{"id":"8278f436-5fab-4054-95f9-19877ecea599","arxiv_id":"2506.01690","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Any two non-cyclic point stabilizers in a hyperbolic-like circle action have an explicit proper ping-pong partition, so their generated subgroup is their free product.","lead":"This paper proves that stabilizers of special points in hyperbolic-like groups of circle homeomorphisms admit explicit ping-pong partitions, implying free product decompositions. It is a technical step toward Bonatti's conjecture about the algebraic structure of such groups.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Proposition A.2's unenumerated case analysis is the load-bearing gap: if one of the four discarded commutator configurations is consistent, Theorem I/II and Corollary 1.3 can fail.","rationale":"The reader's weakest_assumption identifies exactly the same load-bearing concern: Proposition A.2's finite case analysis is summarized rather than enumerated. I find no independent internal inconsistency in the paper; the arguments around it are careful and the omitted cases are finite and explicitly framed in Table 2. Still, Proposition A.2 is used at every stage of both main theorems, so the completeness of its case check is genuinely load-bearing. The proposed test is feasible: it is a finite combinatorial re-derivation from the already stated graph-crossing method, and it would settle whether the missing enumeration hides a genuine counterconfiguration. I therefore keep the reader's CONDITIONAL verdict unchanged; no adjustment is needed beyond insisting on the independent enumeration before accepting the central claim as fully established.","tokens_in":24969,"tokens_out":10289,"duration_ms":114744,"concrete_test":"Independently enumerate the nine combinations of row-3 diagram types in Table 1 applied to the two factorizations in Eq. (A.4), with the fixed-point constraints p<q<p<q imposed. For each combination, use the graph-crossing rule from Proposition A.1 to determine the attracting and repelling fixed points of [f,h], [h,f^{-1}], [h^{-1},f], and [f^{-1},h^{-1}], and verify that exactly the five rows of Table 2 are compatible while the four rejected cases force an element with zero or more than two fixed points, or violate the cyclic-order constraints. If a rejected combination survives, Proposition A.2 and the proof of Theorem 5.2 need revision.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Theorems I and II, and hence Corollary 1.3, rest on Proposition A.2, which asserts that for f in H+(q) and h in H+(p) with p<q<p<q, the fixed points of the four commutators must occur in one geometric configuration or in one of four non-geometric configurations (2.I)-(2.IV). The proof in Appendix A reduces the problem to a 3x3 compatibility test via the two factorizations in Eq. (A.4): [f,h] = f.(hfh^{-1})^{-1} = (fhf^{-1}).h^{-1}. However, the text states that 'a case-by-case analysis shows that only five are compatible' and lists the five survivors in Table 2, without describing the four rejected combinations or the contradiction each produces. This is not a cosmetic omission: Theorem 5.2 invokes Proposition A.2 repeatedly, including the precise orderings of cases (2.I)-(2.IV), to run the contradiction arguments in Claims 1-3 and the Main Claim. If one of the four discarded configurations is actually consistent with a hyperbolic-like action, then an unhandled case could break the proof of geometric linkage and with it the explicit ping-pong partition. The dependency on [BCT24, Prop. 3.6] is external and not circular; the concern is the internal completeness of a finite case check.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper studies subgroups of Homeo_+(S^1) in which every nontrivial element has exactly two fixed points, one attracting and one repelling (hyperbolic-like groups). It proves that for any pair of non-simple points p and q with p not in {q, q̄}, the stabilizer subgroups admit a proper ping-pong partition, provided the corresponding fixed-point pairs are linked (Theorem I) or unlinked (Theorem II), with an explicit partition in each case. The main structural corollary (Corollary 1.3) is that if such a group is generated by two non-cyclic abelian subgroups, it is their free product. The proofs rely on the local-discreteness theorem of Bonatti–Carnevale–Triestino [BCT24] and on a combinatorial classification (Appendix A, Proposition A.2) of the possible positions of fixed points of four commutators built from elements of the two stabilizers.","tokens_in":25208,"tokens_out":5388,"duration_ms":48768,"significance":"The result is a meaningful step toward Bonatti's conjecture for hyperbolic-like actions: it gives dynamical ping-pong partitions in a case where classical ping-pong is not directly available. A notable strength is that the partitions are explicit, and the paper correctly reduces the problem to a finite configuration check. The dependency on [BCT24, Prop. 3.6] is external and published; although it shares an author with the present paper, it is not circular. The correctness of the main theorems is, however, tied to the completeness of the case analysis in Proposition A.2, which is currently asserted rather than shown in detail.","major_comments":[{"comment":"The proof of Proposition A.2 performs a 3×3 compatibility test using the two factorizations in (A.4) and states that 'a case-by-case analysis shows that only five are compatible', listing the five survivors in Table 2. The four discarded combinations are not enumerated, and no contradiction is exhibited for them. Since Theorem 5.2 invokes Proposition A.2, relying on the precise ordering of the non-geometric cases (2.I)–(2.IV), and Theorems I, II and Corollary 1.3 inherit this dependence, the completeness of this case check is load-bearing. Please provide the complete enumeration of all nine candidates with the explicit incompatibility for each rejected one, or a reproducible formal/computational verification.","section":"Appendix A, Proposition A.2"},{"comment":"The proof of the main contradiction is given only for the non-geometric case (2.II), and the other three cases are dismissed as 'similar'. Given that the argument relies on the specific inequalities in each of the cases (2.I)–(2.IV), it is not transparent that the same construction of the interval I and the claims adapt without modification. Please either expand these cases or explain precisely how the argument for (2.II) transfers to (2.I), (2.III), and (2.IV).","section":"Section 5, proof of Theorem 5.2"}],"minor_comments":[{"comment":"The sentence 'Note that when StabG(p), there exists a unique point p ∈ S1∖{p} such that StabG(p)=StabG(p)' is missing a condition; it should presumably read 'when StabG(p) is non-simple' or similar.","section":"Section 2.2"},{"comment":"The definitions 'Up = (Ip∪Ip)∖Iq∪Iq' and 'Uq = (Iq∪Iq)∖Ip∪Ip' are ambiguous; add parentheses, e.g., Up=(Ip∪Ip)∖(Iq∪Iq).","section":"Corollary 5.4"},{"comment":"The displayed expression 'h−1(I3∖I2)∖I2' appears to be a typo; it should be 'h−1(I3∖I2)⊂I2'.","section":"Corollary 6.9, Claim 2"},{"comment":"The text '32 = 9 total possibilities' should read '3^2 = 9 total possibilities'.","section":"Appendix A, Table 2 preamble"},{"comment":"The introduction cites '[Kov99]' but the reference list contains both [Kov99a] and [Kov99b]; please disambiguate the citation.","section":"References"}],"recommendation":"major_revision","confidential_remarks":"The main risk is the unenumerated case analysis in Proposition A.2; the editor may wish to ask the authors to supply an expanded appendix or a computer-verified table. The self-citation to [BCT24] is legitimate, but since one of the authors is a co-author of the cited paper, it would be good to ensure independent scrutiny of that dependency."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The paper proves something concrete: in a hyperbolic-like group, any pair of non-cyclic point stabilizers admits a proper ping-pong partition, with explicit partitions in both the linked and unlinked cases. Corollary 1.3 — that a non-elementary hyperbolic-like group generated by two non-cyclic abelian subgroups is their free product — is new and is exactly the kind of structural consequence Bonatti's conjecture predicts.\n\nWhat the paper does well: the overall strategy is clean. It uses local discreteness (from BCT24) to get gaps, then analyzes how gaps of two non-simple points can sit inside each other, and rules out non-geometric arrangements using fixed-point configurations of commutators. The proof of Theorem 5.2 in the linked case is substantial, and Section 6 handles the unlinked case honestly, explicitly flagging that the non-geometric cases may not actually occur. The dependence on [BCT24, Prop. 3.6] is a self-citation, but the cited result is published and does not assume the ping-pong conclusion, so I do not see circularity.\n\nThe main soft spot is Proposition A.2. The proof reduces the analysis to a 3x3 compatibility test via the two factorizations in (A.4), then says \"a case-by-case analysis shows that only five are compatible\" and lists the five survivors in Table 2. The four discarded combinations are not described, nor are the contradictions they produce. This is load-bearing: Theorem 5.2 and the Main Claim repeatedly invoke the precise orderings of cases (2.I)–(2.IV). If one of the discarded configurations were actually consistent with a hyperbolic-like action, the ping-pong partitions could fail. I cannot point to an actual error — the diagrams and Table 2 make the check plausible — but as written it is a summarized finite check, not a verification a reader can easily reproduce. A referee should ask the authors to enumerate the rejected cases or provide a short machine check. This is a presentation gap, not a demonstrated flaw.\n\nMinor issue: a few typos in Section 6 (e.g., in Corollary 6.9 one inclusion appears to be missing a subset symbol), but nothing that affects the arguments.\n\nThis paper is for people working on groups of circle homeomorphisms, convergence groups, and the Bonatti/Frankel conjectures. It does not resolve the conjecture and does not introduce a new technology, but it gives the first explicit ping-pong partitions for stabilizers of non-simple points and a clean free-product corollary. I would send it to a serious referee, with a request to expand Appendix A.","headline":"Explicit ping-pong partitions for non-simple point stabilizers, a solid step toward Bonatti's conjecture, but the proof leans on a finite case check that should be expanded before publication.","tokens_in":25749,"tokens_out":2068,"would_cite":true,"duration_ms":21577,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["57M60","37C85","37B05"],"pacs":[],"model":"deepseek-v4-flash","headline":"Explicit ping-pong partitions found for pairs of point stabilizers","keywords":["hyperbolic-like group","circle homeomorphisms","ping-pong partition","non-simple points","free products","commutators","point stabilizers","gaps"],"falsifier":"Find a non-elementary hyperbolic-like group containing linked non-simple points $p,q$ and elements $f\\in H_+(q)$, $h\\in H_+(p)$ such that the fixed points of $[f,h]$ occupy one of the four configurations that Proposition A.2 discards, or exhibit an unlinked pair whose gap configuration falls outside the three listed in Theorem II.","tokens_in":24732,"feed_emoji":"🏓","tokens_out":10076,"duration_ms":86214,"temperature":0.7,"pith_summary":"Hyperbolic-like groups are subgroups of the orientation-preserving circle homeomorphisms in which every non-trivial element has exactly two fixed points, one attracting and one repelling. This paper proves that if two points in such a group have non-cyclic stabilizers, then those two stabilizers admit a proper ping-pong partition: disjoint open sets with finitely many components that the two subgroups push into each other. The partitions are constructed explicitly from the gaps of the two points, with the linked and unlinked cases handled separately. A direct consequence is that any hyperbolic-like group without a global fixed point, generated by two non-cyclic abelian subgroups, is isomorphic to their free product, matching the splitting that a conjecture in the field predicts for minimal actions.","feed_headline":"Explicit ping-pong partitions found for pairs of point stabilizers","feed_subtitle":"A proper partition of the circle shows two non-cyclic stabilizers generate a free product.","key_machinery":"The load-bearing object is the proper ping-pong partition itself, defined as a pair of disjoint non-empty open sets $(U_H,U_K)$ with finitely many connected components such that $(H\\setminus\\{\\mathrm{id}\\})(U_K)\\subset U_H$ and $(K\\setminus\\{\\mathrm{id}\\})(U_H)\\subset U_K$. Starting from the stabilizer of a non-simple point $p$, the paper forms the right and left sides of $p$ relative to its companion point $\\bar p$, and defines gaps as the wandering intervals of the stabilizer action on those sides. A key input is a classification, Proposition A.2, of the possible positions of fixed points of the four commutators $[f,h]$, $[h,f^{-1}]$, $[h^{-1},f]$, and $[f^{-1},h^{-1}]$ for $f\\in H_+(q)$ and $h\\in H_+(p)$, where $H_+(p)$ is the semigroup of elements whose attracting fixed point is $p$; among nine candidates from a graph-crossing analysis, only five are compatible, one geometric and four non-geometric. This restriction on commutator fixed points is what forces the gap configurations and ultimately produces the ping-pong partitions.","core_discovery":"The central claim is that the classical ping-pong lemma applies to any pair of stabilizers of non-simple points (points whose stabilizer is neither trivial nor infinite cyclic) in a non-elementary hyperbolic-like group. If $p$ and $q$ are non-simple points with $p\\notin\\{q,\\bar q\\}$, then $\\operatorname{Stab}_G(p)$ and $\\operatorname{Stab}_G(q)$ admit a proper ping-pong partition realized by intervals built from their gaps. For linked pairs, Theorem I gives a partition of the form $(J_p\\cup J_{\\bar p},\\, J_q\\cup J_{\\bar q})$; for unlinked pairs, Theorem II says one of three configurations occurs, a geometric two-interval one or two explicitly non-geometric ones, and the non-geometric cases cannot occur when $p$ and $q$ lie in the same orbit. Corollary 1.3 then yields the free product structure $\\langle H,K\\rangle\\cong H*K$ for any non-elementary hyperbolic-like group generated by non-cyclic abelian subgroups $H$ and $K$.","pith_inferences":["A similar commutator-fixed-point analysis might handle three or more non-simple points and yield free-product splittings for groups generated by finitely many abelian stabilizers, though the paper considers only pairs.","The paper leaves it open whether the non-geometric case in Theorem II actually occurs; settling this would either sharpen the trichotomy or reduce it to the geometric case.","Because the partitions are defined through gaps and monotone maps, they should be invariant under semi-conjugacy, suggesting the free-product splitting is a semi-conjugacy invariant of hyperbolic-like actions—a consequence the paper does not draw.","If the ambient conjecture is true, hyperbolic-like groups that are not semi-conjugate to Fuchsian groups split as amalgams over abelian subgroups; Corollary 1.3 realizes the simplest such splitting, a free product of two abelian pieces."],"forward_implications":["Whenever a non-elementary hyperbolic-like group is generated by two non-cyclic abelian subgroups, the two subgroups are free factors, so the group is exactly their free product.","The ping-pong partitions are explicit and built from gaps, so the free-product splitting comes with a dynamical description rather than an abstract algebra argument.","In the linked case the partition structure canonically yields four intervals covering the circle, pinning down how the two stabilizers move each other's regions.","For unlinked points in the same orbit, the non-geometric configurations are excluded, so the partition is always the simple two-interval geometric one.","The gap and core constraints proven along the way give new restrictions on how the minimal invariant set of a hyperbolic-like action can intersect stabilizers of non-simple points."],"supporting_citations":[{"why":"Provides the local discreteness theorem used to ensure non-simple points have gaps and the stabilizer actions are discrete.","marker":"[BCT24]"},{"why":"Supplies the classical ping-pong lemma that converts a proper ping-pong partition into a free product.","marker":"[Mas88]"},{"why":"Gives the classification of elementary hyperbolic-like groups that underlies the structure of point stabilizers and companion points.","marker":"[Kov99b]"},{"why":"Supplies the standard facts on minimal invariant sets and elementary actions of circle homeomorphism groups used throughout.","marker":"[Ghy01]"}],"fun_headline_variants":["Explicit ping-pong partitions for non-simple point pairs","Ping-pong lemma applies to non-cyclic stabilizers","Free product structure from hyperbolic-like actions","Proper ping-pong partition for stabilizer pairs","Non-simple points admit explicit ping-pong partitions"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The proof depends on the assertion in Proposition A.2 that only five of the nine possible commutator fixed-point configurations are compatible, with the four discarded configurations not individually enumerated; if one of those four were actually possible, the classification and the partitions built on it would fail.","fun_headline_variants_meta":{"raw":{"variants":["Explicit ping-pong partitions for non-simple point pairs","Ping-pong lemma applies to non-cyclic stabilizers","Free product structure from hyperbolic-like actions","Proper ping-pong partition for stabilizer pairs","Non-simple points admit explicit ping-pong partitions"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000159,"raw_usage":{"total_tokens":1175,"prompt_tokens":837,"completion_tokens":338,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":453,"completion_tokens_details":{"reasoning_tokens":262}},"tokens_in":453,"tokens_out":338,"duration_ms":3869,"temperature":1.0,"reasoning_tokens":262,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-07T11:35:38.454261+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Find a non-elementary hyperbolic-like group containing linked non-simple points $p,q$ and elements $f\\in H_+(q)$, $h\\in H_+(p)$ such that the fixed points of $[f,h]$ occupy one of the four configurations that Proposition A.2 discards, or exhibit an unlinked pair whose gap configuration falls outside the three listed in Theorem II.","supporting_citations":[],"review_version":1}