dependence on [LLW22b] very prominent since that preprint is not yet refereed.","headline":"First C0-closedness of Ham inside Symp with nontrivial symplectic mapping class group; the proof is mostly convincing but leans on an unpublished, overlapping preprint for its algebraic backbone.","tokens_in":21431,"tokens_out":1946,"would_cite":true,"duration_ms":20017,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["53D35","53D05","57R17","57S05"],"pacs":[],"model":"deepseek-v4-flash","headline":"On positive symplectic rational surfaces of type D, the Hamiltonian diffeomorphism group is a closed connected component of the symplectomorphism group in the $C^0$-topology.","keywords":["C0-symplectic topology","Hamiltonian diffeomorphism group","symplectic rational surfaces","symplectic mapping class group","pure braid groups","J-holomorphic foliation","inflation","filling divisor"],"falsifier":"Exhibit a type D positive symplectic rational surface $(X,\\omega)$ and a sequence of Hamiltonian diffeomorphisms $f_n$ with $f_n\\to f$ in the $C^0$-topology where $f\\in\\mathrm{Symp}(X,\\omega)\\setminus\\mathrm{Ham}(X,\\omega)$; Theorem 1.2 says no such sequence exists. A more local test of the quantitative engine: find $f$ with $d_{C^0}(f,\\mathrm{id})<\\epsilon$ such that for every Hamiltonian $\\phi$, $\\phi\\circ f$ moves at least one point of a filling divisor $\\Sigma$, contradicting Theorem 4.1.","tokens_in":20059,"feed_emoji":"","tokens_out":11900,"duration_ms":104237,"temperature":0.7,"pith_summary":"This paper establishes that for positive symplectic rational surfaces of type D, the Hamiltonian diffeomorphism group is closed in the $C^0$-topology inside the full symplectomorphism group, and since the manifold is simply connected it is exactly the identity component of that group. This answers the long-standing Question 1.1 affirmatively for the first family of manifolds with nontrivial symplectic mapping class group. The strategy is to prove a quantitative statement: symplectomorphisms sufficiently $C^0$-close to the identity can be adjusted by a Hamiltonian diffeomorphism so that they fix a filling divisor pointwise, and pointwise fixing that divisor forces the map to be Hamiltonian. A non-Hamiltonian symplectomorphism therefore cannot be approximated by Hamiltonian diffeomorphisms, which is the content of the rigidity.","feed_headline":"Hamiltonian diffeomorphisms are C0-closed on type D rational surfaces","feed_subtitle":"First nontrivial family where Ham is a closed connected component of Symp in the C0 topology.","key_machinery":"The central object is a filling divisor $\\Sigma\\subset X$: a union of embedded symplectic spheres whose homology classes are configured as in Figures 2.1 and 2.2, chosen so that the complement is tractable. The proof studies the iterated fibration (2.3) linking compactly supported symplectomorphisms of $X\\setminus\\Sigma$, the stabilizers $\\mathrm{Stab}(\\Sigma)$ and $\\mathrm{Stab}_0(\\Sigma)$, and the symplectic Torelli group $\\mathrm{Symp}_h(X,\\omega)$, whose $\\pi_0$ is identified with the pure braid group quotient $PB_k(S^2)/(\\mathbb{Z}/2)$. Two mechanisms carry the argument: the bihopfian property of that quotient, meaning every epimorphism or monomorphism $G\\to G$ is an isomorphism, which turns the diagram chase of Proposition 3.2 into an equality of image subgroups; and a family inflation procedure along $J$-holomorphic foliations in the fiber class $H-E_1$, which deforms the symplectic form until the previously known convex-complement case (2.4) applies. Theorem 4.1 uses a $J$-holomorphic ruling fibration $\\pi_J:X\\to D_1$ to control the motion of marked points on the divisor and prove that a $C^0$-small symplectomorphism cannot braid them.","core_discovery":"The central discovery is Theorem 1.2: for a positive symplectic rational surface $(X,\\omega)$ of type D, $\\mathrm{Ham}(X,\\omega)$ is a connected component of $\\mathrm{Symp}(X,\\omega)$ in the $C^0$-topology, hence closed. The proof rests on Theorem 4.1, which produces an $\\epsilon>0$ with the property that every symplectomorphism $f$ with $d_{C^0}(f,\\mathrm{id})<\\epsilon$ can be composed with a Hamiltonian diffeomorphism $\\phi$ so that $\\phi\\circ f$ fixes the filling divisor $\\Sigma$ pointwise. Theorem 3.1 then shows that any symplectomorphism fixing $\\Sigma$ pointwise is Hamiltonian isotopic to the identity, so the $C^0$-ball around the identity in $\\mathrm{Symp}(X,\\omega)$ consists entirely of Hamiltonian diffeomorphisms. Since $X$ is simply connected, $\\mathrm{Symp}_0(X,\\omega)=\\mathrm{Ham}(X,\\omega)$, giving the affirmative answer to Question 1.1 for type D.","pith_inferences":["Our inference: if the symplectic mapping class groups of type E rational surfaces can be described with the same fibration data, the inflation-plus-bihopfian strategy of this paper should carry over essentially unchanged, extending Theorem 1.2 to all positive rational surfaces.","Our inference: the paper's Question 5.2, if answered positively, would upgrade the rigidity result to local path-connectedness of $\\mathrm{Symp}(X,\\omega)$ in the $C^0$-topology, since the divisorial decomposition would then give short Hamiltonian paths moving divisors; this is the direction the authors indicate for a sequel.","Our inference: the role of the bihopfian property suggests the mechanism is not specific to rational surfaces; any closed symplectic 4-manifold whose Torelli mapping class group is a finite quotient of a braid group and that admits a filling divisor with a weakly contractible complement should satisfy the same $C^0$-rigidity."],"forward_implications":["On any positive symplectic rational surface of type D, a $C^0$-limit of Hamiltonian diffeomorphisms that happens to be smooth is automatically Hamiltonian, so the Hamiltonian group is closed in $\\mathrm{Symp}(X,\\omega)$.","Question 1.1 has a positive answer for type D: the identity component $\\mathrm{Symp}_0(X,\\omega)$ equals $\\mathrm{Ham}(X,\\omega)$ and is closed in the $C^0$-topology.","This is the first family of examples where $C^0$-closedness of the Hamiltonian group is known even though the symplectic mapping class group $\\pi_0(\\mathrm{Symp}_h(X,\\omega))\\cong PB_k(S^2)/(\\mathbb{Z}/2)$ is nontrivial, so the obstruction to rigidity is genuinely nontrivial.","The proof gives an effective statement at the level of $C^0$-balls: there is a radius $\\epsilon>0$ around the identity in $\\mathrm{Symp}(X,\\omega)$ inside which every element is Hamiltonian after composition with a Hamiltonian diffeomorphism."],"supporting_citations":[{"why":"Supplies Theorem 2.7, the identification of $\\pi_0(\\mathrm{Symp}_h(X,\\omega))$ with $PB_k(S^2)/(\\mathbb{Z}/2)$, and the fibration data (2.3)-(2.6) used in the diagram chase.","marker":"[LLW22b]"},{"why":"Supplies the n=5 case of Proposition 2.8, the bihopfian property recorded as Lemma 2.9, and the Lagrangian $\\mathbb{RP}^2$ construction used in Section 4.2.","marker":"[LLW22a]"},{"why":"Supplies the filling divisor framework, the homotopy equivalence $C^0_\\omega\\simeq C_\\omega$, and the monotone n=5 base case for Stab0($\\Sigma$) being Hamiltonian.","marker":"[Eva11]"},{"why":"Lemma 2.6: the minimal-area exceptional class has an embedded J-holomorphic representative for every compatible J, used to build the ruling fibration.","marker":"[Pin08]"},{"why":"Lemma 2.5: the class $H-E_1$ admits an embedded J-holomorphic representative and hence a foliation for every J, the base of the fibration argument.","marker":"[LZ15]"},{"why":"Provides the inflation machinery along exceptional classes that underpins Lemma 3.4's family of symplectic forms.","marker":"[LU06]"},{"why":"Supplies the singular/negative inflation technology used to deform the symplectic form to the convex-complement regime (2.4).","marker":"[McD15]"},{"why":"Lemma 2.9 in Proposition 2.8: asserts the surjectivity of $\\pi_1((S^1)^{2n-6}\\times \\mathrm{Diff}^+(S^2,n-1))\\to\\pi_0(\\mathrm{Stab}_0(\\Sigma))$, used to prove the Stab0-to-Stab map is trivial.","marker":"[LLW15]"}],"fun_headline_variants":["C0-closed Ham on type D symplectic surfaces","First C0-closed Ham group on type D rational surfaces","Ham is a connected C0-component on type D surfaces","C0-rigidity of Ham on type D rational surfaces"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The proof depends on an imported and not-yet-published classification of the symplectic mapping class group of type D rational surfaces — Theorem 2.7 from [LLW22b] — together with the associated claims that certain compactly supported symplectomorphism groups are weakly contractible and certain connecting maps are surjective; if those inputs are wrong, the chain from pointwise stabilizer to Hamiltonian breaks.","fun_headline_variants_meta":{"raw":{"variants":["C0-closed Ham on type D symplectic surfaces","First C0-closed Ham group on type D rational surfaces","Ham is a connected C0-component on type D surfaces","C0-rigidity of Ham on type D rational surfaces"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001259,"raw_usage":{"total_tokens":5157,"prompt_tokens":948,"completion_tokens":4209,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":564,"completion_tokens_details":{"reasoning_tokens":4139}},"tokens_in":564,"tokens_out":4209,"duration_ms":27344,"temperature":1.0,"reasoning_tokens":4139,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-15T16:51:47.019447+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Exhibit a type D positive symplectic rational surface $(X,\\omega)$ and a sequence of Hamiltonian diffeomorphisms $f_n$ with $f_n\\to f$ in the $C^0$-topology where $f\\in\\mathrm{Symp}(X,\\omega)\\setminus\\mathrm{Ham}(X,\\omega)$; Theorem 1.2 says no such sequence exists. A more local test of the quantitative engine: find $f$ with $d_{C^0}(f,\\mathrm{id})<\\epsilon$ such that for every Hamiltonian $\\phi$, $\\phi\\circ f$ moves at least one point of a filling divisor $\\Sigma$, contradicting Theorem 4.1.","supporting_citations":[],"review_version":2}