Pith. sign in

REVIEW 4 major objections 5 minor 35 references

On combination theorems and Bowditch boundaries of relatively hyperbolic TDLC groups

T0 review · 4 major / 5 minor · reviewed 2026-08-15 · deepseek-v4-flash

Pith's one-line read Combination theorems hold for relatively hyperbolic TDLC groups.

desk verdict Solid combination theorems and a useful equivalence of definitions, but the boundary-rigidity main theorem is not yet proven: Section 5 skips a load-bearing topology verification. read the letter →

arxiv 2508.13038 v1 pith:3IGPTYIF submitted 2025-08-18 math.GR

classification math.GR MSC 20F6520F6722D05
keywords TDLCgroupsrelativehyperbolicityBowditchboundarycombinationtheoremsamalgamatedfreeproductsHNNextensionsCayley-Abelsgraphsroughends
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

This paper establishes combination theorems for relatively hyperbolic totally disconnected locally compact (TDLC) groups. It proves that when two relatively hyperbolic TDLC groups are glued along a compact subgroup, either as an amalgamated free product or as an HNN extension, the resulting group is again relatively hyperbolic, and it constructs its Bowditch boundary from the boundaries of the pieces. A consequence is boundary rigidity: if the rough ends of the amalgam are infinite, the topology of the Bowditch boundary of the amalgam is uniquely determined by the topologies of the Bowditch boundaries of the two factors. The paper also shows that a relatively hyperbolic TDLC group with one rough end has connected Bowditch boundary, and that a hyperbolic TDLC group with totally disconnected Gromov boundary splits as a finite graph of compact groups.

What carries the argument

The argument runs through the relative Cayley-Abels graph: a locally finite graph on which the group acts with compact edge stabilizers and vertex stabilizers that are either compact or parabolic. To build the Bowditch boundary, the paper replaces each parabolic coset's Cayley-Abels graph by a combinatorial horoball, forming the augmented Cayley-Abels graph whose Gromov boundary is the Bowditch boundary. For an amalgam or HNN extension, the graph is assembled as a tree of such augmented graphs, one per vertex of the tree of the splitting, and hyperbolicity follows from a classical combination theorem for trees of hyperbolic spaces. The boundary-rigidity statements are then obtained by recognizing the Bowditch boundary as a dense amalgam of the factors' boundaries—a compact metric space assembled from embedded copies of a finite list of compact metric spaces—which records which connected components come from copies of each vertex group.

What would settle it

Construct an explicit amalgam $G=A\ast_C B$, with $C$ compact, where $A$ is relatively hyperbolic with parabolic family $H_A$ and $B$ is relatively hyperbolic with parabolic family $H_B$, such that some $h\in H_A$ is conjugate in $G$ to a distinct $k\in H_B$; if such an example exists, the pair $(G,H_A\cup H_B)$ is not proper, so the relative Cayley-Abels graph is undefined and Theorem 1.1 cannot hold as stated.

Watch

Extended reading notes

Core claim

The central discovery is that relative hyperbolicity, originally developed for discrete groups, can be carried over to TDLC groups in three equivalent definitions, and that the standard combination and boundary-rigidity results hold in this setting. Theorems 1.1 and 1.2 state that the amalgamated free product and HNN extension of relatively hyperbolic TDLC groups along a compact edge group are relatively hyperbolic TDLC groups. Theorems 1.5 and 1.7 state that, when the rough ends of the resulting group are infinite, the homeomorphism type of its Bowditch boundary is determined by the homeomorphism types of the Bowditch boundaries of the vertex groups. Theorem 1.8 states that one rough end forces connectedness of the Bowditch boundary, and Proposition 1.9 states that a hyperbolic TDLC group with totally disconnected Gromov boundary is a finite graph of compact groups.

Load-bearing premise

The construction needs the collection of parabolic subgroups from both factors to remain a proper pair inside the glued group—no two distinct non-compact parabolic subgroups from different factors may become conjugate—and the paper states this rather than proving it.

Editorial extensions

If this is right

  • Any amalgam or HNN extension of relatively hyperbolic TDLC groups along a compact edge group is itself relatively hyperbolic, and the same holds for finite graphs of such groups with compact edge groups.
  • When the rough ends of the glued group are infinite, its Bowditch boundary is homeomorphic to the dense amalgam of the factors' Bowditch boundaries, so its homeomorphism type depends only on the factors' boundary types.
  • Distinct parabolic subgroups of a relatively hyperbolic TDLC group intersect compactly, and conjugates of distinct parabolics intersect compactly.
  • A relatively hyperbolic TDLC group with exactly one rough end has a connected Bowditch boundary.
  • A hyperbolic TDLC group whose Gromov boundary is totally disconnected splits as a finite graph of groups with compact vertex groups.

Reading between the lines

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

  • The dense-amalgam description suggests that the same boundary rigidity should persist for finite graphs of relatively hyperbolic TDLC groups with compact edge groups, by induction on the graph, although the paper only sketches this in Theorem 5.7.
  • If the properness assumption can fail in a natural construction—for instance, a parabolic subgroup of one factor becoming conjugate to one of the other inside the amalgam—then the relative hyperbolicity statement as written would need a notion that tolerates non-proper collections, or the theorem would be false in that case.
  • Connectedness of the Bowditch boundary for one rough end, combined with accessibility of hyperbolic TDLC groups, suggests an end-theoretic characterization of splittings over compact subgroups, with totally disconnected Gromov boundary as a testable symptom.
  • Because the proof uses dense amalgams rather than the earlier free-product technique, the same boundary decomposition should describe how geodesic rays from the different vertex spaces accumulate on the boundary of the amalgam, which could connect to limit-set theory for such groups.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

4 major / 5 minor

Summary. The paper develops a framework for relative hyperbolicity of totally disconnected locally compact (TDLC) groups. It introduces three definitions (TDRH-I, TDRH-II, TDRH-III), proves their equivalence for compactly generated groups, and uses them to prove combination theorems: if A and B are relatively hyperbolic TDLC groups and the edge group C is compact, then the amalgamated free product A*_C B and the HNN extension A*_C are relatively hyperbolic. The main new results are the construction of the Bowditch boundary for such amalgams and HNN extensions via a tree of augmented Cayley-Abels graphs, and the claim that, when the rough ends are infinite, the topology of the Bowditch boundary is determined by the Bowditch boundaries of the vertex groups (Theorems 1.5 and 1.7). The paper also proves connectedness of the Bowditch boundary for one rough end (Theorem 1.8) and a splitting theorem for hyperbolic TDLC groups with totally disconnected Gromov boundary (Proposition 1.9).

Significance. If the main claims hold, the paper provides a substantial extension of classical combination theorems and boundary-rigidity results from discrete relatively hyperbolic groups to TDLC groups, and it connects the existing notions of relative hyperbolicity in this setting. The use of Świątkowski's dense amalgam technique is appropriate, and the paper is clearly organized, with explicit credit to the external results on which it builds ([BF92], [KM08], [Sis12], [S16], [MS15]). However, several load-bearing steps are asserted rather than proved, most importantly the construction of the topology on the compactification X^h in Section 5.3. For this reason the boundary rigidity theorems should be regarded as conditional on a complete proof.

major comments (4)
  1. [§2.3, Lemma 2.3] The proof of properness of the embedding i: X_H → X_G contains an invalid counting argument. The sentence "for every N there exists h_N with d_{X_H}(U_H,h_N U_H) ≤ M but d_{X_G}(U,h_N U) ≤ N" does not imply that there are infinitely many distinct points in the ball of radius M in X_H: the elements h_N could be the same or could repeat. To derive the required contradiction, one must instead use local finiteness of X_G and the fact that, if d_{X_G} is bounded on infinitely many distinct cosets of U_H, then a finite ball in X_G contains infinitely many vertices. This lemma is used in Section 4 to justify the quasi-isometry invariance of the augmented and coned-off Cayley-Abels graphs, so the argument needs to be corrected even if the statement is true.
  2. [§5.3, Proposition 5.1] The central topological identification is not proved. After defining the sets V_U(ξ) and V_n(η), the text says "We skip a verification that the above collections of sets satisfy the axioms for the basis of open neighborhoods," and Proposition 5.1 is then justified only by a "word on the proof" referring to [MS15, Section 3]. The cited argument applies to trees of hyperbolic spaces whose vertex spaces are Cayley graphs of hyperbolic groups, whereas here the vertex spaces are augmented Cayley-Abels graphs with horoballs attached along peripheral cosets. It is not automatic from [MS15] that the V_U and V_n collections form a neighborhood basis, that δ(X^h) is Hausdorff, or that the natural bijection from ∂_{rel}G to δ(X^h) is continuous and open. Without this verification, the homeomorphism type of the Bowditch boundary has not been established, and Proposition 5.4 and Theorems 1.5 and 1.7 are not justified.
  3. [§3, Theorems 1.1 and 1.2, Lemma 3.1] The proof of Lemma 3.1 asserts that the constructed graph X is a Cayley-Abels graph of G with respect to H_A ∪ H_B, but it does not verify that (G, H_A ∪ H_B) is a proper pair in the sense of Definition 2.7. This requires showing that no two distinct non-compact subgroups from H_A and H_B become conjugate in the amalgam or HNN extension. Since the definition of relative hyperbolicity and the entire boundary construction depend on properness, this missing verification is load-bearing. A proof using the Bass-Serre tree and the compactness of C is plausible, but it needs to be supplied.
  4. [§5.4, Proposition 5.4] The proof of condition (2) (nullness of the family Y) uses the compactness of δ(X^h) to pass to a finite subcover of an arbitrary open cover. Compactness of δ(X^h) is exactly part of the conclusion of Proposition 5.1, which is not proved. Thus the argument is circular unless Proposition 5.1 is established independently. This is not a mere technicality, because the dense-amalgam theorem [Ś16, Theorem 0.2] is applied to a space whose compact metrizability is assumed as a hypothesis.
minor comments (5)
  1. [§1, Theorem 1.7] The statement says that A and A' are hyperbolic relative to H_A and H_B, respectively; the second collection should presumably be H'_A, not H_B.
  2. [§3.1, Lemma 3.1] The notation "H ∈ H" is overloaded, since H is used both for a subgroup and for the collection of subgroups. Writing H ∈ H_A ∪ H_B explicitly would improve readability.
  3. [§6, proof of Theorem 1.8] The sentence "Since G is compact, there exists a subsequence {a_{m_k}} that converges to η ∈ ∂_{rel}G" is not correct as written; G is not compact in general. The intended statement is that the compactification of X^h is compact, so a bounded sequence of vertices has a subsequence converging to a boundary point.
  4. [§5.2] The set δStab(X) should be δStab(X^h) for consistency with the amalgamated case.
  5. [§5.4, proof of Proposition 5.4] In condition (2), the symbol U is used both for an arbitrary finite open cover and for a neighborhood basis of a boundary point; this makes the argument hard to follow.

Circularity Check

0 steps flagged · score 2.0 of 10

No circular derivation: the main theorems are proved from external combination and boundary results; the two self-citations are contextual, and the main gap is an omitted topology verification, not a circular step.

full rationale

The paper's derivation chain is not circular. The combination theorems (Theorems 1.1, 1.2, 3.10) construct a relative Cayley-Abels graph X for the amalgam or HNN extension from the relative Cayley-Abels graphs of the vertex groups, then import hyperbolicity from Bestvina-Feighn [BF92] and fineness from the vertex fineness; no parameter is fitted from the target conclusion and the result is not assumed in the construction. The equivalence Theorem 4.12 is proved by citing Sisto [Sis12] and Dahmani [Dah03a], and Theorem 1.4 follows from fineness of the coned-off graph, not from the conclusion. In Section 5, the authors construct a candidate boundary δ(X^h) and invoke Martin-Swiatkowski [MS15] to identify it with ∂relG. The paper explicitly says 'We skip a verification that the above collections of sets satisfy the axioms for the basis of open neighborhoods,' and Proposition 5.1 is justified only by a 'word on the proof' saying continuity follows from [MS15, Section 3, pp.280-281]. This is an omitted proof and a genuine correctness risk, because [MS15] treats discrete free products rather than augmented Cayley-Abels graphs of TDLC amalgams; however, it is not a circular step, since [MS15] is external and the argument does not reduce ∂relG to its own claimed value. The dense-amalgam regularity proof (Proposition 5.4) certifies the topology constructed in Subsection 5.3, and uniqueness is imported from Swiatkowski [S16]. The self-citations [Tom25] and [Tom22] are contextual: [Tom25] is described as the discrete analogue being generalized and is not used as a lemma, and [Tom22] appears only in Remark 1.6 to illustrate failure when edge groups are non-compact. Neither self-citation is load-bearing for Theorems 1.5, 1.7, 1.8, or Proposition 1.9. Overall, the central claims are derived from independent published results, and the paper's own gaps are omissions rather than circular reductions.

Assumptions & free parameters 0 free parameters · 6 assumptions · 0 invented entities

The paper introduces no fitted parameters. The central claims rest on standard external theorems (Kron-Moller, Bestvina-Feighn, Sisto, Swiatkowski, de Cornulier accessibility) and on the unstated properness of H_A union H_B in the amalgam, which is listed as an ad hoc assumption. The augmented and coned-off Cayley-Abels graphs are definitions introduced by the paper, not unexplained postulated entities.

assumptions (6)
  • standard math Existence and quasiisometry invariance of Cayley-Abels graphs (KM08, Theorems 2.2 and 2.7+)
    Used throughout Sections 2-4 to define hyperbolic and relatively hyperbolic TDLC groups; taken as background from Kron-Moller.
  • standard math Bestvina-Feighn combination theorem for trees of uniformly hyperbolic spaces with bounded edge spaces
    Used in the proofs of Theorems 1.1 and 1.2 and in Section 5 to conclude the constructed graph Xh is hyperbolic.
  • standard math Equivalence of relative hyperbolicity notions for graphs (Sis12, Theorem 1.1)
    Used in Theorem 4.12 to equate TDRH-II and TDRH-III; the paper does not prove this equivalence itself.
  • standard math Dense amalgam characterization of compact metric spaces (S16, Theorem 0.2 and Proposition 0.1)
    Used in Proposition 5.4 and Theorems 1.5 and 1.7 to identify boundaries as dense amalgams and to conclude homeomorphism types.
  • standard math Accessibility of hyperbolic TDLC groups (de Cornulier, Corollary 19.46)
    Used in the proof of Proposition 1.9 to obtain a finite graph of groups splitting with compact edge groups.
  • ad hoc to paper The collection H_A union H_B remains a proper pair in the amalgam G (Definition 2.7)
    Assumed without proof in Section 3 so that the relative Cayley-Abels graph used in Theorems 1.1 and 1.2 exists; the paper does not establish malnormality of the parabolic collection after amalgamation.

how reviews work

0 comments
Cite this review

Pith. "Pith review of On combination theorems and Bowditch boundaries of relatively hyperbolic TDLC groups." pith.science (2026). https://pith.science/paper/3IGPTYIF

@misc{pith2026250813038,
  author       = {Pith},
  title        = {Pith review of: On combination theorems and Bowditch boundaries of relatively hyperbolic TDLC groups},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/3IGPTYIF}},
  note         = {Machine review of arXiv:2508.13038}
}
abstract

Based on the work of Farb, Bowditch, and Groves-Manning on discrete relatively hyperbolic groups, we introduce an approach to relative hyperbolicity for totally disconnected locally compact (TDLC) groups. For compactly generated TDLC groups, we prove that this notion is equivalent to the one introduced by Arora-Pedroza. Let $G=A\ast_C B$ or $G=A\ast_C$ where $A$ and $B$ are relatively hyperbolic TDLC groups and $C$ is compact. We prove that $G$ is a relatively hyperbolic TDLC group and give a construction of the Bowditch boundary of $G$. As a consequence, we prove that if the rough ends of $G$ are infinite, then the topology of the Bowditch boundary of $G$ is uniquely determined by the topology of the Bowditch boundary of $A$ and $B$. Further, we show that if a relatively hyperbolic TDLC group has one rough end, then its Bowditch boundary is connected. Finally, we show that if the Gromov boundary of a hyperbolic TDLC group $G$ is totally disconnected, then $G$ splits as a finite graph of compact groups.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

35 extracted references · 31 canonical work pages

  1. [1]

    Herbert Abels, Specker- K ompaktifizierungen von lokal kompakten topologischen G ruppen , Math. Z. 135 (1973/74), 325--361

  2. [2]

    Algebra 631 (2023), 830--876

    Shivam Arora and Eduardo Mart\' nez-Pedroza, Topological groups with a compact open subgroup, relative hyperbolicity and coherence, J. Algebra 631 (2023), 830--876

  3. [3]

    Bestvina and M

    M. Bestvina and M. Feighn, A C ombination theorem for N egatively C urved G roups , J. Diff. Geom., vol 35 (1992), 85--101

  4. [4]

    Bridson and A Haefliger, Metric spaces of nonpositive curvature, Grundlehren der mathematischen Wissenchaften, Vol 319, Springer-Verlag (1999)

    M. Bridson and A Haefliger, Metric spaces of nonpositive curvature, Grundlehren der mathematischen Wissenchaften, Vol 319, Springer-Verlag (1999)

  5. [5]

    Hadi Bigdely and Eduardo Mart\'inez-Pedroza, Relative D ehn functions, hyperbolically embedded subgroups and combination theorems , Glasg. Math. J. 66 (2024), no. 1, 1--23

  6. [6]

    ognvaldur G. M\

    Udo Baumgartner, R\"ognvaldur G. M\"oller, and George A. Willis, Hyperbolic groups have flat-rank at most 1, Israel J. Math. 190 (2012), 365--388

  7. [7]

    B. H. Bowditch, Relatively hyperbolic groups, Internat. J. Algebra and Computation. 22, 1250016, 66pp (2012)

  8. [8]

    25, European Mathematical Society (EMS), Z\"urich, 2016, Winner of the 2016 EMS Monograph Award

    Yves Cornulier and Pierre de la Harpe, Metric geometry of locally compact groups, EMS Tracts in Mathematics, vol. 25, European Mathematical Society (EMS), Z\"urich, 2016, Winner of the 2016 EMS Monograph Award

Show all 35 references
  1. [9]

    Yves de Cornulier, On the quasi-isometric classification of locally compact groups, London Math. Soc. Lecture Note Ser. 447, 275--342

  2. [10]

    Fran cois Dahmani, Les groupes relativement hyperboliques et leurs bords, Pr\'epublication de l'Institut de Recherche Math\'ematique Avanc\'ee [Prepublication of the Institute of Advanced Mathematical Research], vol. 2003/13, Universit\'e Louis Pasteur, D\'epartement de Math\'...

  3. [11]

    7 (2003), 933--963

    Francois Dahmani, Combination of convergence groups, Geometry and Topology, vol. 7 (2003), 933--963

  4. [12]

    5, 959--1058, With an appendix by Denis Osin and Mark Sapir

    Cornelia Dru t u and Mark Sapir, Tree-graded spaces and asymptotic cones of groups, Topology 44 (2005), no. 5, 959--1058, With an appendix by Denis Osin and Mark Sapir

  5. [13]

    Farb, Relatively hyperbolic groups, Geom

    B. Farb, Relatively hyperbolic groups, Geom. Funct. Anal. 8 (1998), 810--840

  6. [14]

    Ghys and P

    E. Ghys and P. de la Harpe(eds.), Sur les groupes hyperboliques d'apres M ikhael G romov , Progress in Math. vol 83, Birkhauser, Boston Ma. (1990)

  7. [15]

    Daniel Groves and Jason Fox Manning, Dehn filling in relatively hyperbolic groups, Israel J. Math. 168 (2008), 317--429

  8. [16]

    Gromov, Hyperbolic G roups , in Essays in Group Theory, ed

    M. Gromov, Hyperbolic G roups , in Essays in Group Theory, ed. Gersten, MSRI Publ.,vol.8, Springer Verlag (1985), 75--263

  9. [17]

    Ursula Hamenst\" a dt, Hyperbolic relatively hyperbolic graphs and disk graphs, Groups Geom. Dyn. 10 (2016), no. 1, 365--405

  10. [18]

    Christopher Hruska, Cusped spaces and quasi-isometries of relatively hyperbolic groups, arxiv 2010.09876 (2021)

    Brendan Burns Healy and G. Christopher Hruska, Cusped spaces and quasi-isometries of relatively hyperbolic groups, arxiv 2010.09876 (2021)

  11. [19]

    Ilya Kapovich, The combination theorem and quasiconvexity, Internat. J. Algebra Comput. 11 (2001), no. 2, 185--216

  12. [20]

    o n and R\

    Bernhard Kr\" o n and R\" o gnvaldur G. M\" o ller, Analogues of C ayley graphs for topological groups , Math. Z. 258 (2008), no. 3, 637--675. 2369049

  13. [21]

    282, American Mathematical Society, Providence, RI, [2024] 2024

    Michael Kapovich and Pranab Sardar, Trees of hyperbolic spaces, Mathematical Surveys and Monographs, vol. 282, American Mathematical Society, Providence, RI, [2024] 2024

  14. [22]

    Martin, Non-positively curved complexes of groups and boundries, Geom

    A. Martin, Non-positively curved complexes of groups and boundries, Geom. Topol. 18(1) (2014), 31--102

  15. [23]

    Mj and Lawrence Reeves, A Combination Theorem for Strong Relative Hyperbolicity , Geometry and Topology 12 (2008), 1777--1798

    M. Mj and Lawrence Reeves, A Combination Theorem for Strong Relative Hyperbolicity , Geometry and Topology 12 (2008), 1777--1798

  16. [24]

    Mj and P

    M. Mj and P. Sardar, A Combination Theorem for metric bundles , Geom. Funct. Anal. 22, no. 6 (2012), 1636--1707

  17. [25]

    Group Theory 18 (2015), no

    Alexandre Martin and Jacek \' S wi a tkowski, Infinitely-ended hyperbolic groups with homeomorphic G romov boundaries , J. Group Theory 18 (2015), no. 2, 273--289

  18. [26]

    Mackay and Alessandro Sisto, Maps between relatively hyperbolic spaces and between their boundaries, Trans

    John M. Mackay and Alessandro Sisto, Maps between relatively hyperbolic spaces and between their boundaries, Trans. Amer. Math. Soc. 377 (2024), no. 2, 1409--1454

  19. [27]

    Osin, Relatively hyperbolic groups: intrinsic geometry, algebraic properties, and algorithmic problems, Mem

    Denis V. Osin, Relatively hyperbolic groups: intrinsic geometry, algebraic properties, and algorithmic problems, Mem. Amer. Math. Soc. 179 (2006), no. 843, vi+100

  20. [28]

    Abhijit Pal, Cannon-Thurston maps for Trees of (Strong) Relatively Hyperbolic Metric Spaces , Ph. D. Thesis, Indian Statistical Institute, Calcutta (in preparation)

  21. [29]

    Jacek \' S wi a tkowski, The dense amalgam of metric compacta and topological characterization of boundaries of free products of groups, Groups Geom. Dyn. 10 (2016), no. 1, 407--471

  22. [30]

    J.-P. Serre, Trees, Springer Monographs in Mathematics, Springer-Verlag, Berlin, 2003, Translated from the French original by John Stillwell, Corrected 2nd printing of the 1980 English translation. 1954121

  23. [31]

    Alessandro Sisto, On metric relative hyperbolicity, arXiv:1210.8081 (2012)

  24. [32]

    Indian Acad

    Ravi Tomar, Boundaries of graphs of relatively hyperbolic groups with cyclic edge groups, Proc. Indian Acad. Sci. Math. Sci. 132 (2022), no. 2, Paper No. 47, 26

  25. [33]

    Group Theory 28 (2025), no

    , Homeomorphism types of the B owditch boundaries of infinite-ended relatively hyperbolic groups , J. Group Theory 28 (2025), no. 1, 51--65

  26. [34]

    Willis, The structure of totally disconnected, locally compact groups, Math

    G. Willis, The structure of totally disconnected, locally compact groups, Math. Ann. 300 (1994), no. 2, 341--363

  27. [35]

    Group Theory 26 (2023), no

    Stefanie Zbinden, Morse boundaries of graphs of groups with finite edge groups, J. Group Theory 26 (2023), no. 5, 969--1002

Pith tools

Reviewed August 15, 2026 · model on record in the stance chip above.