Ping-pong dynamics of hyperbolic-like actions with non-simple points

KyeongRo Kim; Michele Triestino

REVIEW 2 major objections 5 minor 3 references

Ping-pong dynamics of hyperbolic-like actions with non-simple points

T0 review · 2 major / 5 minor · reviewed 2026-08-07 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read Explicit ping-pong partitions found for pairs of point stabilizers

desk verdict 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. read the letter →

arxiv 2506.01690 v1 pith:O3XE4TXD submitted 2025-06-02 math.GT math.DSmath.GR

KyeongRo Kim , Michele Triestino This is my paper

classification math.GTmath.DSmath.GR MSC 57M6037C8537B05

keywords hyperbolic-likegroupcirclehomeomorphismsping-pongpartitionnon-simplepointsfreeproductscommutatorspointstabilizersgaps

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

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.

What carries the argument

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.

What would settle it

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.

Watch

Extended reading notes

Core claim

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 $K$.

Load-bearing premise

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.

Editorial extensions

If this is right

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.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

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.

Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the 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.

What 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.

The 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.

Minor 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.

This 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.

Referee Report

2 major / 5 minor

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.

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 (2)

[Appendix A, Proposition A.2] 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 5, proof of Theorem 5.2] 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).

minor comments (5)

[Section 2.2] 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.
[Corollary 5.4] The definitions 'Up = (Ip∪Ip)∖Iq∪Iq' and 'Uq = (Iq∪Iq)∖Ip∪Ip' are ambiguous; add parentheses, e.g., Up=(Ip∪Ip)∖(Iq∪Iq).
[Corollary 6.9, Claim 2] The displayed expression 'h−1(I3∖I2)∖I2' appears to be a typo; it should be 'h−1(I3∖I2)⊂I2'.
[Appendix A, Table 2 preamble] The text '32 = 9 total possibilities' should read '3^2 = 9 total possibilities'.
[References] The introduction cites '[Kov99]' but the reference list contains both [Kov99a] and [Kov99b]; please disambiguate the citation.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the ping-pong partitions are derived from the hyperbolic-like axioms and an external published discreteness theorem, not from the conclusion.

full rationale

The derivation chain is self-contained except for two cited external results. Section 2.4 says 'The first important observation, which is a consequence of [BCT24], is that the stabilizer of a non-simple point p cannot act minimally on both connected components of S^1 minus {p,p} (Proposition 4.4)', and Proposition 3.6 is imported as the statement that a non-elementary hyperbolic-like subgroup of Homeo_+(S^1) is discrete, with proof attributing the first statement to [BCT24, Lemma 4.6]. This citation shares an author (Triestino) and is load-bearing for the paper's discreteness arguments, but it is a published, externally checkable result with stated hypotheses that do not include the ping-pong conclusion; under the reviewing rules, such a citation is real evidence and does not raise the circularity score. The main claims (Theorems I and II, Corollary 1.3) are not obtained by fitting parameters or by defining the objects in terms of the target conclusion: the ping-pong partitions are explicitly constructed from gap configurations, and the fixed-point classification in Proposition A.2 is a case analysis from the hyperbolic-like fixed-point property rather than an assumption of the conclusion. The only notable weakness is that Appendix A's proof says 'a case-by-case analysis shows that only five are compatible' without enumerating the four rejected combinations; this is an omitted-case completeness risk, not a circularity, because the surviving classification is then used to derive, not presuppose, the partition. Corollary 1.3 follows from the ping-pong lemma once such a partition exists, and no equation in the paper reduces to its own input. Score 0.

Assumptions & free parameters 0 free parameters · 5 assumptions · 0 invented entities

No fitted constants or invented entities appear. The mathematical objects (gaps, cores, companion points, geometric configurations) are defined within the paper. The listed axioms are the prior theorems and the unverified combinatorial classification that the proofs rest on.

assumptions (5)

standard math Elementary hyperbolic-like groups fix exactly two points and are semi-conjugate to translations on each of the two complementary intervals (Theorem 3.1).
Invoked in Section 3.1 to describe stabilizers, define companion points and gaps, and in Section 4.
standard math Non-elementary hyperbolic-like subgroups of Homeo_+(S^1) are locally discrete (Proposition 3.6 from [BCT24]).
Used in Proposition 4.4 and Lemma 4.3 to guarantee existence of gaps and discreteness on non-wandering intervals; proven in [BCT24], which shares author Triestino.
standard math A non-elementary action on the circle has a unique minimal closed invariant set, the limit set, which is the closure of the fixed points (Lemma 3.3, relying on Ghys).
Used to connect wandering intervals with gaps and to prove discreteness properties.
standard math Classical ping-pong lemma: if H,K act on a set with disjoint nonempty sets X,Y with (H\{id})(Y) subset X and (K\{id})(X) subset Y, then <H,K> is isomorphic to H * K.
Cited in Corollary 1.3 as Maskit [Mas88, Theorem VII.A.10] to turn ping-pong partitions into free products.
domain assumption The fixed point configuration classification for products and commutators of hyperbolic-like elements (Propositions A.1 and A.2) is complete.
This is the paper's own combinatorial backbone; its proof is summarized with a case-by-case analysis that is not fully enumerated, so correctness is assumed by the reader. It is load-bearing for Theorems I and II.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Ping-pong dynamics of hyperbolic-like actions with non-simple points." pith.science (2026). https://pith.science/paper/O3XE4TXD

@misc{pith2026250601690,
  author       = {Pith},
  title        = {Pith review of: Ping-pong dynamics of hyperbolic-like actions with non-simple points},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/O3XE4TXD}},
  note         = {Machine review of arXiv:2506.01690}
}

abstract

A hyperbolic-like group is a subgroup of $\operatorname{Homeo}_+(S^1)$ such that every non-trivial element has exactly two fixed points, one attracting and one repelling. We investigate the ping-pong dynamics of hyperbolic-like groups, inspired by a conjecture of Bonatti. We show the existence of a proper ping-pong partition for any pair of non-cyclic point stabilizers. More precisely, our results explicitly provide such a ping-pong partition.

Figures

Figures reproduced from arXiv: 2506.01690 by the authors.

**Figure 2.3.** Possible configurations of gaps and associated ping-pong partitions for the linked case (left) and the unlinked cases. The red and blue chords represent gaps of StabG(q) and of StabG(p), respectively. The highlighted intervals describe the associated ping-pong partitions [PITH_FULL_IMAGE:figures/full_fig_p003_2_3.png] view at source ↗

**Figure 5.2.** for a sketch of ˆf when ϑ is finite and t = p¯. ˆf(0) ˆf(0) + θ(h) m(q) = m(s) = 0 θ(h) ϑ ϑ + θ(h) [PITH_FULL_IMAGE:figures/full_fig_p010_5_2.png] view at source ↗

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

3 extracted references · 3 canonical work pages

[1]

Z.307 (2024), no

[BCT24] Christian Bonatti, João Carnevale, and Michele Triestino,Non-locally discrete actions on the circle with at most n fixed points, Math. Z.307 (2024), no. 1, paper no

work page 2024
[1988]

Solodov,Topological problems in the theory of dynamical systems, Uspekhi Mat

[Sol91] Victor V. Solodov,Topological problems in the theory of dynamical systems, Uspekhi Mat. Nauk46 (1991), no. 4(280), 93–114,

work page 1991
[2022]

[CJ94] Andrew Casson and Douglas Jungreis,Convergence groups and Seifert fibered3-manifolds, Invent

PhD thesis, Université Bourgogne Franche-Comté. [CJ94] Andrew Casson and Douglas Jungreis,Convergence groups and Seifert fibered3-manifolds, Invent. Math. 118 (1994), no. 3, 441–456. [Fra13] Steven Frankel,Quasigeodesic flows and Möbius-like groups, J. Differential Geom.93 (2013), no. 3, 401–429. [Fra18] , Coarse hyperbolicity and closed orbits for quasig...

work page 1994

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.

[1] [1]

Z.307 (2024), no

[BCT24] Christian Bonatti, João Carnevale, and Michele Triestino,Non-locally discrete actions on the circle with at most n fixed points, Math. Z.307 (2024), no. 1, paper no

work page 2024

[2] [1988]

Solodov,Topological problems in the theory of dynamical systems, Uspekhi Mat

[Sol91] Victor V. Solodov,Topological problems in the theory of dynamical systems, Uspekhi Mat. Nauk46 (1991), no. 4(280), 93–114,

work page 1991

[3] [2022]

[CJ94] Andrew Casson and Douglas Jungreis,Convergence groups and Seifert fibered3-manifolds, Invent

PhD thesis, Université Bourgogne Franche-Comté. [CJ94] Andrew Casson and Douglas Jungreis,Convergence groups and Seifert fibered3-manifolds, Invent. Math. 118 (1994), no. 3, 441–456. [Fra13] Steven Frankel,Quasigeodesic flows and Möbius-like groups, J. Differential Geom.93 (2013), no. 3, 401–429. [Fra18] , Coarse hyperbolicity and closed orbits for quasig...

work page 1994