Jump to content

Ping-pong lemma: Difference between revisions

From Wikipedia, the free encyclopedia
Content deleted Content added
No edit summary
Citation bot (talk | contribs)
Misc citation tidying. | Use this bot. Report bugs. | Suggested by AManWithNoPlan | #UCB_CommandLine
 
(45 intermediate revisions by 35 users not shown)
Line 1: Line 1:
In [[mathematics]], the '''ping-pong lemma''', or '''table-tennis lemma''', is any of several mathematical statements that ensure that several elements in a group [[group action|acting]] on a set freely [[Generating set of a group|generates]] a [[free group|free]] [[subgroup]] of that group.
In [[mathematics]], the '''ping-pong lemma''', or '''table-tennis lemma''', is any of several mathematical statements that ensure that several elements in a [[group (mathematics)|group]] [[group action|acting]] on a set freely [[Generating set of a group|generates]] a [[free group|free]] [[subgroup]] of that group.


==History==
==History==


The ping-pong argument goes back to late 19th century and is commonly attributed<ref name="DH"/> to [[Felix Klein]] who used it to study subgroups of [[Kleinian group]]s, that is, of discrete groups of isometries of the [[hyperbolic 3-space]] or, equivalently [[Möbius transformation]]s of the [[Riemann sphere]]. The ping-pong lemma was a key tool used by [[Jacques Tits]] in his 1972 paper<ref name="T">J. Tits. [http://www.sciencedirect.com/science?_ob=ArticleURL&_udi=B6WH2-4D7K6RV-19X&_user=10&_coverDate=02%2F29%2F1972&_rdoc=3&_fmt=high&_orig=browse&_srch=doc-info(%23toc%236838%231972%23999799997%23518342%23FLP%23display%23Volume)&_cdi=6838&_sort=d&_docanchor=&_ct=12&_acct=C000050221&_version=1&_urlVersion=0&_userid=10&md5=bb8d98f320404accc3525ca235cf821e ''Free subgroups in linear groups.''] [[Journal of Algebra]], vol. 20 (1972), pp. 250–270 </ref> containing the proof of a famous result now known as the [[Tits alternative]]. The result states that a [[finitely generated group|finitely generated]] [[linear group]] is either [[virtually]] [[solvable group|solvable]] or contains a [[free group|free]] [[subgroup]] of rank two. The ping-pong lemma and its variations are widely used in [[geometric topology]] and [[geometric group theory]].
The ping-pong argument goes back to the late 19th century and is commonly attributed<ref name="DH"/> to [[Felix Klein]] who used it to study subgroups of [[Kleinian group]]s, that is, of discrete groups of [[isometry|isometries]] of the [[hyperbolic 3-space]] or, equivalently [[Möbius transformation]]s of the [[Riemann sphere]]. The ping-pong lemma was a key tool used by [[Jacques Tits]] in his 1972 paper<ref name="T">J. Tits. [https://www.sciencedirect.com/science/article/pii/0021869372900580 ''Free subgroups in linear groups.''] [[Journal of Algebra]], vol. 20 (1972), pp. 250–270</ref> containing the [[mathematical proof|proof]] of a famous result now known as the [[Tits alternative]]. The result states that a [[finitely generated group|finitely generated]] [[linear group]] is either [[virtually]] [[solvable group|solvable]] or contains a free subgroup of [[free group|rank]] two. The ping-pong lemma and its variations are widely used in [[geometric topology]] and [[geometric group theory]].


Modern versions of the ping-pong lemma can be found in many books such as Lyndon&Schupp<ref name="LS">[[Roger Lyndon|Roger C. Lyndon]] and Paul E. Schupp. ''Combinatorial Group Theory.'' Springer-Verlag, New York, 2001. "Classics in Mathematics" series, reprint of the 1977 edition. ISBN 978-3-540-41158-1; Ch II, Section 12, pp. 167&ndash;169</ref>, de la Harpe<ref name="DH">Pierre de la Harpe. [http://books.google.com/books?id=cRT01C5ADroC&pg=PA25&dq=ping+pong+lemma+group+theory&sig=_1EZ9oSfAdljZFH1g7uvFiHuI-w#PPA25,M1 ''Topics in geometric group theory.''] Chicago Lectures in Mathematics. University of Chicago Press, Chicago. ISBN 0-226-31719-6; Ch. II.B "The table-Tennis Lemma (Klein's criterion) and examples of free products"; pp. 25&ndash;41.</ref>, Bridson&Haefliger<ref name="BH">Martin R. Bridson, and André Haefliger. [http://books.google.com/books?id=3DjaqB08AwAC&printsec=frontcover&dq=Martin+R.+Bridson,+and+Andr%C3%A9+Haefliger.+%22Metric+spaces+of+non-positive+curvature%22 ''Metric spaces of non-positive curvature.''] Grundlehren der Mathematischen Wissenschaften [Fundamental Principles of Mathematical Sciences], 319. Springer-Verlag, Berlin, 1999. ISBN 3-540-64324-9; Ch.III.&Gamma;, pp. 467&ndash;468</ref> and others.
Modern versions of the ping-pong lemma can be found in many books such as Lyndon & Schupp,<ref name="LS">[[Roger Lyndon|Roger C. Lyndon]] and Paul E. Schupp. ''Combinatorial Group Theory.'' Springer-Verlag, New York, 2001. "Classics in Mathematics" series, reprint of the 1977 edition. {{isbn|978-3-540-41158-1}}; Ch II, Section 12, pp. 167&ndash;169</ref> de la Harpe,<ref name="DH">Pierre de la Harpe. [https://books.google.com/books?id=cRT01C5ADroC&dq=ping+pong+lemma+group+theory&pg=PA25 ''Topics in geometric group theory.''] Chicago Lectures in Mathematics. University of Chicago Press, Chicago. {{isbn|0-226-31719-6}}; Ch. II.B "The table-Tennis Lemma (Klein's criterion) and examples of free products"; pp. 25&ndash;41.</ref> Bridson & Haefliger<ref name="BH">Martin R. Bridson, and André Haefliger. [https://books.google.com/books?id=3DjaqB08AwAC&q=Martin+R.+Bridson,+and+Andr%C3%A9+Haefliger.+%22Metric+spaces+of+non-positive+curvature%22 ''Metric spaces of non-positive curvature.''] Grundlehren der Mathematischen Wissenschaften [Fundamental Principles of Mathematical Sciences], 319. Springer-Verlag, Berlin, 1999. {{isbn|3-540-64324-9}}; Ch.III.&Gamma;, pp. 467&ndash;468</ref> and others.


==Formal statements==
==Formal statements==
Line 11: Line 11:
===Ping-pong lemma for several subgroups===
===Ping-pong lemma for several subgroups===


This version of the ping-pong lemma ensures that several [[subgroup]]s of a group acting on a set generate a [[free product]]. The following statement appears in <ref>Andrij Olijnyk and Vitaly Suchchansky. [http://www.worldscinet.com/cgi-bin/details.cgi?id=pii:S0218196704001931&type=html Representations of free products by infinite unitriangular matrices over finite fields.] International Journal of Algebra and Computation. Vol. 14 (2004), no. 5&ndash;6, pp. 741&ndash;749; Lemma 2.1</ref>, and the proof is from <ref name="DH"/>.
This version of the ping-pong lemma ensures that several subgroups of a group [[group action|acting]] on a set generate a [[free product]]. The following statement appears in Olijnyk and Suchchansky (2004),<ref>Andrij Olijnyk and Vitaly Suchchansky. [http://www.worldscinet.com/cgi-bin/details.cgi?id=pii:S0218196704001931&type=html Representations of free products by infinite unitriangular matrices over finite fields.] International Journal of Algebra and Computation. Vol. 14 (2004), no. 5&ndash;6, pp. 741&ndash;749; Lemma 2.1</ref> and the proof is from de la Harpe (2000).<ref name="DH"/>


Let ''G'' be a group acting on a set ''X'' and let ''H''<sub>1</sub>, ''H''<sub>2</sub>, ..., ''H''<sub>''k''</sub> be subgroups of ''G'' where ''k'' ≥ 2, such that at least one of these subgroups has [[order (group theory)|order]] greater than 2.
Suppose there exist [[Disjoint sets#Generalizations|pairwise disjoint]] [[empty set|nonempty]] subsets {{math|''X''<sub>1</sub>, ''X''<sub>2</sub>, ...,''X''<sub>''k''</sub>}} of {{math|''X''}} such that the following holds:


Let ''G'' be a group acting on a set ''X'' and let ''H''<sub>1</sub>, ''H''<sub>2</sub>,...., ''H''<sub>''k''</sub> be nontrivial subgroups of ''G'' where ''k''≥2, such that at least one of these subgroups has [[order (group theory)|order]] greater than 2.
*For any {{math|''i'' ''s''}} and for any {{math|''h''}} in {{math|''H''<sub>''i''</sub>}}, {{math|''h'' ≠ 1}} we have {{math|''h''(''X''<sub>''s''</sub>) ''X''<sub>''i''</sub>}}.
Then <math display="block">\langle H_1,\dots, H_k\rangle=H_1\ast\dots \ast H_k.</math>
Suppose there exist disjoint nonempty subsets ''X''<sub>1</sub>, ''X''<sub>2</sub>,....,''X''<sub>''k''</sub> of ''X'' such that the following holds:

*For any ''i''≠''j'' and for any ''h''∈''H''<sub>''i''</sub>, ''h''≠1 we have ''h''(''X''<sub>''j''</sub>)⊆''X''<sub>''i''</sub>.

Then
:<math>\langle H_1,\dots, H_k\rangle=H_1\ast\dots \ast H_k.</math>


====Proof====
====Proof====
By the definition of free product, it suffices to check that a given (nonempty) reduced word represents a nontrivial element of <math>G</math>. Let <math>w</math> be such a word of length <math>m\geq 2</math>, and let <math display="block">w = \prod_{i=1}^m w_i,</math> where <math display="inline">w_i \in H_{\alpha_i}</math> for some <math display="inline">\alpha_i \in \{1,\dots,k\}</math>. Since <math display="inline">w</math> is reduced, we have <math>\alpha_i \neq \alpha_{i+1}</math> for any <math>i = 1, \dots, m-1</math> and each <math>w_i</math> is distinct from the [[identity element]] of <math>H_{\alpha_i}</math>. We then let <math>w</math> [[group action|act]] on an element of one of the sets <math display="inline">X_i</math>. As we assume that at least one subgroup <math>H_i</math> has order at least 3, [[without loss of generality]] we may assume that <math>H_1</math> has order at least 3. We first make the assumption that <math>\alpha_1</math>and <math>\alpha_m</math> are both 1 (which implies <math>m \geq 3</math>). From here we consider <math>w</math> acting on <math>X_2</math>. We get the following chain of containments:
By the definition of free product, it suffices to check that a given reduced word is nontrivial. Let ''w'' be such a word, and let
<math display="block">w(X_2) \subseteq \prod_{i=1}^{m-1} w_i(X_1) \subseteq \prod_{i=1}^{m-2} w_i(X_{\alpha_{m-1}}) \subseteq \dots \subseteq w_1(X_{\alpha_2}) \subseteq X_1.</math>


By the assumption that different <math>X_i</math>'s are disjoint, we conclude that <math>w</math> acts nontrivially on some element of <math>X_2</math>, thus <math>w</math> represents a nontrivial element of <math>G</math>.
:<math> w= \prod_{i=1}^m w_{\alpha_i,\beta_i}. </math>


To finish the proof we must consider the three cases:
Where ''w''<sub>''j'',''β''<sub>''k''</sub></sub>∈ ''H''<sub>''j''</sub> for all such ''β''<sub>''k''</sub>, and since ''w'' is fully reduced ''α''<sub>''i''</sub>≠ α<sub>''i''+1 </sub> for any ''i''. We then let ''w'' act on an element of one of the sets ''X''<sub>''i''</sub>. As we assume for at least one subgroup ''H''<sub>i</sub> has order at least 3, without loss we may assume that ''H''<sub>1</sub> is at least 3. We first make the assumption that α<sub>1</sub> and α<sub>m</sub> are both 1. From here we consider ''w'' acting on ''X''<sub>2</sub>. We get the following chain of containments and note that since the ''X''<sub>i</sub> are disjoint that ''w'' acts nontrivially and is thus not the identity element.


*if <math>\alpha_1 = 1,\,\alpha_m \neq 1</math>, then let <math>h\in H_1\setminus \{w_1^{-1},1\}</math> (such an <math>h</math> exists since by assumption <math>H_1</math> has order at least 3);
:<math> w(X_2) \subseteq \prod_{i=1}^{m-1} w_{\alpha_i,\beta_i}(X_1) \subseteq \prod_{i=1}^{m-2} w_{\alpha_i,\beta_i}(X_{\alpha_{m-1}})\subseteq \dots \subseteq w_{1,\beta_1}w_{\alpha_2,\beta_2}(X_{\alpha_3})\subseteq w_{1,\beta_1}(X_{\alpha_2})\subseteq X_1 </math>
*if <math>\alpha_1 \neq 1,\,\alpha_m=1</math>, then let <math>h\in H_1\setminus \{w_m,1\}</math>;
*and if <math>\alpha_1\neq 1,\,\alpha_m\neq 1</math>, then let <math>h\in H_1\setminus \{1\}</math>.


In each case, <math>hwh^{-1}</math> after reduction becomes a reduced word with its first and last letter in <math>H_1</math>. Finally, <math>hwh^{-1}</math> represents a nontrivial element of <math>G</math>, and so does <math>w</math>. This proves the claim.
To finish the proof we must consider the three cases:
*If <math> \alpha_1 = 1;\alpha_m\neq 1 </math>, then let <math> h\in H_1\setminus \{w_{1,\beta_1}^{-1},1\} </math>
*If <math>\alpha_1\neq 1;\alpha_m=1</math>, then let <math> h\in H_1\setminus \{w_{1,\beta_m},1\} </math>
*And if <math> \alpha_1\neq 1;\alpha_m\neq 1</math>, then let <math> h\in H_1\setminus \{1\} </math>
In each case, ''hwh''<sup>-1</sup> is a reduced word with ''α''<sub>1</sub>' and ''α''<sub>''m'' '</sub>' both 1, and thus is nontrivial. Finally, ''hwh''<sup>-1</sup> is not 1, and so neither is ''w''. This proves the claim.


===The Ping-pong lemma for cyclic subgroups===
===The Ping-pong lemma for cyclic subgroups===


Let ''G'' be a group [[group action|acting]] on a set ''X''. Let ''a''<sub>1</sub>,...,''a''<sub>''k''</sub> be elements of ''G'', where ''k'' ≥ 2. Suppose there exist disjoint nonempty subsets
Let ''G'' be a group acting on a set ''X''. Let ''a''<sub>1</sub>, ...,''a''<sub>''k''</sub> be elements of ''G'' of infinite [[order (group theory)|order]], where ''k'' ≥ 2. Suppose there exist disjoint nonempty subsets


:''X''<sub>1</sub><sup>+</sup>,...,''X''<sub>''k''</sub><sup>+</sup> and ''X''<sub>1</sub><sup>&ndash;</sup>,...,''X''<sub>''k''</sub><sup>&ndash;</sup>
{{block indent | em = 1.5 | text = {{math|''X''<sub>1</sub><sup>+</sup>, ..., ''X''<sub>''k''</sub><sup>+</sup> and ''X''<sub>1</sub><sup>&ndash;</sup>, ..., ''X''<sub>''k''</sub><sup>&ndash;</sup>}}}}


of ''X'' with the following properties:
of {{math|''X''}} with the following properties:


*''a''<sub>''i''</sub>(''X''&nbsp;&minus;&nbsp;''X''<sub>''i''</sub><sup>&ndash;</sup>) ⊆ ''X''<sub>''i''</sub><sup>+</sup> for ''i'' = 1, ..., ''k'';
*{{math|''a''<sub>''i''</sub>(''X''''X''<sub>''i''</sub><sup>&ndash;</sup>) ⊆ ''X''<sub>''i''</sub><sup>+</sup>}} for {{math|1=''i'' = 1, ..., ''k''}};
*{{math|''a''<sub>''i''</sub><sup>−1</sup>(''X'' − ''X''<sub>''i''</sub><sup>+</sup>) ⊆ ''X''<sub>''i''</sub><sup>&ndash;</sup>}} for {{math|1=''i'' = 1, ..., ''k''}}.


*''a''<sub>''i''</sub><sup>&minus;1</sup>(''X''&nbsp;&minus;&nbsp;''X''<sub>''i''</sub><sup>+</sup>) ''X''<sub>''i''</sub><sup>&ndash;</sup> for ''i'' = 1, ..., ''k''.
Then the subgroup {{math|1=''H'' = {{angbr|''a''<sub>1</sub>, ..., ''a''<sub>''k''</sub>}} ≤ ''G''}} [[Generating set of a group|generated]] by ''a''<sub>1</sub>, ..., ''a''<sub>''k''</sub> is [[free group|free]] with free basis {{math|{''a''<sub>1</sub>, ..., ''a''<sub>''k''</sub>}<nowiki/>}}.

Then the subgroup ''H'' = <''a''<sub>1</sub>, ..., ''a''<sub>''k''</sub>> ≤ ''G'' [[Generating set of a group|generated]] by ''a''<sub>1</sub>, ..., ''a''<sub>''k''</sub> is [[free group|free]] with free basis {''a''<sub>1</sub>, ..., ''a''<sub>''k''</sub>}.


====Proof====
====Proof====
This statement follows as a corollary of the version for general subgroups if we let ''X''<sub>''i''</sub>= ''X''<sub>''i''</sub><sup>+</sup>∪''X''<sub>''i''</sub><sup>-</sup> and let ''H''<sub>''i''</sub> = ⟨''a''<sub>''i''</sub>⟩.
This statement follows as a [[corollary]] of the version for general subgroups if we let {{math|1=''X''<sub>''i''</sub> = ''X''<sub>''i''</sub><sup>+</sup> ''X''<sub>''i''</sub><sup></sup>}} and let {{math|1=''H''<sub>''i''</sub> = ⟨''a''<sub>''i''</sub>⟩}}.


==Examples==
==Examples==


===[[Special linear group]] example===
===Special linear group example===
One can use the ping-pong lemma to prove<ref name="DH"/> that the subgroup ''H'' = <''A'',''B''>≤SL(2,'''Z'''), generated by the matrices
One can use the ping-pong lemma to prove<ref name="DH"/> that the subgroup {{math|1=''H'' = {{angbr|''A'',''B''}} ≤ [[special linear group|SL]]<sub>2</sub>('''Z''')}}, generated by the [[matrix (mathematics)|matrices]]
<math display="block">A = \begin{pmatrix}1 & 2\\ 0 &1 \end{pmatrix}</math> and <math display="block">B = \begin{pmatrix}1 & 0\\ 2 &1 \end{pmatrix}</math>

is free of [[free group|rank]] two.
:<math>\scriptstyle A=\begin{pmatrix}1 & 2\\ 0 &1 \end{pmatrix} </math> and <math>\scriptstyle B=\begin{pmatrix}1 & 0\\ 2 &1 \end{pmatrix} </math>

is [[free group|free]] of rank two.


====Proof====
====Proof====
Indeed, let ''H''<sub>1</sub> = <''A''> and ''H''<sub>2</sub> = <''B''> be [[cyclic group|cyclic]] [[subgroup]]s of SL(2,'''Z''') generated by ''A'' and ''B'' accordingly. It is not hard to check that A and B are elements of infinite order in SL(2,'''Z''') and that
Indeed, let {{math|1=''H''<sub>1</sub> = {{angbr|''A''}}}} and {{math|1=''H''<sub>2</sub> = {{angbr|''B''}}}} be [[cyclic subgroup]]s of {{math|SL<sub>2</sub>('''Z''')}} generated by {{math|''A''}} and {{math|''B''}} accordingly. It is not hard to check that {{math|''A''}} and {{math|''B''}} are elements of infinite order in {{math|SL<sub>2</sub>('''Z''')}} and that
<math display="block">H_1 = \{A^n \mid n\in \Z\} = \left\{\begin{pmatrix}1 & 2n\\ 0 & 1 \end{pmatrix} : n\in\Z\right\}</math>

:<math>H_1=\{A^n|n\in \mathbb Z\}=\left\{\begin{pmatrix}1 & 2n\\ 0 & 1 \end{pmatrix} : n\in\mathbb Z\right\}</math>

and
and
<math display="block">H_2 = \{B^n \mid n\in \Z\} = \left\{\begin{pmatrix}1 & 0\\ 2n & 1 \end{pmatrix} : n\in\Z\right\}.</math>


Consider the standard [[group action|action]] of {{math|SL<sub>2</sub>('''Z''')}} on {{math|'''R'''<sup>2</sup>}} by [[linear transformation]]s. Put
:<math>H_2=\{B^n|n\in \mathbb Z\}=\left\{\begin{pmatrix}1 & 0\\ 2n & 1 \end{pmatrix} : n\in\mathbb Z\right\}.</math>
<math display="block">X_1 = \left\{ \begin{pmatrix}x \\ y \end{pmatrix}\in \R^2 : |x|>|y|\right\}</math>

Consider the standard action of SL(2,'''Z''') on '''R'''<sup>2</sup> by [[linear transformation]]s. Put

:<math>X_1=\left\{\begin{pmatrix}x \\ y \end{pmatrix}\in \mathbb R^2 : |x|>|y|\right\}</math>

and
and
<math display="block">X_2 = \left\{ \begin{pmatrix}x \\ y \end{pmatrix}\in \mathbb R^2 : |x|<|y|\right\}.</math>


It is not hard to check, using the above explicit descriptions of ''H''<sub>1</sub> and ''H''<sub>2</sub>, that for every nontrivial {{math|''g'' ''H''<sub>1</sub>}} we have {{math|''g''(''X''<sub>2</sub>) ''X''<sub>1</sub>}} and that for every nontrivial {{math|''g'' ''H''<sub>2</sub>}} we have {{math|''g''(''X''<sub>1</sub>) ''X''<sub>2</sub>}}. Using the alternative form of the ping-pong lemma, for two subgroups, given above, we conclude that {{math|1=''H'' = ''H''<sub>1</sub>&thinsp;&thinsp;''H''<sub>2</sub>}}. Since the groups {{math|''H''<sub>1</sub>}} and {{math|''H''<sub>2</sub>}} are infinite [[cyclic group|cyclic]], it follows that ''H'' is a [[free group]] of rank two.
:<math>X_2=\left\{\begin{pmatrix}x \\ y \end{pmatrix}\in \mathbb R^2 : |x|<|y|\right\}.</math>


===Word-hyperbolic group example===
It is not hard to check, using the above explicitly descriptions of ''H''<sub>1</sub> and ''H''<sub>2</sub> that for every nontrivial ''g''&nbsp;&nbsp;''H''<sub>1</sub> we have ''g''(''X''<sub>2</sub>)&nbsp;&nbsp;''X''<sub>1</sub> and that for every nontrivial ''g''&nbsp;&nbsp;''H''<sub>2</sub> we have ''g''(''X''<sub>1</sub>)&nbsp;&nbsp;''X''<sub>2</sub>. Using the alternative form of the ping-pong lemma, for two subgroups, given above, we conclude that ''H''&nbsp;=&nbsp;''H''<sub>1</sub>∗''H''<sub>2</sub>. Since the groups ''H''<sub>1</sub> and ''H''<sub>2</sub> are [[infinite cyclic group|infinite cyclic]], it follows that ''H'' is a [[free group]] of rank two.


Let {{math|''G''}} be a [[word-hyperbolic group]] which is [[torsion-free group|torsion-free]], that is, with no nonidentity elements of finite [[Order (group theory)|order]]. Let {{math|''g'', ''h'' ''G''}} be two non-commuting elements, that is such that {{math|''gh'' ''hg''}}. Then there exists ''M'' ≥ 1 such that for any [[integer]]s {{math|''n'' ''M''}}, {{math|''m'' ''M''}} the subgroup {{math|1=H = {{angbr|''g''<sup>''n''</sup>, ''h''<sup>''m''</sup>}} ''G''}} is free of rank two.
===[[Word-hyperbolic group]] example===


====Sketch of the proof<ref name="Gromov">M. Gromov. ''Hyperbolic groups.'' Essays in group theory, pp. 75&ndash;263, Mathematical Sciences Research Institute Publications, 8, Springer, New York, 1987; {{isbn|0-387-96618-8}}; Ch. 8.2, pp. 211&ndash;219.</ref>====
Let ''G'' be a [[word-hyperbolic group]] which is [[torsion-free group|torsion-free]], that is, with no nontrivial elements of finite [[Order (group theory)|order]]. Let ''g'',&nbsp;''h''&nbsp;&nbsp;''G'' be two non-commuting elements, that is such that ''gh''&nbsp;&nbsp;''hg''. Then there exists ''M''≥1 such that for any integers ''n''&nbsp;&nbsp;''M'', ''m''&nbsp;&nbsp;''M'' the subgroup H&nbsp;=&nbsp;<''g''<sup>''n''</sup>, ''h''<sup>''m''</sup>>&nbsp;&nbsp;''G'' is [[free group|free]] of rank two.
The group ''G'' acts on its ''hyperbolic boundary'' ∂''G'' by [[homeomorphism]]s. It is known that if ''a'' in ''G'' is a nonidentity element then ''a'' has exactly two distinct fixed points, {{math|''a''<sup>∞</sup>}} and {{math|''a''<sup>−∞</sup>}} in {{math|∂''G''}} and that {{math|''a''<sup>∞</sup>}} is an [[attracting fixed point]] while {{math|''a''<sup>−∞</sup>}} is a [[Fixed point (mathematics)|repelling fixed point]].


Since {{math|''g''}} and {{math|''h''}} do not commute, basic facts about word-hyperbolic groups imply that {{math|''g''<sup>∞</sup>}}, {{math|''g''<sup>−∞</sup>}}, {{math|''h''<sup>∞</sup>}} and {{math|''h''<sup>−∞</sup>}} are four distinct points in {{math|∂''G''}}. Take disjoint [[Neighbourhood (mathematics)|neighborhoods]] {{math|''U''<sub>+</sub>}}, {{math|''U''<sub>&ndash;</sub>}}, {{math|''V''<sub>+</sub>}}, and {{math|''V''<sub>&ndash;</sub>}} of {{math|''g''<sup>∞</sup>}}, {{math|''g''<sup>−∞</sup>}}, {{math|''h''<sup>∞</sup>}} and {{math|''h''<sup>−∞</sup>}} in {{math|∂''G''}} respectively.
====Sketch of the proof<ref name="Gromov">M. Gromov. ''Hyperbolic groups.'' Essays in group theory, pp. 75&ndash;263,
Then the attracting/repelling properties of the fixed points of ''g'' and ''h'' imply that there exists {{math|''M'' 1}} such that for any integers {{math|''n'' ''M''}}, {{math|''m'' ''M''}} we have:
Mathematical Sciiences Research Institute Publications, 8, Springer, New York, 1987; ISBN 0-387-96618-8; Ch. 8.2, pp. 211&ndash;219.</ref>====
*{{math|''g''<sup>''n''</sup>(∂''G'' &ndash; ''U''<sub>&ndash;</sub>) ⊆ ''U''<sub>+</sub>}}
The group ''G'' [[Group action|acts]] on its ''hyperbolic boundary'' ∂''G'' by [[homeomorphism]]s. It is known that if ''a''&nbsp;∈&nbsp;''G'' is a nontrivial element then ''a'' has exactly two distinct fixed points, ''a''<sup>∞</sup> and ''a''<sup>&minus;∞</sup> in ∂''G'' and that ''a''<sup>∞</sup> is an [[attracting fixed point]] while ''a''<sup>&minus;∞</sup> is a [[Fixed point (mathematics)|repelling fixed point]].
*{{math|''g''<sup>''n''</sup>(∂''G'' &ndash; ''U''<sub>+</sub>) ⊆ ''U''<sub>&ndash;</sub>}}
*{{math|''h''<sup>''m''</sup>(∂''G'' &ndash; ''V''<sub>&ndash;</sub>) ⊆ ''V''<sub>+</sub>}}
*{{math|''h''<sup>''m''</sup>(∂''G'' &ndash; ''V''<sub>+</sub>) ⊆ ''V''<sub>&ndash;</sub>}}


The ping-pong lemma now implies that {{math|1=''H'' = {{angbr|''g''<sup>''n''</sup>, ''h''<sup>''m''</sup>}} ''G''}} is free of rank two.
Since ''g'' and ''h'' do not commute, the basic facts about [[word-hyperbolic group]]s imply that ''g''<sup>∞</sup>, ''g''<sup>&minus;∞</sup>, ''h''<sup>∞</sup> and ''h''<sup>&minus;∞</sup> are four distinct points in ∂''G''. Take disjoint [[Neighbourhood (mathematics)|neighborhoods]] ''U''<sub>+</sub>, ''U''<sub>&ndash;</sub>, ''V''<sub>+</sub> and ''V''<sub>&ndash;</sub> of ''g''<sup>∞</sup>, ''g''<sup>&minus;∞</sup>, ''h''<sup>∞</sup> and ''h''<sup>&minus;∞</sup> in ∂''G'' respectively.
Then the attracting/repelling properties of the fixed points of ''g'' and ''h'' imply that there exists ''M''&nbsp;&nbsp;1 such that for any integers ''n''&nbsp;&nbsp;''M'', ''m''&nbsp;&nbsp;''M'' we have:
*''g''<sup>''n''</sup>(∂''G'' &ndash; ''U''<sub>&ndash;</sub>) ⊆ ''U''<sub>+</sub>
*''g''<sup>&minus;''n''</sup>(∂''G'' &ndash; ''U''<sub>+</sub>) ⊆ ''U''<sub>&ndash;</sub>
*''h''<sup>''m''</sup>(∂''G'' &ndash; ''V''<sub>&ndash;</sub>) ⊆ ''V''<sub>+</sub>
*''h''<sup>&minus;''m''</sup>(∂''G'' &ndash; ''V''<sub>+</sub>) ⊆ ''V''<sub>&ndash;</sub>

The ping-pong lemma now implies that ''H''&nbsp;=&nbsp;<''g''<sup>''n''</sup>, ''h''<sup>''m''</sup>>&nbsp;&nbsp;''G'' is [[free group|free]] of rank two.


==Applications of the ping-pong lemma==
==Applications of the ping-pong lemma==


*The ping-pong lemma is used in [[Kleinian group]]s to study their so-called [[Schottky group|Schottky subgroups]]. In the Kleinian groups context the ping-pong lemma can be used to show that a particular group of isometries of the [[hyperbolic 3-space]] is not just [[free group|free]] but also [[properly discontinuous]] and [[geometrically finite group|geometrically finite]].
*The ping-pong lemma is used in [[Kleinian group]]s to study their so-called [[Schottky group|Schottky subgroups]]. In the Kleinian groups context the ping-pong lemma can be used to show that a particular group of isometries of the [[hyperbolic 3-space]] is not just [[free group|free]] but also [[properly discontinuous]] and [[geometrically finite group|geometrically finite]].
*Similar Schottky-type arguments are widely used in [[geometric group theory]], particularly for subgroups of [[word-hyperbolic group]]s<ref name="Gromov"/> and for automorphism groups of trees.<ref>[[Alexander Lubotzky]]. [http://www.springerlink.com/content/g374700j2401nl64/ ''Lattices in rank one Lie groups over local fields.''] [[Geometric and Functional Analysis]], vol. 1 (1991), no. 4, pp. 406&ndash;431 </ref>
*Similar Schottky-type arguments are widely used in [[geometric group theory]], particularly for subgroups of [[word-hyperbolic group]]s<ref name="Gromov"/> and for [[automorphism group]]s of trees.<ref>[[Alexander Lubotzky]]. [https://doi.org/10.1007%2FBF01895641 ''Lattices in rank one Lie groups over local fields.''] [[Geometric and Functional Analysis]], vol. 1 (1991), no. 4, pp. 406&ndash;431</ref>
*Ping-pong lemma is also used for studying Schottki-type subgroups of [[mapping class group]]s of [[Riemann surface]]s, where the set on which the mapping class group acts is the Thurston boundary of the [[Teichmüller space]].<ref>Richard P. Kent, and Christopher J. Leininger. ''Subgroups of mapping class groups from the geometrical viewpoint.'' In the tradition of Ahlfors-Bers. IV, pp. 119&ndash;141,
*The ping-pong lemma is also used for studying Schottky-type subgroups of [[mapping class group]]s of [[Riemann surface]]s, where the set on which the mapping class group acts is the Thurston boundary of the [[Teichmüller space]].<ref>Richard P. Kent, and Christopher J. Leininger. ''Subgroups of mapping class groups from the geometrical viewpoint.'' In the tradition of Ahlfors-Bers. IV, pp. 119&ndash;141,
Contemporary Mathematics series, 432, [[American Mathematical Society]], Providence, RI, 2007; ISBN 978-0-8218-4227-0; 0-8218-4227-7</ref> A similar argument is also utilized in the study of subgroups of the [[outer automorphism group]] of a [[free group]].<ref>[[Mladen Bestvina|M. Bestvina]], M. Feighn, and M. Handel. [http://www.springerlink.com/content/50hq64n0l6gpuukk/ ''Laminations, trees, and irreducible automorphisms of free groups.''] [[Geometric and Functional Analysis]], vol. 7 (1997), no. 2, pp. 215&ndash;244. </ref>
Contemporary Mathematics series, 432, [[American Mathematical Society]], Providence, RI, 2007; {{isbn|978-0-8218-4227-0}}; 0-8218-4227-7</ref> A similar argument is also utilized in the study of subgroups of the [[outer automorphism group]] of a free group.<ref>[[Mladen Bestvina|M. Bestvina]], M. Feighn, and M. Handel. [https://doi.org/10.1007%2FPL00001618 ''Laminations, trees, and irreducible automorphisms of free groups.''] [[Geometric and Functional Analysis]], vol. 7 (1997), no. 2, pp. 215&ndash;244.</ref>
*One of the most famous applications of the ping-pong lemma is in the proof of [[Jacques Tits]] of the so-called [[Tits alternative]] for [[linear group]]s. <ref name="T"/> (see also <ref>Pierre de la Harpe. ''Free groups in linear groups.'' L'Enseignement Mathématique (2), vol. 29 (1983), no. 1-2, pp. 129&ndash;144 </ref> for an overview of Tits' proof and an explanation of the ideas involved, including the use of the ping-pong lemma).
*One of the most famous applications of the ping-pong lemma is in the proof of [[Jacques Tits]] of the so-called [[Tits alternative]] for [[linear group]]s.<ref name="T"/> (see also <ref>Pierre de la Harpe. ''Free groups in linear groups.'' L'Enseignement Mathématique (2), vol. 29 (1983), no. 1-2, pp. 129&ndash;144</ref> for an overview of Tits' proof and an explanation of the ideas involved, including the use of the ping-pong lemma).
*There are generalizations of the ping-pong lemma that produce not just [[free product]]s but also [[free product with amalgamation|amalgamated free products]] and [[HNN extension]]s<ref name="LS"/>. These generalizations are used, in particular, in the proof of Maskit's Combination Theorem for [[Kleinian group]]s<ref>Bernard Maskit.
*There are generalizations of the ping-pong lemma that produce not just [[free product]]s but also [[free product with amalgamation|amalgamated free products]] and [[HNN extension]]s.<ref name="LS"/> These generalizations are used, in particular, in the proof of [[Maskit's Combination Theorem]] for Kleinian groups.<ref>[[Bernard Maskit]].
''Kleinian groups.'' Grundlehren der Mathematischen Wissenschaften [Fundamental Principles of Mathematical Sciences], 287. Springer-Verlag, Berlin, 1988. ISBN 3-540-17746-9; Ch. VII.C and Ch. VII.E pp.149&ndash;156 and pp. 160&ndash;167</ref>.
''Kleinian groups.'' Grundlehren der Mathematischen Wissenschaften [Fundamental Principles of Mathematical Sciences], 287. Springer-Verlag, Berlin, 1988. {{isbn|3-540-17746-9}}; Ch. VII.C and Ch. VII.E pp.149&ndash;156 and pp. 160&ndash;167</ref>
*There are also versions of the ping-pong lemma which guarantee that several elements in a group generate a [[free semigroup]]. Such versions are available both in the general context of a [[group action]] on a set<ref name="DH1">Pierre de la Harpe. [http://books.google.com/books?id=cRT01C5ADroC&pg=PA188&vq=semi-group&dq=ping+pong+lemma+group+theory&source=gbs_search_s&sig=ACfU3U2oMEeKTE_pB7Gt_MqNjOaUNZL8yw ''Topics in geometric group theory.''] Chicago Lectures in Mathematics. University of Chicago Press, Chicago. ISBN 0-226-31719-6; Ch. II.B "The table-Tennis Lemma (Klein's criterion) and examples of free products"; pp. 187&ndash;188. </ref>, and for specific types of actions, e.g. in the context of [[linear group]]s<ref>Alex Eskin, Shahar Mozes and Hee Oh. [http://www.springerlink.com/content/3ybuud1bpkkkcxn0/ On uniform exponential growth for linear groups.] [[Inventiones Mathematicae]]. vol. 60 (2005), no. 1, pp.1432&ndash;1297; Lemma 2.2</ref>, groups [[Bass-Serre theory|acting on trees]]<ref>Roger C. Alperin and Guennadi A. Noskov. [http://books.google.com/books?id=w7LO6AkB8Y8C&pg=PA2&lpg=PA2&dq=%22ping-pong+lemma%22+semigroup&source=web&ots=aBPNu6adQ2&sig=7mZjESpp-6Bkekw68RCPEDYJSTM&hl=en&sa=X&oi=book_result&resnum=4&ct=result#PPA2,M1 Uniform growth, actions on trees and GL<sub>2</sub>.] Computational and Statistical Group Theory:AMS Special Session Geometric Group Theory, April 21-22, 2001, Las Vegas, Nevada, AMS Special Session Computational Group Theory, April 28-29, 2001, Hoboken, New Jersey. (Robert H. Gilman, Vladimir Shpilrain, Alexei G. Myasnikov, editors). [[American Mathematical Society]], 2002. ISBN 978-0-8218-3158-8; page 2, Lemma 3.1</ref> and others.<ref>Yves de Cornulier and Romain Tessera. [http://msp.warwick.ac.uk/gt/2008/12-01/p011.xhtml Quasi-isometrically embedded free sub-semigroups.] [[Geometry & Topology]], vol. 12 (2008), pp. 461&ndash;473; Lemma 2.1</ref>
*There are also versions of the ping-pong lemma which guarantee that several elements in a group generate a [[free semigroup]]. Such versions are available both in the general context of a [[group action]] on a set,<ref name="DH1">Pierre de la Harpe. [https://books.google.com/books?id=cRT01C5ADroC&q=semi-group&pg=PA188 ''Topics in geometric group theory.''] Chicago Lectures in Mathematics. University of Chicago Press, Chicago. {{isbn|0-226-31719-6}}; Ch. II.B "The table-Tennis Lemma (Klein's criterion) and examples of free products"; pp. 187&ndash;188.</ref> and for specific types of actions, e.g. in the context of linear groups,<ref>Alex Eskin, Shahar Mozes and Hee Oh. [https://doi.org/10.1007%2Fs00222-004-0378-z On uniform exponential growth for linear groups.] [[Inventiones Mathematicae]]. vol. 60 (2005), no. 1, pp.1432&ndash;1297; Lemma 2.2</ref> groups [[Bass–Serre theory|acting on trees]]<ref>Roger C. Alperin and Guennadi A. Noskov. [https://books.google.com/books?id=w7LO6AkB8Y8C&dq=%22ping-pong+lemma%22+semigroup&pg=PA2 Uniform growth, actions on trees and GL<sub>2</sub>.] Computational and Statistical Group Theory:AMS Special Session Geometric Group Theory, April 21–22, 2001, Las Vegas, Nevada, AMS Special Session Computational Group Theory, April 28–29, 2001, Hoboken, New Jersey. (Robert H. Gilman, Vladimir Shpilrain, Alexei G. Myasnikov, editors). [[American Mathematical Society]], 2002. {{isbn|978-0-8218-3158-8}}; page 2, Lemma 3.1</ref> and others.<ref>Yves de Cornulier and Romain Tessera. [http://msp.warwick.ac.uk/gt/2008/12-01/p011.xhtml Quasi-isometrically embedded free sub-semigroups.] ''[[Geometry & Topology]]'', vol. 12 (2008), pp. 461&ndash;473; Lemma 2.1</ref>


==References==
==References==
{{reflist}}
{{reflist}}




==See also==
==See also==
Line 123: Line 107:
*[[Schottky group]]
*[[Schottky group]]


[[Category:Algebra]]
[[Category:Lemmas in group theory]]
[[Category:Group theory]]
[[Category:Discrete groups]]
[[Category:Discrete groups]]
[[Category:Lie groups]]
[[Category:Lie groups]]
[[Category:Combinatorics on words]]
[[Category:Combinatorics on words]]

[[tr:Klein kriteri]]

Latest revision as of 14:59, 20 March 2023

In mathematics, the ping-pong lemma, or table-tennis lemma, is any of several mathematical statements that ensure that several elements in a group acting on a set freely generates a free subgroup of that group.

History

[edit]

The ping-pong argument goes back to the late 19th century and is commonly attributed[1] to Felix Klein who used it to study subgroups of Kleinian groups, that is, of discrete groups of isometries of the hyperbolic 3-space or, equivalently Möbius transformations of the Riemann sphere. The ping-pong lemma was a key tool used by Jacques Tits in his 1972 paper[2] containing the proof of a famous result now known as the Tits alternative. The result states that a finitely generated linear group is either virtually solvable or contains a free subgroup of rank two. The ping-pong lemma and its variations are widely used in geometric topology and geometric group theory.

Modern versions of the ping-pong lemma can be found in many books such as Lyndon & Schupp,[3] de la Harpe,[1] Bridson & Haefliger[4] and others.

Formal statements

[edit]

Ping-pong lemma for several subgroups

[edit]

This version of the ping-pong lemma ensures that several subgroups of a group acting on a set generate a free product. The following statement appears in Olijnyk and Suchchansky (2004),[5] and the proof is from de la Harpe (2000).[1]

Let G be a group acting on a set X and let H1, H2, ..., Hk be subgroups of G where k ≥ 2, such that at least one of these subgroups has order greater than 2. Suppose there exist pairwise disjoint nonempty subsets X1, X2, ...,Xk of X such that the following holds:

  • For any is and for any h in Hi, h ≠ 1 we have h(Xs) ⊆ Xi.

Then

Proof

[edit]

By the definition of free product, it suffices to check that a given (nonempty) reduced word represents a nontrivial element of . Let be such a word of length , and let where for some . Since is reduced, we have for any and each is distinct from the identity element of . We then let act on an element of one of the sets . As we assume that at least one subgroup has order at least 3, without loss of generality we may assume that has order at least 3. We first make the assumption that and are both 1 (which implies ). From here we consider acting on . We get the following chain of containments:

By the assumption that different 's are disjoint, we conclude that acts nontrivially on some element of , thus represents a nontrivial element of .

To finish the proof we must consider the three cases:

  • if , then let (such an exists since by assumption has order at least 3);
  • if , then let ;
  • and if , then let .

In each case, after reduction becomes a reduced word with its first and last letter in . Finally, represents a nontrivial element of , and so does . This proves the claim.

The Ping-pong lemma for cyclic subgroups

[edit]

Let G be a group acting on a set X. Let a1, ...,ak be elements of G of infinite order, where k ≥ 2. Suppose there exist disjoint nonempty subsets

X1+, ..., Xk+ and X1, ..., Xk

of X with the following properties:

  • ai(XXi) ⊆ Xi+ for i = 1, ..., k;
  • ai−1(XXi+) ⊆ Xi for i = 1, ..., k.

Then the subgroup H = a1, ..., akG generated by a1, ..., ak is free with free basis {a1, ..., ak}.

Proof

[edit]

This statement follows as a corollary of the version for general subgroups if we let Xi = Xi+Xi and let Hi = ⟨ai.

Examples

[edit]

Special linear group example

[edit]

One can use the ping-pong lemma to prove[1] that the subgroup H = A,BSL2(Z), generated by the matrices and is free of rank two.

Proof

[edit]

Indeed, let H1 = A and H2 = B be cyclic subgroups of SL2(Z) generated by A and B accordingly. It is not hard to check that A and B are elements of infinite order in SL2(Z) and that and

Consider the standard action of SL2(Z) on R2 by linear transformations. Put and

It is not hard to check, using the above explicit descriptions of H1 and H2, that for every nontrivial gH1 we have g(X2) ⊆ X1 and that for every nontrivial gH2 we have g(X1) ⊆ X2. Using the alternative form of the ping-pong lemma, for two subgroups, given above, we conclude that H = H1 ∗ H2. Since the groups H1 and H2 are infinite cyclic, it follows that H is a free group of rank two.

Word-hyperbolic group example

[edit]

Let G be a word-hyperbolic group which is torsion-free, that is, with no nonidentity elements of finite order. Let g, hG be two non-commuting elements, that is such that ghhg. Then there exists M ≥ 1 such that for any integers nM, mM the subgroup H = gn, hmG is free of rank two.

Sketch of the proof[6]

[edit]

The group G acts on its hyperbolic boundaryG by homeomorphisms. It is known that if a in G is a nonidentity element then a has exactly two distinct fixed points, a and a−∞ in G and that a is an attracting fixed point while a−∞ is a repelling fixed point.

Since g and h do not commute, basic facts about word-hyperbolic groups imply that g, g−∞, h and h−∞ are four distinct points in G. Take disjoint neighborhoods U+, U, V+, and V of g, g−∞, h and h−∞ in G respectively. Then the attracting/repelling properties of the fixed points of g and h imply that there exists M ≥ 1 such that for any integers nM, mM we have:

  • gn(∂GU) ⊆ U+
  • gn(∂GU+) ⊆ U
  • hm(∂GV) ⊆ V+
  • hm(∂GV+) ⊆ V

The ping-pong lemma now implies that H = gn, hmG is free of rank two.

Applications of the ping-pong lemma

[edit]

References

[edit]
  1. ^ a b c d Pierre de la Harpe. Topics in geometric group theory. Chicago Lectures in Mathematics. University of Chicago Press, Chicago. ISBN 0-226-31719-6; Ch. II.B "The table-Tennis Lemma (Klein's criterion) and examples of free products"; pp. 25–41.
  2. ^ a b J. Tits. Free subgroups in linear groups. Journal of Algebra, vol. 20 (1972), pp. 250–270
  3. ^ a b Roger C. Lyndon and Paul E. Schupp. Combinatorial Group Theory. Springer-Verlag, New York, 2001. "Classics in Mathematics" series, reprint of the 1977 edition. ISBN 978-3-540-41158-1; Ch II, Section 12, pp. 167–169
  4. ^ Martin R. Bridson, and André Haefliger. Metric spaces of non-positive curvature. Grundlehren der Mathematischen Wissenschaften [Fundamental Principles of Mathematical Sciences], 319. Springer-Verlag, Berlin, 1999. ISBN 3-540-64324-9; Ch.III.Γ, pp. 467–468
  5. ^ Andrij Olijnyk and Vitaly Suchchansky. Representations of free products by infinite unitriangular matrices over finite fields. International Journal of Algebra and Computation. Vol. 14 (2004), no. 5–6, pp. 741–749; Lemma 2.1
  6. ^ a b M. Gromov. Hyperbolic groups. Essays in group theory, pp. 75–263, Mathematical Sciences Research Institute Publications, 8, Springer, New York, 1987; ISBN 0-387-96618-8; Ch. 8.2, pp. 211–219.
  7. ^ Alexander Lubotzky. Lattices in rank one Lie groups over local fields. Geometric and Functional Analysis, vol. 1 (1991), no. 4, pp. 406–431
  8. ^ Richard P. Kent, and Christopher J. Leininger. Subgroups of mapping class groups from the geometrical viewpoint. In the tradition of Ahlfors-Bers. IV, pp. 119–141, Contemporary Mathematics series, 432, American Mathematical Society, Providence, RI, 2007; ISBN 978-0-8218-4227-0; 0-8218-4227-7
  9. ^ M. Bestvina, M. Feighn, and M. Handel. Laminations, trees, and irreducible automorphisms of free groups. Geometric and Functional Analysis, vol. 7 (1997), no. 2, pp. 215–244.
  10. ^ Pierre de la Harpe. Free groups in linear groups. L'Enseignement Mathématique (2), vol. 29 (1983), no. 1-2, pp. 129–144
  11. ^ Bernard Maskit. Kleinian groups. Grundlehren der Mathematischen Wissenschaften [Fundamental Principles of Mathematical Sciences], 287. Springer-Verlag, Berlin, 1988. ISBN 3-540-17746-9; Ch. VII.C and Ch. VII.E pp.149–156 and pp. 160–167
  12. ^ Pierre de la Harpe. Topics in geometric group theory. Chicago Lectures in Mathematics. University of Chicago Press, Chicago. ISBN 0-226-31719-6; Ch. II.B "The table-Tennis Lemma (Klein's criterion) and examples of free products"; pp. 187–188.
  13. ^ Alex Eskin, Shahar Mozes and Hee Oh. On uniform exponential growth for linear groups. Inventiones Mathematicae. vol. 60 (2005), no. 1, pp.1432–1297; Lemma 2.2
  14. ^ Roger C. Alperin and Guennadi A. Noskov. Uniform growth, actions on trees and GL2. Computational and Statistical Group Theory:AMS Special Session Geometric Group Theory, April 21–22, 2001, Las Vegas, Nevada, AMS Special Session Computational Group Theory, April 28–29, 2001, Hoboken, New Jersey. (Robert H. Gilman, Vladimir Shpilrain, Alexei G. Myasnikov, editors). American Mathematical Society, 2002. ISBN 978-0-8218-3158-8; page 2, Lemma 3.1
  15. ^ Yves de Cornulier and Romain Tessera. Quasi-isometrically embedded free sub-semigroups. Geometry & Topology, vol. 12 (2008), pp. 461–473; Lemma 2.1

See also

[edit]