Arxiv:1907.10166V2 [Math.GR] 15 Jan 2020 57M07

arXiv:1907.10166v2 [math.GR] 15 Jan 2020 57M07. n yteSvr co rgam o ete fEclec i Excellence of Centres for Programme Ochoa Severo the by and hyperboli relatively con groups, automatic of graph group, of class group mapping fundamental group, Hausdorff compact BTATHMMRHSSFO OETOPOLOGICAL SOME FROM HOMOMORPHISMS ABSTRACT .Pofo togvrino hoe 23 32 24 16 18 6 8 continuity automatic for theorems References Combination Groups applications 11. Class General A Mapping Theorem the of to 10. version Homomorphisms strong a of 9. Proof A radical Theorem elliptic of 8. the Proof A and Theorem actions of acylindrical 7. version Universal weak a of 6. Proof B spaces Theorem hyperbolic of 5. on Proof groups of group actions 4. fundamental Acylindric its and earring 3. Hawaiian 2. Introduction 1. 2010 e od n phrases. and words Key h oko h eodato sspotdb uoenResear European by supported is author second the of work The RUST CLNRCLYHPROI GROUPS HYPERBOLIC ACYLINDRICALLY TO GROUPS ahmtc ujc Classification. Subject Mathematics ϕ ufc .I hscs hr xssa pnnra subgroup normal open an exists there case this In Σ. surface H ϕ o di n nvra clnrclato nahyperbolic a by- on hyp a action acylindrically As acylindrical is universal deduced. group any admit also earring not Hawaiian are the groups of that prove graphs of groups damental otnos eas eciehmmrhssfo h Hawaii the from to homomorphisms describe also We continuous. hw ht nacransne ihrteiaeof image the either sense, compact certain topologica countably a a in locally is that, or shown domain metrizable the completely and either hyperbolic acylindrically is Abstract. ( : V G saoeadMdΣ stempigcasgopo once c connected a of group class mapping the is Mod(Σ) and above as epoeamr rcs eutfrhomomorphisms for result precise more a prove We oeatmtccniut eut o eaieyhyperbol relatively for results continuity automatic Some H sfiie eas rv h nlgu ttmn o homomor for statement analogous the prove also We finite. is ) sabove. as → Out( LGBGPLK N AULM CORSON M. SAMUEL AND BOGOPOLSKI OLEG edsrb homomorphisms describe We G ,where ), clnrclyhproi ru,cmltl erzbeg metrizable completely group, hyperbolic acylindrically G saoeeddhproi group. hyperbolic one-ended a is Contents rmr 06,2F7 eodr 07,54H11, 20F70, Secondary 20F67; 20F65, Primary 1 ϕ : H → ru,uieslayidia action. acylindrical universal group, c G ϕ ϕ o hc h codomain the which for iut,Hwia arn group, earring Hawaiian tinuity, ssalor small is & SEV-20150554. R&D n : H hCuclgatPCG-336983 grant Council ch → cgop n fun- and groups ic space. roi,btdoes but erbolic, ru hc is which group l V asoff tis It Hausdorff. neriggroup earring an o() where Mod(Σ), 6 H rdc,we product, ϕ uhthat such salmost is ompact phisms op locally roup, 33 28 21 14 2 2 OLEGBOGOPOLSKIANDSAMUELM.CORSON

1. Introduction It is natural to ask, for which commonly interesting classes A and B of topological groups any abstract homomorphism ϕ : A → B, where A ∈ A and B ∈ B, is continuous, has ﬁnite image, or satisﬁes a weaker variation of these properties. In many interesting cases one can see that, informally speaking,

either the image of ϕ is small or ϕ is almost continuous.

So the famous superrigidity theorem of Margulis [40] says that if G is a real semisimple Lie group G of rank at least 2 with finite center and no nontrivial compact factors, Γ is an irreducible lattice in G, and H is a simple Lie group, then any homomorphism ϕ :Γ → H has finite image or uniquely extends to a continuous homomorphism ϕ : G → H. Some variations of this theorem for non-Lie type groups are given by Farb and Masur [28], Fujiwara [29], Bader and Furman [4]. Another importante result in this direction is a theorem of Nikolov and Segal [44, Theorem 1.1], which says that every abstract homomorphism from a finitely generated in topological sense profinite group to any profinite group is continuous. Papers [23, 42, 14, 20, 48, 38, 17, 20] deal with automatic continuity of abstract homomorphisms from locally compact Hausdorff groups to some discrete groups; papers [17, 20] also deal with completely metrizable groups as domains. One of the novelties of this paper is that we describe homomorphisms ϕ whose domain is an arbitrary completely metrizable or countably locally compact Haus- dorff group and whose target group is an arbitrary acylindrically hyperbolic group (in fact we consider larger classes of domains and targets). Recall that a space is countably compact if every countable open cover admits a finite subcover, and a space X is locally countably compact if for each point x ∈ X there exists a countably compact neighborhood of x in X. Clearly the class LCC of locally countably compact Hausdorff groups includes the class LC of locally compact Hausdorff groups, but the inclusion is strict [35]. The class of locally compact groups has been studied for several decades and is well understood (see, for example, the book [51]). By contrast, almost nothing is known regarding the algebraic-topological structure of a locally countably compact Hausdorff group. It seems that the structure of such groups can be very complicated: there exist two countably compact Hausdorff topological groups whose direct product is not countably compact [35]. The acylindrically hyperbolic groups- those groups which admit a non-elementary acylindrical action on a hyperbolic space- have become a focus of attention in geo- metric group theory. The class of such groups is large enough to admit a great number of groups of classical interest. Examples include non-(virtually cyclic) groups that are hyperbolic relative to proper subgroups, many 3-manifold groups, many groups acting on trees, non-(virtually cyclic) groups acting properly on proper CAT(0)-spaces and containing rank-one elements, non-cyclic directly indecompos- able right-angled Artin groups, all but finitely many mapping class groups, groups of deficiency at least 2, Out(Fn) for n > 2 (see [46, 47] for references and historical remarks). Suppose that a group G acts isometrically on a metric space S. An element g ∈ G is called elliptic for this action if some (equivalently every) orbit of hgi is bounded. We introduce the elliptic radical of an arbitrary group G. ABSTRACT HOMOMORPHISMS FROM SOME TOPOLOGICAL GROUPS 3

Definition 1.1. The elliptic radical of G, denoted by RadEll(G), is the set of all elements g ∈ G which are elliptic for any acylindrical action of G on any hyperbolic space. Note that the elliptic radical of an acylindrically hyperbolic group is, in a certain sense, relatively small (see Remark 6.3). For example, the elliptic radical of a relatively hyperbolic group is contained in the union of the set of elements of finite order and the set of parabolic elements (see Example 6.2). Informally, the main Theorem A says that abstract homomorphisms from certain topological groups to acylindrically hyperbolic groups either have relatively small images or are almost continuous. Theorem A. Let H be a topological group which is either completely metrizable or locally countably compact Hausdorff. Let ϕ : H → G be an abstract homomorphism, where G is an arbitrary group. Then either ϕ(H) is contained in the elliptic radical RadEll(G), or there exists a normal open subgroup V 6 H with finite ϕ(V ). Theorem 8.3, generalizing Theorem A, says that an analogous statement is valid for all finite index subgroups in completely metrizable or locally countably compact Hausdorff topological groups. The second main theorem is Theorem B. It describes homomorphisms from the Hawaiian earring group to acylindrically hyperbolic groups, and its proof is used in the proof of Theorem A. Definitions of the Hawaiian earring group HEG and its subgroups HEGn are given in Section 2. Theorem B. Let G be a group which acts coboundedly and acylindrically on a hyperbolic space S. Then for any abstract group homomorphism ϕ : HEG → G, there exists a natural number n such that ϕ(HEGn) is a subgroup of G consisting of only elliptic elements with respect to this action. Results like Theorems A and B are called atomic properties in the literature (see [25] and [43]). Note that in Theorems A and B we are not requiring the group G to be acyllindrically hyperbolic. For example any group acts coboundedly and acylindrically on a point, but the set of elliptic elements with respect to this action is all of G. Remark 1.2. The conclusions in Theorems B and A cannot be strengthened to say that ker(ϕ) is open, as seen by easy examples below. The free product G = F2 ∗ (Z/2Z) of the free group of rank 2 and the group of order 2 is acylindrically hyperbolic. There exists a homomorphism from HEG to Z/2Z whose kernel does not contain any HEGn (see, for example, [18]), hence the kernel is not open, and so such a homomorphism exists to G. Also, by endowing the direct product N(Z/2Z) with the Tychonov topology (each factor being discrete) we obtain a topological group which is completely Q metrizable and compact Hausdorff. Let ei be the element of N(Z/2Z) with supp(ei) = {i}. Then any homomorphism ϕ : N(Z/2Z) → Z/2Z which sends all Q elements ei to the nontrivial element of Z/2Z, does not have open kernel. Therefore Q there is a homomorphism from N(Z/2Z) to G which does not have open kernel. Remark 1.3. Theorem A impliesQ that if G is an acylindrically hyperbolic group, then G cannot be given a nondiscrete completely metrizable or locally compact Hausdorff group topology. In particular, HEG cannot be given a nondiscrete topology of this kind. 4 OLEGBOGOPOLSKIANDSAMUELM.CORSON

An exact computation of the elliptic radical is, in general, a diﬃcult problem. However, one can compute this radical in some important cases which we list below. Recall that an acylindrical action of a group G on a hyperbolic space S is called universal acylindrical if the set of elliptic elements for this action coincides with the elliptic radical RadEll(G). The universal (cobounded) acylindrical actions of groups on hyperbolic spaces were ﬁrst introduced by Osin in [46] (using another terminology, see Section 6) and studied in [1, 2, 3]. Many interesting groups admit such an action. Among them are the following. (a) Hyperbolic groups. (b) Relatively hyperbolic groups with respect to a collection of subgroups {Hλ}λ∈Λ such that none of the Hλ’s is virtually cyclic or acylindrically hyperbolic. (c) Mapping class groups of compact surfaces. (d) Right-angled Artin groups. (e) Fundamental groups of compact orientable 3-manifolds with empty or toroidal boundary. (f) Hierarchically hyperbolic groups, see [3, Theorem A]. This includes groups that act properly and cocompactly on a proper CAT (0)-complex, in particular, right-angled Artin groups and right-angled Coxeter groups.

Cases (a)-(e) were considered in [46, Example 6.5] and in [2, Theorem 2.18]. Case (f) was considered in [3, Theorem A]. The history and the sequence of appearance of these results are illuminated in [2, 3].

Approaching to Theorem C, we focus on the case (c). Let Σg,b be an orientable surface of genus g > 0 with b > 0 boundary components. In Subsection 9.1, we recall definitions of the mapping class group Mod(Σg,b) and of the curve graph C(Σg,b). It is known that if 3g+b−4 > 0 then the curve graph C(Σg,b) is hyperbolic (Masur- Minsky [41]), the action of Mod(Σg,b) on C(Σg,b) is acylindrical (Bowditch [11]), and that the elliptic elements of this action are precisely elements of finite order and reducible elements (Masur-Minsky [41]). In this case Theorem A says that for any homomorphism ϕ : H → Mod(Σg,b), where H as above, there is an open normal subgroup V of H such that ϕ(V ) consists of elements of finite order and reducible elements. Theorem C strengthens this statement. Theorem C. Let H be a topological group which is either completely metrizable or locally countably compact Hausdorff. Let Σ be a connected compact surface (possibly non-orientable and possibly with boundary components). Then for any homomorphism ϕ : H → Mod(Σ) there exists an open normal subgroup V 6 H such that ϕ(V ) is finite.

The proof of this theorem uses a generalization of Theorem A (see Theorem 8.3) and a theorem of Ivanov (see [36, Theorem 1]) on the structure of subgroups of the mapping class groups.

Remark 1.4. A homomorphism between topological groups ϕ : H → G is called virtually continuous if there exists a ﬁnite index subgroup H1 of H such that the restriction ϕ ↾ H1 is continuous. Using the fact that Mod(Σ) has a ﬁnite index torsion-free subgroup (see [36, Theorem 3]), one can prove that the statement of ABSTRACT HOMOMORPHISMS FROM SOME TOPOLOGICAL GROUPS 5

Theorem C is equivalent to the assertion that any homomorphism ϕ : H → Mod(Σ) is virtually continuous, where Mod(Σ) is considered as a discrete topological group.

± ± Note that Mod (Σ) = Out(π1(Σ)) if Σ is orientable and closed (here Mod (Σ) is the full mapping class group of Σ). Therefore it is natural to ask, whether Theorem C holds for ϕ : H → Out(G), where G is a one-ended hyperbolic group. Combining Theorem C with a result of Levitt [39] about the structure of Out(G), we deduce the following theorem. Theorem D. Let H be a topological group which is either completely metrizable or locally countably compact Hausdorﬀ. Let G be a one-ended hyperbolic group. Then for any homomorphism ϕ : H → Out(G) there exists an open normal subgroup V 6 H such that ϕ(V ) is ﬁnite.

Remark 1.5. It is interesting to compare Theorems C and D with the following results of Farb and Masur, and Fujiwara. In [28, Theorem 1.1] Farb and Masur proved the following. Let H be a real semisimple Lie group of rank at least 2 with finite center and no nontrivial compact factors, let Γ be an irreducible lattice in H, and S be a compact connected orientable surface. Then any homomorphism ϕ : Γ → Mod(S) has finite image. Fujiwara proved the same for nonorientable surfaces and then extended this result to homomorphisms ϕ :Γ → Out(G), where G is a one-ended hyperbolic group [29, Theorems 1 and 3]. Theorems A and B have many other applications. As a corollary of Theorem B, we show that HEG is an acylindrically hyperbolic group which does not admit any universal acylindrical action (see Proposition 6.6). This seems to be the first torsion- free example of this kind. In [1] Abbott proved that Dunwoody’s inaccessible finitely generated group does not have universal acylindrical actions. Her proof extensively uses the fact that this group has unbounded torsion. Problem 1. Does there exist a finitely generated (countable) torsion-free acylindrically hyperbolic group which does not admit universal acylindrical actions? Note that since free groups admit universal acylindrical actions and HEG does not, this gives still another proof that HEG is not a free group. As another corollary of Theorem A, we recover a result of Eda from [24] that any endomorphism of the Hawaiian earring group is induced, up to conjugacy, by a continuous map from the Hawaiian earring to itself preserving the basepoint (0, 0) (see Corollary 10.3). Other applications include automatic continuity results for homomorphisms whose domain is as in Theorem A or B and whose codomain is relatively hyperbolic or the fundamental group of a graph of groups. To present some of them in a short form, we recall the following terminology. A group G is cm-slender if every abstract homomorphism from a completely metrizable topological group H to G has open kernel [17]. Similarly one defines G to be lcH-slender (respectively n-slender) by replacing H with a locally compact Hausdorff group (respectively HEG). The abbreviation lccH-slender stands for locally countably compact slenderness. Free (abelian) groups were classically shown to be cm-slender, lcH-slender and n-slender (see [23] and [34]). More recent work has shown that torsion-free word hyperbolic groups, right-angled Artin groups, braid groups and many other groups 6 OLEGBOGOPOLSKIANDSAMUELM.CORSON satisfy various of these slenderness conditions (see [43], [17], [20], [38]). Note that a group which is either n-slender, cm-slender, or lccH-slender must be torsion-free. We prove the following combination theorems. Theorem E. If G is torsion-free and relatively hyperbolic with respect to a collection {Hλ}λ∈Λ of n-slender (respectively cm-slender, lcH-slender, lccH-slender) subgroups, then G is also n-slender (respectively cm-slender, lcH-slenderm, lccH- slender). Theorem F. Suppose that G is the fundamental group of a (possibly infinite) graph of groups (Γ, G) where each vertex group Gv is n-slender (respectively cm- slender, lcH-slender, lccH-slender) and that the action of G on the Bass-Serre tree associated with (Γ, G) is (k,n)-acylindrical for some k,n ∈ N. Then G is n-slender (respectively cm-slender, lcH-slender, lccH-slender). Remark 1.6. In [38, Theorem D] Kramer and Varghese proved that given an abstract homomorphism ϕ from a locally compact Hausdorff group H to a group G in which torsion subgroups are finite and abelian subgroups are a direct sum of cyclic groups, it is either the case that ker(ϕ) is open or that im(ϕ) lies in the normalizer of a finite non-trivial subgroup of G. Note that the technique they use cannot be applied in the case of completely metrizable or locally countably compact Hausdorff groups. The paper is organized as follows. We first provide some background in Sections 2 and 3. In Section 4 we give a proof of Theorem B, which serves as a basis for the proof of Theorem A. A weak version of Theorem A is proved in Section 5. Universally acylindrical actions and the elliptic radical are discussed in Section 6. Theorem A and its strong version are proved in Sections 7 and 8, respectively. Theorems C and D are proved Section 9. Some general applications are given in Section 10. Theorems E and F are proved in Section 11. We thank Andreas Thom for a useful discussion about a result of Koubi from [37].

2. Hawaiian earring and its fundamental group

We deﬁne N = {1, 2,... } and N0 = N ∪ {0}. The Hawaiian earring is the 2 subspace E = n∈N Cn of the Euclidean plane R where Cn is the circle with center (− 1 , 0) and radius 1 , see Fig. 1. n S n

. . .

Fig. 1. Hawaiian earring We stress that E is a topological space with the topology induced by that of the Euclidean plane. The fundamental group of E with respect to the basepoint (0,0) is denoted by HEG in the present paper. The unusual properties of this group have ABSTRACT HOMOMORPHISMS FROM SOME TOPOLOGICAL GROUPS 7 been explored over the last several decades. Higman showed that HEG is not free (see [34] and [30]). Much later the abelianization of HEG was explicitly computed by Eda and Kawamura [26] and was shown to include exotic direct summands, as well as the group R. We recall a combinatorial description of this group, which somewhat resembles that for free groups (see [24]). For a totally ordered set X let X∗ be the same set with the reverse order. For two totally ordered sets X and Y we define their concatenation as the set X ⊔ Y together with the order which preserves that of both X and Y and places elements in X below those of Y . ±1 Let A = {an }n∈N be a countably infinite set with formal inverses. By W (A) we denote the class of all maps w : X → A with the following two properties: (1) X is a totally ordered set; (2) for each a ∈ A the set {i ∈ X | w(i)= a} is finite.

The elements of W (A) are called words. For a word w : X → A we deﬁne the inverse word w−1 : X∗ → A by setting w−1(i) = (w(i))−1. For two words w : X → A and u : Y → A we deﬁne their concatenation wu : X ⊔ Y → A by

w(i) if i ∈ X wu(i)= . (u(i) if i ∈ Y

Note that W (A) is not a group since W (A) is a class (not a set yet) and since the concatenation ww−1 does not coincide with the trivial map ∅ : ∅ → A. To define a group, we first define an equivalence relation ≡ on W (A): Given two words w : X → A and u : Y → A, we write w ≡ u provided there exists an order isomorphism ϕ : X → Y such that w(i) = u(ϕ(i)). For brevity, we denote the set W (A)/ ≡ again by W (A) and we will term the ≡ classes as words. Now we may assume that W (A) is a set. There are two ways to transform W (A) to a group. The first way utilizes an appropriate definition of (infinite) cancellation in words. The second way (which we describe below) uses restriction maps pn : W (A) → Fn, where n ∈ N0 and Fn is the free group with basis {a1,...,an}. For a word w : X → A we define

±1 ±1 pn(w)= w ↾ {i ∈ X | w(i) ∈{a1 ,...,an }}.

For w,u ∈ W (A) we write w ∼ u provided that for every n ∈ N the words pn(w) and pn(u) are equal as elements in the free group Fn. The quotient set HEG = W (A)/ ∼ possesses a group structure given by concatenation: [w][u] = [wu] and [u]−1 = [u−1]. n For each n ∈ N0, we deﬁne subgroups HEGn and HEG of the group HEG:

• The subgroup HEGn consists of those elements of HEG which have a rep- ±1 ±1 resentative w : X → A satisfying w(X) ⊆{a1 ,...,an }.

• The subgroup HEGn consists of those elements of HEG which have a rep- ±1 ±1 resentative w : X → A satisfying w(X) ∩{a1 ,...,an } = ∅. n Then, for each n ∈ N0, there is a natural decomposition HEG ≃ HEGn ∗ HEG . n Moreover, there are natural isomorphisms HEGn ≃ Fn and HEG ≃ HEG. 8 OLEGBOGOPOLSKIANDSAMUELM.CORSON

The group HEG can be considered as a topological group with respect to the topology T for which the set {ker(pn) | n ∈ N0} is a basis of open sets of neighborhoods of the neutral element. Note that ker(pn) coincides with the normal closure of the subgroup HEGn in HEG.

3. Acylindric actions of groups on hyperbolic spaces In this paper, all actions of groups on metric spaces are supposed to be isometric. Recall that a group G acts coboundedly on a metric space S if there exists a bounded subset B ⊂ S such that S = G · B. We work only with rectifiable paths in metric spaces. In most cases the paths under consideration are just finite concatenations of geodesics. At some point we consider a bi-infinite path, whose construction is given precisely. For a path p, let p− and p+ denote the initial and the terminal points of p and ℓ(p) the length of p. For κ > 1 and ε > 0, we call the path p a (κ,ε)-quasi-geodesic if for any subpath q of p holds 1 d(q−, q+) > ℓ(q) − ε. κ Let M > 0. A path p is called an M-local (κ,ε)-quasi-geodesic if for any subpath q of p with ℓ(q) 6 M the above inequality holds. For two subsets A, B in a metric space S let dHau(A, B) be the Hausdorff distance between A and B. 3.1. Hyperbolic spaces. Let A, B, C be three points in a metric space S. Recall that the Gromov product of A, B with respect to C is the number d(C, A)+ d(C,B) − d(A, B) (A, B) := . C 2 We use the following definition of a δ-hyperbolic space (see [15, Chapter III.H, Definition 1.16 and Proposition 1.17]). For δ > 0, we say that a geodesic triangle ABC in S is δ-thin at the vertex C if for any two points A1 and B1 on the sides [C, A] and [C,B] with d(C, A1) = d(C,B1) 6 (A, B)C , we have d(A1,B1) 6 δ. We say that a metric space S is δ- hyperbolic if it is geodesic and every geodesic triangle in S is δ-thin at each of its vertices. Lemma 3.1. (see [19, Chapitre 3, Théorème 3.1]) For all δ > 0, κ > 1, ε > 0, there exists a constant µ = µ(δ, κ,ε) > 0 with the following property: If S is a δ-hyperbolic space and p and q are infinite (κ,ε)-quasi-geodesics in S with the same limit points on the Gromov boundary ∂S, then the Hausdorff distance between p and q is at most µ(δ, κ,ε). Lemma 3.2. (see [19, Chapitre 3, Théorème 1.4]) Let S be a δ-hyperbolic space. Let k > 1, k′ > k and l > 0. There exists a real number M = M(δ, k, k′,l) > 0 such that every M-local (k,l)-quasi-geodesic is a global (k′,l)-quasi-geodesic.

3.2. Acylindrical actions. Definition 3.3. (see [11]) An action G y S of a group G on a metric space (S, d) is acylindrical if for each ǫ > 0 there exist R,N > 0 such that for any two points p, q ∈ S satisfying d(p, q) > R the set {g ∈ G | d(p,gp) 6 ǫ and d(q,gq) 6 ǫ} ABSTRACT HOMOMORPHISMS FROM SOME TOPOLOGICAL GROUPS 9 is of cardinality at most N. Recall that an action of a group G on a hyperbolic space S is called elementary if the limit set of G on the Gromov boundary ∂G contains at most 2 points. Definition 3.4. (see [46, Definition 1.3]) A group G is called acylindrically hyperbolic if it satisfies one of the following equivalent conditions:

(AH1) There exists a generating set X of G such that the corresponding Cayley graph Γ(G, X) is hyperbolic, |∂Γ(G, X)| > 2, and the natural action of G on Γ(G, X) is acylindrical.

(AH2) G admits a non-elementary acylindrical action on a hyperbolic space.

In the case (AH1), we also write that G is acylindrically hyperbolic with respect to X. Definition 3.5. Given a group G acting on a metric space S, an element g ∈ G is called elliptic if some (equivalently, any) orbit of g is bounded, and loxodromic if the map Z → S defined by n 7→ gnx is a quasi-isometric embedding for some (equivalently, any) x ∈ S. That is, for x ∈ S, there exist κ > 1 and ε > 0 such that for any n,m ∈ Z we have 1 d(gnx, gmx) > |n − m|− ε. κ The set of loxodromic (respectively elliptic) elements of G with respect to this action is denoted by Lox(G y S) (respectively by Ell(G y S)). Bowditch [11, Lemma 2.2] proved that every element of a group G acting acylindrically on a hyperbolic space is either elliptic or loxodromic (see a more general statement in [46, Theorem 1.1]). Every loxodromic element g of such group G is contained in a unique maximal virtually cyclic subgroup [22, Lemma 6.5]. This subgroup, denoted by EG(g), is called the elementary subgroup associated with g; it can be described as follows (see equivalent definitions in [22, Corollary 6.6]): −1 n ±n EG(g)= {f ∈ G | ∃n ∈ N : f g f = g }. Lemma 3.6. (see [46, Lemma 6.8]) Suppose that a group G acts acylindrically on a hyperbolic space S. Then there exists L ∈ N such that for every loxodromic element g ∈ G, EG(g) contains a cyclic subgroup of index at most L.

Deﬁnition 3.7. Suppose that G acts on a metric space (S, d). Let s0 be a point in S. Then we deﬁne a G-equivariant pseudo-metric ds0 on G by the formula ds0 (g,h) = d(gs0,hs0), where g,h ∈ G. We set |g|s0 = ds0 (1,g). We call the function | · |s0 : G → R>0 the length function on G induced by the action of G on the based space (S,s0). Note that if X is a generating set of G, then the length function on G induced by the action of G on the based Cayley graph (Γ(G, X), 1) is the usual length function on G with respect to X. Lemma 3.8. Let G be a group which acts coboundedly and acylindrically on a hyperbolic space S. Let s0 be a point in S and | · |s0 : G → R>0 be the length function on G induced by the action of G on the based space (S,s0). Then there exists a constant K > 0 such that the following holds: For any g ∈ Ell(G y S) −1 n there exists h ∈ G such that |h g h|s0 6 K for any n ∈ Z. 10 OLEGBOGOPOLSKIANDSAMUELM.CORSON

Proof. For any point s ∈ S and any real r > 0 let Br(s) be the open ball in S of radius r with the center at s. Since the action of G on S is cobounded, there exists a constant L> 0 such that S = G · BL(s0). (3.1) Let δ > 0 be a constant such that S is δ-hyperbolic and let g be an element from Ell(G y S). Then there exists a point s ∈ S such that the diameter of the orbit hgis is less than 4δ + 1 (see [46, Corollary 6.7]; the proof can be also extracted from [10]). Thus, hgis ⊆ B4δ+1(s). By (3.1), there exist an element h ∈ G and −1 a point s1 ∈ BL(s0) such that s = hs1. Then h hgihs1 ⊆ B4δ+1(s1). Thus −1 h hgihs0 ⊆ BK (s0) for K =2L +4δ + 1.

In the following we use variants of Lemma 2.10 and Lemma 2.11 from [8]. These lemmas were formulated for groups G and generating sets X of G such that the Cayley graph Γ(G, X) is hyperbolic and G acts acylindrically on Γ(G, X). They can be easily adapted to the case where G acts coboundedly and acylindrically on a hyperbolic space. Lemma 3.9. (see [8, Lemma 2.10]) Let G be a group which acts coboundedly and acylindrically on a hyperbolic space S. Let s0 be a point in S and | · |s0 : G → R>0 be the length function on G induced by the action of G on the based space (S,s0). Then there exists a constant N0 ∈ N such that the following holds: For any a,b ∈ Lox(G y S) with EG(a) 6= EG(b) and for any n,m ∈ N we have that

n m min{n,m} |a b |s0 > . N0 Lemma 3.10. (see [8, Lemma 2.11]) Let G be a group which acts coboundedly and acylindrically on a hyperbolic space S. Let s0 be a point in S and | · |s0 : G → R>0 be the length function on G induced by the action of G on the based space (S,s0). Then there exists a constant N1 ∈ N such that the following holds: For any a ∈ Lox(G y X), for any b ∈ Ell(G y X) \ EG(a), and for any n ∈ N we have that n n |a b|s0 > . N1 Corollary 3.11. Let G be a group which acts coboundedly and acylindrically on a hyperbolic space S. Then there exists a constant C0 ∈ N such that the following holds: For any a,b ∈ Lox(G y S) with EG(a) 6= EG(b) and for any n,m > C0 we have that anbm ∈ Lox(G y S).

Proof. Let s0 be a point in S and | · |s0 : G → R>0 be the length function on G induced by the action of G on the based space (S,s0). We set C0 = ⌈K⌉N0, where K and N0 are the constants from Lemmas 3.8 and 3.9, respectively. Suppose to the contrary that there exist a,b ∈ Lox(G y S) and n,m > C0 such that anbm ∈ Ell(G y S). By Lemma 3.8, there exists g ∈ G such that −1 n m −1 −1 |g (a b )g|s0 6 K. We set a1 = g ag, b1 = g bg. Then a1,b1 ∈ Lox(G y S) n m and |a1 b1 |s0 6 K. This contradicts Lemma 3.9, which says that

n m min{n,m} |a1 b1 |s0 > > K. N0

The proof of the following corollary follows analogously from Lemma 3.10. ABSTRACT HOMOMORPHISMS FROM SOME TOPOLOGICAL GROUPS 11

Corollary 3.12. Let G be a group which acts coboundedly and acylindrically on a hyperbolic space S. There exists a constant C1 ∈ N such that the following holds: n If a ∈ Lox(G y S) and b ∈ Ell(G y S) \ EG(a), then a b ∈ Lox(G y S) for any n > C1.

3.3. Translation lengths and stable norms. The aim of this subsection is to prove Corollary 3.23 that will be used in the proof of Theorem A. Definition 3.13. (see [19] and [31]) Suppose that G is a group acting on a metric space S. Let g be an element of G. The translation length of g is the number [g] defined by [g] = inf d(s,gs). s∈S The stable norm of g is the number ||g|| defined by 1 ||g|| = lim d(s,gns), n→∞ n where s is an arbitrary point of S. One can show that this definition does not depend on the choice of s. This limit exists since the function f : N → R>0, n 7→ d(s,gns), is subadditive. The following lemma is quite trivial. Lemma 3.14. The stable norm satisfies the following properties. d(s,gns) (1) [g] > ||g|| = inf . n∈N n (2) ||x−1gx|| = ||g|| for any x ∈ G. (3) ||gk|| = |k| · ||g|| for any k ∈ Z. Lemma 3.15. [19, Chapitre 10, Proposition 6.4] Let S be a δ-hyperbolic space and g a loxodromic isometry of S. Then ||g|| 6 [g] 6 ||g|| + 32δ. Definition 3.16. Let G be a group acting on a metric space S and let g be a loxodromic element in G and s be a point in S. We choose a geodesic path p0 with i (p0)− = s and (p0)+ = gs and set pi = g p0 for any i ∈ Z. The bi-infinite path ...p−1p0p1 ... is denoted by L(s,g). For ε > 0 we call the path L(s,g) an ε-line of g if d(s,gs) 6 [g]+ ε. We define an ε-axis of g as follows:

Aε(g)= {x ∈ S | d(x, gx) 6 [g]+ ε}. Clearly, any ε-line of g lies in the ε-axis of g. Lemma 3.17. (see [21, Proposition 2.28]) Let g be a loxodromic isometry of a δ-hyperbolic space S and s be an arbitrary point of S. Then

d(s,gs) > 2d(s, A8δ(g)) + [g] − 6δ. Deﬁnition 3.18. Let G be a group acting on a metric space. The injectivity radius for the action of G on S is the number inj (G y S) = inf{||g|| : g ∈ Lox(G y S)}. Bowditch proved the following important lemma. 12 OLEGBOGOPOLSKIANDSAMUELM.CORSON

Lemma 3.19. ([11, Lemma 2.2]) If G acts acylindrically on a hyperbolic space S, then inj (G y S) > 0. We deduce from this lemma the following corollary. Corollary 3.20. Let G be a group which acts acylindrically on a δ-hyperbolic space S. Then for any ε > 0, there exist constants κ = κ(ε) > 1 and τ = τ(ε) > 0 such that for any g ∈ Lox(G y S) we have that any ε-line of g is a (κ, τ)-quasi- geodesic. Proof. We set M = M(δ, 1, 2,ε+δ), see Lemma 3.2. Let g be a loxodromic element i of G and let L(s,g) be an ε-line of g. We call the points Ai = g s the phase points of L(s,g). Let pi be the subpath of L(s,g) with (pi)− = Ai and (pi)+ = Ai+1. Clearly, pi is a geodesic. Case 1. Suppose that [g] >M. We claim that L(s,g) is an M-local (1,ε + δ)-quasi-geodesic. Indeed, let q be a subpath of L(s,g) of length less than or equal to M. If q is contained in some pn, then q is a geodesic and the claim holds. Suppose that q is not contained in any pn. Since ℓ(pn) >M for any n and ℓ(q) 6 M, the path q decomposes as q = q1q2, where q1 is a terminal subpath of some pi−1 and q2 is an initial subpath of some pi. Let q3 be a geodesic path connecting q− and q+, see Fig. 2.

Ai−1 Ai Ai+1 g−1Y X Y q− q+

Fig. 2. A part of the ε-line L(s,g).

Let ∆ be the geodesic triangle with the sides q1, q2, q3 and let X ∈ q1 and Y ∈ q2 be the points such that d(Ai,X) = d(Ai, Y ) = (q−, q+)Ai . Since ∆ is δ-thin, we have d(X, Y ) 6 δ. ε+δ Subcase 1. Suppose that d(Ai,X) > 2 . Then d(g−1Y, Y ) 6 d(g−1Y,X)+ d(X, Y )

= d(Ai−1, Ai) − 2d(X, Ai)+ d(X, Y ) < ([g]+ ε) − (ε + δ)+ δ = [g], a contradiction. ε+δ Subcase 2. Suppose that d(Ai,X) 6 2 . Then

d(q−, q+)= ℓ(q1)+ ℓ(q2) − 2(q−, q+)Ai = ℓ(q) − 2d(Ai,X) > ℓ(q) − (ε + δ). Therefore L(s,g) is an M-local (1,ε + δ)-quasi-geodesic. By Lemma 3.2, L(s,g) is a global (2,ε + δ)-quasi-geodesic. Case 2. Suppose that [g] 6 M. Denote θ = inj (G y S). By Lemmas 3.14 and 3.19 we have [gn] > ||gn|| = n||g|| > nθ > 0 (3.2) for all natural n. In particular, M > [g] > θ. (3.3) ABSTRACT HOMOMORPHISMS FROM SOME TOPOLOGICAL GROUPS 13

3(M+ε) We prove that L(s,g) is an (α, β)-quasi-geodesic for α = θ and β = 2(M +ε). Let p be a subpath of L(s,g). If p does not contain phase points of L(s,g), then p is a geodesic path, and hence a (1, 0)-quasi-geodesic. Suppose that p contains at least i one phase point of L(s,g). Let q be the maximal subpath of p such that q− = g s i+n and q+ = g s for some i ∈ Z and n ∈ N0. If n = 0 then ℓ(p) < 2([g]+ ε) and it is easy to check that 1 d(p−,p ) > 0 > ℓ(p) − β. + α If n> 0 then d(p−,p+) > d(q−, q+) − 2([g]+ ε) > [gn] − 2([g]+ ε) > nθ − 2([g]+ ε) 1 > α (3n(M + ε)) − β 1 > α (n + 2)([g]+ ε) − β 1 > α ℓ(p) − β. Thus, L(s,g) is an (α, β)-quasi-geodesic in this case. In both cases L(s,g) is a (κ, τ)-quasi-geodesic for κ = max{2, α} and τ = max{ε + δ, β}.

Lemma 3.21. Let G be a group which acts acylindrically on a δ-hyperbolic space S. For any ε > 0, there exists a constant ν = ν(δ, ε) > 0 such that for any loxodromic element g ∈ G and any nonzero integers n,m, we have n m dHau(Aε(g ), Aε(g )) 6 ν(δ, ε).

n m Proof. Let s ∈ Aε(g ) and t ∈ Aε(g ) be arbitrary points. By Corollary 3.20, the bi-inﬁnite paths L(s,gn) and L(t,gm) are (κ(ε), τ(ε))-quasi-geodesics. Since these quasi-geodesics have the same limit points on the Gromov boundary of S, then, by Lemma 3.1, there exists a number µ = µ(δ, κ(ε), τ(ε)) such that n m dHau(L(s,g ),L(t,g )) 6 µ. n m m Since s ∈ L(s,g ) and L(t,g ) ⊆ Aε(g ), we have m d(s, Aε(g )) 6 µ, and the proof is completed.

Proposition 3.22. Let G be a group which acts acylindrically on a δ-hyperbolic space S. There exists a constant c0 > 0 with the following property: Let s0 be a point in S and | · |s0 : G → R>0 be the length function on G induced by the action of G on the based space (S,s0). Then for any loxodromic element g ∈ G and any nonzero integers n and m we have n m n m |g |s0 − |g |s0 > [g ] − [g ] − c0. In particular, n m |g |s0 − |g |s0 > (|n| − |m|) ||g|| − (c0 + 32δ). 14 OLEGBOGOPOLSKIANDSAMUELM.CORSON

n Proof. Let λ> 0 be an arbitrary real number. Then there exists a point u ∈ A8δ(g ) such that n d(s0,u) < d(s0, A8δ(g )) + λ. Using Lemma 3.17, we have n n n n |g |s0 = d(s0,g s0) > 2d(s0, A8δ(g )) + [g ] − 6δ (3.4) n > 2d(s0,u) + [g ] − (6δ +2λ). m By Lemma 3.21, there exists a point v ∈ A8δ(g ) such that d(u, v) 6 ν(δ, 8δ). Then m m m m m |g |s0 = d(s0,g s0) 6 d(s0, v)+ d(v,g v)+ d(g v,g s0) m =2d(s0, v)+ d(v,g v) (3.5) m 6 2(d(s0,u)+ d(u, v)) + d(v,g v) m 6 2 d(s0,u)+ ν(δ, 8δ) + [g ]+8δ . It follows from (3.4) and (3.5) that n m n m |g |s0 − |g |s0 > [g ] − [g ] − 14δ +2ν(δ, 8δ)+2λ .

Since λ > 0 was arbitrary, we can set c0 = 14δ +2ν(δ, 8δ) to satisfy the ﬁrst inequality in the proposition. The second inequality follows from the ﬁrst one with the help of Lemmas 3.14 and 3.15, which give the following estimations: [gn] > ||gn|| = |n| · ||g||, [gm] 6 ||gm|| + 32δ = |m| · ||g|| + 32δ.

Using the fact that ||g|| > inj (G y S) > 0, we deduce from Proposition 3.22 the following corollary. Corollary 3.23. Let G be a group which acts acylindrically on a hyperbolic space S. There exist constants α > 1 and β > 0 such that the following holds: Let s0 be a point in S and | · |s0 : G → R>0 be the length function on G induced by the action of G on the based space (S,s0). Then for any loxodromic element g ∈ G and any natural number n we have 1 |gn| > |g| + n − β. s0 s0 α

4. Proof of Theorem B

Let s0 be a point in S and | · |s0 : G → R>0 be the length function on G induced by the action of G on the based space (S,s0). To shorten the proof, we write |g| instead of |g|s0 , Lox(G) instead of Lox(G y S), and Ell(G) instead of Ell(G y S). n Suppose the contrary. Then for any n ∈ N, there exists un ∈ HEG such that n mi ϕ(un) ∈ Lox(G). We deﬁne w1 ∈ HEG inductively by the formula wi = ui wi+1, where n and mi are appropriate numbers, which we choose below. Thus,

n n m2 m1 w1 = u1 (u2 (... ) ) . n mi Denote Ui = ϕ(ui) and Wi = ϕ(wi). We have Wi = Ui Wi+1. We have that Ui ∈ Lox(G), but we cannot assert that Wi ∈ Lox(G). ABSTRACT HOMOMORPHISMS FROM SOME TOPOLOGICAL GROUPS 15

• By Lemma 3.6, EG(Ui) contains a cyclic subgroup of index at most L. Let zi be a generator of such a subgroup. Then there exists an integer Ki such that L! Ki Ui = zi . (4.1) −1 Replacing zi by zi if necessary, we may assume that Ki ∈ N. In the following deﬁnition of numbers n and mi for i ∈ N we use • the numbers C0 and C1 from Corollaries 3.11 and 3.12, and • the numbers α > 1 and β > 0 from Corollary 3.23. Increasing α and β if necessary, we may assume that α, β ∈ N.

We set N = max{C0, C1} and n n = NL!,Mi = max{|Ui |,i +1} , mi =2αβnMiKi for i ∈ N. (4.2) l m Claim 1. For all i ∈ N, we have Wi ∈ Lox(G).

n mi Proof. Recall that Wi = Ui Wi+1. We consider four cases. Case 1. Suppose that Wi+1 ∈ Lox(G) and EG(Ui) 6= EG(Wi+1). Then Wi ∈ Lox(G) by Lemma 3.11.

Case 2. Suppose that Wi+1 ∈ Lox(G) and EG(Ui) = EG(Wi+1). Then there exists an integer Pi such that L! Pi Wi+1 = zi . (4.3) From (4.1) and (4.3), and using (4.2), we deduce

n mi NKi+2αβNMiKiPi Wi = Ui Wi+1 = zi .

Hence Wi is a nontrivial power of a loxodromic element. Therefore Wi ∈ Lox(G). mi Case 3. Suppose that Wi+1 ∈ Ell(G) and Wi+1 ∈/ EG(Ui). Then Wi ∈ Lox(G) by Corollary 3.12.

• Thus, if Wi+1 satisﬁes assumptions of one of the Cases 1-3, then Wi ∈ Lox(G). mi Case 4. Suppose that Wi+1 ∈ Ell(G) and Wi+1 ∈ EG(Ui). It follows that n mi 1) Wi = Ui Wi+1 ∈ EG(Ui), mi 2) Wi+1 has a ﬁnite order dividing L!.

• The element Wi+2 cannot satisfy assumptions of Cases 1-3, since Wi+1 ∈ Ell(G). Thus, Wi+2 satisﬁes assumptions of Case 4, and we have

1’) Wi+1 ∈ EG(Ui+1). Since Wi+1 is an elliptic element in EG(Ui+1), its order divides L!. Since L! is mi n a divisor of mi, we have Wi+1 = 1. Then Wi = Ui ∈ Lox(G). The proof of the claim is completed.

By Corollary 3.23, we have m |W mi | > |W | + i − β. i+1 i+1 α n mi From (4.2) we deduce |Ui | 6 Mi 6 2α . Then m |W | = |U nW mi | > |W mi | − |U n| > |W | + i − β > |W | + i. i i i+1 i+1 i i+1 2α i+1 It follows that i(i + 1) |W | > |W | +1 > |W | +2+1 > ··· > |W | + i + (i − 1)+ ··· +1 > 1 2 3 i+1 2 16 OLEGBOGOPOLSKIANDSAMUELM.CORSON for all i ∈ N. A contradiction.

5. Proof of a weak version of Theorem A

In this section we prove the following weak version of Theorem A. Proposition 5.1. Let H be a topological group which is either completely metrizable or locally countably compact Hausdorff, G be a group which acts coboundedly and acylindrically on a hyperbolic space S, and ϕ : H → G be a group homomorphism. Then there is an open neighborhood V of identity 1H such that ϕ(V ) ⊆ Ell(G y S). Preparing to the proof, we give the following definition, notation and define some constants.

Deﬁnition 5.2. We call an element h ∈ H pre-loxodromic if ϕ(h) ∈ Lox(G y S).

Notation 5.3. For a subset U ⊆ H and a natural number k we denote U k := {hk | h ∈ U},U 1/k := {h ∈ H | hk ∈ U}.

Constants 5.4. Select universal constants L, C0, C1 for the group G from Lemma 3.6, Corollary 3.11, and Corollary 3.12. Also select numbers α > 1 and β > 0 from Corollary 3.23, without loss of generality both being natural numbers. As in the proof of Theorem B we let N = max{C0, C1} and n = NL!.

Proof of Proposition 5.1. The proof is a modiﬁcation of the proof of Theorem A. (a) Assume ﬁrst that H is completely metrizable. Suppose for contradiction that the conclusion of Theorem B fails, i.e. that any open neighborhood V of 1H contains a pre-loxodromic element. This implies the following claim, which will be used several times.

Claim 1. For any open neighborhood O of 1H and any natural k, the subset O1/k contains a pre-loxodromic element.

k Proof. There is an open neighborhood V of 1H such that V ⊆ O. By assumption, V contains a pre-loxodromic element.

Let d be a complete metric which gives the topology on H. We will construct a sequence (ui)i∈N of pre-loxodromic elements of H and a sequence (Oi)i∈N of open neighborhoods of 1H with special properties, which will ﬁnally lead to a contradiction. Let O1 be an arbitrary open neighborhood of 1H , say O1 = H. 1/n Step 1. Select a pre-loxodromic element u1 ∈ O1 . We set U1 = ϕ(u1) and deﬁne the numbers K1,M1 and m1 as in (4.1) and (4.2). Select an open neighborhood O2 of 1H such that 1 d(un, un Om1 ) 6 . 1 1 2 2 ABSTRACT HOMOMORPHISMS FROM SOME TOPOLOGICAL GROUPS 17

1/n Step 2. Select a pre-loxodromic element u2 ∈ O2 . We set U2 = ϕ(u2) and deﬁne the numbers K2,M2 and m2 as in (4.1) and (4.2). Select an open neighborhood O3 of 1H such that n n m1 n n m2 m1 1 d(u1 (u2 ) , u1 (u2 O3 ) ) 6 22 , n n m2 1 d(u2 , u2 O3 ) 6 22 . Step i. Suppose that in previous steps we have already selected pre-loxodromic elements u1,...,ui−1, numbers K1,...,Ki−1 and M1,...,Mi−1 and m1,...,mi−1, 1/n and neighborhoods O1,..., Oi. Select a pre-loxodromic element ui ∈ Oi . We set Ui = ϕ(ui) and deﬁne the numbers Ki,Mi and mi as in (4.1) and (4.2). Select an open neighborhood Oi+1 of 1H such that

n n n mi−1 m2 m1 n n n mi mi−1 m2 m1 1 d u1 (u2 (··· (ui ) ··· ) ) , u1 (u2 (··· (ui Oi+1) ··· ) ) 6 2i ,

n n mi−1 m2 n n mi mi−1 m2 1 d u2 (··· (ui ) ··· ) , u2 (··· (ui Oi+1) ··· ) 6 2i , . . n n mi 1 d ui , ui Oi+1 6 2i . For any r, i ∈ N, where i > r, we deﬁne the element n n n n mi−1 mr+1 mr wr,i = ur (ur+1(··· ui−1(ui ) ··· ) ) .

Claim 2. For any r ∈ N, the sequence (wr,i)i>r is Cauchy and hence converges.

n Proof. Using ui+1 ∈ Oi+1, we deduce from the last group of inequations that 1 d(w1,i, w1,i+1) 6 2i , 1 d(w2,i, w2,i+1) 6 2i , . . 1 d(wi,i, wi,i+1) 6 2i . Then, for any j>i > r, we have 1 1 1 d(w , w ) 6 d(w , w )+ ··· + d(w − , w ) 6 + ··· + < . r,i r,j r,i r,i+1 r,j 1 r,j 2i 2j−1 2i−1

n mr Let wr = lim wr,i. Passing to limit in the equation wr,i = ur w +1 , which holds i→∞ r ,i for any i > r + 1, we obtain n mr wr = ur wr+1. Now perform the same argument as in the proof of Theorem B on the images under ϕ of the ur and wr for a contradiction. (b) Now we assume that H is locally countably compact Hausdorﬀ. Suppose for contradiction that the conclusion of Theorem B fails, i.e. that any open neighborhood V of 1H contains a pre-loxodromic element. We will construct a sequence (ui)i∈N of pre-loxodromic elements of H and a sequence (Oi)i∈N of open neighborhoods of 1H with special properties, which will ﬁnally lead to a contradiction. Let O1 be an open neighborhood of 1H with J ⊇ O1 a countably compact subspace of X. 18 OLEGBOGOPOLSKIANDSAMUELM.CORSON

1/n Step 1. Select a pre-loxodromic element u1 ∈ O1 . We set U1 = ϕ(u1) and deﬁne the numbers K1,M1 and m1 as in (4.1) and (4.2). Select an open neighborhood O2 of 1H for which n m1 O1 ⊇ u1 O2 .

Step i. Suppose that in previous steps we have already selected pre-loxodromic elements u1,...,ui−1, numbers K1,...,Ki−1 and M1,...,Mi−1 and m1,...,mi−1, 1/n and neighborhoods O1,..., Oi. Select a pre-loxodromic element ui ∈ Oi . We set Ui = ϕ(ui) and deﬁne the numbers Ki,Mi and mi as in (4.1) and (4.2). Select an open neighborhood Oi+1 of 1H for which

n mi Oi ⊇ ui Oi+1.

n m1 Let J1 = O1 ∩ J, J2 = u1 O2 ∩ J and for i> 2 letting

i− n n n m 1 mi−2 m2 m1 Ji = u1 (u2 (··· (ui−1Oi ) ··· ) ) ∩ J, it is clear that J ⊇ J1 ⊇ J2 ⊇ · · · and so the sequence {Ji}i∈N is a nesting sequence of nonempty closed sets in the countably compact Hausdorﬀ space J. Let w1 ∈ i∈N Ji. Select i ∈ N large enough that i > |ϕ(w1)|. Select wi+2 ∈ Oi+2 for which T n n n mi+1 mi m2 m1 w1 = u1 (u2 (··· (ui+1wi+2 ) ··· ) ) . n mi+1 n mi n m2 m1 Let wi+1 = ui+1wi+2 , wi = ui wi+1,...,w2 = u2 w3 . Clearly w1 = u1w2 . As in Claim 1 of the proof of Theorem A we see that ϕ(w1),...,ϕ(wi+1) are loxodromic and arguing as in the conclusion of the proof of Theorem A we get |ϕ(w1)| > i, a contradiction.

6. Universal acylindrical actions and the elliptic radical

In this section we introduce the notion of the elliptic radical RadEll(G) of a group G. This notion is important for two reasons. The first one is that it naturally appears in Theorem A. The second one is that an important class of actions - universal acylindrical actions (introduced by Osin in [46]) - can be alternatively defined via the elliptic radical. Knowing that G admits such an action can help in computing the elliptic radical of G. As it was mentioned in the introduction, many interesting groups admit such an action. In Proposition 6.6 we show that the the Hawaiian earring group HEG is acylindrically hyperbolic, the elliptic radical of HEG is trivial, and that HEG does not admit any universal acylindrical action. Note that the latter claim is deduced from Theorem B. We also prove Lemma 6.9 which is used in the proof of Theorem A. Definition 6.1. Let G be a group. An element g ∈ G is called universally elliptic if for any hyperbolic space S and any acylindrical action of G on S the element g acts elliptically on S. The set

RadEll(G)= {g ∈ G | g is universally elliptic} is called the elliptic radical of G.

Example 6.2. We consider the case where G is relatively hyperbolic. ABSTRACT HOMOMORPHISMS FROM SOME TOPOLOGICAL GROUPS 19

(a) If G is a relatively hyperbolic group with respect to a collection of subgroups {Hλ}λ∈Λ, then U1 ⊆ RadEll(G) ⊆U1 ∪U2,

where U1 is the set of elements of finite order of G and U2 is the union of conjugates of all Hλ’s. We prove the second inclusion; the first one is obvious. Let X be a finite relative generating set of G with respect to {Hλ}λ∈Λ. We set H = ⊔ Hλ. λ∈Λ Then the Cayley graph Γ(G, X ∪ H) is hyperbolic (see [45, Corollary 2.54]) and G acts acylindrically on Γ(G, X ∪H) (see [46, Proposition 5.2]). By [45, Theorem 4.25], the set of elliptic elements with respect to this action lies in U1 ∪U2. This proves the second inclusion. (b) If G is a hyperbolic group, then

RadEll(G)= U1. This follows directly from (a). (c) If G is a relatively hyperbolic group with respect to a collection of subgroups {Hλ}λ∈Λ, where none of the Hλ’s is virtually cyclic or acylindrically hyperbolic, then RadEll(G)= U1 ∪U2. Indeed, if G acts acylindrically on a hyperbolic space S, then the restriction of this action of each Hλ is elliptic by [46, Theorem 1.1]. Together with (a) this completes the proof. Remark 6.3. The following shows that, in a certain sense, the elliptic radical of an acylindrically hyperbolic group is relatively small. (a) In [50] Sisto proved that for any acylindrically hyperbolic group G generated by a ﬁnite symmetrized set X, the probability that the simple random walk on the Cayley graph Γ(G, X) arrives at an element of the elliptic radical of G in n steps is O(εn) for some ε ∈ (0, 1). We don’t know how to generalize this statement for the case where G is not ﬁnitely generated. (b) Suppose that G acts acylindrically on a δ-hyperbolic space S. Then any subgroup H which is contained in the elliptic radical of G has bounded orbits (see [46, Theorem 1.1]) and therefore there exists a point s ∈ S such that the diameter of the orbit Hx is at most 4δ + 1 (see [46, Corollary 6.7]).

To go further, we recall some relevant definitions from [46]. Definition 6.4. In [46, Definition 6.4] Osin gives the following definition: Let G be a group. An element g ∈ G is called generalized loxodromic if it satisfies either of the four equivalent conditions in [46, Theorem 1.4]. Two of these conditions are the following:

(L1) There exists a generating set X of G such that the corresponding Cayley graph Γ(G, X) is hyperbolic, the natural action of G on Γ(G, X) is acylindrical, and g is loxodromic.

(L2) There exists an acylindrical action of G on a hyperbolic space such that g is loxodromic. Because of quantifiers, we prefer to call these elements existentially loxodromic. Comparing with our Definition 6.1, we conclude that any group G is the disjoint 20 OLEGBOGOPOLSKIANDSAMUELM.CORSON union of the set of universally elliptic and the set of existentially loxodromic elements. Osin notes that, in general, generalized loxodromic elements of a group G may be elliptic for some acylindrical actions of G on unbounded hyperbolic spaces. This happens, for instance, for G = F2. However for F2 there exists a single acylindrical action on a hyperbolic space with respect to which all generalized loxodromic elements of F2 are loxodromic. This justifies the following definition. Definition 6.5. (see [46, after Definition 6.4]) An action of a group G on a hyperbolic space is called universal acylindrical if it is acylindrical and all generalized loxodromic elements of G are loxodromic with respect to this action. In other words, an action of a group G on a hyperbolic space is called universal acylindrical if it is acylindrical and the set of elliptic elements of this action coincides with RadEll(G). Many interesting groups admit such an action, see groups in (a)-(f) from in the introduction; these groups were considered in [2, 3, 46]. In [46, Example 6.10], Osin gives a countable (but not finitely) generated group which does not admit any universal acylindrical action. In [1], Abbott shows that Dunwoody’s finitely generated inaccessible group also does not admit such an action. Both proofs ultimately use the presence of elements of increasing finite order and Lemma 3.6. The following seems to be the first torsion-free example of this kind. Proposition 6.6. The group HEG is acylindrically hyperbolic, the elliptic radical RadEll(HEG) is trivial, and HEG does not admit any universal acylindrical action. Proof. The group HEG is acylindrically hyperbolic. To see this, we use the de- n composition HEG = HEGn ∗ HEG , which holds for any n ∈ N as mentioned in Section 2. Let Γn be the Cayley graph of HEG with respect to the generating n set Yn = {a1,a2,...,an} ∪ HEG . This Cayley graph is tree-like, and hence hyperbolic, and it is easy to check that HEG acts acylindrically and nonelementary on Γn. n More particularly the group HEG has relative presentation HEG = hYn ∪HEG | ∅ ∪ Rel(HEGn)i, where Rel(HEGn) is the set of all relations of HEGn with respect to the generating set consisting of all elements of HEGn. This relative presentation has trivial relative Dehn function. Therefore HEG is relatively hyperbolic with respect to the subgroup HEGn (see [45, Definition 2.35]). Hence, HEG is n acylindrically hyperbolic with respect to Yn ∪ HEG (see Definition (AH1)) by [46, Proposition 5.2]. Moreover,

n Ell (HEG y Γn)= g (HEG ). ∈ g [HEG To prove that RadEll(HEG) = 1, it suﬃces tob prove that

∩ Ell (HEG y Γn)=1. n∈N To the contrary, let x be a nontrivial element from this intersection. Then, −1 n for every n ∈ N, there exists gn ∈ HEG such that gn xgn ∈ HEG . We denote −1 −1 −1 hn = gn xgn and zn = gn gn+1. Then zn hnzn = hn+1. By malnormality of n n n HEG in HEG = HEGn ∗ HEG , we have zn ∈ HEG . Now we consider the element ABSTRACT HOMOMORPHISMS FROM SOME TOPOLOGICAL GROUPS 21

∞ n ∞ w = zi. For any n we can write w = unvn, where un = zi, vn = zi. i=1 i=1 i=n+1 Then Q Q Q −1 −1 −1 −1 w h1w = vn un h1unvn = vn hn+1vn n+1 n lies in HEG for any n, hence it lies in the intersection of all HEG , i.e. h1 = 1. This contradicts the assumption that x 6= 1. Thus, RadEll(HEG) = 1. Suppose that HEG admits a universal acylindrical action HEG y S on a hyperbolic space S. Then Ell (HEG y S) = 1. This contradicts Theorem B applied to this action and the identity homomorphism id : HEG → HEG.

Remark 6.7. Since free groups admit universal acylindrical actions, this gives still another proof that HEG is not a free group.

Sometimes it is more convenient to work with cobounded actions. So, we give a variation of Deﬁnition 6.1 for cobounded actions. Deﬁnition 6.8. Let G be a group. An element g ∈ G is called c-universally elliptic if for any hyperbolic space S and any cobounded acylindrical action of G on S the element g acts elliptically on S. We set

CRadEll(G)= {g ∈ G | g is c-universally elliptic}. The following lemma is used in the proof of Theorem A.

Lemma 6.9. For any group G holds RadEll(G) = CRadEll(G).

Proof. We show that CRadEll(G) ⊆ RadEll(G); the inverse inclusion is obvious. Suppose that g ∈ G \ RadEll(G). Since g is not universally elliptic, g is existentially loxodromic and by condition (L1) there exists a generating set X of G such that the corresponding Cayley graph Γ(G, X) is hyperbolic, the natural action of G on Γ(G, X) is acylindrical, and g is loxodromic. Since G acts coboundedly on Γ(G, X), we have g ∈ G \ CRadEll(G).

7. Proof of Theorem A To complete the proof of Theorem A, we need some additional statements about groups which act acylindrically on hyperbolic spaces.

7.1. Normalizers of elliptic subgroups of acylindrically hyperbolic groups. We use the following two results of Osin. Theorem 7.1. (see [46, Theorem 1.1]) Let G be a group acting acylindrically on a hyperbolic space. Then G satisfies exactly one of the following three conditions. (a) G has bounded orbits. (b) G is virtually cyclic and contains a loxodromic element. (c) G contains infinitely many loxodromic elements whose limit sets are pair- wise disjoint. In this case the action of G is non-elementary and G is acylindrically hyperbolic. Recall [46] that a subgroup E of G is called s-normal in G if |Eg ∩ E| = ∞ for every g ∈ G. In particular, if E is infinite, then E is s-normal in NG(E). 22 OLEGBOGOPOLSKIANDSAMUELM.CORSON

Lemma 7.2. (see [46, Lemma 7.1]) Let G be a group acting acylindrically and non-elementary on a hyperbolic space S. Then every s-normal subgroup of G acts non-elementary on S. Lemma 7.3. Let G be a group acting acylindrically on a hyperbolic space S. Sup- pose that E is an inﬁnite subgroup of G which is contained in the set Ell(G y S). Then the normalizer NG(E) is also contained in Ell(G y S).

Proof. Suppose that NG(E) is not contained in Ell(G y S). Applying Theo- rem 7.1 to NG(E), we conclude that either

1) NG(E) contains a finite index cyclic subgroup Z generated by a loxodromic element, or 2) NG(E) acts acylindrically and non-elementary on S. Consider the first case. Since E is elliptic, we have E ∩ Z = 1. But two infinite subgroups of a virtually cyclic group cannot intersect trivially. A contradiction. Consider the second case. Since E is infinite, it is s-normal in NG(E). Then, by Lemma 7.2, E acts non-elementary on S. A contradiction. Recall that a subset X of a group G is called symmetrized if X = X−1. Koubi [37] proved the following proposition.

Proposition 7.4. (see [37, Proposition 3.2]) Let G be a group which acts by isometries on a δ-hyperbolic metric space S. Let X be a finite symmetrized subset of G such that each element of X ∪X ·X acts elliptically on S. Then there exists a point s ∈ S such that d(s,xs) 6 100δ for every x ∈ X. Now we strengthen the proposition of Koubi in the case of acylindrical actions. Proposition 7.5. Let G be a group which acts acylindricaly by isometries on a δ-hyperbolic metric space S. Then there exists µ> 0 such that the following holds. Let X be a (possibly infinite) symmetrized subset of G such that each element of X ∪ X · X acts elliptically on S. Then there exists a point s ∈ S such that d(s,xs) 6 µ for every x ∈ X. Proof. We assume that X is infinite. Let R,N > 0 be the numbers for ε = 100δ +1 from Definition 3.3 of an acylindrical action. We may assume that N is a natural number. Let Y be a finite subset of X such that |Y | = N + 1. By Proposition 7.4, there exists s ∈ S such that d(s,ys) 6 ε for every y ∈ Y . Now we take an arbitrary x ∈ X. Again there exists s1 ∈ S such that d(s1,ys1) 6 ε for every y ∈ Y ∪{x}. By acylindrical hyperbolicity, we have d(s,s1) < R. Then

d(s, x(s)) 6 d(s,s1)+ d(s1, xs1)+ d(xs1, xs) 6 ε +2R.

7.2. Proof of Theorem A.. Suppose that ϕ(H) is not contained in RadEll(G). Then there exists a hyperbolic space S and a cobounded acylindrical action G y S such that ϕ(H) is not contained in Ell(G y S), see Lemma 6.9. It follows from Proposition 5.1 that there exists an open symmetrized neighborhood O of 1H with the property ϕ(O(2)) ⊆ Ell(G y S). (7.1) ABSTRACT HOMOMORPHISMS FROM SOME TOPOLOGICAL GROUPS 23

Let µ be the constant from Proposition 7.5. By this proposition, there exists a point s ∈ S such that d(s, ϕ(x)s) 6 µ for every x ∈ O. Let R,N > 0 be the numbers such that for any two points p, q ∈ S satisfying d(p, q) > R the set {g ∈ G | d(p,gp) 6 µ and d(q,gq) 6 µ} (7.2) is of cardinality at most N. Since ϕ(H) is not contained in Ell(G y S), there exists a loxodromic element ϕ(h) ∈ ϕ(H) (see Theorem 7.1). We take a natural n such that d(s, ϕ(h)ns) > R. Applying (7.2) to p = s and q = ϕ(hn)s, we obtain n |ϕ(O) ∩ ϕ(O)ϕ(h )| 6 N. n h We set U = O ∩O . Then U is an open neighborhood of 1H satisfying |ϕ(U)| 6 N and ϕ(U) ⊆ Ell(G y S). We may assume that |ϕ(U)| is minimal possible. Then

′ ′ ′ for any open neighborhood U of 1H with U ⊆ U we have ϕ(U )= ϕ(U). (7.3)

We show that the subgroup V = hUi · ker(ϕ) satisﬁes the conclusion of Theo- rem A. Clearly V is open. Condition (7.3) implies that ϕ(U) is a subgroup. Then, for any z ∈ H, we have ϕ(V )= ϕ(U)= ϕ(U ∩ V z) ⊆ ϕ(V z)= ϕ(V )ϕ(z). This implies that ϕ(V ) is normal in ϕ(H); hence V is normal in H. Moreover, ϕ(V ) = ϕ(U) ⊆ Ell(G y S). Recall that we have assumed ϕ(H) * Ell(G y S). Then, by Lemma 7.3, the group ϕ(V ) is ﬁnite.

8. Proof of a strong version of Theorem A Notation 8.1. For any subset X of a group G and any natural n, we denote the product X · X · ... · X by X(n). A subset X of G is called symmetrized if X = X−1. n Lemma| 8.2. {zLet G be} a group and H a subgroup of G of ﬁnite index n. Then for any symmetrized subset X of G with 1 ∈ X, we have hH ∩ X(2n)i = H ∩ hXi.

(2n) Proof. We prove only the nontrivial inclusion H ∩ hXi ⊆ hH ∩ X i. Let Hg1, Hg2, ... , Hgn be all right cosets of H in G, where g1 = 1. For any x ∈ X, we connect Hgi to Hgj by a directed edge with label x if Hgix = Hgj . We denote the resulting graph by Γ. Let h ∈ H ∩hXi. We consider h as a word in the alphabet X. Then there exists a path p in Γ that starts and ends at H and has the label h. This path is homotopic in Γ to a product of paths p1,...,pk, where each pi starts and ends at H and has length at most 2n. Let hi be the element of H corresponding (2n) to the label of pi. Then h = h1 ...hk and each hi lies in H ∩ X .

We need the following slight generalization of Theorem A for the proof of The- orem C. Theorem 8.3. Let H be a topological group which is either completely metrizable, or locally countably compact Hausdorff, and let H0 be a finite index subgroup of H. Let ϕ : H0 → G an abstract homomorphism, where G is an arbitrary group. Then either ϕ(H0) is contained in the elliptic radical RadEll(G), or there exists an open subgroup V 6 H such that H0 ∩ V normal in H0 and ϕ(H0 ∩ V ) is finite. 24 OLEGBOGOPOLSKIANDSAMUELM.CORSON

Proof. Suppose that ϕ(H0) is not contained in RadEll(G). Then there exists a hyperbolic space S and a cobounded acylindrical action G y S such that ϕ(H0) is not contained in Ell(G y S), see Lemma 6.9. First we prove that there exists an open neighborhood V of identity 1H such that ϕ(H0 ∩V) ⊆ Ell(G y S). The proof is basically the same as the proof of Proposition 5.1. We need only two following changes. Assuming the contrary, we can choose ui’s so that they additionally lie in H0. Moreover, we can choose mi’s so that they additionally become multiplies of |H : H0|!. Then, because of equalities n mi wi = ui wi+1, all constructed wi’s lie in H0, which enables us to apply ϕ to this n mi equality and get Wi = Ui Wi+1 in G. Analyzing these inequalities as in the case H = HEG of Theorem B, we get a contradiction. Thus, there exists such a neighborhood V. Then there exists an open neighbor- (2) (2) hood O⊆V of identity 1H such that O ⊆V. Then ϕ((H0 ∩O) ) ⊆ Ell(G y S). Arguing as in the proof of Theorem A starting from (7.1), we obtain that there exists a relatively open subgroup A 6 H0 such that ϕ(A) is ﬁnite and A is normal in H0. Let A be an open subset of H such that H0 ∩ A = A.

Let k be the index of H0 in H. Let A1 ⊆ H be a symmetrized open neighborhood e (2k) e of 1H such that A1 ⊆ A. Then, using Lemma 8.2, we have H ∩ hA i = hH ∩ A(2k)i ⊆ hH ∩ Ai = A. 0 e 1 0 1 0 We set V = hA1i. Then V is an open subgroup of H such that ϕ(H0 ∩ V ) is ﬁnite e and H0 ∩ V is normal in H0.

9. Homomorphisms to the Mapping Class Groups 2 For (g,b) ∈ N0, let Σ = Σg,b be the compact orientable surface of genus g with b boundary components. Let Mod(Σ) be the mapping class group of Σ (we recall the definition and some relevant results about this group below). We will prove the following slight generalization of Theorem B. Note that we do not know how to prove Theorem B without using this generalization. Theorem 9.1. Let H be a topological group which is either completely metrizable or locally countably compact Hausdorff. Let H0 be a finite index subgroup of H. Let Σ be a connected compact surface (possibly non-orientable and possibly with boundary components). Then for any homomorphism ϕ : H0 → Mod(Σ) there exists an open subgroup V of H such that ϕ(H0 ∩ V ) is finite and normal in ϕ(H0). We note that, in case H is locally compact and one is not concerned with finite index subgroups, a slight refinement of the conclusion of Theorem C holds. A proof of Kramer and Varghese [38, Proposition C, Theorem D] shows that for any homomorphism ϕ : H → Mod(Σg,b) one has an open subgroup V of H, with V containing the connected component of 1H , such that ϕ(V ) is finite. Their argument uses the structure theory for locally compact groups. This approach cannot be applied in the more general situation involving finite indices and completely metrizable and locally countably compact groups. 9.1. The mapping class group and the curve complex of Σ. Some authors use slightly different definitions of the mapping class group. Since we use results of Masur, Minsky, Ivanov, and Bowditch, we also use their definition of this group. ABSTRACT HOMOMORPHISMS FROM SOME TOPOLOGICAL GROUPS 25

The mapping class group Mod(Σ) is the group of isotopy classes of orientation preserving self-homeomorphisms of Σ, where admissible isotopies fix each component of ∂Σ setwise. In particular, the homeomorphisms are allowed to permute the boundary components, and the classes of Dehn twists around curves parallel to boundary components are trivial. Let PMod(Σ) be the subgroup of Mod(Σ) consisting of the classes of homeomorphisms which induce the identity permutation of the boundary components and the identity permutation of the punctures. Note that this definition slightly differs from that given in the book of Farb and Margalit [27]. A powerful tool for studying the mapping class group is the curve complex of Σ defined by Harvey in [32]. We need only the 1-skeleton of this complex, which is called the curve graph of Σ and is denoted by C(Σ) in our paper. Recall that the vertices of C(Σ) are isotopy classes of essential (i.e., neither homotopically trivial nor peripheral) simple closed curves on Σ, and two isotopy classes are joined by an edge if they contain disjoined representatives. The mapping class group Mod(Σ) naturally acts on the curve graph C(Σ).

Theorem 9.2. Let Σ= Σg,b be an orientable surface of genus g with b > 0 boundary components such that 3g + b − 4 > 0. Then the following statements hold: (a) (ﬁrst proved by Masur-Minsky [41]; Bowditch [13] gave an alternative proof) The curve graph C(Σ) is hyperbolic. (b) (Bowditch [11]) The action Mod(Σ) y C(Σ) is acylindrical. (c) (Masur-Minsky [41]) An element of the mapping class group Mod(Σ) acts on the curve graph C(Σ) loxodromically, i.e., with positive translation length, if and only if it is pseudo-Anosov. (d) The action Mod(Σ) y C(Σ) is cobounded. Statement (d) follows from the fact that Mod(Σ) acts transitively (respectively, has ﬁnitely many orbits) acting on the set of isotopy classes of nonseparating (respectively, separated) simple closed curves of Σ.

Remark 9.3. If 3g + b − 4 6 0, then the groups PMod(Σg,b) are very simple: PMod(Σ0,b) = 1 for b 6 3 and PMod(Σ0,4) = F2 (see [27, Chapter 4]), PMod(Σ1,1) =∼ SL2(Z), see [27, Theorem 3.6]. We also need the following theorem of Ivanov, which is a generalization of Thurston’s theorem (see [53]) on elements of the mapping class group. Theorem 9.4. (see [36, Theorem 1]) Let Σ be a compact orientable surface. For any subgroup H of the mapping class group Mod(Σ) one of the following holds: (a) H is ﬁnite. (b) H is reducible (that is there is an essential 1-submanifold of Σ whose isotopy class is ﬁxed by each element of H). (c) H contains a pseudo-Anosov element. The following lemma is due to Szepeitowski (in the case b = 0 it easily follows from [7]).

Lemma 9.5. (see [52, Lemma 3]) Suppose that g +2b > 3. Let Ng,b be a non- orientable surface of genus g with b boundary components. Let Σg−1,2b be an orientable surface which is a double cover of Ng,b. Then there exists a natural embedding of Mod(Ng,b) into Mod(Σg−1,2b). 26 OLEGBOGOPOLSKIANDSAMUELM.CORSON

9.2. Preparation to the proof of Theorem 9.1. Definition 9.6. Let H be a topological group which is either completely metrizable or locally countably compact Hausdorff. A group G is called H-admissible if for any finite index subgroup H0 of H and any homomorphism ϕ : H0 → G, there exists an open subgroup V of H such that ϕ(H0 ∩ V ) is finite. Proposition 9.7. Let H be a topological group which is either completely metrizable, or locally countably compact Hausdorff. Then the following holds. (1) The class of H-admissible groups is closed under taking subgroups and under taking of overgroups of finite index. In particular, the class of H-admissible groups is closed under the commensurability relation. (2) The class of H-admissible groups is closed under taking extensions, i.e. if G/A = B and the groups A and B are H-admissible, then G is H-admissible as well. (3) The class of H-admissible groups contains all finite groups and all acylindrically hyperbolic groups G whose elliptic radical RadEll(G) contains only H-admissible subgroups of G. In particular, all hyperbolic groups are H- admissible. (4) Let H1 be an open subgroup of H. Then H1 is itself either completely metrizable or locally countably compact Hausdorff. Moreover, the class of H1-admissible groups is contained in the class of H-admissible groups.

Proof. Statements (1) and (4) are obvious, statement (3) follows from Theo- rem 8.3. We prove statement (2). Let H0 be a finite index subgroup of H and let ϕ : H0 → G be an arbitrary homomorphism. Let ϕ1 : G → B be the canonical homomorphism. Consider the homomorphism ϕ1 ◦ ϕ : H0 → B. Since B is H-admissible, there exists an open subgroup U of H such that ϕ1 ◦ϕ (H0 ∩U) is finite. Then there exists a finite index subgroup H1 of H0 ∩ U such that ϕ1 ◦ ϕ (H1) = 1, i.e. ϕ(H1) ⊆ A. Clearly H1 has finite index in U and U (as an open subgroup of H) is itself either completely metrizable or locally countably compact Hausdorff. Therefore there exists an open subgroup V in U such that ϕ(H1 ∩ V ) is finite. But

|(H0 ∩ V ) : (H1 ∩ V )| = |((H0 ∩ U) ∩ V ) : (H1 ∩ V )| 6 |(H0 ∩ U): H1| < ∞.

Therefore ϕ(H0 ∩ V ) is finite. 9.3. Proof of Theorem 9.1. Due to Lemma 9.5, we may assume that Σ is orientable, say Σ = Σg,b for some g and b. We set G = Mod(Σg,b). First we prove by induction on the complexity of Σ (defined by the parameters (g,b) under the lexicographic ordering) the following statement: The group G is H-admissible for each H which is either completely metrizable or locally countably compact Hausdorff. We fix such an H, a finite index subgroup H0 6 H, and a homomorphism ϕ : H0 → G. We shall show that there exists an open subgroup V 6 H such that ϕ(H0 ∩ V ) is finite. If 3g + b − 4 6 0, then G is virtually free by Remark 9.3, and the statement follows from Proposition 9.7. Now we consider the case 3g + b − 4 > 0. By Theorem 9.2, Mod(Σ) acts acylindrically and coboundedly on the curve graph C(Σ), which is hyperbolic. Therefore, ABSTRACT HOMOMORPHISMS FROM SOME TOPOLOGICAL GROUPS 27 by Theorem 8.3, either ϕ(H0) is elliptic, or there exists an open subgroup U 6 H such that H0 ∩ U is normal in H0 and ϕ(H0 ∩ U) is finite. In the latter case we are done. Therefore suppose that ϕ(H0) is elliptic. If ϕ(H0) is finite, we are again done. Thus, we may assume that ϕ(H0) is infinite. Then, by Theorem 9.4 of Ivanov, there exists an essential 1-submanifold M of Σ whose isotopy class is fixed by each element of ϕ(H0). We introduce the following notations.

(a) Let Σ1,..., Σm be the closures of all components of Σ \ M. (b) Let PMod(Σ,M) denote the subgroup of PMod(Σ) which ﬁxes the isotopy class of each component of M and induces the trivial permutation of Σi’s.

Then there exists a finite index subgroup K of H0 such that ϕ(K) 6 PMod(Σ,M). There is an exact sequence m {1} → A → PMod(Σ,M) → PMod(Σi) →{1}, i=1 Y where A is the free abelian group generated by Dehn-twists around the components of M. Since the complexity of each Σi is smaller than the complexity of Σ, all the groups Mod(Σi) are H-admissible. By Proposition 9.7, PMod(Σ,M) is H- admissible. Since the index |H : K| is finite and ϕ maps K to PMod(Σ,M), we conclude that there exists an open subgroup V 6 H such that ϕ(K ∩ V ) is finite. Then ϕ(H0 ∩ V ) is also finite, since |(H0 ∩ V ) : (K ∩ V | 6 |H0 : K| < ∞. We may assume that |ϕ(H0 ∩ V )| is minimal possible over all open subgroups h V 6 H. We prove that ϕ(H0 ∩ V ) is normal in ϕ(H0). Let h ∈ H0. Then V ∩ V is open in H and, by minimality, we have h h ϕ(h) ϕ(H0 ∩ V )= ϕ(H0 ∩ (V ∩ V )) 6 ϕ(H0 ∩ V )= ϕ(H0 ∩ V ) .

Theorem 9.8. Let H be a topological group which is either completely metrizable, or locally countably compact Hausdorff. Let H0 be a finite index subgroup of H. Let G be a one-ended hyperbolic group. Then for any homomorphism ϕ : H → Out(G) there exists an open subgroup V of H such that ϕ(H0 ∩ V ) is finite and normal in ϕ(H0).

Proof. In [12], Bowditch described a canonical JSJ-decomposition of a one- ended hyperbolic group G. This is a special finite graph of groups Γ such that π1(Γ) = G. We need only to know that some vertex subgroups Gv of Γ are called “MHF subgroups” and that each such subgroup Gv is the fundamental group of a compact orbifold Σv. Let V2 be the set of vertices v of Γ corresponding to the MHF subgroups. Using this decomposition, Levitt proved in [39, Theorem 5.3] that Out(G) has a finite index subgroup of the form Zq × M, where q ∈ N and M fits in the exact sequence

{1} → A → M → Mod(Σv) →{1} ∈ vYV2 with A virtually Zs for some s ∈ N. 28 OLEGBOGOPOLSKIANDSAMUELM.CORSON

By [33], generalized to the non-orientable case in [29], each group Mod(Σv) is ′ commensurable with the the mapping class groups Mod(Σv) of a regular com- ′ ′ pact surface Σv (see also [39, Section 6]). By Theorem 9.1, each Mod(Σv) is H-admissible. Using Proposition 9.7, we conclude that Out(G) is H-admissible. Therefore there exists an open subgroup V 6 H such that ϕ(H0 ∩ V ) is ﬁnite. We may assume that |ϕ(H0 ∩ V )| is minimal possible. Then, arguing as in the last paragraph of the proof of Theorem 9.1, we conclude that ϕ(H0 ∩ V ) is normal in ϕ(H0). 10. General applications Given a group G and an element g ∈ G, let g : G 7→ G be the map given by g(x)= g−1xg, x ∈ G. 10.1. Applications to endomorphisms of theb group HEG. The following lemmab can be easily deduced from the Bass – Serre theory. It can be also proved directly by using normal forms. Lemma 10.1. Let H be a subgroup of a free product A ∗ B. If every element of H is conjugate to an element of B, then the whole group H is conjugate to a subgroup of B.

Let ai ∈ HEG be the element corresponding the i-th loop of the Hawaiian earring. For any natural n, the subgroup HEGn generated by a1,...,an is free of rank n. n We use the decomposition HEG = HEGn ∗ HEG from Section 2. Corollary 10.2. For any endomorphism ϕ : HEG → HEG there exists g ∈ HEG such that the following holds: For any natural n, there exists a natural m such that (g ◦ ϕ)(HEGm) ⊆ HEGn .

In particular, for i > m +1, the elements (g ◦ ϕ)(ai) considered as reduced words do not contain the letters a1,...,ab n. b Proof. Let ϕ : HEG → HEG be an endomorphism and let n be an arbitrary natural number. In the ﬁrst paragraph of the proof of Proposition 6.6, we established that HEG acts acylindrically on the hyperbolic Cayley graph Γn. By Theorem A, applied to G = HEG acting on its Cayley graph Γn, there exists a natural number f(n) such that f(n) ϕ(HEG ) ⊆ Ell (HEG y Γn). (10.1) Without loss of generality, we may assume that the function f : N → N is monotone increasing. Using (10.1), we deduce ϕ(HEGf(n)) ⊆ g (HEGn). ∈ g [HEG Thus, each element of ϕ(HEGf(n)) is conjugateb to an element of the second factor n of the free product HEG = HEGn ∗ HEG . By Lemma 10.1, there exists gn ∈ HEG such that f(n) n (gn ◦ ϕ)(HEG ) 6 HEG . (10.2)

It remains to show that gn can be chosen independently of n. Since the function f is monotone increasing, we havec f(n+1) −1 n −1 n ϕ(HEG ) ⊆ gn (HEG ) g[n+1 (HEG ). \ c ABSTRACT HOMOMORPHISMS FROM SOME TOPOLOGICAL GROUPS 29

Case 1. Suppose that ϕ(HEGf(n+1)) 6= 1. Then, since HEGn is malnormal in n −1 n the free product HEG = HEGn ∗ HEG , we have gn+1gn ∈ HEG . Therefore n there exists zn ∈ HEG such that gn+1 = zngn. f(n+1) Case 2. Suppose that ϕ(HEG ) = 1. Then we redefine gn+1 by setting gn+1 = gn. n Thus, we may assume that for any n, there exists zn ∈ HEG such that gn+1 = zngn. We set g = (...z2z1)g1. Then (10.2) implies that (g ◦ ϕ)(HEGf(n)) 6 HEGn for any n. b The following corollary follows straightforwardly from Corollary 10.2. Recall that HEG is a topological group with respect to the topology T described in Section 2. Corollary 10.3. Any endomorphism of the abstract group HEG is continuous with respect to the topology T of the topological group HEG. The following corollary also follows straightforward from Corollary 10.2. It was first proved by Eda in [24, Corollary 2.11]. Corollary 10.4. (see [24]) For any endomorphism ϕ : HEG → HEG there exists g ∈ HEG such that the endomorphism g ◦ ϕ is induced by a continuous map from the Hawaiian earring E to itself preserving the basepoint (0, 0). 10.2. Applications to homomorphismsb to relatively hyperbolic groups. We use the terminology of the monograph [45]. Let G be a relatively hyperbolic group with respect to a collection of subgroups {Hλ}λ∈Λ. Let X be a finite relative generating set of G with respect to {Hλ}λ∈Λ and let Γ = Γ(G, X ∪ H) be the right Cayley graph of G with respect to the generating set X ∪ H, where

H = ∪ Hλ. λ∈Λ It is proved in [45] that the elements of G are either elliptic or loxodromic with respect to the left action of G on Γ(G, X ∪ H). This classification of elements of G does not depend on the choice of the finite set X. Moreover, the elliptic elements of G are precisely those which have finite order or conjugate to elements of H. We call a subgroup H of G parabolic if it is conjugate to a subgroup of Hλ for some λ ∈ Λ. Otherwise we call H non-parabolic. Lemma 10.5. (see [9, Lemma 2.9]) Let H be a subgroup of a relatively hyperbolic group G. Then H contains a loxodromic element if and only if H is infinite and non-parabolic. Proposition 10.6. (see [46, Proposition 5.2]) Let G be a relatively hyperbolic group with respect to a collection of subgroups {Hλ}λ∈Λ. Let X be a finite relative generating set of G with respect to {Hλ}λ∈Λ. Then the canonical action of G on the Cayley graph Γ(G, X ∪ H) is acylindrical. Corollary 10.7. Let G be a relatively hyperbolic group with respect to a collection of subgroups {Hλ}λ∈Λ. Let ϕ : HEG → G be a homomorphism. Then there exists n ∈ N such that the subgroup ϕ(HEGn) is finite or parabolic. 30 OLEGBOGOPOLSKIANDSAMUELM.CORSON

Proof. Let X be a ﬁnite relative generating set of G with respect to {Hλ}λ∈Λ. By Proposition 10.6, G acts acylindrically on the Cayley graph Γ = Γ(G, X ∪ H). By Theorem B, there exists a natural number m such that ϕ(HEGm) ⊆ Ell(G y Γ). Thus, ϕ(HEGm) does not contain loxodromic elements. By Lemma 10.5, ϕ(HEGm) is ﬁnite or parabolic.

In the following corollary we obtain a stronger conclusion for a broader class of groups H. The proof directly follows from Proposition 10.6, Lemma 10.5 and Theorem A. Corollary 10.8. Let G be a relatively hyperbolic group with respect to a collection of subgroups {Hλ}λ∈Λ. Let ϕ : H → G be a homomorphism with H either completely metrizable or locally countably compact Hausdorff. Then either ϕ(H) is parabolic, or there exists a normal open subgroup V 6 H with finite ϕ(V ). 10.3. Applications to homomorphisms to fundamental groups of graphs of groups. For a graph Γ we denote by Γ0 the set of its vertices and by Γ1 the set 1 of its edges. For any edge e ∈ Γ , let e− be the initial vertex of e and e+ be the terminal vertex of e. Let e−1 be the edge inverse to e. We recall some terminology from the book [49]. A graph of groups is a pair G = (Γ, G), where Γ is a connected graph and G is a triple consisting of groups 0 1 Gv, where v ∈ Γ , of groups Ge where e ∈ Γ (we require that Ge = Ge−1 ), and 1 of embeddings αe : Ge → Ge− and ωe : Ge → Ge+ where e ∈ Γ (we require that αe = ωe−1 ). A graph of groups (Γ, G) is called finite if Γ is finite. Let ∆ be a maximal subtree of Γ. The fundamental group π1(Γ, G, ∆) of the 1 graph of groups (Γ, G) is the group generated by the elements of Gv (v ∈ Γ ) and 1 the elements te (e ∈ Γ ), subject to the relations −1 0 te αe(g)te = ωe(g), g ∈ Ge, e ∈ Γ , −1 0 te−1 = te , e ∈ Γ , 1 te =1, e ∈ ∆ .

The group π1(Γ, G, ∆) is independent, up to isomorphism, of the choice of the maximal tree ∆. The subgroups Gv of the group G = π1(Γ, G, ∆) are called the vertex subgroups of G. Particular cases of fundamental groups of graph of groups are amalgamated products and HNN extensions. The group G acts by left multiplication on the Bass-Serre tree T associated with (Γ, G), see [49]. Recall that the vertices of T are the cosets gGv, where g runs over G and Gv runs over vertex subgroups of G. It is well known that each element of a group acting on a tree has either a nonempty fixed points set (which is a subtree), or an invariant axis where it acts by translation through a nonzero distance. In the first case the element is called elliptic, in the second one loxodromic. Our nearest aim is to give a condition which provides the acylindricity of a group action on a simplicial tree. The following definition is very similar to the definition of Sela of k-acylindricity and to the definition of Bowditch of acylindricity. Definition 10.9. Let G be a group acting on a simplicial tree T . Let k and n be natural numbers. We say that G acts on T (k,n)-acylindrically if the pointwise stabilizer of any segment of T of length k contains at most n elements. ABSTRACT HOMOMORPHISMS FROM SOME TOPOLOGICAL GROUPS 31

Lemma 10.10. Let G be a group that acts (k,n)-acylindrically on a simplicial tree T . Then for any loxodromic element a ∈ G and any nonzero integer m, there exist at most n elements b satisfying a = bm.

m m Proof. Let b and b1 be two loxodromic elements satisfying b = b1 . Then the axes of b and b1 coincide, they have the same translation length and translation −1 −1 direction. Then b1b ﬁxes this axis pointwise, hence the number of b1b is at most n.

Proposition 10.11. An action of a group G on a simplicial tree T is (k,n)- acylindrical for some k,n if and only if it is acylindrical in the sense of Defini- tion 3.3. Proof. Suppose that G acts on T (k,n)-acylindrically. We prove that G acts on T acylindrically in the sense of Definition 3.3. For that we check that for any integer σ> 1 and any two vertices u, v ∈ T 0 with d(u, v) > k +2σ +2σ2 the set M = {g ∈ G | d(u,gu) < σ, d(v,gv) < σ} contains at most 2σn2 + n elements. 2 Let u1,u2, v1, v2 be vertices on [u, v] such that d(u,u1) = σ, d(u1,u2) = σ , 2 d(v1, v) = σ, d(v2, v1) = σ . Thus, [u1, v1] is the central subsegment of [u, v] of 2 length at least k +2σ and [u2, v2] is the central subsegment of [u1, v1] of length at least k. Let g be an element of the set M. Case 1. Suppose that g acts elliptically on T . Then g has fixpoints on the segments [u,gu] and [v,gv]. Since the fixpoint set of g is connected, it contains the segment [u1, v1]. Since the length of this segment is larger than k, the number of such g is at most n. Case 2. Now suppose that g acts loxodromically on T . Then the translation length t(g) of g is at most σ. Moreover, the axis of g has edges in [u,gu] and in [v,gv]. Then [u1, v1] lies on the axis of g. Let f be another loxodromic element from M. Then [u1, v1] lies on the axes of g and f and 0

Lemma 10.13. Suppose that G is a group acting acylindrically on a simplicial tree T . If G does not contain loxodromic elements, then there exists a vertex or an edge of T which, up to inversion, is ﬁxed by G.

Proof. By Theorem 7.1, G has an orbit of finite diameter, say Gx, where x is a vertex of T . Let Γ be the tree spanned by Gx. Then Γ has finite diameter and is G-invariant. A vertex of Γ is called outer if it has degree 1. An edge of Γ is called 32 OLEGBOGOPOLSKIANDSAMUELM.CORSON outer if at least one of its endpoints is outer. All other vertices and edges of Γ are called inner. Let Γ1 be the subset of Γ consisting of all inner vertices and edges. If Γ contains at least 2 edges, then Γ1 is a nonempty subtree of Γ. Moreover, Γ1 has smaller diameter than Γ and is G-invariant. Inducting on the diameter of Γ, we complete the proof. The following two corollaries follow directly from Theorem A (respectively, from Theorem B) with the help of Proposition 10.11 and Lemma 10.13. Corollary 10.14. Let H be a topological group which is either completely metrizable, or locally countably compact Hausdorff. Let G be the fundamental group of a (possibly infinite) graph of groups (Γ, G). Suppose that the action of G on the Bass-Serre tree associated with (Γ, G) is (k,n)-acylindrical for some k,n ∈ N. Then for any abstract group homomorphism ϕ : H → G either ϕ(H) is conjugate into a vertex subgroup of G or there exists a normal open subgroup V 6 H such that ϕ(V ) is finite. Corollary 10.15. Let H be a topological group which is either completely metrizable, or locally countably compact Hausdorff. Let G be the fundamental group of a (possibly infinite) graph of groups (Γ, G). Suppose that the action of G on the Bass-Serre tree associated with (Γ, G) is (k,n)-acylindrical for some k,n ∈ N. Then for any abstract group homomorphism ϕ : HEG → G there exists a natural number m such that ϕ(HEGm) is a subgroup of G consisting of only elliptic elements with respect to this action.

11. Combination theorems for automatic continuity From Corollaries 10.7 and 10.8 one obtains new results regarding automatic continuity. Recall that a group G is n-slender if for every abstract homomorphism ϕ : HEG → G there exists n ∈ N such that HEGn 6 ker(ϕ). Similarly G is cm- slender (respectively lcH-slender) if every abstract group homomorphism from a completely metrizable (respectively locally compact Hausdorﬀ) topological group to G has open kernel [17]. The abbreviation lccH-slender stands for locally countably compact slenderness. These can be considered as a strong type of automatic continuity: one can endow a cm-slender group G with the discrete topology and assert that any abstract group homomorphism from a completely metrizable topological group to G is continuous. Free groups, Thompson’s group F , Baumslag-Solitar 1 groups, Z[ m ], torsion-free word hyperbolic groups, graph products of such groups and numerous other groups are now known to satisfy these notions of slenderness. A group which is either n-slender, cm-slender, or lccH-slender must be torsion-free. Theorem E. If G is torsion-free and relatively hyperbolic with respect to a collection {Hλ}λ∈Λ of n-slender (respectively cm-slender, lcH-slender, lccH-slender) subgroups, then G is also n-slender (respectively cm-slender, lcH-slenderm, lccH- slender). Proof. Suppose G is torsion-free, relatively hyperbolic to a collection of n-slender subgroups. Let ϕ : HEG → G be a homomorphism. By Corollary 10.7, there exists some n ∈ N for which ϕ(HEGn) is parabolic. Then the restriction ϕ ↾ HEGn is a map from an isomorphic copy of HEG to an n-slender group, so there exists an m > n for which ϕ ↾ HEGm is the trivial map. Thus G is n-slender. The proof for other types of slenderness is similar. ABSTRACT HOMOMORPHISMS FROM SOME TOPOLOGICAL GROUPS 33

We obtain the following from Corollaries 10.14 and 10.15: Theorem F. Suppose that G is the fundamental group of a (possibly inﬁnite) graph of groups (Γ, G) where each vertex group Gv is n-slender (respectively cm-slender, lcH-slender, lccH-slender) and that the action of G on the Bass-Serre tree associated with (Γ, G) is (k,n)-acylindrical for some k,n ∈ N. Then G is n-slender (respectively cm-slender, lcH-slender, lccH-slender).

References

[1] Carolyn R. Abbott, Not all finitely generated groups have universal acylindrical actions, Proc. Amer. Math. Soc., 144, 10 (2016), 4151-4155. [2] C. Abbott, S. Balasubramanya, and D. Osin, Hyperbolic structures on groups, Algebr. Geom. Topol. 19 (2019), 1747-1835. [3] Carolyn R. Abbott, Jason Behrstock, and Matthew Centry Durham, Largest acylindrical actions and stability in hierarchically hyperbolic groups, ArXiv, 2019 [4] U. Bader and A. Furman, Margulis’ super-rigidity theorem for non-lattices, arXiv:1810.01608v1, 2018. [5] Jason Behrstock, Mark F. Hagen, and Alessandro Sisto, Hierarchically hyperbolic spaces I: curve complexes for cubical groups, Geometry & Topology, 21 (2017), 1731-1804. [6] J. Behrstock, M.F. Hagen, and A. Sisto, Hierarchically hyperbolic spaces II: combination theorems and the distance formula, ArXiv: 1509.0063v2, 2015. [7] J.S. Birman and D.R.J. Chillingworth, On the homeotopy group of a non-orientable surface, Proc. Cambridge Philos. Soc., 71 (1972), 437-448. [8] Oleg Bogopolski, Equations in acylindrically hyperbolic groups and verbal closedness, ArXiv, 2019. Available at https://arxiv.org/pdf/1805.08071.pdf [9] Oleg Bogopolski and Kai-Uwe Bux, From local to global conjugacy of subgroups of relatively hyperbolic groups, Intern. J. of Algebra and Computation 27, no. 3 (2017), 299-314. [10] O. V. Bogopolskii, V.N. Gerasimov, Finite subgroups of hyperbolic groups, Algebra Logic 34 (1995), 343-345. [11] B. Bowditch, Tight geodesics in the curve complex, Invent. Math. 171 (2008) 281-300. [12] B. Bowditch, Cut points and canonical splittings of hyperbolic groups, Acta Math, 180 (1998), 145-186. [13] B. Bowditch, Intersection numbers and the hyperbolicity of the curve complex, J. Reine Angew. Math., 598 (2006), 105-129. [14] O. Braun, K.H. Hofmann, L. Kramer, Automatic continuity of abstract homomorphisms between locally compact and polish groups, arXiv: 1611.05801, (2016). To appear in Trans- formation Groups. [15] M. Bridson, A. Haeffliger, Metric spaces of non-positive curvature, Springer, 1999. [16] J. Cannon, G. Conner, The combinatorial structure of the Hawaiian earring group, Topol. Appl., 106 (2000), 225-271. [17] G. Conner, S. Corson, A note on automatic continuity, Proc. Amer. Math. Soc. 147 (2019), 1255-1268. [18] G. Conner, K. Spencer, Anomalous behavior of the Hawaiian Earring group, J. Group Theory 8 (2005), 225-227. [19] M. Coornaert, T. Delzant, A. Papadopoulos, Geometrie et theorie des groupes. Les groupes hyperboliques de Gromov. Lecture Notes in Mathematics, 1441. Springer-Verlag, Berlin, 1990. x+165 pp. [20] S. Corson, I. Kazachkov, On preservation of automatic continuity, Monatsh Math (2019) https://doi.org/10.1007/s00605-019-01281-x. [21] Rémi Coulon, On the geometry of burnside quotients of torsion free hyperbolic groups, Int. J. of Algebra and Computation 24, no. 3 (2014), 251-345. [22] F. Dahmani, V. Guirardel, D. Osin, Hyperbolically embedded subgroups and rotating families in groups acting on hyperbolic spaces, Mem. Am. Math. Soc. 245 (2017), no. 1156. [23] R. Dudley, Continuity of homomorphisms, Duke Math. J. 28 (1961), 587-594. [24] K. Eda, Free σ-products and fundamental groups of subspaces of the plane, Topol. Appl 84 (1998), 283-306. 34 OLEGBOGOPOLSKIANDSAMUELM.CORSON

[25] K. Eda, Atomic property of the fundamental groups of the Hawaiian earring and wild locally path-connected spaces, Jour. Math. Soc. Japan 63 (2011), 769-787. [26] K. Eda, K. Kawamura, The singular homology of the Hawaiian earring, J. London Math. Soc. 62 (2000), 305-310. [27] B. Farb and D. Margalit, A primer on mapping class groups, Princeton Mathematical Series 49, Princeton University Press, 2011. [28] B. Farb and H. Masur, Superrigidity and mapping class groups, Topology, 37 (1998), 1169- 1176. [29] K. Fujiwara, On the outer automorphism group of a hyperbolic group, Israel J. of Math., 131 (2002), 277-284. [30] H. Griffiths, Infinite products of semi-groups and local connectivity, Proc. London Math. Soc. 3 (1956), 455-456. [31] M. Gromov, Hyperbolic groups. In Essays in group theory, pages 75-263, Springer, New York, 1987. [32] W.J. Harvey, Boundary structure of the modular group. In Riemann surfaces and related topics, Proceedings of the 1979 Stony Brook Conference (State Univ. New York, Stony Brook, N.Y., 1978), v. 97 of Ann. of Math. Stud., pages 245-251, Princeton, N.J., 1981. Princeton Univ. Press. [33] W.J. Harvey and C. Maclachlan, On mapping class groups and Teichmüller spaces, Proc. London Math. Soc., 30 (1975), 496-512. [34] G. Higman, Unrestricted free products and varieties of topological groups, J. London Math. Soc. 27 (1952), 73-81. [35] M. Hruˇsák, J. van Mill, U. Ramos-Garc´ıa, S. Shelah, Countably compact groups without non-trivial convergent sequences, preprint available at https://www.matmor.unam.mx/ ˜michael/preprints files/Countably compact.pdf [36] N. V. Ivanov, Subgroups of Teichmüller modular groups, volume 115 of Translations of Mathematical Monographs, American Mathematical Society, Providence, RI, 1992. Trans- lated from the Russian by E.J.F. Primrose and revised by the author. [37] M. Koubi, Croissance uniforme dans les groupes hyperboliques, Ann. Inst. Fourier (Greno- ble), 48 (1998), no. 5, 1441-1453. [38] L. Kramer, O. Varghese, Abstract homomorphisms from locally compact groups to discrete groups, J. Algebra 538 (2019), 127-139. [39] G. Levitt, Automorphisms of hyperbolic groups and graphs of groups, [40] G.A. Margulis, Discrete subgroups of semisimple Lie groups, Ergebnisse der Mathematik und ihrer Grenzgebiete (3) [Results in Mathematics and Related Areas (3)], vol. 17, Springer Verlag, Berlin, 1991. [41] H.A. Masur and Y.N. Minsky, Geometry of the complex of curves I. Hyperbolicity, Invent. Math., 138, (1) (1999), 103-149. [42] S.A. Morris and P. Nickolas, Locally compact group topologies on an algebraic free product of groups, J. Algebra, 38, no. 2 (1976), 393-397. [43] J. Nakamura, Atomic properties of the Hawaiian earring group for HNN extensions, Com- mun. Algebra 43 (2015) 4138-4147. [44] N. Nikolov and D. Segal, On finitely generated profinite groups, I: strong completeness and uniform bounds, Annals of Math., 165 (2007), 171-238. [45] D. Osin, Relatively hyperbolic groups: Intrinsic geometry, algebraic properties, and algo- rithmic problems, Memoirs Amer. Math. Soc., v. 179 (2006), no. 843. [46] D. Osin, Acylindrically hyperbolic groups, Trans. Amer. Math. Soc., v. 368 (2016), 851-888. [47] D. Osin, Groups acting acylindrically on hyperbolic spaces, ArXiv, 2017. Available at https://arxiv.org/pdf/1712.00814.pdf [48] G. Paolini, S. Shelah, Polish topologies for graph products of groups, J. London Math. Soc., (2019), https://doi.org/10.1112/jlms.12219. [49] J.P. Serre, Trees, Springer-Verlag, 1980. [50] A. Sisto, Contracting elements and ranfom walks, Journal für die reine und angewandte Mathematik (Crelles Journal), Volume 2018, Issue 742, 79-114. [51] M. Stroppel, Locally compact groups, EMS textbooks in Math., EMS, 2006. [52] B. Szepietowski, Embedding the braid group in mapping class groups, Publ. Mat., 54 (2010), 359-368. ABSTRACT HOMOMORPHISMS FROM SOME TOPOLOGICAL GROUPS 35

[53] W.P. Thurston, On the geometry and dynamics of diﬀeomorphisms of surfaces, Bull. Amer. Math. Soc., (N.S.), 19, no. 2 (1988), 417-431.

Sobolev Institute of Mathematics of Siberian Branch of Russian Academy of Sci- ences, Novosibirsk, Russia and Dusseldorf¨ University, Germany E-mail address: Oleg [email protected]

Instituto de Ciencias Matematicas´ CSIC-UAM-UC3M-UCM, 28049 Madrid, Spain. E-mail address: [email protected]