A note on the Burnside problem for homeomorphism groups of manifolds
Donggyun Seo · 2026
Read A note on the Burnside problem for homeomorphism groups of manifolds by Donggyun Seo (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2604-15627.txt.
Introduction
Seo takes up the Burnside problem for groups of homeomorphisms of compact connected manifolds. For surfaces, the identity part has no torsion except on the sphere, torus, projective plane, and Klein bottle, and a Tits-alternative extension step makes every finite-built periodic subgroup finite elsewhere. For the circle, such subgroups are finite and cyclic, stretching a Guelman-Liousse theorem to non-orientable surfaces.
Catalog record
| CREATOR | Donggyun Seo |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 20585 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2604.15627
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
A note on the Burnside problem for homeomorphism groups of manifolds Donggyun Seo Abstract . This note studies the Burnside problem for homeomorphism groups of compact connected manifolds. For surfaces, we prove that the identity component of the homeomorphism group is torsion-free precisely when the surface is not the sphere, torus, projective plane, or Klein bottle. An extension argument based on the Tits alternative for mapping class groups then implies that every finitely generated periodic subgroup of the full homeomorphism group is finite for all surfaces outside this exceptional list, recovering and extending a theorem of Guelman and Liousse to non-orientable surfaces. For the circle, we prove that every finitely generated periodic subgroup of its homeomorphism group is finite and cyclic. We close with remarks on manifolds with boundary and open questions on the Burnside problem for hyperbolic three-manifolds and doubled handlebodies. Burnside problem for surface homeomorphisms Let \Sigma be a real, compact, connected surface (orientable or non-orientable, possibly with boundary). The main object of study in this note is the group \Homeo(\Sigma)=\Homeo(\Sigma,\partial\Sigma) of homeomorphisms of \Sigma fixing the boundary pointwise. If \Sigma is orientable, let \Homeo^{+}(\Sigma) denote the subgroup of orientation-preserving elements; if \Sigma is non-orientable, set \Homeo^{+}(\Sigma)=\Homeo(\Sigma) . Let \Homeo_{0}(\Sigma) denote the identity component of \Homeo^{+}(\Sigma) . We write \mathbb{S}^{2} , T^{2} , \mathbb{R}P^{2} , and K for the sphere, torus, real projective plane, and Klein bottle, respectively. Proposition 1 . For a compact connected surface \Sigma , the following are equivalent: (1) \Sigma\notin\{\mathbb{S}^{2},T^{2},\mathbb{R}P^{2},K\} , (2) \Homeo_{0}(\Sigma) is torsion-free, and (3) \Homeo_{0}(\Sigma) is contractible. The equivalence of conditions ( 1 ) and ( 3 ) was established by Kneser [ 20 ] , Hamstrom [ 16 , 15 , 17 ] , and Luke–Mason [ 22 ] . Although the equivalence of ( 1 ) and ( 2 ) is likely well known, the author could not discover a proof in the literature. This gap motivated the present note. The implication ( 2 ) \Rightarrow ( 1 ) can be derived by exhibiting a nontrivial torsion element in \Homeo_{0}(\Sigma) for each \Sigma\in\{\mathbb{S}^{2},\mathbb{R}P^{2},T^{2},K\} . (1) Sphere. Identify \mathbb{S}^{2}=\{v\in\mathbb{R}^{3}\mid\|v\|=1\} . The continuous, injective homomorphism \mathrm{SO}(3)\to\Homeo_{0}(\mathbb{S}^{2}) given by matrix multiplication is well defined since \mathrm{SO}(3) is connected. For \theta=q\pi with q\in\mathbb{Q}\setminus 2\mathbb{Z} , the rotation v\mapsto\begin{pmatrix}1&0&0\\ 0&\cos\theta&-\sin\theta\\ 0&\sin\theta&\cos\theta\end{pmatrix}v is a nontrivial element of finite order in \Homeo_{0}(\mathbb{S}^{2}) ; writing q=p/n in lowest terms, its order is 2n if p is odd and n if p is even. (2) Real projective plane. Identify \mathbb{R}P^{2}=\mathbb{S}^{2}/(v\sim-v) . Since every A\in\mathrm{SO}(3) commutes with the antipodal map, it descends to a homeomorphism \bar{A}([v])=[Av] of \mathbb{R}P^{2} . This defines a continuous, injective homomorphism \mathrm{SO}(3)\to\Homeo_{0}(\mathbb{R}P^{2}) , where injectivity follows from -I\notin\mathrm{SO}(3) . The same rotation as in the spherical case, \bar{A}_{\theta}([v])=\left[\begin{pmatrix}1&0&0\\ 0&\cos\theta&-\sin\theta\\ 0&\sin\theta&\cos\theta\end{pmatrix}v\right] with \theta=q\pi and q\in\mathbb{Q}\setminus 2\mathbb{Z} , therefore yields a nontrivial torsion element of \Homeo_{0}(\mathbb{R}P^{2}) . (3) Torus. Identify T^{2}=\mathbb{R}^{2}/\mathbb{Z}^{2} . Since \mathbb{R}^{2} is connected, the translation action v\cdot[x]=[x+v] maps into \Homeo_{0}(T^{2}) . For any v=(p/q,\,r/s)\in\mathbb{Q}^{2}\setminus\mathbb{Z}^{2} , the translation \tau_{v}\colon[x]\mapsto[x+v] is a nontrivial homeomorphism of finite order \mathrm{lcm}(q,s) , hence a nontrivial torsion element of \Homeo_{0}(T^{2}) . (4) Klein bottle. Identify K=\mathbb{R}^{2}/\Gamma , where \Gamma is generated by \gamma_{1}\colon(x,y)\mapsto(x+1,\,y),\qquad\gamma_{2}\colon(x,y)\mapsto(-x,\,y+1). The map \varphi\colon K\to K induced by (x,y)\mapsto(x,\,y+\frac{1}{2}) is well-defined, satisfies \varphi^{2}=id , and is isotopic to the identity via the homotopy \varphi_{t}\colon(x,y)\mapsto(x,\,y+\tfrac{t}{2}) , t\in[0,1] . Hence \varphi is a nontrivial torsion element of \Homeo_{0}(K) . In each case \Homeo_{0}(\Sigma) contains a nontrivial torsion element. To prove the implication ( 1 ) \Rightarrow ( 2 ) in Proposition 1 , we recall several basic properties of surface braid groups and mapping class groups. Most of the material between Proposition 1 and its proof is adapted from the survey paper of Guaschi–Juan-Pineda [ 12 ] . Let D_{n}(\Sigma) denote the space of unordered n -tuples of distinct points in \Sigma . More precisely, D_{n}(\Sigma) is the quotient of \tilde{D}_{n}(\Sigma):=\{(p_{1},\dots,p_{n})\in\Sigma^{n}\mid p_{i}\neq p_{j}\text{ for }i\neq j\} under the map (p_{1},\dots,p_{n})\mapsto\{p_{1},\dots,p_{n}\} . We equip D_{n}(\Sigma) with the quotient topology induced from \tilde{D}_{n}(\Sigma) . Fix a positive integer n and a finite subset P\subset\Sigma with |P|=n . The n -strand braid group of \Sigma based at P , denoted B_{n}(\Sigma) or B_{n}(\Sigma,P) , is defined as the fundamental group \pi_{1}(D_{n}(\Sigma),P) . It is known from Fadell–Neuwirth [ 10 ] and van Buskirk [ 32 ] that B_{n}(\Sigma) is torsion-free if and only if \Sigma\notin\{\mathbb{S}^{2},\mathbb{R}P^{2}\} . The natural diagonal action of \Homeo^{+}(\Sigma) on D_{n}(\Sigma) induces a map \Psi:\Homeo^{+}(\Sigma)\to D_{n}(\Sigma),\quad f\mapsto f(P), which is a fiber bundle with fiber \Homeo^{+}(\Sigma,P) , the subgroup fixing P setwise. This gives rise to the following long exact sequence: {\lx@inpgf@ignorespaces\cdots} {\lx@inpgf@ignorespaces\pi_{2}(D_{n}(\Sigma),P)} {\lx@inpgf@ignorespaces\pi_{1}(\Homeo^{+}(\Sigma,P),id)} {\lx@inpgf@ignorespaces\pi_{1}(\Homeo^{+}(\Sigma),id)} {\lx@inpgf@ignorespaces\pi_{1}(D_{n}(\Sigma),P)} {\lx@inpgf@ignorespaces\pi_{0}(\Homeo^{+}(\Sigma,P),id)} {\lx@inpgf@ignorespaces\pi_{0}(\Homeo^{+}(\Sigma),id)} {\lx@inpgf@ignorespaces 1} On the other hand, Hamstrom [ 17 ] proved that \pi_{1}(\Homeo^{+}(\Sigma),id) is trivial whenever \Sigma\notin\{\mathbb{S}^{2},T^{2},\mathbb{R}P^{2},K\} . This yields the short exact sequence 1\to B_{n}(\Sigma,P)\to\Mod(\Sigma,P)\to\Mod(\Sigma)\to 1. For n=1 , this coincides with the classical Birman exact sequence. We summarize these facts in the following lemma. Lemma 2 . For all n\geq 1 , the following hold: (1) B_{n}(\Sigma) is torsion-free if and only if \Sigma\notin\{\mathbb{S}^{2},\mathbb{R}P^{2}\} . (2) If \Sigma\notin\{\mathbb{S}^{2},T^{2},\mathbb{R}P^{2},K\} , then the following exact sequence holds: 1\to B_{n}(\Sigma,P)\to\Mod(\Sigma,P)\to\Mod(\Sigma)\to 1. Proof of Proposition 1 . Suppose there exists a non-identity torsion element f\in\Homeo_{0}(\Sigma) . Choose a point x\in\Sigma in the support of f , and let P be the \langle f\rangle -orbit of x . Then P contains finitely many points, with at least two elements. Since f cyclically permutes the points of P , it represents a nontrivial element of \Mod(\Sigma,P) . However, f represents the identity element in \Mod(\Sigma) . By Lemma 2 ( 2 ), f therefore corresponds to a finite-order element in B_{n}(\Sigma,P) . This contradicts Lemma 2 ( 1 ), which asserts that B_{n}(\Sigma,P) is torsion-free. Hence, no such f exists, and \Homeo_{0}(\Sigma) is torsion-free. ∎ An element of a group is called periodic if it has finite order, and a group is called periodic if all of its elements are periodic. Burnside asked whether every finitely generated periodic group is finite. Many counterexamples have since been constructed by Golod–Shafarevich [ 11 ] , Adian–Novikov [ 27 , 28 , 29 ] , and others. Nevertheless, one can often reduce Burnside-type questions to subgroups of specific groups. For instance, Hurtado–Kocsard–Rodriguez-Hertz [ 18 ] proved that every finitely generated periodic subgroup of smooth area-preserving homeomorphisms of \mathbb{S}^{2} is finite, and Guelman–Liousse [ 13 ] showed Burnside property holds for measure-preserving homeomorphisms on T^{2} . By Tits alternative for mapping class groups of orientable surfaces [ 23 ] , every periodic subgroup is virtually solvable. In the non-orientable case (excluding \mathbb{R}P^{2} and K ), the orientable double cover induces an embedding of the mapping class group into that of an orientable surface [ 4 , 31 , 19 ] , implying that Tits alternative still holds. Since the mapping class groups of \mathbb{R}P^{2} and K are finite, it follows that all mapping class groups of surfaces satisfy Tits alternative. Lemma 3 . Let G be a group satisfying the Tits alternative, i.e., every finitely generated subgroup of G is either virtually solvable or contains a free subgroup of rank 2 . Then every finitely generated periodic subgroup of G is finite. Proof. Let K\leq G be a finitely generated