Conjugacy in fibre products, distortion, and the geometry of cyclic subgroups

Martin R Bridson

REVIEW 2 major objections 5 minor 1 cited by

Conjugacy in fibre products, distortion, and the geometry of cyclic subgroups

T0 review · 2 major / 5 minor · reviewed 2026-08-06 · deepseek-v4-flash

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

Pith's one-line read The conjugator length of a fibre product of a torsion-free hyperbolic group is governed, up to the standard equivalence, by the rel-cyclics Dehn function of the quotient.

desk verdict Solid quantitative machinery, but Theorem B's torsion-free clause is false as stated—Q=Z breaks it; fixable with a non-virtually-cyclic hypothesis. read the letter →

arxiv 2507.17598 v2 pith:22ZNDWIT submitted 2025-07-23 math.GR

Martin R Bridson This is my paper · ORCID

classification math.GR MSC 20F6720F1020F65

keywords fibreproductsconjugatorlengthrel-cyclicsDehnfunctioncyclicsubgroupshyperbolicgroupsdistortionconjugacyproblemuniformlymonotonecyclics

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 gives quantitative answers to the question: how hard is it to conjugate elements in a fibre product $P

What carries the argument

The load-bearing tool is the rel-cyclics Dehn function $\delta_Q^c(n)$, which records the maximum, over words $w,u$ with $|w|+|u|\le n$ and $w=u^{-p}$ with $p$ minimal, of the area of $wu^p$ plus $|p|n$; it simultaneously controls the ordinary Dehn function and the distortion of cyclic subgroups of $Q$. The second ingredient is a new geometric fact, Proposition 4.7: torsion-free hyperbolic groups have uniformly monotone cyclics, meaning there is a constant $k$ such that $|g^i|\le k|g^p|$ for every non-trivial $g$ and all $0<i<p$. The proof uses uniqueness of roots in torsion-free hyperbolic groups and a thin-quadrilateral argument in the Cayley graph to control the length of the path from the identity to a geodesic axis. These two pieces, combined with the fibre product description $P=(N\times 1)\rtimes\Delta$ (with $\Delta$ the diagonal copy of $G$), carry the chain of inequalities in Theorem B.

What would settle it

A concrete counterexample would be a torsion-free hyperbolic group containing elements $g$ and integers $0<i<p$ with $|g^i|$ much larger than $|g^p|$: specifically, a sequence $(g_n,i_n,p_n)$ with $|g_n^{i_n}|/|g_n^{p_n}|\to\infty$. Since Proposition 4.7 rules this out, finding such a sequence in any torsion-free hyperbolic group, for instance by inspecting the Cayley graph of a surface group, would invalidate the upper-bound half of Theorem B.

Watch

Extended reading notes

Core claim

The central discovery is Theorem B: for a fibre product $P$ of a torsion-free hyperbolic group $G$ over a finitely presented quotient $Q$, the Dehn function of $Q$ is a lower bound for the conjugator length of $P$, the conjugator length of $P$ is bounded above by the rel-cyclics Dehn function $\delta_Q^c$, and under the geometric hypothesis that $Q$ has uniformly quasigeodesic cyclic subgroups, $\delta_Q^c$ is itself bounded above by a constant multiple of the relative conjugator length $CL_P^{G\times G}(n)$ when that function is at least quadratic. The paper also proves that the distortion of $P$ in $G\times G$ is exactly the Dehn function $\delta_Q$ (Theorem E) and provides an algorithmic construction (Proposition D) replacing any finite presentation of $Q$ by one for a group $Q^\dagger$ with uniformly quasigeodesic cyclics whose Dehn function lies between $n\delta_Q(n)$ and $n(\delta_Q(n))^2$. Together with the standard small-cancellation construction that produces hyperbolic preimages of arbitrary finitely presented quotients, and the finite-presentation criterion for fibre products, these results yield finitely presented groups whose conjugator length functions lie between $n\delta(n)$ and $n^2(\delta(n^2))^2$ for every standard Dehn function $\delta$ (Theorem F).

Load-bearing premise

Everything rests on the new claim that in a torsion-free hyperbolic group the length of $g^i$ is always bounded by a fixed multiple of the length of $g^p$ for $0<i<p$; if this uniform monotonicity of cyclic subgroups failed, the upper bound tying conjugator length to the rel-cyclics Dehn function would not go through.

Editorial extensions

If this is right

If Theorem B is correct, then for a fibre product $P$ over a quotient $Q$ with uniformly quasigeodesic cyclics, the relative conjugator length $CL_P^{G\times G}(n)$ determines the rel-cyclics Dehn function $\delta_Q^c(n)$ up to the usual equivalence, so the complexity of the conjugacy problem in $P$ is read directly off the cyclic geometry of $Q$.
The distortion theorem (Theorem E) makes the fibre product a sharp device for transferring the Dehn function of $Q$ into the word-metric geometry of $P$: any finitely presented quotient's filling function appears as the distortion of an explicit subgroup of a hyperbolic group.
Corollary C places $CL_P(n)$ between $\delta_Q(n)$ and $\delta_Q(n^2)$ whenever $Q$ has uniformly quasigeodesic cyclics, so any super-polynomial Dehn function of $Q$ forces a correspondingly fast-growing conjugator length in $P$.
The dagger construction (Proposition D) gives an algorithmic way to impose the geometric hypothesis 'uniformly quasigeodesic cyclics' on any given quotient without losing finite presentability, so the hypotheses of the main theorem are realised for every finitely presented $Q$.
Theorem F yields finitely presented (indeed residually finite) groups with conjugator length functions of the order $n\delta(n)$ for any standard Dehn function $\delta$, within a factor $n$ and a quadratic shift of the argument, so the known range of conjugator length functions is enlarged.

Reading between the lines

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

One could test whether the rel-cyclics Dehn function plays the same governing role in other families of groups with unique roots, such as torsion-free CAT(0) groups; if uniform monotonicity fails there, the proof would break precisely at the step where hyperbolicity is used, isolating the geometric content needed.
The $n$ versus $n^2$ gap between the bounds in Corollary C appears to be the only obstruction to closing the gap in Theorem F; Remark 8.1 suggests that uniformly monotone cyclics would give sharper area estimates, so proving that property for wider classes would immediately tighten the engineerable conjugator lengths.
The estimates introduced here mark $\delta_Q^c$ as a natural complexity measure for cyclic-subgroup distortion; one could define analogous rel-cyclic isoperimetric functions for semigroups or modules and expect similar bridge theorems to invariants of those structures.
A plausible strengthening of Theorem F would replace the quadratic shift by a linear one by proving the improved bound $\delta_Q^c(n)\preceq\delta_Q(n)$ for quotients with uniformly monotone cyclics; the paper leaves this gap open, and closing it would settle the exact range of feasible conjugator length growth rates.

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

Martin's new paper is worth engaging with, but Theorem B as stated overreaches. The stress-test is right: take G=Q=Z with the identity map. The fibre product is the diagonal copy of Z, conjugacy is trivial, so CL^{G×G}_P(n)=0. But δ^c_Z(n)≃n^2 (take u=x, w=x^{-(n-1)}; the area is 0 but the |pn| term gives ~n^2). The asserted δ^c_Q ⪯ n CL^{G×G}_P then reads n^2 ⪯ 0. That kills the torsion-free clause of Theorem B as stated.

Where exactly does it break? Proposition 6.2's second assertion, m_Q ⪯ CL^{G×G}_P, is false for virtually cyclic Q. The proof actually obtains m ≤ c_2(CL(6n)+n), which only yields m ⪯ CL+n. Corollary 6.3 then drops the n, and the torsion-free case of Theorem B inherits the error. The fix is straightforward: add the hypothesis that Q is not virtually cyclic, or impose CL^{G×G}_P(n) ⪰ n. The intended Rips-construction applications satisfy that, so the main results survive.

That said, there is real value here. Proposition 4.7—uniformly monotone cyclics for torsion-free hyperbolic groups—is a genuinely new structural fact with a delicate, careful proof. Theorem E gives a clean distortion calculation. The upper bound CL^{G×G}_P ⪯ δ^c_Q is proved in full detail. Proposition D and Theorem F deliver the promised tools for realising conjugator length functions, and the 1-2-3 theorem hypothesis is handled properly. The paper is mostly careful; the mistake is an overstatement in the final statements, not a broken central mechanism.

Minor: the introduction claims the constructed P is virtually special without proof in the body, and Section 7's use of Scholium 3.4 is slightly non-literal though repairable with an O(n) term.

Recommendation: send it to a serious referee. The core is sound and publishable after the theorem statements and Corollary 6.3 are corrected and the extra hypothesis is stated. I would take the corrected version seriously and cite it. As it stands, the false torsion-free clause should not stand.

Referee Report

2 major / 5 minor

Summary. The paper studies the conjugator length function of fibre products P < G × G associated to epimorphisms G → Q with G torsion-free hyperbolic, and relates it to the Dehn function and rel-cyclics Dehn function of Q, as well as to the distortion of P in G × G. The main result, Theorem B, claims inequalities δ_Q ⪯ CL_P ⪯ CL_P^{G×G} ⪯ δ_Q^c, plus a torsion-free lower bound δ_Q^c ⪯ n CL_P^{G×G} and a sharpened equivalence under a uniform-quasigeodesic-cyclics assumption. The paper also proves a distortion theorem (Theorem E), a monotonicity result for cyclic subgroups of torsion-free hyperbolic groups (Proposition 4.7), a 'dagger' construction forcing uniformly quasigeodesic cyclics (Proposition D), and an application to realising large conjugator length functions (Theorem F). The proofs combine Rips constructions, the 1-2-3 Theorem, hyperbolic geometry, and estimates from van Kampen diagram areas.

Significance. If the stated inequalities held in the intended generality, the paper would provide a quantitative companion to Theorem A and a useful tool for constructing finitely presented groups with prescribed conjugator length functions. The manuscript contains several genuinely interesting components: the distortion calculation Theorem E, the new monotonicity Proposition 4.7, and the dagger construction of Proposition D. However, a central statement is false as written: the torsion-free part of Theorem B fails for the elementary example G = Q = Z. Because this example is not an isolated pathology but exposes a dropped term in the proof of Proposition 6.2, the significance of the paper is conditional on a corrected statement and proof. The intended non-elementary applications are likely salvageable, but the current formulation overclaims.

major comments (2)

[Theorem B (Section 8), Proposition 6.2, Corollary 6.3] The torsion-free part of Theorem B is false as stated. Take G = Q = Z with the identity epimorphism and the standard presentation <x | >. The kernel is trivial, so the fibre product is the diagonal subgroup generated by (x,x). Since P is abelian, two elements are conjugate exactly when they are equal, and the identity conjugator has length 0; hence CL_P(n) = CL_P^{G×G}(n) = 0 for all n. But δ^c_Z(n) ≃ n²: for u = x and w = x^{-p} with p = n-1, we have |w| + |u| = n and w = u^{-p} in Q, with Area(wu^p) = 0 and contribution |p|n = n(n-1). Thus the asserted inequality δ^c_Q(n) ⪯ n CL_P^{G×G}(n) would give n² ⪯ 0. The source of the error is Proposition 6.2: the proof derives m ≤ c₂(CL(6n) + n), but the proposition states m_Q(n) ⪯ CL_P^{G×G}(n), dropping the additive +n. That drop is not justified when CL is sublinear, and exactly the same failure propagates to Corollary 6.3 and to the torsion-free clause of Theorem B. The theorem needs a hypothesis such as CL_P^{G×G}(n) ⪰ n, or an explicit exclusion of virtually cyclic quotients, and the proof of Proposition 6.2 must retain the additive term unless such a hypothesis is added.
[Proposition 7.1 and Scholium 3.4] The application of Scholium 3.4 to ζ'' = (y_1^{p''}γ, y_2^{r_2}) is not literally valid as written: Scholium 3.4 assumes |(γ_1,γ_2)|_{G×G} ≤ n, and the proof has not yet bounded |y_1^{p''}|, so |ζ''|_{G×G} may exceed the n appearing in (7.4). The repair is straightforward: the underlying estimate from the proof of Theorem E gives |(γ_1,γ_2)|_P ≤ (L+1)Area_Q(w) + |w| + |γ_2|_G, and here |γ_2|_G = |y_2^{r_2}|_G ≤ kn. Using this form changes only constants. Also, in the first paragraph of the same proof, the claim |(γ,γ)|_P = (1/2)|γ|_G is incorrect; the earlier argument in Theorem E establishes |(γ,γ)|_P = |γ|_G. These are repairable technical slips rather than obstacles to the intended result.

minor comments (5)

[Proof of Theorem B, Section 8] The sentence 'the inequality CL_P^{G×G}(n) ≃ δ^c_Q(n) was established in Proposition 7.1' should say '⪯' rather than '≃', since Proposition 7.1 only proves the upper bound; the reverse equivalence is obtained later only under additional hypotheses.
[Theorem F and Section 9.2] The assertion that the constructed group is 'virtually special' is stated without proof or a concrete reference to an argument establishing it. If this property is part of the theorem's claim, it should be proved; if it is an aside, it should be labelled as such.
[Notation, Sections 4.1 and 4.2] The abbreviation for 'uniformly quasigeodesic cyclics' alternates between UQC and UQG (Definition 4.1, Lemma 4.2, Lemma 4.4); please standardise.
[Proposition 6.2, proof] There is a typographical error in the line '|U|_{G×G}+|V|)G×G' where a parenthesis is misplaced; it should read |U|_{G×G}+|V|_{G×G}.
[Definition 2.2] The remark that the condition |p| ≤ o(u)/2 is equivalent to |p| being the least non-negative integer with w = u^{±p} deserves a short explanation, especially in the infinite-order case, where the condition is vacuous but uniqueness of p is used later.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the paper's quantitative bounds are derived from explicit estimates and standard external results, not from its own conclusions.

full rationale

The derivation chain in Theorems B, C, E, and F is self-contained in the relevant sense. The rel-cyclics Dehn function and the return-of-cyclics function are defined independently of conjugator length; they are functions of the quotient Q and are not constructed so as to encode the target inequality. The upper bound CL_P^{G×G} ⪯ δ^c_Q (Proposition 7.1) is proved by taking an actual conjugator in P, using the cyclic structure of centralizers in the torsion-free hyperbolic group G, and bounding its P-length by an area in Q; it does not assume Theorem B. The lower bounds in Propositions 5.1 and 6.2 and Corollary 6.3 construct explicit conjugate pairs from elements of large area or large return exponent and bound the minimal conjugator from below; they do not fit a parameter to the conclusion. The proof of Proposition 4.7 is new and uses only standard hyperbolic geometry and uniqueness of roots, with no appeal to the theorem it supports. Self-citations appear as context or as published external theorems (e.g. Theorem A from [9] and Theorem 9.1 from [10]); the quantitative arguments do not reduce to these citations. Even if a correctness objection about the virtually cyclic case (G=Q=Z) were sustained, that would be a mathematical error in the statement or proof, not a circularity of the kind defined here. Accordingly, the circularity score is 0.

Assumptions & free parameters 0 free parameters · 7 assumptions · 0 invented entities

No free parameters are fitted. The new invariants δ^c, δ^z, δ^o, m, t are definitions, not parameters. The constants λ, k, C, L are existential uniform constants from the hypotheses. The paper depends on standard external theorems, all clearly cited.

assumptions (7)

standard math Torsion-free hyperbolic groups have uniformly quasigeodesic cyclics.
Used in Sections 5-7 to bound exponents; follows from [10], pp. 464-466.
standard math Hyperbolic groups satisfy a linear isoperimetric inequality.
Used in Theorem E proof and Proposition 6.2; standard property of hyperbolic groups.
standard math Centralizers of non-trivial elements in torsion-free hyperbolic groups are infinite cyclic.
Used in Props 5.1, 6.2, 7.1 to describe conjugating elements.
standard math Roots are unique in torsion-free hyperbolic groups.
Used in Proposition 4.7 to compare conjugators of powers; flagged with footnote 1.
standard math The Rips construction associates to any finite presentation Q a torsion-free hyperbolic G and an epimorphism with 2-generated kernel.
Used in Section 9 to build Q† and P; standard algorithm from [20].
standard math The 1-2-3 Theorem: if the kernel is finitely generated and Q is type F3, the fibre product is finitely presented.
Used to ensure P† is finitely presented; from [2].
standard math Theorem III.Γ.6.20 of [10] bounding the Dehn function of a trivial HNN extension in terms of the Dehn function and distortion of the base.
Used in Proposition D to estimate δ_{Q†}; external result.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Conjugacy in fibre products, distortion, and the geometry of cyclic subgroups." pith.science (2026). https://pith.science/paper/22ZNDWIT

@misc{pith2026250717598,
  author       = {Pith},
  title        = {Pith review of: Conjugacy in fibre products, distortion, and the geometry of cyclic subgroups},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/22ZNDWIT}},
  note         = {Machine review of arXiv:2507.17598}
}

abstract

We investigate the complexity of the conjugacy problem for fibre products in torsion-free hyperbolic groups. Let $G$ be a torsion-free hyperbolic group and let $P<G\times G$ be the fibre product associated to an epimorphism $G\twoheadrightarrow Q$. We establish inequalities that relate the conjugator length function of $P$ to the geometry of cyclic subgroups in $Q$, the Dehn function of $Q$, the {\em rel-cyclics Dehn function} of $Q$, and the distortion of $P$ in $G\times G$. These estimates provide tools for extending the library of (large) functions that are known to arise as the conjugator length functions of finitely generated and finitely presented groups.

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

Conjugator length in finitely presented groups
math.GR 2026-07 accept novelty 6.0 of 10

The conjugator length function is quadratic for the integral Heisenberg group and Stallings' group, and cyclic-subgroup distortion can be promoted to conjugator length.

Reference graph

Works this paper leans on

20 extracted references · 19 canonical work pages · cited by 1 Pith paper

[1]

Alibegovic,Translation Lengths inOut(F n), Geom

E. Alibegovic,Translation Lengths inOut(F n), Geom. Dedicata92(2002), 87–93

work page 2002
[2]

Baumslag, M.R

G. Baumslag, M.R. Bridson, C.F. Miller III, H. Short,Fibre products, non-positive curvature, and decision problems, Comment. Math. Helv.75(2000), 457–477

work page 2000
[3]

Birget, A

J-C. Birget, A. Yu. Ol’shanskii, E. Rips, M.V . Sapir,Isoperimetric and isodiametric functions of groups, Ann. of Math.156(2002), 345–466

work page 2002
[4]

Brady and M.R

N. Brady and M.R. Bridson,There is only one gap in the isoperimetric spectrum, Geom. Funct. Anal. (GAFA),10(2000), 1053-1070

work page 2000
[5]

The geometry of the word problem for finitely generated groups

N. Brady and T.R. Riley and H. Short, “The geometry of the word problem for finitely generated groups”, Adv. Courses Math., CRM Barcelona, Birkh¨auser–Verlag, Basel 2007

work page 2007
[6]

Invitations to geometry and topology

M.R. Bridson,The geometry of the word problem, in “Invitations to geometry and topology”, (M.R. Bridson and S.M. Salamon, eds.), pp. 29–91, OUP, Oxford 2002. 20 MARTIN R. BRIDSON

work page 2002
[7]

Bridson,Non-positive curvature and complexity for finitely presented groups, Intern

M.R. Bridson,Non-positive curvature and complexity for finitely presented groups, Intern. Congr. Math. (Madrid 2006), V ol. II, pp. 961–987, Eur. Math. Soc. Publ., Z¨urich 2006

work page 2006
[8]

Bridson,On the subgroups of right-angled Artin groups and mapping class groups, Math

M.R. Bridson,On the subgroups of right-angled Artin groups and mapping class groups, Math. Res. Lett.,20(2013), 203–212

work page 2013

Show all 20 references

[9]

Bridson,On the conjugacy problem for subdirect products of hyperbolic groups, preprint, Stan- ford, April 2025

M.R. Bridson,On the conjugacy problem for subdirect products of hyperbolic groups, preprint, Stan- ford, April 2025

2025
[10]

Metric Spaces of Non-Positive Curvature

M.R. Bridson and A. Haefliger, “Metric Spaces of Non-Positive Curvature”, Grundl. Math. Wiss., V ol. 319, Springer-Verlag, Heidelberg-Berlin, 1999

1999
[11]

Bridson, J

M.R. Bridson, J. Howie, C.F. Miller III, H. Short,On the finite presentation of subdirect products and the nature of residually free groups, (with J. Howie, C.F. Miller III, and H. Short), American J. Math., 135(2013), 891–933

2013
[12]

Bridson and C.F

M.R. Bridson and C.F. Miller III,Structure and finiteness properties of subdirect products of groups. Proc. London Math. Soc. (3)98(2009), 631–651

2009
[13]

Bridson and T.R

M.R. Bridson and T.R. Riley,Groups with fast-growing conjugator length functions, in preparation
[14]

Bridson, T.R

M.R. Bridson, T.R. Riley and A.W. Sale,Conjugator length in finitely presented groups, in preparation
[15]

Dison,Isoperimetric functions for subdirect products and Bestvina-Brady groups, PhD thesis, Im- perial College London, 2008

W. Dison,Isoperimetric functions for subdirect products and Bestvina-Brady groups, PhD thesis, Im- perial College London, 2008. arXiv:0810.4060

2008 arXiv
[16]

Essays in Group Theory

M. Gromov,Hyperbolic groups, in “Essays in Group Theory”, pp. 75–263, Math. Sci. Res. Inst. Publ. vol. 8, Springer, New York 1987

1987
[17]

Haglund and D.T

F. Haglund and D.T. Wise,Special cube complexes, Geom. Funct. Anal.17(2008), 1551–1620

2008
[18]

Mihailova,The occurrence problem for direct products of groups, Dokl

K.A. Mihailova,The occurrence problem for direct products of groups, Dokl. Akad. Nauk SSSR119 (1958) 1103–1105

1958
[19]

On group-theoretic decision problems and their classification

C. F. Miller III, “On group-theoretic decision problems and their classification”, Ann. Math. Studies, No. 68, Princeton University Press (1971)

1971
[20]

Rips,Subgroups of small cancellation groups, Bull

E. Rips,Subgroups of small cancellation groups, Bull. London Math. Soc.14(1982), 45–47. MathematicalInstitute, AndrewWilesBuilding, WoodstockRoad, Oxford, OX2 6GG Email address:bridson@maths.ox.ac.uk

1982

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.

[1] [1]

Alibegovic,Translation Lengths inOut(F n), Geom

E. Alibegovic,Translation Lengths inOut(F n), Geom. Dedicata92(2002), 87–93

work page 2002

[2] [2]

Baumslag, M.R

G. Baumslag, M.R. Bridson, C.F. Miller III, H. Short,Fibre products, non-positive curvature, and decision problems, Comment. Math. Helv.75(2000), 457–477

work page 2000

[3] [3]

Birget, A

J-C. Birget, A. Yu. Ol’shanskii, E. Rips, M.V . Sapir,Isoperimetric and isodiametric functions of groups, Ann. of Math.156(2002), 345–466

work page 2002

[4] [4]

Brady and M.R

N. Brady and M.R. Bridson,There is only one gap in the isoperimetric spectrum, Geom. Funct. Anal. (GAFA),10(2000), 1053-1070

work page 2000

[5] [5]

The geometry of the word problem for finitely generated groups

N. Brady and T.R. Riley and H. Short, “The geometry of the word problem for finitely generated groups”, Adv. Courses Math., CRM Barcelona, Birkh¨auser–Verlag, Basel 2007

work page 2007

[6] [6]

Invitations to geometry and topology

M.R. Bridson,The geometry of the word problem, in “Invitations to geometry and topology”, (M.R. Bridson and S.M. Salamon, eds.), pp. 29–91, OUP, Oxford 2002. 20 MARTIN R. BRIDSON

work page 2002

[7] [7]

Bridson,Non-positive curvature and complexity for finitely presented groups, Intern

M.R. Bridson,Non-positive curvature and complexity for finitely presented groups, Intern. Congr. Math. (Madrid 2006), V ol. II, pp. 961–987, Eur. Math. Soc. Publ., Z¨urich 2006

work page 2006

[8] [8]

Bridson,On the subgroups of right-angled Artin groups and mapping class groups, Math

M.R. Bridson,On the subgroups of right-angled Artin groups and mapping class groups, Math. Res. Lett.,20(2013), 203–212

work page 2013