{"id":"8f06e3d8-fd45-4bf6-af57-bae28734fe1c","arxiv_id":"2411.15471","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"For any genus-three reducible Heegaard splitting of a connected sum of two lens spaces, the Goeritz group is finitely generated and the reducing sphere complex is connected.","lead":"This paper proves that the Goeritz group, the symmetry group of a genus-three Heegaard splitting of a connected sum of two lens spaces, is finitely generated. It also shows the reducing sphere complex is connected, and that stabilizers of reducing curves inherit finite generation.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The load-bearing gap is Lemma 5.1: the unproved 'induction on genus' assertion that a surface diffeomorphism fixing all oriented essential curves is a boundary twist is what makes the exact sequences (1)-(3) and finite generation of the stabilizers work.","rationale":"I read the proof as reducing Theorem 1.1 to two subtheorems: G=H (Theorem 4.5) and finite generation of the stabilizers H_i (Theorem 1.3 plus Lemma 5.3). The reduction uses standard tools (strong Haken, Birman exact sequence) and is coherent. The most exposed point is Lemma 5.1: the one-line induction on genus is the only justification that the capping homomorphism ρ is well-defined and that the common kernel of the two capping maps is the boundary twist. Without this, the exact sequences (1)-(3) have no proven kernel, and finite generation of H_i is not established. I do not see a counterexample; the assertion is a known rigidity property of the curve complex action, so the theorem may well be correct. But the paper as written leaves a key lemma as an exercise, and the reader's conditional verdict is appropriate. I also considered whether the trivial bubble S3 would lead to a circular genus-3 hypothesis; capping the ball actually leaves a genus-2 splitting of the same connected-sum manifold, so the cited genus-at-most-2 results do apply. Thus the concern is not that the strategy is circular but that Lemma 5.1 needs a complete proof.","tokens_in":12361,"tokens_out":37347,"duration_ms":342792,"concrete_test":"Write out the proof of Lemma 5.1 explicitly for the case Σ_B = a genus-2 surface with one boundary component, by computing the kernel of the action of Mod(Σ_{2,1}) on the set of oriented isotopy classes of essential simple closed curves; verify it is exactly the cyclic group generated by the boundary twist. If a nontrivial element beyond ⟨τ_μ⟩ appears, the exact sequences (1)-(3) and Theorem 1.1 are unsupported. Also check that [FM12, Lemma 3.16] applies verbatim to oriented curves in Σ_B when the isotopy in Σ is orientation-preserving, since the proof of Lemma 5.1 depends on that application.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Lemma 5.1 is the hinge of Section 5: it is used to define the capping homomorphism ρ and to identify I(ρ') ∩ I(ρ) with ⟨τ̃_μ⟩, which then yields the finite-generation of every stabilizer H_i via the exact sequences (1)-(3) and Lemma 5.3. The proof has two steps. First, from ρ1(f)=id it infers, via [FM12, Lemma 3.16], that f fixes every oriented essential curve in the bordered surface Σ_B. Second, it asserts without proof: 'Then we can prove by induction on genus that such a diffeomorphism is isotopic to a power of Dehn twist along the boundary curve μ.' This is a strong rigidity statement about the action of the mapping class group of a bordered surface on oriented simple closed curves; it is not a formal consequence of the cited lemma, and the induction is not supplied. If the kernel of this action is larger than ⟨τ_μ⟩ (for example, if some orientation-preserving finite-order element such as a hyperelliptic-type involution fixed all oriented classes, or if a point-push subgroup survived in low genus), then the kernel of ρ would be larger, the intersection I(ρ')∩I(ρ) would not be just the boundary twist, and the finite-generation argument for H_i would collapse. The same pattern of omitted justification appears in Lemma 4.1 ('The proof for the other cases are similar. So we omit it') and in Claim 4.3, but Lemma 5.1 is the keystone for Theorem 1.3, which in turn gives Theorem 1.1.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies genus-three reducible Heegaard splittings of connected sums of two lens spaces. Its main claim, Theorem 1.1, is that the Goeritz group of any such splitting is finitely generated, and Corollary 1.2 asserts connectedness of the corresponding reducing sphere complex. The proof strategy is modular: introduce three reducing curves μ_i, define their stabilizers H_i inside the Goeritz group, prove via the strong Haken theorem of Freedman-Scharlemann and eyeglass-twist reductions that the whole Goeritz group equals the subgroup H generated by the H_i (Theorem 4.5), and then prove each H_i is finitely generated using capping homomorphisms and known finite generation of genus at most two Goeritz groups (Section 5). The paper also states a more general result, Theorem 1.3, that finite generation of the two summand Goeritz groups implies finite generation of the stabilizer G_μ of the reducing curve μ.","tokens_in":12710,"tokens_out":9328,"duration_ms":87409,"significance":"If the proof can be completed, Theorem 1.1 would give the first finite-generation result for Goeritz groups of weakly reducible genus-three Heegaard splittings outside the genus-two cases already treated by Cho and Koda, and Corollary 1.2 would provide a new family of connected reducing sphere complexes. The overall architecture is attractive and potentially reusable: it separates the problem into a generation statement for the whole Goeritz group and a finite-generation statement for stabilizers of reducing curves. The paper is also honest in its dependence on external anchors—Freedman-Scharlemann's strong Haken theorem, Cho-Koda's lower-genus results, and standard mapping class group sequences—and it introduces no free parameters. However, several load-bearing arguments are only sketched or delegated to omitted cases, most importantly the rigidity statement in Lemma 5.1. The significance is therefore conditional on filling these gaps.","major_comments":[{"comment":"Lemma 5.1 is the keystone of the finite-generation argument, but its proof is not supplied. From ρ1(f)=id and [FM12, Lemma 3.16] the authors conclude that f fixes every oriented essential simple closed curve in the bordered surface Σ_B, and then state: 'Then we can prove by induction on genus that such a diffeomorphism is isotopic to a power of Dehn twist along the boundary curve μ.' This is a strong rigidity statement and is not a formal consequence of the cited lemma. The omitted induction is the only justification for identifying I(ρ') ∩ I(ρ) with ⟨τ̃_μ⟩, which in turn makes the exact sequences (1)-(3) and Lemma 5.3 valid. If the kernel of the action on oriented essential curves of Σ_B is larger than the subgroup generated by the boundary twist, then the finite-generation conclusion for G_μ, and hence for each H_i, collapses. Please provide a complete proof of Lemma 5.1 or a precise reference that contains this statement.","section":"Lemma 5.1, Section 5"},{"comment":"The proof of Lemma 4.1 is incomplete in the induction step of Case 3. After applying the visional bubble move h, the authors assert that an outermost-disk compression of D along D′ produces two essential disks whose boundary curves α1 ∪ α2 intersect h(λ) in at most one point; no argument for this bound is given, and the subsequent split into Subcases 3.1 and 3.2 depends on it. Moreover, the induction step only records I(h(α), μ3) ≤ I(α, μ3), not a strict decrease, so it is unclear why the induction hypothesis applies when equality occurs. Finally, the sentence 'The proof for the other cases are similar. So we omit it' covers several remaining configurations, including the analogues for i=2 and i=3. Since Lemma 4.1 is used in Theorem 4.5 to show every eyeglass twist lies in H, these gaps are load-bearing.","section":"Lemma 4.1, Section 4"},{"comment":"The compression argument in Lemma 4.2 is not fully justified. Claim 4.3 only establishes that the assumption that S contains no scar of D leads to a contradiction when α is separating, but it does not prove the converse direction needed for the construction: that the curves ℓ1 and ℓ2 chosen on the compressed sphere S are reducing curves for the original Heegaard surface Σ. The sentence 'It is not hard to see that both ℓ1 and ℓ2 are reducing curves' is essential, because these curves are used to produce the complete sphere triplet T′ and to apply Lemma 2.2. This lemma is needed for Subcase 2.2 of Theorem 4.5, so the missing justification should be supplied.","section":"Lemma 4.2 and Claim 4.3, Section 4"},{"comment":"The proof of Theorem 4.5, Case 1, contains an unjustified bridge. After choosing a reducing sphere S with μ̄ ∈ O1, the text says: 'By an innermost argument, these two essential spheres S and S1 are isotopic.' Under Definition 2.1, isotopy of reducing spheres is required to preserve the Heegaard surface Σ, so if μ is not isotopic to μ1 as a curve on Σ, this statement cannot hold. The subsequent appeal to Theorem 4.4 requires the sphere sets to be properly isotopic. The proof should either justify this isotopy claim or reformulate the argument so that only the aligned sphere sets associated with S and S1 are compared after applying a suitable element of the Goeritz group. As written, this step is the main bridge from arbitrary elements of G(N,Σ) to bubble moves and eyeglass twists, and it is not established.","section":"Theorem 4.5, Section 4"},{"comment":"The case N1 = N2 is dismissed with 'By the same argument as above ... with only slight modifications.' This case is needed for the full statement of Theorem 1.1, since a connected sum of two lens spaces may have equal summands. In particular, the construction of a diffeomorphism f with f(S1,S2,S3) = (S2,S1,S3) is asserted, and it is claimed that this implies O′_1 = O′_2. But O′_1 and O′_2 are orbits under the subgroup H, not under the full Goeritz group, so it is unclear why existence of f ∈ Diff^+(N,Σ) swapping the spheres implies equality of the H-orbits. Please supply the details of this case, including why the constructed f can be taken in H or why the orbit equality follows differently.","section":"Theorem 4.5, Section 4, Case 2"}],"minor_comments":[{"comment":"There are numerous typographical and grammatical errors, including 'Futhermore' before Definition 3.2, 'buddle move' in Theorem 4.5, 'Geschlect' in the reference [Goe33], 'squence' after equation (2), and 'M od' in Lemma 5.3. These should be corrected.","section":"Throughout"},{"comment":"The notation E_k(S_i) is overloaded: the same symbol is used for the generating set and for the subgroup it generates, and the displayed chain E_1(S_i) = E_2(S_i) = ... appears immediately after proving E_{k+1}(S_i) ≤ E_k(S_i). The intended meaning is clear, but the notation should be disambiguated and the inclusions stated consistently.","section":"Lemma 3.3 and following display"},{"comment":"In the proof of Lemma 5.2, the phrase 'Then We extend f|B by the identity' is ambiguous; the intended operation is to extend the identity on B to obtain a diffeomorphism of N, not to extend the restriction f|B . Rewording would avoid confusion.","section":"Lemma 5.2, Section 5"},{"comment":"Several arguments in Lemma 4.1 and Lemma 4.2 rely on Figures 5, 6, 9, and 10, but the written text does not state which facts are being illustrated and which are being proved. The induction in Lemma 4.1 would be much easier to verify if the figure captions explicitly identified the curve h(λ), the arc λ3, and the new eyeglasses in each subcase.","section":"Figure references, Section 4"}],"recommendation":"major_revision","confidential_remarks":"The manuscript addresses an important question and the proposed mechanism is plausible, but the current version is not yet suitable for publication because Lemma 5.1 and several case analyses in Section 4 are essential and under-proved. The authors should be encouraged to revise with complete proofs of these points. I do not see a circularity problem: the arguments are appropriately anchored in external results."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nThe headline is: this paper is a genuine step forward for the Goeritz group question, but it needs a serious revision before the main theorem can be taken as proved. The new content is real—Theorem 1.1 gives the first finite-generation result for genus-three reducible splittings of connected sums of two lens spaces, Corollary 1.2 shows connectedness of the reducing sphere complex, and Theorem 1.3 is a general stabilizer theorem that may have legs beyond this setting. The visional bubble move is a reasonable extension of Scharlemann's bubble moves and does useful work.\n\nThe proof architecture is coherent. It reduces G(N,Σ) to the subgroups H_i stabilizing the three reducing curves, uses the strong Haken theorem to show every reducing sphere orbit is reached by bubble moves and eyeglass twists, and then proves each H_i finitely generated by capping off the bubble and applying exact sequences to known finite generation of genus-two Goeritz groups. No circular dependencies, no fitted parameters.\n\nThe soft spots are where the proofs are sketched rather than written. Lemma 5.1 is the load-bearing one. It asserts that a diffeomorphism of Σ_B fixing every oriented essential simple closed curve is isotopic to a power of the boundary twist, with the proof 'by induction on genus' and no details. This is exactly what identifies the kernel of ρ with the Dehn twist of μ, and the exact sequences (1)–(3) and Lemma 5.3 depend on it. The statement is plausible—it is the standard rigidity of the curve complex for bordered surfaces, with the boundary twist as the only kernel—but the paper neither proves it nor gives a reference for this precise form. If the kernel were larger in low genus, the finite generation argument for the stabilizers would break. Also, Lemma 4.1 and 4.2 leave several cases as 'similar', and the base/induction details in Lemma 4.1 Case 3 are abbreviated. These are likely fillable, but they are real gaps in a proof that otherwise is quite clear.\n\nWho should read this: anyone working on Goeritz groups, reducing sphere complexes, or mapping class groups of reducible splittings. It deserves a serious referee. I'd send it out, with a clear instruction to the authors to supply the missing proofs for Lemma 5.1 and the omitted cases; if those hold up, the paper is a solid contribution.","headline":"Real, useful progress on finite generation of Goeritz groups, but the keystone lemma is under-proved and needs to be filled before the theorem is fully established.","tokens_in":13221,"tokens_out":5062,"would_cite":true,"duration_ms":46351,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["57K30","57K20","20F05"],"pacs":[],"model":"deepseek-v4-flash","headline":"The Goeritz groups of genus-three reducible Heegaard splittings of connected sums of two lens spaces are finitely generated, and their reducing sphere complexes are connected.","keywords":["Goeritz group","Heegaard splitting","reducing sphere complex","finite generation","lens spaces","eyeglass twist","visional bubble move","mapping class group"],"falsifier":"Check Lemma 5.1 directly: on a genus-two surface with one boundary component, explicitly search for a diffeomorphism in the kernel of the action on isotopy classes of oriented essential simple closed curves that is not a power of the boundary Dehn twist. The existence of such a diffeomorphism would invalidate Lemma 5.1 and with it the exact sequence (1), so the finite-generation conclusion for Gμ would no longer follow.","tokens_in":12161,"feed_emoji":"🔵","tokens_out":8659,"duration_ms":65458,"temperature":0.7,"pith_summary":"The paper proves that for every genus-three reducible Heegaard splitting of a connected sum of two lens spaces, the Goeritz group — the group of orientation-preserving diffeomorphisms of the manifold that preserve the two handlebodies — is finitely generated. This is the first finite-generation result for this family of weakly reducible splittings (splittings whose two handlebodies contain disjoint essential disks). The same argument shows that the reducing sphere complex, a simplicial complex whose vertices are reducing curves, is connected for these splittings. The proof works by pinning the Goeritz group down to three stabilizer subgroups, proving each stabilizer is finitely generated by a reduction to genus at most two, and then assembling them.","feed_headline":"Genus-3 Goeritz groups of lens-space sums are finitely generated","feed_subtitle":"This also proves the reducing sphere complex is connected, a key check for weakly reducible splittings.","key_machinery":"The central objects are eyeglass twists, automorphisms built from a weakly reducing pair of disks joined by an arc, and visional bubble moves, automorphisms that push a singular bubble (a submanifold bounded by a reducing sphere) along a path; together they generate enough of the Goeritz group. The load-bearing identity is $G(N,\\Sigma) = \\langle H_1, H_2, H_3\\rangle$, where $H_i$ fixes the isotopy class of the reducing curve $\\mu_i$. Finite generation of each $H_i$ is obtained from exact sequences that cap off one side of the reducing sphere; the kernel of the capping map is generated by a single Dehn twist about the reducing curve, and the image is a Goeritz group of the capped manifold, which has smaller genus.","core_discovery":"On the paper's own terms, the central discovery is that the Goeritz group $G(N,\\Sigma)$ of a genus-three reducible splitting $N = V \\cup_\\Sigma W$ with $N$ a connected sum of two lens spaces coincides with the subgroup $H$ generated by the stabilizers $H_1, H_2, H_3$ of the three curves $\\mu_1, \\mu_2, \\mu_3$ cut out by a complete sphere triplet. Each $H_i$ is finitely generated via an exact-sequence argument that caps off the bubble bounded by the reducing sphere and compares the stabilizer to the Goeritz group of the capped manifold; the base cases are the genus at most two Goeritz groups of lens spaces and their connected sums, which were previously known to be finitely generated. Consequently $G(N,\\Sigma)$ is finitely generated, and the reducing sphere complex is connected.","pith_inferences":["A plausible extension is that the same exact-sequence scheme proves finite presentability of these Goeritz groups, not just finite generation, if the genus at most two base cases are finitely presented; the paper only states finite generation for the genus-three case.","The identity $G(N,\\Sigma)=H$ suggests a normal-generating set: the whole Goeritz group of such a splitting might be generated by the stabilizers of the three curves of a complete sphere triplet; this could serve as an algorithm to compute presentations for explicit lens-space sums.","The connectedness of the reducing sphere complex may open the door to studying its higher homotopy type, in analogy with curve complexes; the paper only proves connectedness.","Because the proof of Lemma 5.1 is sketched by induction on genus, a fully written inductive proof of that lemma would be needed before adapting the same strategy to arbitrary genus reducible splittings."],"forward_implications":["If Theorem 1.1 is right, every genus-three reducible splitting of a connected sum of two lens spaces has a finitely generated Goeritz group, resolving the finiteness question for that family.","The reducing sphere complex $R$ is connected for such splittings (Corollary 1.2), so any two reducing spheres can be connected by a chain of pairwise disjoint reducing spheres.","Theorem 1.3 gives a general transfer principle: whenever the two summand splittings have finitely generated (or finitely presented) Goeritz groups, the stabilizer $G_\\mu$ of the reducing curve is finitely generated (or finitely presented).","Combining Theorem 1.3 with known finite-generation results for genus at most two yields finite generation of the stabilizers $H_i$, and hence of $G(N,\\Sigma)$."],"supporting_citations":[{"why":"Supplies Theorem 4.4 (uniqueness in strong Haken) used to show every reducing sphere can be transformed to a base sphere by bubble moves and eyeglass twists, establishing G(N,Σ)=H.","marker":"[FS24]"},{"why":"Establishes finite generation of Goeritz groups of reducible genus-two Heegaard splittings, the base case for the reduction in Theorem 1.3.","marker":"[CK19]"},{"why":"Establishes finite generation of genus-two Goeritz groups of lens spaces, needed for the summand splittings.","marker":"[Cho13]"},{"why":"Provides Lemma 3.16 (isotopy of curves in a subsurface) and Proposition 3.19 (kernel of the capping homomorphism), used in Lemma 5.1 and the exact sequences.","marker":"[FM12]"},{"why":"Gives the formula Tη = τα τβ τγ^{-1} for eyeglass twists, the basic algebraic relation used throughout the subgroup arguments.","marker":"[Zup20]"},{"why":"Introduces bubble moves, which the paper extends to visional bubble moves to construct the stabilizer subgroups Hi.","marker":"[Sch22]"}],"fun_headline_variants":["Lens-sum genus-3: Goeritz groups finite, sphere complex connected","Finitely generated Goeritz groups for lens sums, sphere complexes connected","Genus-3 lens-sum Goeritz groups finite and sphere complex connected","Lens-sum Goeritz groups: finite gen, sphere complex connected (genus 3)","Goeritz groups of lens sums: finite generation and sphere complex connectivity"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The whole proof depends on the claim that the only diffeomorphisms of a capped surface that fix the isotopy class of every oriented essential simple closed curve are powers of the Dehn twist about the reducing curve, and that this twist acts trivially in the capped manifold; the paper sketches this by induction on genus without writing out the induction.","fun_headline_variants_meta":{"raw":{"variants":["Lens-sum genus-3: Goeritz groups finite, sphere complex connected","Finitely generated Goeritz groups for lens sums, sphere complexes connected","Genus-3 lens-sum Goeritz groups finite and sphere complex connected","Lens-sum Goeritz groups: finite gen, sphere complex connected (genus 3)","Goeritz groups of lens sums: finite generation and sphere complex connectivity"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001115,"raw_usage":{"total_tokens":4545,"prompt_tokens":747,"completion_tokens":3798,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":363,"completion_tokens_details":{"reasoning_tokens":3708}},"tokens_in":363,"tokens_out":3798,"duration_ms":24052,"temperature":1.0,"reasoning_tokens":3708,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T14:15:48.954411+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Check Lemma 5.1 directly: on a genus-two surface with one boundary component, explicitly search for a diffeomorphism in the kernel of the action on isotopy classes of oriented essential simple closed curves that is not a power of the boundary Dehn twist. The existence of such a diffeomorphism would invalidate Lemma 5.1 and with it the exact sequence (1), so the finite-generation conclusion for Gμ would no longer follow.","supporting_citations":[],"review_version":1}