Isometry groups of non-positively curved spaces: structure theory

We develop the structure theory of full isometry groups of locally compact non-positively curved metric spaces. Amongst the discussed themes are de Rham decompositions, normal subgroup structure and characterising properties of symmetric spaces and B…

Authors: P.-E. Caprace, N. Monod

Isometry groups of non-positively curved spaces: structure theory
ISOMETR Y GR OUPS OF NON-POSITIVEL Y CUR VED SP A CES: STR UCTURE THEOR Y PIERRE-EMMANUEL CAPRA CE* AND NICOLAS MONOD ‡ Abstra t. W e dev elop the struture theory of full isometry groups of lo ally ompat non-p ositiv ely urv ed metri spaes. Amongst the disussed themes are de Rham deom- p ositions, normal subgroup struture and  haraterising prop erties of symmetri spaes and BruhatTits buildings. Appliations to disrete groups and further dev elopmen ts on non-p ositiv ely urv ed latties are exp osed in a ompanion pap er [ CM08b ℄. 1. Intr odution Non-p ositiv ely urv ed metri spaes w ere in tro dued b y A. D. Alexandro v [ Ale57 ℄ and p opularised b y M. Gromo v, who alled them CA T(0) spaes. Their theory oers a wide gatew a y to a form of generalised dieren tial geometry , whose ob jets enompass Riemannian manifolds of non-p ositiv e setional urv ature as w ell as large families of singular spaes inluding Eulidean buildings and man y other p olyhedral omplexes. It has found a wide range of appliations to v arious elds, inluding semi-simple algebrai and arithmeti groups, and geometri group theory . A reurren t theme in this area is the in terpla y b et w een the geometry of a lo ally ompat CA T(0) spae X and the algebrai prop erties of a disrete group Γ ating prop erly on X b y isometries. This in teration is exp eted to b e esp eially ri h and tigh t when the Γ -ation is o ompat; the pair ( X, Γ) is then alled a CA T(0) group . The purp ose of the presen t pap er and its ompanion [ CM08b ℄ is to highligh t the rle of a third en tit y through whi h the in teration b et w een X and Γ transits: namely the full isometry group Is( X ) of X . The top ology of uniform on v ergene on ompata mak es Is( X ) a lo ally ompat seond oun table group whi h is th us anonially endo w ed with Haar measures. It therefore mak es sense to onsider latties in Is( X ) , i.e. disrete subgroups of nite in v arian t o v olume; w e all su h pairs ( X, Γ) CA T(0) latties (th us CA T(0) groups are preisely uniform CA T(0) latties). This immediately suggests the follo wing t w o-step programme: (I) T o dev elop the basi struture theory of the lo ally ompat group Is( X ) and dedue onsequenes on the o v erall geometry of the underlying prop er CA T(0) spae X . This is the main purp ose of the presen t pap er. (I I) T o study CA T(0) latties and th us in partiular CA T(0) groups b y building up on the struture results of the presen t pap er, using new geometri densit y and sup er- rigidit y te hniques. This is arried out in the subsequen t pap er [ CM08b ℄. W e no w pro eed to desrib e the main results of this rst part in more detail. First, in  1.A , w e presen t results in the sp eial ase of geo desially omplete CA T(0) spaes, i.e. spaes in whi h ev ery geo desi segmen t an b e extended to a bi-innite geo desi line  Key wor ds and phr ases. Non-p ositiv e urv ature, CA T(0) spae, lo ally ompat group, lattie. *F.N.R.S. Resear h Asso iate. ‡ Supp orted in part b y the Swiss National Siene F oundation. 1 2 PIERRE-EMMANUEL CAPRA CE AND NICOLAS MONOD whi h need not b e unique. Imp ortan t examples of geo desially omplete spaes are pro vided b y BruhatTits buildings and of ourse Hadamard manifolds, e.g. symmetri spaes. The seond and longer part of the In tro dution,  1.B , will presen t results v alid for arbi- trary lo ally ompat CA T(0) spaes. In either ase, the en tire on ten ts of the In tro dution rely on more general, more detailed but probably also more um brous statemen ts pro v ed in the ore of the text. 1.A. Spaes with extensible geo desis. The onlusions of sev eral results b eome esp e- ially lear and p erhaps more striking in the sp eial ase of geo desially omplete CA T(0) spaes. Bey ond Eulidean buildings and Hadamard manifolds, w e reall that a omplete CA T(0) spae that is also a homology manifold has automatially extensible geo desis [ BH99 , I I.5.12℄. Note also that it is alw a ys p ossible to artiially mak e a CA T(0) spae geo desially omplete b y gluing ra ys, though it is not alw a ys p ossible to preserv e prop erness (onsider a ompat but total set in an innite-dimensional Hilb ert spae). Deomp osing CA T(0) spaes in to pro duts of symmetri spaes and lo ally nite ell omplexes. Protot ypial examples of lo ally ompat CA T(0) spaes are mainly pro vided b y the follo wing t w o soures.  Riemannian manifolds of non-p ositiv e setional urv ature, whose most prominen t represen tativ es are the Riemannian symmetri spaes of non-ompat t yp e. These spaes are regular in the sense that an y t w o geo desi segmen ts in terset in at most one p oin t. The full isometry group of su h a spae is a Lie group.  P olyhedral omplexes of pieewise onstan t non-p ositiv e urv ature, su h as trees or Eulidean buildings. These spaes are singular in the sense that geo desis do bran h. The subgroup of the isometry group whi h preserv es the ell struture is totally disonneted. The follo wing result seems to indiate that a CA T(0) spae often splits as a pro dut of spaes b elonging to these t w o families. Theorem 1.1. L et X b e a pr op er ge o desi al ly  omplete CA T(0) sp a e whose isometry gr oup ats  o  omp atly without xe d p oint at innity. Then X admits an Is( X ) -e quivariant split- ting X = M × R n × Y , wher e M is a symmetri sp a e of non- omp at typ e and the isometry gr oup Is( Y ) is total ly dis onne te d and ats by semi-simple isometries on Y (e ah fator may b e trivial). F urthermor e, the sp a e Y admits an Is( Y ) -e quivariant lo  al ly nite de  omp osition into  onvex  el ls, wher e the  el l supp orting a p oint y ∈ Y is dene d as the xe d p oint set of the isotr opy gr oup Is( Y ) y . If X is regular, then Is( Y ) is disrete. In other w ords, the spae Y has bran hing geo desis as so on as Is( Y ) is non-disrete. W e refer to Theorem 1.6 and A ddendum 1.8 b elo w for a v ersion of the ab o v e without the assumption of extensibilit y of geo desis. W e emphasize that the `ells' pro vided b y Theorem 1.1 need not b e ompat; in fat if Is( Y ) ats freely on Y then the deomp osition in question b eomes trivial and onsists of a single ell, namely the whole of Y . Con v ersely the ell deomp osition is non-trivial pro vided Is( Y ) do es not at freely . The most ob vious w a y for the Is( Y ) -ation not to b e free is if Is( Y ) is not disrete. A strong v ersion of the latter ondition is that no op en sub gr oup of xes a p oint at innity ; this holds notably for symmetri spaes and Bruhat Tits buildings. A quite immediate onsequene of this ondition is that the ab o v e ells are ISOMETR Y GR OUPS OF NON-POSITIVEL Y CUR VED SP A CES: STR UCTURE THEOR Y 3 then neessarily ompat. W e shall sho w that m u h additional struture an b e deriv ed from it (see Setion 7.D b elo w). . Smo othness. The ell deomp osition of the third fator in Theorem 1.1 is deriv ed from the follo wing smo othness result for isometri ations of totally disonneted groups. Theorem 1.2. L et X b e a ge o desi al ly  omplete pr op er CA T(0) sp a e X and G < I s ( G ) a total ly dis onne te d (lose d) sub gr oup ating minimal ly. The the p ointwise stabiliser in G of every b ounde d set is op en. This prop ert y , whi h is familiar from lassial examples, do es in gener al fail without geo desi ompleteness (see Remark 6.10 in [ CM08b ℄). It is an imp ortan t ingredien t for the onsiderations of Setion 7.D alluded to ab o v e, as w ell as for angle rigidit y results regarding b oth the Alexandro v angle (Prop osition 6.8 ) and the Tits angle (Prop osition 7.15 ). . A  haraterisation of symmetri spaes and Eulidean buildings. In symmetri spaes and BruhatTits buildings, the stabilisers of p oin ts at innit y are exatly the para- b oli subgroups; as su h, they are o ompat. This o ompatness holds further for all Bass Serre trees, namely bi-regular trees. Com bining our results with w ork of B. Leeb [ Lee00 ℄ and A. Lyt hak [ Lyt05 ℄, w e establish a orresp onding  haraterisation. Theorem 1.3. L et X b e a ge o desi al ly  omplete pr op er CA T(0) sp a e. Supp ose that the stabiliser of every p oint at innity ats  o  omp atly on X . Then X is isometri to a pr o dut of symmetri sp a es, Eulide an buildings and BassSerr e tr e es. The Eulidean buildings app earing in the preeding statemen t admit an automorphism group that is strongly transitiv e, i.e. ats transitiv ely on pairs ( c, A ) where c is a  ham b er and A an apartmen t on taining c . This prop ert y  haraterises the BruhatTits buildings, exept p erhaps for some t w o-dimensional ases where this is a kno wn op en question. The ab o v e  haraterisation is of a dieren t nature and indep enden t of the  haraterisa- tions using latties that will b e presen ted in [ CM08b ℄. . A tions of simple algebrai groups. Both for the general theory and for the geometri sup errigidit y/arithmetiit y statemen ts of the ompanion pap er [ CM08b ℄, it is imp ortan t to understand ho w algebrai groups at on CA T(0) spaes. Theorem 1.4. L et k b e a lo  al eld and G b e an absolutely almost simple simply  onne te d k -gr oup. L et X b e a CA T(0) sp a e (not r e du e d to a p oint) on whih G = G ( k ) ats  ontinuously and  o  omp atly by isometries. Then X is isometri to X model , the R iemannian symmetri sp a e or BruhatTits building asso iate d with G . A stronger and m u h more detailed statemen t is pro vided b elo w as Theorem 7.4 . In partiular, a mo diation of the statemen t holds without extensibilit y of geo desis and the 4 PIERRE-EMMANUEL CAPRA CE AND NICOLAS MONOD o ompatness assumption an b e relaxed. Ho w ev er, w e also sho w there b y means of t w o examples that some assumptions remain neessary . (As a p oin t of terminology , w e do not  ho ose a partiular saling fator on X model , so that the isometry of our statemen t ould b eome a homothet y for another mo del.) . 1.B. General ase. When dealing with CA T(0) spae in the highest p ossible lev el of gen- eralit y , one has to fae sev eral te hnial diulties aused b y the exibilit y of the CA T(0) ondition. F or example, giv en a CA T(0) spae X , there are man y w a ys to deform it in order to onstrut another spae Y , non-isometri to X , but with the prop ert y that X and Y ha v e isomorphi isometry groups or/and iden tial b oundaries. Amongst the simplest on- strutions, one an form (p ossibly w arp ed) pro duts with ompat CA T(0) spaes or gro w hair equiv arian tly along a disrete orbit. Mu h wilder (non-quasi-isometri) examples an b e onstruted for instane b y onsidering w arp ed pro duts with the v ery v ast family of CA T(0) spaes ha ving no isometries and a unique p oin t at innit y . In order to address these issues, w e in tro due the follo wing terminology . Minimalit y . . . . ís a ti   n eÒpoi sfara âgk¸ m ia, aÎt t a ˜ ut a kaÈ falkra âgk¸ m ia diexèr q et ai. Sunèsi o Kurena ou, Falkra âgk¸ m i o n . 1 An isometri ation of a group G on a CA T(0) spae X is said to b e minimal if there is no non-empt y G -in v arian t losed on v ex subset X ′ ( X ; the spae X is itself alled minimal if its full isometry group ats minimally . A CA T(0) spae X is alled b oundary-minimal if it p ossesses no losed on v ex subset Y ( X su h that ∂ Y = ∂ X . Here is ho w these notions relate to one another. Prop osition 1.5. L et X b e a pr op er CA T(0) sp a e. (i) Assume ∂ X nite-dimensional. If X is minimal, then it is b oundary-minimal. (ii) Assume Is( X ) has ful l limit set. If X is b oundary-minimal, then it is minimal. (iii) If X is  o  omp at and ge o desi al ly  omplete, then it is b oth minimal and b oundary- minimal. W e emphasize that it is not true in general that a minimal CA T(0) spae is geo desially omplete, ev en if one assumes that the isometry group ats o ompatly and without global xed p oin t at innit y . . Group deomp ositions. W e no w turn to struture results on the lo ally ompat isometry group Is( X ) of a prop er CA T(0) spae X in a broad generalit y; w e shall mostly ask that no p oin t at innit y b e xed sim ultaneously b y all isometries of X . This non-degeneray assumption will b e sho wn to hold automatially in the presene of latties (see Theorem 3.11 in [ CM08b ℄). The the follo wing result w as the starting p oin t of this w ork. 1 Synesius of Cyrene, Falkra âgk¸ m i o n (kno wn as Calvitii en omium ), end of Chapter 8 (at 72A in the page n um b ering from Denis P étau's 1633 edition). The En omium w as written around 402; w e used the 1834 edition b y J. G. Krabinger (Ch. G. Löund, Stuttgart). The ab o v e exerpt translates roughly to: as muh pr aise as is given to the spher es is due to the b ald he ad to o . ISOMETR Y GR OUPS OF NON-POSITIVEL Y CUR VED SP A CES: STR UCTURE THEOR Y 5 Theorem 1.6. L et X b e a pr op er CA T(0) sp a e with nite-dimensional Tits b oundary. Assume that Is( X ) has no glob al xe d p oint in ∂ X . Then ther e is a  anoni al lose d  onvex Is( X ) -stable subset X ′ ⊆ X suh that G = Is( X ′ ) has a nite index op en har ateristi sub gr oup G ∗ ✁ G whih admits a  anoni al de  omp osition (1.i) G ∗ ∼ = S 1 × · · · × S p ×  R n ⋊ O ( n )  × D 1 × · · · × D q ( p, q , n ≥ 0) wher e S i ar e almost  onne te d simple Lie gr oups with trivial  entr e and D j ar e total ly dis-  onne te d irr e duible gr oups with trivial amenable r adi al. A ny pr o dut de  omp osition of G ∗ is a r e gr ouping of the fators in ( 1.i ) . Mor e over, al l non-trivial normal, subnormal or as ending sub gr oups N < D j ar e stil l irr e duible with trivial amenable r adi al and trivial  entr aliser in D j . (These pr op erties also hold for latti es in N and their normal, subnormal or as ending sub gr oups, se e [ CM08b ℄ .) (A top ologial group is alled irreduible if no nite index op en subgroup splits non- trivially as a diret pro dut of losed subgroups. The amenable radial of a lo ally ompat group is the largest amenable normal subgroup; it is indeed a r adi al sine the lass of amenable lo ally ompat groups is stable under group extensions.) Remarks 1.7. (i) The nite-dimensionalit y assumption holds automatially when X has a o ompat group of isometries [ Kle99 , Theorem C℄. It is also automati for uniquely geo desi spaes, e.g. manifolds (Prop osition 7.11 ). (ii) The onlusion fails in v arious w a ys if G xes a p oin t in ∂ X . (iii) The quotien t G/G ∗ is just a group of p erm utations of p ossibly isomorphi fators in the deomp osition. In partiular, G = G ∗ ⋊ G/G ∗ . (iv) The anonial on tin uous homomorphism Is( X ) → Is( X ′ ) = G is prop er, but its image sometimes has innite o v olume. In Theorem 1.6 , the ondition that Is( X ) has no global xed p oin t at innit y ensures the existene of a losed on v ex Is( X ) -in v arian t subset Y ⊆ X on whi h Is( X ) ats minimally (see Prop osition 4.1 ). The set of these minimal on v ex subsets p ossesses a anonial elemen t, whi h is preisely the spae X ′ whi h app ears in Theorem 1.6 . Prop osition 1.5 explains wh y the distintion b et w een X and X ′ did not app ear in Theorem 1.1 . . De Rham deomp ositions. It is kno wn that pro dut deomp ositions of isometry groups ating minimally and without global xed p oin t at innit y indue a splitting of the spae (for o ompat Hadamard manifolds, this is the La wsonY au [ L Y72 ℄ and GromollW olf [ GW71 ℄ theorem; in general and for more referenes, see [ Mon06 ℄). It is therefore natural that Theorem 1.6 is supplemen ted b y a geometri statemen t. A ddendum 1.8. In The or em 1.6 , ther e is a  anoni al isometri de  omp osition (1.ii) X ′ ∼ = X 1 × · · · × X p × R n × Y 1 × · · · × Y q wher e G ∗ ats  omp onentwise a  or ding to ( 1.i ) and G/G ∗ p ermutes any isometri fators. A l l X i and Y j ar e irr e duible and minimal. As it turns out, a geometri deomp osition is the rst of t w o indep enden t steps in the pro of of Theorem 1.6 . In fat, w e b egin with an analogue of the de Rham deomp osition [ dR52 ℄ 6 PIERRE-EMMANUEL CAPRA CE AND NICOLAS MONOD whose pro of uses (a mo diation of ) argumen ts from the generalised de Rham theorem of F o erts hLyt hak [ FL06 ℄. In purely geometrial terms, w e ha v e the follo wing statemen t. Theorem 1.9. L et X b e a pr op er b oundary-minimal CA T(0) sp a e with ∂ X nite-dimensional. Then X admits a  anoni al maximal isometri splitting X ∼ = R n × X 1 × · · · × X m ( n, m ≥ 0) with e ah X i irr e duible and 6 = R 0 , R 1 . Every isometry of X pr eserves this de  omp osition up on p ermuting p ossibly isometri fators X i . Mor e over, if X is minimal, so is every X i . T o apply this theorem, it is desirable to kno w onditions ensuring b oundary-minimalit y . In addition to the onditions pro vided b y Prop osition 1.5 , w e sho w that a anonial b oundary- minimal subspae exists as so on as the b oundary has irumradius > π / 2 (Corollary 3.10 ). In the seond part of the pro of of Theorem 1.6 , w e analyse the irreduible ase where X admits no isometri splitting, resulting in Theorem 1.10 to whi h w e shall no w turn. Com bining these t w o steps, w e then pro v e the unique deomp osition of the gr oups , using also the splitting theorem from [ Mon06 ℄. . Geometry of normal subgroups. In É. Cartan's orresp ondene b et w een symmetri spaes and semi-simple Lie groups as w ell as in BruhatTits theory , irreduible fators of the spae orresp ond to simple groups. F or general CA T(0) spaes and groups, simpliit y fails of ourse v ery dramatially (free groups are p erhaps the simplest, and y et most non- simple, CA T(0) groups). Nonetheless, w e establish a geometri w eak ening of simpliit y . Theorem 1.10. L et X 6 = R b e an irr e duible pr op er CA T(0) sp a e with nite-dimensional Tits b oundary and G < I s ( X ) any sub gr oup whose ation is minimal and do es not have a glob al xe d p oint in ∂ X . Then every non-trivial normal sub gr oup N ✁ G stil l ats minimal ly and without xe d p oint in ∂ X . Mor e over, the amenable r adi al of N and the  entr aliser Z Is( G ) ( N ) ar e b oth trivial; N do es not split as a pr o dut. This result an for instane b e om bined with the solution to Hilb ert's fth problem in order to understand the onneted omp onen t of the isometry group. Corollary 1.11. Is( X ) is either total ly dis onne te d or an almost  onne te d simple Lie gr oup with trivial  entr e. The same holds for any lose d sub gr oup ating minimal ly and without xe d p oint at in- nity. A more elemen tary appliation of Theorem 1.10 uses the fat that elemen ts with a disrete onjugay lass ha v e op en en traliser. Corollary 1.12. If G is non-disr ete, N  annot b e a nitely gener ate d disr ete sub gr oup. A feature of Theorem 1.10 is that is an b e iterated and th us applies to subnormal subgroups. Reall that more generally a subgroup H < G is asending if there is a (p ossibly transnite)  hain of normal subgroups starting with H and abutting to G . Using limiting argumen ts, w e b o otstrap Theorem 1.10 and sho w: Theorem 1.13. L et N < G b e any non-trivial as ending sub gr oup. Then al l  onlusions of The or em 1.10 hold for N . ISOMETR Y GR OUPS OF NON-POSITIVEL Y CUR VED SP A CES: STR UCTURE THEOR Y 7 . A few ases of sup errigidit y . Com bining the preeding general struture results with some of Margulis' theorems, w e obtain the follo wing sup errigidit y statemen t. Theorem 1.14. L et X b e a pr op er CA T(0) sp a e whose isometry gr oup ats  o  omp atly and without glob al xe d p oint at innity. L et Γ = SL n ( Z ) with n ≥ 3 and G = SL n ( R ) . F or any isometri Γ -ation on X ther e is a non-empty Γ -invariant lose d  onvex subset Y ⊆ X on whih the Γ -ation extends uniquely to a  ontinuous isometri ation of G . ( The  orr esp onding statement applies to al l those latti es in semi-simple Lie gr oups that have virtual ly b ounde d gener ation by unip otents. ) Observ e that the ab o v e theorem has no assumptions whatso ev er on the ation; o om- patness is an assumption on the giv en CA T(0) spae. It an happ en that Γ xes p oin ts in ∂ X , but its ation on Y is without xed p oin ts at innit y and minimal (as w e shall establish in the pro of ). The assumption on b ounded generation holds onjeturally for all non-uniform irreduible latties in higher rank semi-simple Lie groups (but alw a ys fails in rank one). It is kno wn to hold for arithmeti groups in split or quasi-split algebrai groups of a n um b er eld K of K - rank ≥ 2 b y [ T a v90 ℄, as w ell as in a few ases of isotropi but non-quasi-split groups [ ER06 ℄; see also [ WM07 ℄. More generally , Theorem 1.14 holds for (S-)arithmeti groups pro vided the arithmeti subgroup (giv en b y in tegers at innite plaes) satises the ab o v e b ounded generation prop- ert y . F or instane, the SL n example is as follo ws: Theorem 1.15. L et X b e a pr op er CA T(0) sp a e whose isometry gr oup ats  o  omp atly and without glob al xe d p oint at innity. L et m b e an inte ger with distint prime fators p 1 , . . . p k and set Γ = SL n ( Z [ 1 m ]) , G = SL n ( R ) × SL n ( Q p 1 ) × · · · × SL n ( Q p k ) , wher e n ≥ 3 . Then for any isometri Γ -ation on X ther e is a non-empty Γ -invariant lose d  onvex subset Y ⊆ X on whih the Γ -ation extends uniquely to a  ontinuous isometri ation of G . W e p oin t out that a xed p oin t prop ert y for similar groups ating on lo w-dimensional CA T(0) ell omplexes w as established b y B. F arb [ F ar08 ℄. Some of our general results also allo w us to impro v e on the generalit y of the CA T(0) su- p errigidit y theorem for irreduible latties in arbitrary pro duts of lo ally ompat groups pro v ed in [ Mon06 ℄. F or ations on prop er CA T(0) spaes, the results of lo . it. establish an unrestrited sup errigidit y on the b oundary but require, in order to dedue sup errigidit y on the spae itself, the assumption that the ation b e r e du e d (or alternativ ely indeom- p osable). W e pro v e that, as so on as the b oundary is nite-dimensional, an y ation without global xed p oin t at innit y is alw a ys redued after suitably passing to subspaes and diret fators. It follo ws that the sup errigidit y theorem for arbitrary pro duts holds in that generalit y , see Theorem 8.4 b elo w. 8 PIERRE-EMMANUEL CAPRA CE AND NICOLAS MONOD Contents 1. In tro dution 1 1.A. Spaes with extensible geo desis 2 1.B. General ase 4 2. Notation and preliminaries 9 3. Con v ex subsets of the Tits b oundary 10 3.A. Boundary subsets of small radius 10 3.B. Subspaes with b oundary of large radius 12 3.C. Minimal ations and b oundary-minimal spaes 13 4. Minimal in v arian t subspaes for subgroups 14 4.A. Existene of a minimal in v arian t subspae 14 4.B. Di hotom y 15 4.C. Normal subgroups 15 5. Algebrai and geometri pro dut deomp ositions 16 5.A. Preliminary deomp osition of the spae 16 5.B. Pro of of Theorem 1.6 and A ddendum 1.8 18 5.C. CA T(0) spaes without Eulidean fator 19 6. T otally disonneted group ations 20 6.A. Smo othness 20 6.B. Lo ally nite equiv arian t partitions and ellular deomp ositions 21 6.C. Alexandro v angle rigidit y 22 6.D. Algebrai struture 23 7. Co ompat CA T(0) spaes 25 7.A. Fixed p oin ts at innit y 25 7.B. A tions of simple algebrai groups 26 7.C. No bran hing geo desis 31 7.D. No op en stabiliser at innit y 32 7.E. Co ompat stabilisers at innit y 33 8. A few ases of CA T(0) sup errigidit y 35 8.A. CA T(0) sup errigidit y for some lassial non-uniform latties 35 8.B. CA T(0) sup errigidit y for irreduible latties in pro duts 37 Referenes 37 ISOMETR Y GR OUPS OF NON-POSITIVEL Y CUR VED SP A CES: STR UCTURE THEOR Y 9 2. Not a tion and preliminaries A metri spae is prop er if ev ery losed ball is ompat. W e refer to Bridson and Haeiger [ BH99 ℄ for ba kground on CA T(0) spaes. W e reall that the omparison angle ∠ p ( x, y ) determined b y three p oin ts p, x, y in an y metri spae is dened purely in terms of the orresp onding three distanes b y lo oking at the orresp ond- ing Eulidean triangle. In other w ords, it is dened b y d 2 ( x, y ) = d 2 ( p, x ) + d 2 ( p, y ) − 2 d ( p, y ) d ( p, y ) cos ∠ p ( x, y ) . The Alexandro v angle ∠ p ( x, y ) in a CA T(0) spae X is the non-inreasing limit of the omparison angle near p along the geo desi segmen ts [ p, x ] and [ p, y ] , see [ BH99 , I I.3.1℄. In partiular, ∠ p ( x, y ) ≤ ∠ p ( x, y ) . Lik ewise, geo desi ra ys from p determine the Alexandro v angle ∠ p ( ξ , η ) for ξ , η ∈ ∂ X . The Tits angle ∠ T ( ξ , η ) is dened as the suprem um of ∠ p ( ξ , η ) o v er all p ∈ X and has sev eral useful  haraterisations giv en in Prop osition I I.9.8 of [ BH99 ℄. Reall that to an y p oin t at innit y ξ ∈ ∂ X is asso iated the Busemann funtion B ξ : X × X → R : ( x, y ) 7→ B ξ ,x ( y ) dened b y B ξ ,x ( y ) = lim t →∞ ( d (  ( t ) , y ) − d (  ( t ) , x )) , where  : [0 , ∞ ) → X is an y geo desi ra y p oin ting to w ards ξ . The Busemann funtion do es not dep end on the  hoie of  and satises the follo wing: B ξ ,x ( y ) = − B ξ ,y ( x ) B ξ ,x ( z ) = B ξ ,x ( y ) + B ξ ,y ( z ) (the o yle relation) B ξ ,x ( y ) ≤ d ( x, y ) . Com bining the denition of the Busemann funtion and of the omparison angle, w e nd that if r is the geo desi ra y p oin ting to w ards ξ with r (0) = x , then for an y y 6 = x w e ha v e lim t →∞ cos ∠ x ( r ( t ) , y ) = − B ξ ,x ( y ) d ( x, y ) (the asymptoti angle form ula) . By abuse of language, one refers to a Busemann funtion when it is more on v enien t to onsider the on v ex 1 -Lips hitz funtion b ξ : X → R dened b y B ξ ,x for some (usually impliit)  hoie of base-p oin t x ∈ X . W e shall simply denote su h a funtion b y b ξ in lo w er ase; they all dier b y a onstan t only in view of the o yle relation. The b oundary at innit y ∂ X is endo w ed with the ne top ology [ BH99 , I I.8.6℄ as w ell as with the (m u h ner) top ology dened b y the Tits angle. The former is often impliitly understo o d, but when referring to dimension or radius, the top ology and distane dened b y the Tits angle are onsidered (this is sometimes emphasised b y referring to the Tits b oundary). The later distane is not to b e onfused with the asso iated length metri alled Tits distane in the literature; w e will not need this onept (exept in the disussions at the b eginning of Setion 7 ). Reall that an y omplete CA T(0) spae splits o a anonial maximal Hilb ertian fator (Eulidean in the prop er ase studied here) and an y isometry deomp oses aordingly , see Theorem I I.6.15(6) in [ BH99 ℄. Normalisers and en tralisers in a group G are resp etiv ely denoted b y N G and Z G . When some group G ats on a set and x is a mem b er of this set, the stabiliser of x in G is denoted b y Stab G ( x ) or b y the shorthand G x . F or the notation regarding algebrai groups, w e follo w the standard notation as in [ Mar91 ℄. 10 PIERRE-EMMANUEL CAPRA CE AND NICOLAS MONOD Finally , w e presen t t w o remarks that will nev er b e used b elo w but giv e some on text on ertain frequen t assumptions. The rst remark is the follo wing CA T(0) v ersion of the HopfRino w theorem: Every ge o desi al ly  omplete lo  al ly  omp at CA T(0) sp a e is pr op er . Surprisingly , w e ould not nd this statemen t in the literature (though a dieren t statemen t is often referred to as the HopfRino w theorem, see [ BH99 , I.3.7℄). As p oin ted out orally b y A. Lyt hak, the ab o v e result is readily established b y follo wing the strategy of pro of of [ BH99 , I.3.7℄ and extending geo desis. The seond fat is that if a prop er CA T(0) spae is nite-dimensional (in the sense of [ Kle99 ℄), then so is its Tits b oundary (generalising for instane Prop osition 7.11 b elo w). The argumen t is giv en in [ CL08 , Prop osition 2.1℄ and ma y b e outlined as follo ws. F or an y sphere S in the spae X , the visual map ∂ X → S is Tits-on tin uous; if it w ere injetiv e, the result w ould follo w. Ho w ev er, it b eomes injetiv e after replaing S with the ultrapro dut of spheres of un b ounded radius b y the v ery denition of the b oundary; the ultrapro dut onstrution preserv es the b ound on the dimension, nishing the pro of. 3. Convex subsets of the Tits bound ar y 3.A. Boundary subsets of small radius. Giv en a metri spae X and a subset Z ⊆ X , one denes the irumradius of Z in X as inf x ∈ X sup z ∈ Z d ( x, z ) . A p oin t x realising the inm um is alled a irumen tre of Z in X . The in trinsi ir- umradius of Z is its irumradius in Z itself; one denes similarly an in trinsi irum- en tre . It is alled anonial if it is xed b y ev ery isometry of X whi h stabilises Z . W e shall mak e frequen t use of the follo wing onstrution of irumen tres, due to A. Balser and A. Lyt hak [ BL05 , Prop osition 1.4℄: Prop osition 3.1. L et X b e a  omplete CA T(1) sp a e and Y ⊆ X b e a nite-dimensional lose d  onvex subset. If Y has intrinsi ir umr adius ≤ π / 2 , then the set C ( Y ) of intrinsi ir um entr es of Y has a unique ir um entr e, whih is ther efor e a  anoni al (intrinsi) ir um entr e of Y .  Let no w X b e a prop er CA T(0) spae. Prop osition 3.2. L et X 0 ⊃ X 1 ⊃ . . . b e a neste d se quen e of non-empty lose d  onvex subsets of X suh that T n X n is empty. Then the interse tion T n ∂ X n is a non-empty lose d  onvex subset of ∂ X of intrinsi ir umr adius at most π / 2 . In p artiular, if the Tits b oundary is nite-dimensional, then T n ∂ X n has a  anoni al intrinsi ir um entr e. Pr o of. Pi k an y x ∈ X and let x n b e its pro jetion to X n . The assumption T n X n = ∅ implies that x n go es to innit y . Up on extrating, w e an assume that it on v erges to some p oin t ξ ∈ ∂ X ; observ e that ξ ∈ T n ∂ X n . W e laim that an y η ∈ T n ∂ X n satises ∠ T ( ξ , η ) ≤ π / 2 . The prop osition then follo ws b eause (i) the b oundary of an y losed on v ex set is losed and π -on v ex [ BH99 , I I.9.13℄ and (ii) ea h ∂ X n is non-empt y sine otherwise X n w ould b e b ounded, on traditing T n X n = ∅ . When ∂ X has nite dimension, there is a anonial in trinsi irumen tre b y Prop osition 3.1 . F or the laim, observ e that there exists a sequene of p oin ts y n ∈ X n on v erging to η . It sues to pro v e that the omparison angle ∠ x ( x n , y n ) is b ounded b y π / 2 for all n , ISOMETR Y GR OUPS OF NON-POSITIVEL Y CUR VED SP A CES: STR UCTURE THEOR Y 11 see [ BH99 , I I.9.16℄. This follo ws from ∠ x n ( x, y n ) ≥ ∠ x n ( x, y n ) ≥ π / 2 , where the seond inequalit y holds b y the prop erties of the pro jetion on a on v ex set [ BH99 , I I.2.4(3)℄.  The om bination of the preeding t w o prop ositions has the follo wing onsequene, whi h impro v es the results established b y F ujiw ara, Nagano and Shio y a (Theorems 1.1 and 1.3 in [ FNS06 ℄). Corollary 3.3. L et g b e a p ar ab oli isometry of X . The fol lowing assertions hold: (i) The xe d p oint set of g in ∂ X has intrinsi ir umr adius at most π / 2 . (ii) If ∂ X nite-dimensional, then the  entr aliser Z Is( X ) ( g ) has a  anoni al glob al xe d p oint in ∂ X . (iii) F or any sub gr oup H < Is( X )  ontaining g , the (p ossibly empty) xe d p oint set of H in ∂ X has ir umr adius at most π / 2 .  Here is another immediate onsequene. Corollary 3.4. L et G b e a top olo gi al gr oup with a  ontinuous ation by isometries on X without glob al xe d p oint. Supp ose that G is the union of an inr e asing se quen e of  omp at sub gr oups and that ∂ X is nite-dimensional. Then ther e is a  anoni al G -xe d p oint in ∂ X , xe d by al l isometries normalising G . Pr o of. Consider the sequene of xed p oin t sets X K n of the ompat subgroups K n . Its in tersetion is empt y b y assumption and th us Prop osition 3.2 applies.  Finally , w e reord the follo wing elemen tary fat, whi h ma y also b e dedued b y means of Prop osition 3.2 : Lemma 3.5. L et ξ ∈ ∂ X . Given any lose d hor ob al l B  entr e d at ξ , the b oundary ∂ B  oinides with the b al l of Tits r adius π / 2  entr e d at ξ in ∂ X . Pr o of. An y t w o horoballs en tred at the same p oin t at innit y lie at b ounded Hausdor dis- tane from one another. Therefore, they ha v e the same b oundary at innit y . In partiular, the b oundary ∂ B of the giv en horoball oinides with the in tersetion of the b oundaries of all horoballs en tred at ξ . By Prop osition 3.2 , this is of irumradius at most π / 2 ; in fat the pro of of that prop osition sho ws preisely that the set is on tained in the ball of radius at most π / 2 around ξ . Con v ersely , let η ∈ ∂ X b e a p oin t whi h do es not b elong to ∂ B . W e laim that ∠ T ( ξ , η ) ≥ π / 2 . This sho ws that ev ery p oin t of ∂ X at Tits distane less than π / 2 from ξ b elongs to ∂ B . Sine the latter is losed, it follo ws that ∂ B on tains the losed ball of Tits radius π / 2 W e turn to the laim. Let b ξ b e a Busemann funtion en tred at ξ . Sine ev ery geo desi ra y p oin ting to w ards η esap es ev ery horoball en tred at ξ , there exists a ra y  : [0 , ∞ ) → X p oin ting to η su h that b ξ (  (0)) = 0 and b ξ (  ( t )) > 0 for all t > 0 (atually , this inreases to innit y b y on v exit y). Let c : [0 , ∞ ) → X b e the geo desi ra y emanating from  (0) and p oin ting to ξ . W e ha v e ∠ T ( ξ , η ) = lim t,s →∞ ∠  (0) (  ( t ) , c ( s )) , see [ BH99 , I I.9.8℄. Therefore the laim follo ws from the asymptoti angle form ula (Setion 2 ) b y taking y = c ( s ) with s large enough.  12 PIERRE-EMMANUEL CAPRA CE AND NICOLAS MONOD 3.B. Subspaes with b oundary of large radius. As b efore, let X b e a prop er CA T(0) spae. The follo wing result impro v es Prop osition 2.2 in [ Lee00 ℄: Prop osition 3.6. L et Y ⊆ X b e a lose d  onvex subset suh that ∂ Y has intrinsi ir- umr adius > π / 2 . Then ther e exists a lose d  onvex subset Z ⊆ X with ∂ Z = ∂ Y whih is minimal for these pr op erties. Mor e over, the union Z 0 of al l suh minimal subsp a es is lose d,  onvex and splits as a pr o dut Z 0 ∼ = Z × Z ′ . Pr o of. If no minimal su h Z existed, there w ould b e a  hain of su h subsets with empt y in tersetion. The distane to a base-p oin t m ust then go to innit y and th us the  hain on- tains a oun table sequene to whi h w e apply Prop osition 3.2 , on traditing the assumption on the irumradius. Let Z ′ denote the set of all su h minimal sets and Z 0 = S Z ′ b e its union. As in [ Lee00 , p. 10℄ one observ es that for an y Z 1 , Z 2 ∈ Z ′ , the distane z 7→ d ( z , Z 2 ) is onstan t on Z 1 and that the nearest p oin t pro jetion p Z 2 restrited to Z 1 denes an isometry Z 1 → Z 2 . By the Sandwi h Lemma [ BH99 , I I.2.12℄, this implies that Z 0 is on v ex and that the map Z ′ × Z ′ → R + : ( Z 1 , Z 2 ) 7→ d ( Z 1 , Z 2 ) is a geo desi metri on Z ′ . As in [ Mon06 , Setion 4.3℄, this yields a bijetion α : Z 0 → Z × Z ′ : x 7→ ( p Z ( x ) , Z x ) , where Z x is the unique elemen t of Z ′ on taining x . The pro dut of metri spaes Z × Z ′ is giv en the pro dut metri. In order to establish that α is an isometry , it remains as in [ Mon06 , Prop osition 38℄, to trivialise holonom y; it the urren t setting, this is a hiev ed b y Lemma 3.7 , whi h th us onludes the pro of of Prop osition 3.6 . (Notie that Z 0 is indeed losed sine otherwise w e ould extend α − 1 to the ompletion of Z × Z ′ .)  Lemma 3.7. F or al l Z 1 , Z 2 , Z 3 ∈ Z ′ , we have p Z 1 ◦ p Z 3 ◦ p Z 2 | Z 1 = Id Z 1 . Pr o of of L emma 3.7 . Let ϑ : Z 1 → Z 1 b e the isometry dened b y p Z 1 ◦ p Z 3 ◦ p Z 2 | Z 1 and let f b e its displaemen t funtion. Then f : Z 1 → R is a non-negativ e on v ex funtion whi h is b ounded ab o v e b y d ( Z 1 , Z 2 ) + d ( Z 2 , Z 3 ) + d ( Z 3 , Z 1 ) . In partiular, the restrition of f to an y geo desi ra y in Z 1 is non-inreasing. Therefore, a sublev el set of f is a losed on v ex subset Z of Z 1 with full b oundary , namely ∂ Z = ∂ Z 1 . By denition, the subspae Z 1 is minimal with resp et to the prop ert y that ∂ Z 1 = ∂ Y and hene w e dedue Z = Z 1 . It follo ws that the on v ex funtion f is onstan t. In other w ords, the isometry ϑ is a Cliord translation. If it is not trivial, then Z 1 w ould on tain a ϑ -stable geo desi line on whi h ϑ ats b y translation. But b y [ BH99 , Lemma I I.2.15℄, the restrition of ϑ to an y geo desi line is the iden tit y . Therefore ϑ is trivial, as desired.  Let Γ b e a group ating on X b y isometries. F ollo wing [ Mon06 , Denition 5℄, w e sa y that the Γ -ation is redued if there is no un b ounded losed on v ex subset Y ( X su h that g .Y is at nite Hausdor distane from Y for all g ∈ Γ . Corollary 3.8. L et X b e a pr op er irr e duible CA T(0) sp a e with nite-dimensional Tits b oundary, and Γ < Is( X ) b e a sub gr oup ating minimal ly without xe d p oint at innity. Then the Γ -ation is r e du e d. Pr o of. Supp ose for a on tradition that the Γ -ation on X is not redued. Then there exists an un b ounded losed on v ex subset Y ( X su h that g .Y is at nite Hausdor distane from Y for all g ∈ Γ . In partiular ∂ Y is Γ -in v arian t. By Prop osition 3.1 , it m ust ha v e in trinsi irumradius > π / 2 . Prop osition 3.6 therefore yields a anonial losed on v ex subset Z 0 = Z × Z ′ with ∂ ( Z × { z ′ } ) = ∂ Y for all z ′ ∈ Z ′ ; learly Z 0 is Γ -in v arian t and hene w e ha v e Z 0 = X b y minimalit y . Sine X is irreduible b y assumption, w e dedue X = Z and hene X = Y , as desired.  ISOMETR Y GR OUPS OF NON-POSITIVEL Y CUR VED SP A CES: STR UCTURE THEOR Y 13 3.C. Minimal ations and b oundary-minimal spaes. Boundary-minimalit y and min- imalit y , as dened in the In tro dution, are t w o p ossible w a ys for a CA T(0) spae to b e non-degenerate, as illustrated b y the follo wing. Lemma 3.9. L et X b e a  omplete CA T(0) sp a e. (i) A gr oup G < Is( X ) ats minimal ly if and only if any  ontinuous  onvex G -invariant funtion on X is  onstant. (ii) If X is b oundary-minimal then any b ounde d  onvex funtion on X is  onstant. Pr o of. Neessit y in the rst assertion follo ws immediately b y onsidering sub-lev el sets (see [ Mon06 , Lemma 37℄). Suieny is due to the fat that the distane to a losed on v ex set is a on v ex on tin uous funtion [ BH99 , I I.2.5℄. The seond assertion w as established in the pro of of Lemma 3.7 .  Prop osition 3.6 has the follo wing imp ortan t onsequene: Corollary 3.10. L et X b e a pr op er CA T(0) sp a e. If ∂ X has ir umr adius > π / 2 , then X p ossesses a  anoni al lose d  onvex subsp a e Y ⊆ X suh that Y is b oundary-minimal and ∂ Y = ∂ X . Pr o of. Let Z 0 = Z × Z ′ b e the pro dut deomp osition pro vided b y Prop osition 3.6 . The group Is( X ) p erm utes the elemen ts of Z ′ and hene ats b y isometries on Z ′ . Under the presen t h yp otheses, the spae Z ′ is b ounded sine ∂ Z = ∂ X . Therefore it has a irumen tre z ′ , and the bre Y = Z × { z ′ } is th us Is( X ) -in v arian t.  Prop osition 3.11. L et X b e a pr op er CA T(0) sp a e whih is minimal. Assume that ∂ X has nite dimension. Then ∂ X has ir umr adius > π / 2 (unless X is r e du e d to a p oint). In p artiular, X is b oundary-minimal. The pro of of Prop osition 3.11 requires some preliminaries. Giv en a p oin t at innit y ξ , onsider the Busemann funtion B ξ ; the o yle prop ert y (realled in Setion 2 ) implies in partiular that for an y isometry g ∈ Is( X ) xing ξ and an y x ∈ X the real n um b er B ξ ,x ( g .x ) is indep enden t on the  hoie of x and yields a anonial homomorphism β ξ : Is( X ) ξ − → R : g 7− → B ξ ,x ( g .x ) alled the Busemann  harater en tred at ξ . Giv en an isometry g , it follo ws b y the CA T(0) prop ert y that inf n ≥ 0 d ( g n x, x ) /n oinides with the translation length of g indep enden tly of x . W e all an isometry ballisti when this n um b er is p ositiv e. An imp ortan t fat ab out a ballisti isometry g of an y omplete CA T(0) spae X is that for an y x ∈ X the sequene { g n .x } n ≥ 0 on v erges to a p oin t η g ∈ ∂ X indep enden t of x ; η g is alled the (anonial) attrating xed p oin t of g in ∂ X . Moreo v er, this on v ergene holds also in angle, whi h means that lim ∠ x ( g n x, r ( t )) v anishes as n, t → ∞ when r : R + → X is an y ra y p oin ting to η g . This is a (v ery) sp eial ase of the results in [ KM99 ℄. Lemma 3.12. L et ξ ∈ X and g ∈ Is( X ) ξ b e an isometry whih is not annihilate d by the Busemann har ater  entr e d at ξ . Then g is b al listi. F urthermor e, if β ξ ( g ) > 0 then ∠ T ( ξ , η g ) > π / 2 . Pr o of. W e ha v e β ξ ( g ) = B ξ ,x ( g .x ) ≤ d ( x, g .x ) for all x ∈ X . Th us g is ballisti as so on as β ξ ( g ) is non-zero. Assume β ξ ( g ) > 0 and supp ose for a on tradition that ∠ T ( ξ , η g ) ≤ π / 2 . Cho ose x ∈ X and let , σ b e the ra ys issuing from x and p oin ting to w ards ξ and η g resp etiv ely . Reall 14 PIERRE-EMMANUEL CAPRA CE AND NICOLAS MONOD from [ BH99 , I I.9.8℄ that ∠ T ( ξ , η g ) = lim t,s →∞ ∠ x (  ( t ) , σ ( s )) . The on v ergene in diretion of g n x implies that this angle is also giv en b y lim t,n →∞ ∠ x (  ( t ) , g n x ) . Sine β ξ ( g ) > 0 w e an x n large enough to ha v e cos lim in f t →∞ ∠ x (  ( t ) , g n x ) > − β ξ ( g ) d ( g x, x ) . W e no w apply the asymptoti angle form ula from Setion 2 with y = g n x and dedue that the left hand side is − β ξ ( g n x ) /d ( g n x, x ) . Sine β ξ ( g n x ) = nβ ξ ( g x ) and d ( g n x, x ) ≤ nd ( g x, x ) , w e ha v e a on tradition.  Pr o of of Pr op osition 3.11 . W e an assume that ∂ X is non-empt y sine otherwise X is a p oin t b y minimalit y . Supp ose for a on tradition that its irumradius is ≤ π / 2 . Then Is( X ) p ossesses a global xed p oin t ξ ∈ ∂ X and ξ is a irumen tre of ∂ X , see Prop osition 3.1 . Lemma 3.12 implies that Is( X ) = I s ( X ) ξ is annihilated b y the Busemann  harater en tred at ξ . Th us Is( X ) stabilises ev ery horoball, on traditing minimalit y .  W e shall use rep eatedly the follo wing elemen tary fat. Lemma 3.13. L et G b e a gr oup with an isometri ation on a pr op er ge o desi al ly  omplete CA T(0) sp a e X . If G ats  o  omp atly or mor e gener al ly has ful l limit set, then the ation is minimal. (This holds mor e gener al ly when ∆ G = ∂ X in the sense of Se tion 4.B b elow.) Pr o of. Let Y ⊆ X b e a a non-empt y losed on v ex in v arian t subset,  ho ose y ∈ Y and supp ose for a on tradition that there is x / ∈ Y . Let r : R + → X b e a geo desi ra y starting at y and going through x . By on v exit y [ BH99 , I I.2.5(1)℄, the funtion d ( r ( t ) , Y ) tends to innit y and th us r ( ∞ ) / ∈ ∂ Y . This is absurd sine ∆ G ⊆ ∂ Y .  Pr o of of Pr op osition 1.5 . (i) See Prop osition 3.11 . (ii) Sine Is( X ) has full limit set, an y Is( X ) -in v arian t subspae has full b oundary . Minimalit y follo ws, sine b oundary-minimalit y ensures that X p ossesses no prop er subspae with full b oundary . (iii) X is minimal b y Lemma 3.13 , hene b oundary-minimal b y (i), sine an y o ompat spae has nite-dimensional b oundary b y [ Kle99 , Theorem C℄.  4. Minimal inv ariant subsp a es f or subgr oups 4.A. Existene of a minimal in v arian t subspae. F or the reord, w e reall the follo wing elemen tary di hotom y; a renemen t will b e giv en in Theorem 4.3 b elo w: Prop osition 4.1. L et G b e a gr oup ating by isometries on a pr op er CA T(0) . Then either G has a glob al xe d p oint at innity, or any ltering family of non-empty lose d  onvex G -invariant subsets has non-empty interse tion. (Reall that a family of sets is ltering if it is direted b y on tainmen t ⊇ .) Pr o of. (Remark 36 in [ Mon06 ℄.) Supp ose Y is su h a family ,  ho ose x ∈ X and let x Y b e its pro jetion on ea h Y ∈ Y . If the net { x Y } Y ∈ Y is b ounded, then T Y ∈ Y Y is non- empt y . Otherwise it go es to innit y and an y aum ulation p oin t in ∂ X is G -xed in view of d ( g x Y , x Y ) ≤ d ( g x, x ) .  ISOMETR Y GR OUPS OF NON-POSITIVEL Y CUR VED SP A CES: STR UCTURE THEOR Y 15 4.B. Di hotom y. Let G b e a group ating b y isometries on a omplete CA T(0) spae X . Lemma 4.2. Given any two x, y ∈ X , the  onvex losur es of the r esp e tive G -orbits of x and y in X have the same b oundary in ∂ X . Pr o of. Let Y b e the on v ex losure of the G -orbit of x . In partiular Y is the minimal losed on v ex G -in v arian t subset on taining x . Giv en an y losed on v ex G -in v arian t subset Z , let r = d ( x, Z ) . Reall that the tubular losed neigh b ourho o d N r ( Z ) is on v ex [ BH99 , I I.2.5(1)℄. Sine it is also G -in v arian t and on tains x , the minimalit y of Y implies Y ⊆ N r ( Z ) .  This yields a anonial losed on v ex G -in v arian t subset of the b oundary ∂ X , whi h w e denote b y ∆ G . It on tains the limit set Λ G but is sometimes larger. Com bining what w e established th us far with the splitting argumen ts from [ Mon06 ℄, w e obtain a di hotom y: Theorem 4.3. L et G b e a gr oup ating by isometries on a  omplete CA T(0) sp a e X and H < G any sub gr oup. If H admits no minimal non-empty lose d  onvex invariant subset and X is pr op er, then: (A.i) ∆ H is a non-empty lose d  onvex subset of ∂ X of intrinsi ir umr adius at most π / 2 . (A.ii) If ∂ X is nite-dimensional, then the normaliser N G ( H ) of H in G has a glob al xe d p oint in ∂ X . If H admits a minimal non-empty lose d  onvex invariant subset Y ⊆ X , then: (B.i) The union Z of al l suh subsets is a lose d  onvex N G ( H ) -invariant subset. (B.ii) Z splits H -e quivariantly and isometri al ly as a pr o dut Z ≃ Y × C , wher e C is a  omplete CA T(0) sp a e whih admits a  anoni al N G ( H ) /H -ation by isometries. (B.iii) If the H -ation on X is non-evanes ent, then C is b ounde d and ther e is a  anoni al minimal non-empty lose d  onvex H -invariant subset whih is N G ( H ) -stable. (When X is prop er, the non-evanes en e ondition of (iii) simply means that H has no xed p oin t in ∂ X ; see [ Mon06 ℄.) Pr o of. In view of Lemma 4.2 , the set ∆ H is on tained in the b oundary of an y non-empt y losed on v ex H -in v arian t set and is N G ( H ) -in v arian t. Th us the assertions (A.i) and (A.ii) follo w from Prop osition 3.2 , notiing that in a prop er spae ∆ H is non-empt y unless H has b ounded orbits, in whi h ase it xes a p oin t, pro viding a minimal subspae. F or (B.i), (B.ii) and (B.iii), see Remarks 39 in [ Mon06 ℄.  4.C. Normal subgroups. Pr o of of The or em 1.10 . W e adopt the notation and assumptions of the theorem. By (A.ii), N admits a minimal non-empt y losed on v ex in v arian t subset Y ⊆ X . This set is un- b ounded, sine otherwise N xes a p oin t and th us b y G -minimalit y X N = X , hene N = 1 . Sine X is irreduible, p oin ts (B.i) and (B.ii) sho w Y = X and th us N ats indeed minimally . Sine the displaemen t funtion of an y g ∈ Z G ( N ) is a on v ex N -in v arian t funtion, it is onstan t b y minimalit y . Hene g is a Cliord translation and m ust b e trivial sine otherwise X splits o a Eulidean fator, see [ BH99 , I I.6.15℄. The deriv ed subgroup N ′ = [ N , N ] is also normal in G and therefore ats minimally b y the previous disussion, notiing that N ′ is non-trivial sine otherwise N ⊆ Z G ( N ) . If N xed a p oin t at innit y , N ′ w ould preserv e all orresp onding horoballs, on traditing minimalit y . 16 PIERRE-EMMANUEL CAPRA CE AND NICOLAS MONOD Ha ving established that N ats minimally and without xed p oin t at innit y , w e an apply the splitting theorem (Corollary 10 in [ Mon06 ℄) and dedue from the irreduibilit y of X that N do es not split. Finally , let R ✁ N b e the amenable radial and observ e that it is normal in G . The theorem of A damsBallmann [ AB98a ℄ states that R either (i) xes a p oin t at innit y or (ii) preserv es a Eulidean at in X . (Although their result is stated for amenable groups without men tioning an y top ology , the pro of applies indeed to ev ery top ologial group that preserv es a probabilit y measure whenev er it ats on tin uously on a ompat metrisable spae.) If R is non-trivial, w e kno w already from the ab o v e disussion that (i) is imp ossible and that R ats minimally; it follo ws that X is a at. By irreduibilit y and sine X 6 = R , this fores X to b e a p oin t, on traditing R 6 = 1 .  Corollary 1.11 will b e pro v ed in Setion 5.B . F or Corollary 1.12 , it sues to observ e that the en traliser of an y elemen t of a disrete normal subgroup is op en. Next, w e reall the follo wing denition. A subgroup N of a group G is asending if there is a family of subgroups N α < G indexed b y the ordinals and su h that N 0 = N , N α ✁ N α +1 , N α = S β <α N β if α is a limit ordinal and N α = G for α large enough. The smallest su h ordinal is the order . Prop osition 4.4. Consider a gr oup ating minimal ly by isometries on a pr op er CA T(0) sp a e. Then any as ending sub gr oup without glob al xe d p oint at innity stil l ats minimal ly. Pr o of. W e argue b y transnite indution on the order ϑ of asending subgroups N < G , the ase ϑ = 0 b eing trivial. Let X b e a spae as in the statemen t. By Prop osition 4.1 , ea h N α has a minimal set. If ϑ = ϑ ′ + 1 , it follo ws from (B.iii) that N ϑ ′ ats minimally and w e are done b y indution h yp othesis. Assume no w that ϑ is a limit ordinal. F or all α , w e denote as in (B.i) b y Z α ⊆ X the union of all N α -minimal sets. The indution h yp othesis implies that for all α ≤ β < ϑ , an y N β -minimal set is N α -minimal. Th us, if Z 0 = Y 0 × C 0 is a splitting as in (B.ii) with a N -minimal set Y 0 , w e ha v e a nested family of deomp ositions Z α = Y 0 × C α for a nested family of losed on v ex subspaes C α of the ompat CA T(0) spae C 0 , indexed b y α < ϑ . Th us, for an y c ∈ T α<ϑ C α , the spae Y 0 × { c } is G -in v arian t and hene Y 0 = X indeed.  Remark 4.5. Prop osition 4.4 holds more generally for omplete CA T(0) spaes if N is non-ev anesen t. Indeed Prop osition 4.1 hold in that generalit y (Remark 36 in [ Mon06 ℄) and C remains ompat in a w eak er top ology (Theorem 14 in [ Mon06 ℄). Pr o of of The or em 1.13 . In view of Theorem 1.10 , it sues to pro v e that an y non-trivial asending subgroup N < G as in that statemen t still ats minimally and without global xed p oin t at innit y . W e argue b y indution on the order ϑ and w e an assume that ϑ is a limit ordinal b y Theorem 1.10 . Then T α<ϑ ( ∂ X ) N α is empt y and th us b y ompatness there is some α < ϑ su h that ( ∂ X ) N α is empt y . No w N α ats minimally on X b y Prop osition 4.4 and th us w e onlude using the indution h yp othesis.  5. Algebrai and geometri pr odut deompositions 5.A. Preliminary deomp osition of the spae. W e shall prepare our spaes b y means of a geometri deomp osition. F or an y geo desi metri spae with nite ane r ank , F o erts h Lyt hak [ FL06 ℄ established a anonial deomp osition generalising the lassial theorem of de Rham [ dR52 ℄. Ho w ev er, su h a statemen t fails to b e true for CA T(0) spaes that ISOMETR Y GR OUPS OF NON-POSITIVEL Y CUR VED SP A CES: STR UCTURE THEOR Y 17 are merely prop er, due notably to ompat fators that an b e innite pro duts. Nev er- theless, using asymptoti CA T(0) geometry and Setion 3.A , w e an adapt the argumen ts from [ FL06 ℄ and obtain: Theorem 5.1. L et X b e a pr op er CA T(0) sp a e with ∂ X nite-dimensional and of ir- umr adius > π / 2 . Then ther e is a  anoni al lose d  onvex subset Z ⊆ X with ∂ Z = ∂ X , invariant under al l isometries, and admitting a  anoni al maximal isometri splitting (5.i) Z ∼ = R n × Z 1 × · · · × Z m ( n, m ≥ 0) with e ah Z i irr e duible and 6 = R 0 , R 1 . Every isometry of Z pr eserves this de  omp osition up on p ermuting p ossibly isometri fators Z i . Remark 5.2. It is w ell kno wn that in the ab o v e situation the splitting ( 5.i ) indues a deomp osition Is( Z ) = Is( R n ) ×   Is( Z 1 ) × · · · × Is( Z m )  ⋊ F  , where F is the p erm utation group of { 1 , . . . , d } p erm uting p ossible isometri fators amongst the Y j . Indeed, this follo ws from the statemen t that isometries preserv e the splitting up on p erm utation of fators, see e.g. Prop osition I.5.3(4) in [ BH99 ℄. Of ourse, this do es not a priori mean that w e ha v e a unique, nor ev en anonial, splitting in the  ate gory of gr oups ; this shall ho w ev er b e established for Theorem 1.6 . The h yp otheses of Theorem 5.1 are satised in some naturally o urring situations: Corollary 5.3. L et X b e a pr op er CA T(0) sp a e with nite-dimensional b oundary. (i) If Is( X ) has no xe d p oint at innity, then X p ossesses a subsp a e Z satisfying al l the  onlusions of The or em 5.1 . (ii) If Is( X ) ats minimal ly, then X admits a  anoni al splitting as in 5.i . Pr o of of Cor ol lary 5.3 . By Prop osition 3.1 , if Is( X ) has no xed p oin t at innit y , then ∂ X has irumradius > π / 2 . By Prop osition 3.11 , the same onlusion holds is Is( X ) ats minimally .  Pr o of of The or ems 1.9 and 5.1 . F or Theorem 5.1 , w e let Z ⊆ X b e the anonial b oundary- minimal subset with ∂ Z = ∂ X pro vided b y Corollary 3.10 ; w e shall not use the irumradius assumption an y more. F or Theorem 1.9 , w e let Z = X . The remainder of the argumen t is ommon for b oth statemen ts. Realling that in omplete generalit y all isometries preserving the Eulidean fator de- omp osition [ BH99 , I I.6.15℄, w e an assume that Z has no Eulidean fator and shall obtain the deomp osition ( 5.i ) with n = 0 . Sine Z is minimal amongst losed on v ex subsets with ∂ Z = ∂ X , it has no non-trivial ompat fator. On the other hand, an y prop er geo desi metri spae admits some maximal pro dut deomp osition in to non-ompat fators. In onlusion, Z admits some maximal splitting Z = Z 1 × · · · × Z m with ea h Z i irreduible and 6 = R 0 , R 1 . (This an fail in presene of ompat fators). It remains to pro v e that an y other su h deomp osition Z = Z ′ 1 × · · · × Z ′ m ′ oinides with the rst one after p ossibly p erm uting the fators (in partiular, m ′ = m ). W e no w b orro w from the argumen tation in [ FL06 ℄, indiating the steps and the neessary  hanges. It is assumed that the reader has a op y of [ FL06 ℄ at hand but k eeps in mind that our spaes migh t la k the nite ane rank ondition assumed in that pap er. W e shall replae the notion of ane subspaes with a large-sale partiular ase: a ne shall b e an y subspae 18 PIERRE-EMMANUEL CAPRA CE AND NICOLAS MONOD isometri to a losed on v ex ne in some Eulidean spae. This inludes the partiular ases of a p oin t, a ra y or a full Eulidean spae. Whenev er a spae Y has some pro dut deomp osition and Y ′ is a fator, write Y ′ y ⊆ Y for the orresp onding bre Y ′ y ∼ = Y ′ through y ∈ Y . The follo wing is an analogue of Corollary 1.2 in [ FL06 ℄. Lemma 5.4. L et Y b e a pr op er CA T(0) sp a e with nite-dimensional b oundary and without  omp at fators. Supp ose given two de  omp ositions Y = Y 1 × Y 2 = S 1 × S 2 with al l four ( Y i ) y ∩ ( S j ) y r e du e d to { y } for some y ∈ Y . Then Y is a Eulide an sp a e. Pr o of of L emma 5.4 . An y y ∈ Y is on tained in a maximal ne based at y sine ∂ Y has nite dimension; b y abuse of language w e all su h nes maximal. The argumen ts of Setions 3 and 4 in [ FL06 ℄ sho w that an y maximal ne is r e tangular , whi h means that it inherit a pro dut struture from an y pro dut deomp osition of the am bien t CA T(0) spae. Sp eially , it sues to observ e that the pro dut of t w o nes is a ne and that the pro jetion of a ne along a pro dut deomp osition of CA T(0) spaes remains a ne. (In fat, the equalit y of slop es of Setion 4.2 in [ FL06 ℄, namely the fat that parallel geo desi segmen ts in a CA T(0) spae ha v e iden tial slop es in pro dut deomp ositions, is a general fat for CA T(0) spaes. It follo ws from the on v exit y of the metri, see for instane [ Mon06 , Prop osition 49℄ for a more general statemen t.) The dedution of the statemen t of Lemma 5.4 from the retangularit y of maximal nes follo wing [ FL06 ℄ is partiularly short sine all prop er CA T(0) Bana h spaes are Eulidean.  Lemma 5.5. F or a given z ∈ Z and any pr o dut de  omp osition Z = S × S ′ , the interse tion ( Z i ) z ∩ S z is either { z } or ( Z i ) z . Pr o of of L emma 5.5 . W rite P S : Z → S and P Z i : Z → Z i for the pro jetions and set F z = S z ∩ ( Z i ) z . F ollo wing [ FL06 ℄, dene T ⊆ Z b y T = P S ( F z ) × S ′ . W e on tend that P Z i ( T ) has full b oundary in Z i . Indeed, giv en an y p oin t in ∂ ( Z i ) z , w e represen t is b y a ra y r originating from z . W e an  ho ose a maximal ne in Z based at z and on taining r . W e kno w already that this ne is retangular, and therefore the pro of of Lemma 5.2 in [ FL06 ℄ sho ws that P Z i ( r ) lies in P Z i ( T ) , justifying our on ten tion. W e observ e that Z i inherits from Z the prop ert y that it has no losed on v ex prop er subset of full b oundary . In onlusion, sine P Z i ( T ) is a on v ex set, it is dense in Z . Ho w ev er, aording to Lemma 5.1 in [ FL06 ℄, it splits as P Z i ( T ) = P Z i ( F z ) × P Z i ( S ′ ) . Up on p ossibly replaing P Z i ( S ′ ) b y its ompletion (whilst P Z i ( F z ) is already losed in Z i sine P Z i is isometri on ( Z i ) z ), w e obtain a splitting of the losure of P Z i ( T ) , and hene of Z i . This ompletes the pro of of the lemma sine Z i is irreduible.  No w the main argumen t runs b y indution o v er m ≥ 2 . Lemma 5.5 iden ties b y indution Z i with some Z ′ j . Indeed, Lemma 5.4 exludes that all pairwise in tersetions redue to a p oin t sine Z has no Eulidean fator.  5.B. Pro of of Theorem 1.6 and A ddendum 1.8 . The follo wing onsequene of the solution to Hilb ert's fth problem b elongs to the mathematial lore. Theorem 5.6. L et G b e a lo  al ly  omp at gr oup with trivial amenable r adi al. Then G p ossesses a  anoni al nite index op en normal sub gr oup G † suh that G † = L × D , wher e L is a  onne te d semi-simple Lie gr oup with trivial  entr e and no  omp at fators, and D is total ly dis onne te d. ISOMETR Y GR OUPS OF NON-POSITIVEL Y CUR VED SP A CES: STR UCTURE THEOR Y 19 Pr o of. This follo ws from the GleasonMon tgomeryZippin solution to Hilb ert's fth prob- lem and the fat that onneted semi-simple Lie groups ha v e nite outer automorphism groups. More details ma y b e found for example in [ Mon01 ,  11.3℄.  Com bining Theorem 5.6 with Theorem 1.10 , w e nd the statemen t giv en as Corollary 1.11 in the In tro dution. Theorem 5.7. L et X 6 = R b e an irr e duible pr op er CA T(0) sp a e with nite-dimensional Tits b oundary and G < Is( X ) any lose d sub gr oup whose ation is minimal and do es not have a glob al xe d p oint in ∂ X . Then G is either total ly dis onne te d or an almost  onne te d simple Lie gr oup with trivial  entr e. Pr o of. By Theorem 1.10 , G has trivial amenable radial. Let G † b e as in Theorem 5.6 . Applying Theorem 1.10 to this normal subgroup of G , dedue that w e ha v e either G † = L with L simple or G † = D .  W e an no w omplete the pro of of Theorem 1.6 and A ddendum 1.8 and w e adopt their notation. Sine Is( X ) has no global xed p oin t at innit y , there is a anonial minimal non-empt y losed on v ex Is( X ) -in v arian t subset X ′ ⊆ X (Remarks 39 in [ Mon06 ℄). W e apply Corollary 5.3 to Z = X ′ and Remark 5.2 to G = Is( Z ) , setting G ∗ = Is( R n ) × Is( Z 1 ) × · · · × Is( Z m ) . All the laimed prop erties of the resulting fator groups are established in Theorem 1.10 , Theorem 1.13 and Theorem 1.1 in [ CM08b ℄ (the pro of of whi h is ompletely indep enden t from the presen t onsiderations). Finally , the laim that an y pro dut deomp osition of G ∗ is a regrouping of the fators in ( 1.i ) is established as follo ws. Notie that the G ∗ -ation on Z is still minimal and without xed p oin t at innit y (this is almost b y denition but alternativ ely also follo ws from Theorem 1.10 ). Therefore, giv en an y pro dut deomp osition of G ∗ , w e an apply the splitting theorem (Corollary 10 in [ Mon06 ℄) and obtain a orresp onding splitting of Z . No w the uniqueness of the deomp osition of the spae Z (a w a y from the Eulidean fator) implies that the giv en deomp osition of G ∗ is a regrouping of the fators o urring in Remark 5.2 .  5.C. CA T(0) spaes without Eulidean fator. F or the sak e of future referenes, w e reord the follo wing onsequene of the results obtained th us far: Corollary 5.8. L et X b e a pr op er CA T(0) sp a e with nite-dimensional b oundary and no Eulide an fator, suh that G = Is ( X ) ats minimal ly without xe d p oint at innity. Then G has trivial amenable r adi al and any sub gr oup of G ating minimal ly on X has triv- ial  entr aliser. F urthermor e, given a non-trivial normal sub gr oup N ✁ G , any N -minimal N -invariant lose d subsp a e of X is a r e gr ouping of fators in the de  omp osition of A d- dendum 1.8 . In p artiular, if e ah irr e duible fator of G is non-disr ete, then G has no non-trivial nitely gener ate d lose d normal sub gr oup. Pr o of. The trivialit y of the amenable radial omes from the orresp onding statemen t in irreduible fators of X , see Theorem 1.10 . By the seond paragraph of the pro of of Theo- rem 1.10 , an y subgroup of G ating minimally has trivial en traliser. The fat that minimal in v arian t subspaes for normal subgroups are bres in the pro dut deomp osition ( 1.ii ) fol- lo ws sine an y pro dut deomp osition of X is a regrouping of fators in ( 1.ii ) and sine an y normal subgroup of G yields su h a pro dut deomp osition b y Theorem 4.3 (B.i) and (B.ii). Assume nally that ea h irreduible fator in ( 1.i ) is non-disrete and let N < G b e a nitely 20 PIERRE-EMMANUEL CAPRA CE AND NICOLAS MONOD generated losed normal subgroup. Then N is disrete b y Baire's ategory theorem, and N ats minimally on a bre, sa y Y , of the spae deomp osition ( 1.ii ). Therefore, the pro jetion of N to Is( Y ) has trivial en traliser, unless N is trivial. Sine N is disrete, normal and nitely generated, its en traliser is op en. Sine Is( Y ) is non-disrete b y assumption, w e dedue that N is trivial, as desired.  6. Tot all y disonneted gr oup a tions 6.A. Smo othness. When onsidering ations of totally disonneted groups, a desirable prop ert y is smo othness , namely that p oin ts ha v e op en stabilisers. This ondition is imp or- tan t in represen tation theory , but also in our geometri on text, see p oin t ( ii ) of Corollary 6.3 b elo w and [ Cap07 ℄. In general, this ondition do es not hold, ev en for ations that are o ompat, minimal and without xed p oin t at innit y . An example will b e onstruted in Setion 6.C in [ CM08b ℄. Ho w ev er, w e establish it under a rather ommon additional h yp othesis. Reall that a metri spae X is alled geo desially omplete (or said to ha v e extensible geo desis ) if ev ery geo desi segmen t of p ositiv e length ma y b e extended to a lo ally isometri em b edding of the whole real line. The follo wing on tains Theorem 1.2 from the In tro dution. Theorem 6.1. L et G b e a total ly dis onne te d lo  al ly  omp at gr oup with a minimal,  on- tinuous and pr op er ation by isometries on a pr op er CA T(0) sp a e X . If X is ge o desi al ly  omplete, then the ation is smo oth. In fat, the p ointwise stabiliser of every b ounde d set is op en. Remark 6.2. In partiular, the stabiliser of a p oin t ats as a nite group of isometries on an y giv en ball around this p oin t in the setting of Theorem 6.1 . Corollary 6.3. L et X b e a pr op er CA T(0) sp a e and G b e a total ly dis onne te d lo  al ly  omp at gr oup ating  ontinuously pr op erly on X by isometries. Then: (i) If the G -ation is  o  omp at, then every element of zer o tr anslation length is el lipti. (ii) If the G -ation is  o  omp at and every p oint x ∈ X has an op en stabiliser, then the G -ation is semi-simple. (iii) If the G -ation is  o  omp at and X is ge o desi al ly  omplete, then the G -ation is semi-simple. Pr o of of Cor ol lary 6.3 . P oin ts ( i ) and ( ii ) follo w readily from Theorem 6.1 , see [ Cap07 , Corollary 3.3℄. ( iii ) In view of Lemma 3.13 , this follo ws from Theorem 6.1 and ( ii ) .  The follo wing is a k ey fat for Theorem 6.1 : Lemma 6.4. L et X b e a ge o desi al ly  omplete pr op er CA T(0) sp a e. L et ( C n ) n ≥ 0 b e an inr e asing se quen e of lose d  onvex subsets whose union C = S n C n is dense in X . Then every b ounde d subset of X is  ontaine d in some C n ; in p artiular, C = X . Pr o of. Supp ose for a on tradition that for some r > 0 and x ∈ X the r -ball around x on- tains an elemen t x n not in C n for ea h n . W e shall onstrut indutiv ely a sequene { c k } k ≥ 1 of pairwise r -disjoin t elemen ts in C with d ( x, c k ) ≤ 2 r + 2 , on traditing the prop erness of X . If c 1 , . . . , c k − 1 ha v e b een onstruted,  ho ose n large enough to that C n on tains them all and d ( x, C n ) ≤ 1 . Consider the (non-trivial) geo desi segmen t from x n to its nearest p oin t pro jetion x n on C n ; b y geo desi ompleteness, it is on tained in a geo desi line and ISOMETR Y GR OUPS OF NON-POSITIVEL Y CUR VED SP A CES: STR UCTURE THEOR Y 21 w e  ho ose y at distane r + 1 from C n on this line. Notie that x n ∈ [ x n , y ] and hene d ( y , x ) ≤ 2 r + 1 . Moreo v er, d ( y , c i ) ≥ r + 1 for all i < k . Sine C is dense, w e an  ho ose c k lose enough to y to ensure d ( c k , x ) ≤ 2 r + 2 and d ( c k , c i ) ≥ r for all i < k , ompleting the indution step.  End of pr o of of The or em 6.1 . The subset C ⊆ X onsisting of those p oin ts x ∈ X su h that the stabiliser G x is op en is learly on v ex and G -stable. By [ Bou71 , I I I  4 No 6℄, the group G on tains a ompat op en subgroup and hene C is non-empt y . Th us C is dense b y minimalit y of the ation. Sine Is( X ) is seond oun table, w e an  ho ose a desending  hain Q n < G of ompat op en subgroups whose in tersetion ats trivially on X . Therefore, C ma y b e written as the union of an asending family of losed on v ex subsets C n ⊆ X , where C n is the xed p oin t set of Q n . No w the statemen t of the theorem follo ws from Lemma 6.4 .  6.B. Lo ally nite equiv arian t partitions and ellular deomp ositions. Let X b e a lo ally nite ell omplex and G b e its group of ellular automorphisms, endo w ed with the top ology of p oin t wise on v ergene on b ounded subsets. Then G is a totally disonneted lo ally ompat group and ev ery b ounded subset of X has an op en p oin t wise stabiliser in G . One of the in terest of Theorem 6.1 is that it allo ws for a partial on v erse to the latter statemen t: Prop osition 6.5. L et X b e a pr op er CA T(0) sp a e and G b e a total ly dis onne te d lo  al ly  omp at gr oup ating  ontinuously pr op erly on X by isometries. Assume that the p ointwise stabiliser of every b ounde d subset of X is op en in G . Then we have the fol lowing: (i) X admits a  anoni al lo  al ly nite G -e quivariant p artition. (ii) Denoting by σ ( x ) the pie  e supp orting the p oint x ∈ X in that p artition, we have Stab G ( σ ( x )) = N G ( G x ) and N G ( G x ) /G x ats fr e ely on σ ( x ) . (iii) If G \ X is  omp at, then so is Stab G ( σ ( x )) \ σ ( x ) for al l x ∈ X . Pr o of. Consider the equiv alene relation on X dened b y x ∼ y ⇔ G x = G y . This yields a anonial G -in v arian t partition of X . W e need to sho w that it is lo ally nite. Assume for a on tradition that there exists a on v erging sequene { x n } n ≥ 0 su h that the subgroups G x n are pairwise distint. Let x = lim n x n . W e laim that G x n < G x for all suien tly large n . Indeed, up on extrating there w ould otherwise exist a sequene g n ∈ G x n su h that g n .x 6 = x for all n . Up on a further extration, w e ma y assume that g n on v erges to some g ∈ G . By onstrution g xes x . Sine G x is op en b y h yp othesis, this implies that g n xes x for suien tly large n , a on tradition. This pro v es the laim. By h yp othesis the p oin t wise stabiliser of an y ball en tred at x is op en. Th us G x p ossesses a ompat op en subgroup U whi h xes ev ery x n . This implies that w e ha v e the inlusion U < G x n < G x for all n . Sine the index of U in G x is nite, there are only nitely man y subgroups of G x on taining U . This nal on tradition nishes the pro of of (i). (ii) Straigh tforw ard in view of the denitions. (iii) Supp ose for a on tradition that H \ σ ( x ) is not ompat, where H = Stab G ( σ ( x )) . Let then y n ∈ σ ( x ) b e a sequene su h that d ( y n , H .x ) > n . Let no w g n ∈ G b e su h that { g n .y n } is b ounded, sa y of diameter C . By (i), the set { g n G y n g − 1 n } is th us nite. Up on extrating, w e shall assume that it is onstan t. No w, for all n < k , the elemen t g − 1 n g k 22 PIERRE-EMMANUEL CAPRA CE AND NICOLAS MONOD normalises G y k = G x and maps y k to a p oin t at distane ≤ C from y n . In view of (ii), this is absurd.  Remark 6.6. The partition of X onstruted ab o v e is non-trivial whenev er G do es not at freely . This is for example the ase whenev er G is non-disrete and ats faithfully . The piees in the ab o v e partition are generally neither b ounded (ev en if G \ X is ompat), nor on v ex, nor ev en onneted. Ho w ev er, if one assumes that the spae admits a suien tly large amoun t of symmetry , then one obtains a partition whi h deserv es to b e view ed as an equiv arian t ellular deomp osition. Corollary 6.7. L et X b e a pr op er CA T(0) sp a e and G b e a total ly dis onne te d lo  al ly  omp at gr oup ating  ontinuously pr op erly on X by isometries. Assume that the p ointwise stabiliser of every b ounde d subset of X is op en in G , and that no op en sub gr oup of G xes a p oint at innity. Then X admits admits a  anoni al lo  al ly nite G -e quivariant de  omp osition into  omp at  onvex pie  es. Pr o of. F or ea h x ∈ X , let τ ( x ) b e the xed-p oin t-set of G x . Then τ ( x ) is learly on v ex; it is ompat b y h yp othesis. F urthermore the map x 7→ τ ( x ) is G -equiv arian t. The fat that the olletion { τ ( x ) | x ∈ X } is lo ally nite follo ws from Prop osition 6.5 .  6.C. Alexandro v angle rigidit y. A further onsequene of Theorem 6.1 is a phenomenon of angle rigidit y . Giv en an ellipti isometry g of omplete a CA T(0) spae X and a p oin t x ∈ X , w e denote b y c g ,x the pro jetion of x on the losed on v ex set of g -xed p oin ts. Prop osition 6.8. L et G b e a total ly dis onne te d lo  al ly  omp at gr oup with a  ontinuous and pr op er  o  omp at ation by isometries on a ge o desi al ly  omplete pr op er CA T(0) sp a e X . Then ther e is ε > 0 suh that for any el lipti g ∈ G and any x ∈ X with g x 6 = x we have ∠ c g,x ( g x, x ) ≥ ε . (W e will later also pro v e an angle rigidit y for the Tits angle, see Prop osition 7.15 .) Pr o of. First w e observ e that this b ound on the Alexandro v angle is really a lo al prop ert y at c g ,x of the germ of the geo desi [ c g ,x , x ] sine for an y y ∈ [ c g ,x , x ] w e ha v e c g ,y = c g ,x . Next, w e laim that for an y n ∈ N , an y isometry of order ≤ n of an y omplete CA T(0) spae B satises ∠ c g,x ( g x, x ) ≥ 1 /n for all x ∈ B that are not g -xed. Indeed, it follo ws from the denition of Alexandro v angles (see [ BH99 , I I.3.1℄) that for an y y ∈ [ c g ,x , x ] w e ha v e d ( g y, y ) ≤ d ( c g ,x , y ) ∠ c g,x ( g x, x ) . Therefore, if ∠ c g,x ( g x, x ) < 1 /n , the en tire g -orbit of y w ould b e on tained in a ball around y not on taining c g ,x = c g ,y . This is absurd sine the irumen tre of this orbit is a g -xed p oin t. In order to pro v e the prop osition, w e no w supp ose for a on tradition that there are sequenes { g n } of ellipti elemen ts in G and { x n } in X with g n x n 6 = x n and ∠ c n ( g n x n , x n ) → 0 , where c n = c g n ,x n . Sine the G -ation is o ompat, there is (up on extrating) a sequene { h n } in G su h that h n c n on v erges to some c ∈ X . Up on onjugating g n b y h n , replaing x n b y h n x n and c n b y h n c n , w e an assume c n → c without lo osing an y of the onditions on g x , x n and c n , inluding the relation c n = c g n ,x n . Sine d ( g n c, c ) ≤ 2 d ( c n , c ) , w e an further extrat and assume that { g n } on v erges to some limit g ∈ G ; notie also that g xes c . By Lemma 3.13 , the ation is minimal and hene Theorem 6.1 applies. Therefore, w e an assume that all g n oinide with g on some ball B around c and in partiular preserv e B . Using Remark 6.2 , this pro vides a on tradition.  ISOMETR Y GR OUPS OF NON-POSITIVEL Y CUR VED SP A CES: STR UCTURE THEOR Y 23 A rst onsequene is an analogue of a result that E. Sw enson pro v ed for disrete groups (Theorem 11 in [ Sw e99 ℄). Corollary 6.9. L et G b e a total ly dis onne te d lo  al ly  omp at gr oup with a  ontinuous and pr op er  o  omp at ation by isometries on a ge o desi al ly  omplete pr op er CA T(0) sp a e X not r e du e d to a p oint. Then G  ontains hyp erb oli elements (thus in p artiular elements of innite or der). Bey ond the totally disonneted ase, w e an app eal to Theorem 1.6 and A ddendum 1.8 and state the follo wing. Corollary 6.10. L et G b e any lo  al ly  omp at gr oup with a  ontinuous and pr op er  o  omp at ation by isometries on a ge o desi al ly  omplete pr op er CA T(0) sp a e X not r e du e d to a p oint. Then G  ontains elements of innite or der; if mor e over ( ∂ X ) G = ∅ , then G  ontains hyp erb oli elements. Pr o of of Cor ol lary 6.9 . Prop osition 6.8 allo ws us use the argumen t form [ Sw e99 ℄: W e an  ho ose a geo desi ra y r : R + → X , an inreasing sequene { t i } going to innit y in R + and { g i } in G su h that the funtion t 7→ g i r ( t + t i ) on v erges uniformly on b ounded in terv als (to a geo desi line). F or i < j large enough, the angle ∠ h ( r ( t i )) ( r ( t i ) , h 2 ( r ( t i ))) dened with h = g − 1 i g j is arbitrarily lose to π . In order to pro v e that h is h yp erb oli, it sues to sho w that this angle will ev en tually equal π . Supp ose this do es not happ en; b y Corollary 6.3 ( iii ), w e an assume that h is ellipti. W e set x = r ( t i ) and c = c h,x . Considering the ongruen t triangles ( c, x, hx ) and ( c, hx, h 2 x ) , w e nd that ∠ c ( x, hx ) is arbitrarily small. This is in on tradition with Prop osition 6.8 .  Pr o of of Cor ol lary 6.10 . If the onneted omp onen t G ◦ is non-trivial, then it on tains ele- men ts of innite order; if it is trivial, w e an apply Corollary 6.9 . Assume no w ( ∂ X ) G = ∅ . Then Theorem 1.6 and A ddendum 1.8 apply . Therefore, w e obtain h yp erb oli elemen ts either from Corollary 6.9 or from the fat that an y non-ompat semi-simple group on tains elemen ts that are algebraially h yp erb oli, om bined with the fat that the latter at as h yp erb oli isometries. That fat is established in Theorem 7.4 ( i ) b elo w, the pro of of whi h is indep enden t of Corollary 6.10 .  6.D. Algebrai struture. Giv en a top ologial group G , w e dene its so le so c( G ) as the subgroup generated b y all minimal non-trivial losed normal subgroups of G . Notie that G migh t ha v e no minimal non-trivial losed normal subgroup, in whi h ase its so le is trivial. W e also reall that the quasi-en tre of a lo ally ompat group G is the subset Q Z ( G ) onsisting of all those elemen ts p ossessing an op en en traliser. Clearly Q Z ( G ) is a (top o- logially)  harateristi subgroup of G . Sine an y elemen t with a disrete onjugay lass p ossesses an op en en traliser, it follo ws that the quasi-en tre on tains all disrete normal subgroups of G . Prop osition 6.11. L et X b e a pr op er CA T(0) sp a e without Eulide an fator and G < Is( X ) b e a lose d sub gr oup ating minimal ly  o  omp atly without xe d p oint at innity. If G has trivial quasi- entr e, then so c( G ∗ ) is dir e t pr o dut of r non-trivial har ateristi al ly simple gr oups, wher e r is the numb er of irr e duible fators of X and G ∗ is the  anoni al nite index op en normal sub gr oup ating trivial ly on the set of fators of X . The pro of will use the follo wing general fat inspired b y a statemen t for tree automor- phisms, Lemma 1.4.1 in [ BM00 ℄. 24 PIERRE-EMMANUEL CAPRA CE AND NICOLAS MONOD Prop osition 6.12. L et G b e a  omp atly gener ate d total ly dis onne te d lo  al ly  omp at gr oup without non-trivial  omp at normal sub gr oups. Then any ltering family of non- disr ete lose d normal sub gr oups has non-trivial (thus non- omp at) interse tion. A v arian t of this prop osition is pro v ed in [ CM08a ℄; sine the pro of is short, w e giv e it for the sak e of ompleteness. Pr o of of Pr op osition 6.12 . Let g b e a S hreier graph for G . W e reall that it onsists in  ho osing an y op en ompat subgroup U < G (whi h exists b y [ Bou71 , I I I  4 No 6℄), dening the v ertex set of g as G/U and dra wing edges aording to a ompat generating set whi h is a union of double osets mo dulo U ; see [ Mon01 ,  11.3℄. Sine G has no non-trivial ompat normal subgroup, the on tin uous G -ation on g is faithful. Let v 0 b e a v ertex of g and denote b y v ⊥ 0 the set of neigh b ouring v erties. Sine G is v ertex-transitiv e on g , it follo ws that for an y normal subgroup N ✁ G , the N v 0 -ation on v ⊥ 0 denes a nite p erm utation group F N < Sym( v ⊥ 0 ) whi h, as an abstrat p erm utation group, is indep enden t of the  hoie of v 0 . Therefore, if N is non-disrete, this p erm utation group F N has to b e non-trivial sine U is op en and g onneted. No w a ltering family F of non-disrete normal subgroups yields a ltering family of non-trivial nite subgroups of Sym( v ⊥ 0 ) . Th us the in tersetion of these nite groups is non-trivial. Let g b e a non-trivial elemen t in this in tersetion. F or an y N ∈ F , let N g b e the in v erse image of { g } in N v 0 . Th us N g is a non-empt y ompat subset of N for ea h N ∈ F . Sine the family F is ltering, so are { N v 0 | N ∈ F } and { N g | N ∈ F } . The result follo ws, sine a ltering family of non-empt y losed subsets of the ompat set G v 0 has a non-empt y in tersetion.  Eviden tly op en normal subgroups form a ltering family; w e an th us dedue: Corollary 6.13. L et G b e a  omp atly gener ate d lo  al ly  omp at gr oup without any non- trivial  omp at normal sub gr oup. If G is r esidual ly disr ete, then it is disr ete.  Pr o of of Pr op osition 6.11 . W e rst observ e that G ∗ has no non-trivial disrete normal sub- group. Indeed, su h a subgroup has nitely man y G -onjugates, whi h implies that ea h of its elemen ts has disrete G -onjugay lass and hene b elongs to Q Z ( G ) , whi h w as assumed trivial. Let no w { N i } b e a  hain of non-trivial losed normal subgroups of G ∗ . If N i is totally disonneted for some i , then the in tersetion T i N i is non-trivial b y Prop osition 6.12 . Otherwise N ◦ i is non-trivial and normal in ( G ∗ ) ◦ for ea h i , and the in tersetion T i N i is non-trivial b y Theorem 1.6 (sine the latter desrib es in partiular the p ossible normal onneted subgroups of G ∗ ). In all ases, Zorn's lemma implies that the ordered set of non-trivial losed normal subgroups of G ∗ p ossesses minimal elemen ts. Giv en t w o minimal losed normal subgroups M , M ′ , the in tersetion M ∩ M ′ is th us trivial and, hene, so is [ M , M ′ ] . Th us minimal losed normal subgroups of G ∗ en tralise one another. W e dedue from Corollary 5.8 that the n um b er of minimal losed normal subgroups is at most r . Consider no w an irreduible totally disonneted fator H of G ∗ . W e laim that the olletion of non-trivial losed normal subgroups of H forms a ltering family . Indeed, giv en t w o su h normal subgroup N 1 , N 2 , then N 1 ∩ N 2 is again a losed normal subgroup of H . It is is trivial, then the omm utator [ N 1 , N 2 ] is trivial and, hene, the en traliser of N 1 in H is non-trivial, on traditing Theorem 1.10 . This onrms the laim. Th us the in tersetion of all non-trivial losed normal subgroups of H is non-trivial b y Prop osition 6.12 . Clearly this in tersetion is the so le of H ; it is lear w e ha v e just established that it is on tained in ev ery non-trivial losed normal subgroup of H . In partiular so c( H ) is  harateristially ISOMETR Y GR OUPS OF NON-POSITIVEL Y CUR VED SP A CES: STR UCTURE THEOR Y 25 simple. The desired result follo ws, sine so c( H ) is learly a minimal losed normal subgroup of G ∗ .  Theorem 6.14. L et X b e a pr op er irr e duible ge o desi al ly  omplete CA T(0) sp a e. L et G < Is( X ) b e a lose d total ly dis onne te d sub gr oup ating  o  omp atly, in suh a way that no op en sub gr oup xes a p oint at innity. Then we have the fol lowing: (i) Every  omp at sub gr oup of G is  ontaine d in a maximal one; the maximal  omp at sub gr oups fal l into nitely many  onjugay lasses. (ii) Q Z ( G ) = 1 . (iii) so c( G ) is a non-disr ete har ateristi al ly simple gr oup. Pr o of. (i) By Lemma 3.13 , the ation is minimal and hene Theorem 6.1 applies. In par- tiular, w e an apply Corollary 6.7 and onsider the resulting equiv arian t deomp osition. Let Q < G b e a ompat subgroup and x b e a Q -xed p oin t. If G x is not on tained in a maximal ompat subgroup of G , then there is an innite sequene ( x n ) n ≥ 0 su h that x 0 = x and G x n ⊆ G x n +1 . By Corollary 6.7 , the sequene x n lea v es ev ery b ounded subset. Sine the xed p oin ts X G x n form a nested sequene, it follo ws that X G x is un b ounded. In partiular its visual b oundary ∂ ( X G x ) is non-empt y and the op en subgroup G x has a xed p oin t at innit y . This on tradits the h yp otheses, and the laim is pro v ed. Notie that a similar argumen t sho ws that for ea h x ∈ X , there are nitely man y maximal ompat subgroups Q i < G on taining G x . The fat that G p ossesses nitely man y onjugay lasses of maximal ompat subgroups no w follo ws from the ompatness of G \ X . (ii) W e laim that Q Z ( G ) is top ologially lo ally nite , whi h means that ev ery nite subset of it is on tained in a ompat subgroup. The desired result follo ws sine it is then amenable but G has trivial amenable radial b y Theorem 1.10 . Let S ⊆ Q Z ( G ) b e a nite subset. Then G p ossesses a ompat op en subgroup U en tralising S . By h yp othesis the xed p oin t set of U is ompat. Sine h S i stabilises X U , it follo ws that h S i is ompat, whene the laim. (iii) F ollo ws from (ii) and Prop osition 6.11 .  7. Coomp a t CA T(0) sp a es 7.A. Fixed p oin ts at innit y. W e b egin with a simple observ ation. W e reall that t w o p oin ts at innit y are opp osite if they are the t w o endp oin ts of a geo desi line. W e denote b y ξ op the set of p oin ts opp osite to ξ . Reall from [ Bal95 , Theorem 4.11(i)℄ that, if X is prop er, then t w o p oin ts ξ , η ∈ ∂ X at Tits distane > π are neessarily opp osite. (Reall that Tits distane is b y denition the length metri asso iated to the Tits angle.) Ho w ev er, it is not true in general that t w o p oin ts at Tits distane π are opp osite. Prop osition 7.1. L et X b e a pr op er CA T(0) sp a e and H < Is( X ) a lose d sub gr oup ating  o  omp atly. If H xes a p oint ξ at innity, then ξ op 6 = ∅ and H ats tr ansitively on ξ op . Pr o of. First w e laim that there is a geo desi line σ : R → X with σ ( ∞ ) = ξ . Indeed, let r : R + → X ′ b e a ra y p oin ting to ξ and { g n } a sequene in H su h that g n r ( n ) remains b ounded. The ArzelàAsoli theorem implies that g n r ( R + ) sub on v erges to a geo desi line in X . Sine ξ is xed b y all g n , this line has an endp oin t at ξ . Let no w σ ′ : R → X b e an y other geo desi with σ ′ ( ∞ ) = ξ and  ho ose a sequene { h n } n ∈ N in H su h that d ( h n σ ( − n ) , σ ′ ( − n )) remains b ounded. By on v exit y and sine all h n x ξ , d ( h n σ ( t ) , σ ′ ( t )) is b ounded for all t and th us sub on v erges (uniformly for t in 26 PIERRE-EMMANUEL CAPRA CE AND NICOLAS MONOD b ounded in terv als). On the one hand, it implies that { h n } has an aum ulation p oin t h . On the other hand, it follo ws that hσ ( −∞ ) = σ ′ ( −∞ ) .  Reall that an y omplete CA T(0) spae X admits a anonial splitting X = X ′ × V preserv ed b y all isometries, where V is a (maximal) Hilb ert spae alled the Eulidean fator of X , see [ BH99 , I I.6.15(6)℄. F urthermore, there is a anonial em b edding X ′ ⊆ X ′′ × V ′ , where V ′ is a Hilb ert spae generated b y all diretions in X ′ p oin ting to at p oin ts at innit y , namely p oin ts for whi h the Busemann funtions are ane on X ′ ; moreo v er, ev ery isometry of X ′ extends uniquely to an isometry of X ′′ × V ′ whi h preserv es that splitting. This is a result of A damsBallmann [ AB98a , Theorem 1.6℄, who all V ′ the pseudo-Eulidean fator (one ould also prop ose Eulidean pseudo-fator). Corollary 7.2. L et X b e a pr op er CA T(0) sp a e with a  o  omp at gr oup of isometries. Then the pseudo-Eulide an fator of X is trivial. Pr o of. In view of the ab o v e disussion, X ′ is also a prop er CA T(0) spae with a o ompat group of isometries. The set of at p oin ts in ∂ X ′ admits a anonial (in trinsi) irumen tre ξ b y Lemma 1.7 in [ AB98a ℄. In partiular, ξ is xed b y all isometries and therefore, b y Prop osition 7.1 , it has an opp osite p oin t, whi h is imp ossible for a at p oin t unless it lies already in the Eulidean fator (see [ AB98a ℄).  Prop osition 7.3. L et G b e a gr oup ating  o  omp atly by isometries on a pr op er CA T(0) sp a e X without Eulide an fator and assume that the stabiliser of every p oint at innity ats minimal ly on X . Then G has no xe d p oint at innity. Pr o of. If G has a global xed p oin t ξ , then the stabiliser G η of an opp osite p oin t η ∈ ξ op (whi h exists b y Prop osition 7.1 ) preserv es the union Y ⊆ X of all geo desi lines onneting ξ to η . By [ BH99 , I I.2.14℄, this spae is on v ex and splits as Y = R × Y 0 . Sine G η ats minimally , w e dedue Y = X whi h pro vides a Eulidean fator.  7.B. A tions of simple algebrai groups. Let X b e a CA T(0) spae and G b e an alge- brai group dened o v er the eld k . An isometri ation of G ( k ) on X is alled algebrai if ev ery (algebraially) semi-simple elemen t g ∈ G ( k ) ats as a semi-simple isometry . When G is semi-simple, w e denote b y X model the Riemannian symmetri spae or Bruhat Tits building asso iated with G ( k ) . Theorem 7.4. L et k b e a lo  al eld and G b e an absolutely almost simple simply  onne te d k -gr oup. L et X b e a non- omp at pr op er CA T(0) sp a e on whih G = G ( k ) ats  ontinuously by isometries. Assume either: (a) the ation is  o  omp at; or: (b) it has ful l limit set, is minimal and ∂ X is nite-dimensional. Then: (i) The G -ation is algebr ai. (ii) Ther e is a G -e quivariant bije tion ∂ X ∼ = ∂ X model whih is an isometry with r e- sp e t to Tits' metri and a home omorphism with r esp e t to the  ne top olo gy. This bije tion extends to a G -e quivariant r ough isometry β : X model → X . (iii) If X is ge o desi al ly  omplete, then X is isometri to X model . (iv) F or any semi-simple k -sub gr oup L < G , ther e non-empty lose d  onvex subsp a e Y ⊆ X minimal for L = L ( k ) ; mor e over, ther e is no L -xe d p oint in ∂ Y . In the ab o v e p oin t ( ii ), a rough isometry refers to a map β : X model → X su h that there is a onstan t C with d X mod el ( x, y ) − C ≤ d X ( β ( x ) , β ( y )) ≤ d X mod el ( x, y ) + C ISOMETR Y GR OUPS OF NON-POSITIVEL Y CUR VED SP A CES: STR UCTURE THEOR Y 27 for all x, y ∈ X model and su h that β ( X model ) has nite o diameter in X . Su h a map is also alled a (1 , C ) -quasi-isometry . Remarks 7.5. (i) Notie that there is no assumption on the k -rank of G in this result. (ii) W e reall for (b) that minimalit y follo ws from full limit set in the geo desially omplete ase (Lemma 3.13 ). (iii) In the on text of ( ii ), w e reall that in general t w o CA T(0) spaes with the same o ompat isometry group need not ha v e homeomorphi b oundaries [ CK00 ℄. (iv) A p osteriori , p oin t ( ii ) sho ws in partiular that the ation is also o ompat under the assumption (b). Before pro eeding to the pro of, w e giv e t w o examples sho wing that the assumptions made in Theorem 7.4 are neessary . Example 7.6 . Without the assumption of geo desi ompleteness, it is not true in general that, in the setting of the theorem, the spae X on tains a losed on v ex G -in v arian t subspae whi h is isometri to X model . A simple example of this situation ma y obtained as follo ws. Consider the ase where k is non-Ar himedean and G has k -rank one. Let 0 < r < 1 / 2 and let X b e the spae obtained b y replaing the r -ball en tred at ea h v ertex in the tree X model b y an isometri op y of a giv en Eulidean n -simplex, where n + 1 is v alene of the v ertex. In this w a y , one obtains a CA T(0) spae whi h is still endo w ed with an isometri G -ation that is o ompat and minimal, but learly X is not isometri to X model . W e do not kno w whether su h a onstrution ma y also b e p erformed in the Ar himedean ase (see Problem 7.2 in [ CM08b ℄). Example 7.7 . Under the assumptions (b), minimalit y is needed. Indeed, w e laim that for an y CA T(0) spae X 0 there is a anonial CA T( − 1 ) spae X (in partiular X is a CA T(0) spae) together with a anonial map i : Is( X 0 ) ֒ → Is( X ) with the follo wing prop erties: The b oundary ∂ X is redued to a single p oin t; X non-ompat; X is prop er if and only if X 0 is so; the map i is an isomorphism of top ologial groups on to its image. This laim justies that minimalit y is needed sine w e an apply it to the ase where X 0 is the symmetri spae or BruhatTits building asso iated to G ( k ) . (In that ase the ation has indeed full limit set, a  heap feat as the isometry group is non-ompat and the b oundary rather inapaious.) T o pro v e the laim, onsider the parab oli ne Y asso iated to X 0 . This is the metri spae with underlying set X 0 × R ∗ + where the distane is dened as follo ws: giv en t w o p oin ts ( x, t ) and ( x ′ , t ′ ) of Y , iden tify the in terv al [ x, x ′ ] ⊆ X 0 with an in terv al of orresp onding length in R and measure the length from the resulting p oin ts ( x, t ) and ( x ′ , t ′ ) in the upp er half-plane mo del for the h yp erb oli plane. This is a partiular ase of the syn theti v ersion ([ Che99 ℄, [ AB98b ℄) of the BishopO'Neill w arp ed pro duts [ BO69 ℄ and its prop erties are desrib ed in [ BGP92 ℄, [ AB04 , 1.2(2A)℄ and [ HLS00 ,  2℄. In partiular, Y is CA T( − 1 ). W e no w let ξ ∈ ∂ Y b e the p oin t at innit y orresp onding to t → ∞ and dene X ⊆ Y to b e an asso iated horoball; for deniteness, set X = X 0 × [1 , ∞ ) . W e no w ha v e ∂ X = { ξ } b y the CA T( − 1 ) prop ert y or alternativ ely b y the expliit desription of geo desi ra ys ( e.g. 2(iv) in [ HLS00 ℄). The remaining prop erties follo w readily . Pr o of of The or em 7.4 . W e start with a few preliminary observ ations. Finite-dimensionalit y of the b oundary alw a ys holds sine it is automati in the o ompat ase. Sine X is non- ompat, the ation is non-trivial, b eause it has full limit set. It is w ell kno wn that ev ery non-trivial on tin uous homomorphism of G to a lo ally ompat seond oun table group is prop er [ BM96 , Lemma 5.3℄. Th us the G -ation on X is prop er. 28 PIERRE-EMMANUEL CAPRA CE AND NICOLAS MONOD W e laim that the stabiliser of an y p oin t ξ ∈ ∂ X on tains the unip oten t radial of some prop er parab oli subgroup of G . Indeed, x a p olar deomp osition G = K T K . Let x 0 ∈ X b e a K -xed p oin t. Cho ose a sequene { g n } n ≥ 0 of elemen ts of G su h that g n .x 0 on v erges to ξ . W rite g n = k n .a n .k ′ n with k n , k ′ n ∈ K and a n ∈ T . W e ma y furthermore assume, up on replaing { g n } b y a subsequene, that { k n } on v erges to some k ∈ K , that { a n .x 0 } on v erges in X ∪ ∂ X and that { a n .p } on v erges in X model ∪ ∂ X model , where p ∈ X model is some base p oin t. Let η = lim n →∞ a n .x 0 and observ e η = k − 1 ξ . F urthermore, the stabiliser of η on tains the group U = { g ∈ G | li m n →∞ a − 1 n g a n = 1 } . The on v ergene in diretion of { a n } in T implies that U on tains the unip oten t radial U Q of the parab oli subgroup Q < G orresp onding to lim n →∞ a n .p ∈ ∂ X model . (In fat, the argumen ts for Lemma 2.4 in [ Pra77 ℄ probably sho w U = U Q ; this follo ws a p osteriori from (ii) b elo w.) Therefore, the stabiliser of ξ = k .η in G on tains the unip oten t radial of k Qk − 1 , pro ving the laim. Notie that w e ha v e seen in passing that an y p oin t at innit y lies in the limit set of some torus; in the ab o v e notation, ξ is in the limit set of k T k − 1 . ( i ) Ev ery elemen t of G whi h is algebraially ellipti ats with a xed p oin t in X , sine it generates a relativ ely ompat subgroup. W e need to sho w that ev ery non-trivial elemen t of a maximal split torus T < G ats as a semi-simple isometry . Assume for a on tradition that some elemen t t ∈ T ats as a parab oli isometry . Sine X has nite-dimensional b oundary and w e an apply Corollary 3.3 ( ii ) . It follo ws that the Ab elian group T has a anonial xed p oin t at innit y ξ xed b y the normaliser N G ( T ) . By the preeding paragraph, w e kno w furthermore that the stabiliser of ξ in G also on tains the unip oten t radial of some parab oli subgroup of G . Reall that G is generated b y N G ( T ) together with an y su h unip oten t radial: this follo ws from the fat that N G ( T ) has no xed p oin t at innit y in X model and that G is generated b y the unip oten t radials of an y t w o distint parab oli subgroups. Therefore ξ is xed b y the en tire group G . Sine G has trivial Ab elianisation, its image under the Busemann  harater en tred at ξ v anishes, thereb y sho wing that G m ust stabilise ev ery horoball en tred at ξ . This is absurd b oth in the minimal and the o ompat ase. ( ii ) Let T < G b e a maximal split torus. Let F model ⊆ X model b e the (maximal) at stabilised b y T . In view of ( i ) and the prop erness of the T -ation, w e kno w that T also stabilises a at F ⊆ X with dim F = dim T , see [ BH99 , I I.7.1℄. Cho ose a base p oin t p 0 ∈ F model in su h a w a y that its stabiliser K := G p 0 is a maximal ompat subgroup of G . The union of all T -in v arian t ats whi h are parallel to F is N G ( T ) -in v arian t. Therefore, up on replaing F b y a parallel at, w e ma y  and shall  assume that F on tains a p oin t x 0 whi h is stabilised b y N K := N G ( T ) ∩ K . Note that, sine N G ( T ) = h N K ∪ T i , the at F is N G ( T ) -in v arian t. Therefore, there is a w ell dened N G ( T ) -equiv arian t map α of the N G ( T ) -orbit of p 0 to F , dened b y α ( g .p 0 ) = g .x 0 for all g ∈ N G ( T ) . W e laim that, up to a saling fator, the map α is isometri and indues an N G ( T ) - equiv arian t isometry α : F model → F . In order to establish this, remark that the W eyl group W := N G ( T ) / Z G ( T ) ats on F , sine W = N K /T K , where T K := Z G ( T ) ∩ K ats trivially on F . The group N K normalises the oro ot lattie Λ < T . F urthermore N K . Λ ats on F model as an ane W eyl group sine N K . Λ /T K ∼ = W ⋉ Λ . Moreo v er, sine an y reetion in W en tralises an Ab elian subgroup of orank 1 in Λ , it follo ws that N K . Λ ats on F as a disrete reetion group. But a giv en ane W eyl group has a unique (up to saling fator) disrete o ompat ation as a reetion group on Eulidean spaes, as follo ws from [ Bou68 , ISOMETR Y GR OUPS OF NON-POSITIVEL Y CUR VED SP A CES: STR UCTURE THEOR Y 29 Ch. VI,  2, Prop osition 8℄. Therefore the restrition of α to Λ .x 0 is a homothet y . Sine Λ is a uniform lattie in T , the laim follo ws. A t this p oin t, it follo ws that α indues an N G ( T ) -equiv arian t map ∂ α : ∂ F model → ∂ F , whi h is isometri with resp et to Tits' distane. W e reall that N G ( T ) is the stabiliser of ∂ F model in G . Moro v er, for an y η ∈ ∂ F model , the stabilisers in G of η is on tained in that of α ( η ) b eause of the geometri desription of parab oli subgroups alluded to in the preliminary observ ation: see the argumen t for Lemma 2.4 in [ Pra77 ℄. Therefore, ∂ α extends to a w ell dened G -equiv arian t map ∂ X model → ∂ X , whi h w e denote again b y ∂ α . Sine an y t w o p oin ts of ∂ X model are on tained in a ommon maximal sphere (i.e. an apartmen t), and sine G ats transitiv ely on these spheres, the map ∂ α is isometri, b eause so is its restrition to the sphere ∂ F model . Note that ∂ α is surjetiv e: indeed, this follo ws from the last preliminary observ ation, whi h, om bined with ( i ), sho ws in partiular that ∂ X = K.∂ F . W e no w sho w that ∂ α is a homeomorphism with resp et to the ne top ology . Sine ∂ X model is ompat, it is enough to sho w that ∂ α is on tin uous. No w an y on v ergen t sequene in ∂ X model ma y b e written as { k n .ξ n } n ≥ 0 , where { k n } n ≥ 0 (resp. { ξ n } n ≥ 0 ) is a on v ergen t sequene of elemen ts of K (resp. ∂ F model ). On the sphere ∂ F model , the ne top ology oinides with the one indued b y Tits' metri. Therefore, the equiv ariane of the Tits' isometry ∂ α sho ws that { ∂ α ( k n .ξ n ) } n ≥ 0 is a on v ergen t sequene in ∂ X , as w as to b e pro v ed. W e next laim that that G -ation on X is o ompat ev en under the assumption (b). T o w ards a on tradition, assume otherwise. Cho ose a sequene { y n } in X with y 0 a K - xed p oin t and su h that d ( y n , g .y 0 ) ≥ n for all g ∈ G . Up on replaing y n ( n ≥ 1 ) b y an appropriate G -translate, w e an and shall assume that moreo v er (7.i) d ( y n , y 0 ) ≤ d ( y n , g .y 0 ) + c ∀ g ∈ G, n ≥ 1 , where c is some onstan t. Up on extrating a subsequene, the sequene { y n } on v erges to some p oin t η ∈ ∂ X . It w as established ab o v e that ∂ X = K.∂ F ; in partiular there exists k ∈ K su h that k .η ∈ ∂ F . No w, up on replaing y n b y k .y n , w e obtain a sequene { y n } whi h still satises all ab o v e onditions but whi h on v erges to a b oundary p oin t η ′ of the at F . Let r : R + → F b e a geo desi ra y p oin ting to w ards η ′ . Sine N G ( T ) ats o ompatly on the at F , it follo ws from ( 7.i ) that for some onstan t c ′ , w e ha v e (7.ii) d ( y n , y 0 ) ≤ d ( y n , r ( t )) + c ′ ∀ t ≥ 0 , n ≥ 1 . Fix no w s > d ( y 0 , r (0)) + c ′ . F or n suien tly large, let z n b e the p oin t on [ r (0) , y n ] at distane s of r (0) . W e ha v e d ( y n , r ( s )) ≤ d ( y n , z n ) + d ( z n , r ( s )) = d ( y n , r (0)) − s + d ( z n , r ( s )) < d ( y n , r (0)) − d ( y 0 , r (0)) − c ′ + d ( z n , r ( s )) ≤ d ( y n , y 0 ) − c ′ + d ( z n , r ( s )) . As n go es to innit y , this pro vides a on tradition to ( 7.ii ) sine z n on v erges to r ( s ) ; th us o ompatness is established. It remains for ( ii ) to pro v e that ∂ α extends to a G -equiv arian t rough isometry β : X model → X . The orbital map g 7→ g .y 0 asso iated to y 0 yields a map β : G/K → X ; when k is Ar himedean, X model = G/K whereas w e extend β linearly to ea h  ham b er of the building X model in the non-Ar himedean ase. It is a w ell-kno wn onsequene of o om- patness that the G -equiv arian t map β : X model → X is a quasi-isometry (see e.g. the pro of 30 PIERRE-EMMANUEL CAPRA CE AND NICOLAS MONOD of the Milnorv ar lemma giv en in [ BH99 , I.8.19℄). F or our stronger statemen t, it sues, in view of the K T K deomp osition and of equiv ariane, to pro v e that there is a onstan t C ′ su h that d X mod el ( a.p 0 , p 0 ) − C ′ ≤ d X ( a.y 0 , y 0 ) ≤ d X mod el ( a.p 0 , p 0 ) + C ′ for all a ∈ T . This follo ws from the fat that β and α are at b ounded distane from ea h other on F model (indeed, at distane d ( y 0 , x 0 ) ) and that β is isometri on F model . ( iii ) In the higher rank ase, assertion ( iii ) follo ws from ( ii ) and the main result of [ Lee00 ℄. Ho w ev er, the full strength of lo . it. is really not needed here, sine the main diult y there is preisely the absene of an y group ation, whi h is part of the h yp otheses in our setting. F or example, when the ground eld k is the eld of real n um b ers, the argumen ts ma y b e dramatially shortened as follo ws; they are v alid without an y rank assumption. Giv en an y ξ ∈ ∂ X , the unip oten t radial of the parab oli subgroup G ξ ats sharply transitiv ely on the b oundary p oin ts opp osite to ξ . In view of this and of the prop erness of the G -ation, the argumen ts of [ Lee00 , Prop osition 4.27℄ sho w that geo desi lines in X do not bran h; in other w ords X has uniquely extensible geo desis. F rom this, it follo ws that the group N K = N G ( T ) ∩ K onsidered in the pro of of ( ii ) has a unique xed p oin t in X , sine otherwise it w ould x p oin t wise a geo desi line, and hene, b y ( ii ), opp osite p oin ts in ∂ X model . The fat that this is imp ossible is purely a statemen t on the lassial symmetri spae X model ; w e giv e a pro of for the reader's on v eniene: Let F model b e the at orresp onding to T and p 0 ∈ F model b e the K -xed p oin t. If N K xed a p oin t ξ ∈ ∂ X model , then the ra y [ p 0 , ξ ) w ould b e p oin t wise xed and, hene, the group N K w ould x a non-zero v etor in the tangen t spae of X model at p 0 . A Cartan deomp osition g = k ⊕ p of the Lie algebra g of G yields an isomorphism b et w een the isotrop y represen tation of N K on T p 0 X model and the represen tation of the W eyl group W on p . An easy expliit omputation sho ws that the latter represen tation has no non-zero xed v etor. Sine N K has a unique xed p oin t, the latter is stabilised b y the en tire group K . Hene K xes a p oin t lying on a at F stabilised b y T . F rom the K T K -deomp osition, it follo ws that the G -orbit of this xed p oin t is on v ex. Sine the G -ation on X is minimal b y geo desi ompleteness (Lemma 3.13 ), w e dedue that G is transitiv e on X . In partiular X is o v ered b y ats whi h are G -onjugate to F , and the existene of a G -equiv arian t homothet y X model → X follo ws from the existene of a N G ( T ) -equiv arian t homothet y F model → F , whi h has b een established ab o v e. It remains only to  ho ose the righ t sale on X model to mak e it an isometry . In the non-Ar himedean ase, w e onsider only the rank one ase, referring to [ Lee00 ℄ for higher rank. Let K b e a maximal ompat subgroup of G and x 0 ∈ X b e a K -xed p oin t. By ( ii ), the group K ats transitiv ely on ∂ X . Sine X is geo desially omplete, it follo ws that the K -translates of an y ra y emanating from x 0 o v er X en tirely . On the other hand ev ery p oin t in X has an op en stabiliser b y Theorem 6.1 , an y p oin t in X has a nite K -orbit. This implies that the spae of diretions at ea h p oin t p ∈ X is nite. In other w ords X is 1 -dimensional. Sine X is CA T(0) and lo ally ompat, it follo ws that X is a lo ally nite metri tree. As w e ha v e just seen, the group K ats transitiv ely on the geo desi segmen ts of a giv en length emanating from x 0 . One dedues that G is transitiv e on the edges of X . In partiular all edges of X ha v e the same length, whi h w e an assume to b e as in X model , G has at most t w o orbits of v erties, and X is either regular or bi-regular. The v alene of an y v ertex p equals min g 6∈ N G ( G p )  G p : G p ∩ g G p g − 1  , ISOMETR Y GR OUPS OF NON-POSITIVEL Y CUR VED SP A CES: STR UCTURE THEOR Y 31 and oinides therefore with the v alene of X model . It nally follo ws that X and X model are isometri, as w as to b e pro v ed. ( iv ) Let P b e a k -parab oli subgroup of G that is minimal amongst those on taining L . W e ma y assume P 6 = G sine otherwise L has no xed p oin t in ∂ X and the onlusion holds in view of Prop osition 4.1 . It follo ws that L en tralises a k -split torus T of p ositiv e dimension d . It follo ws from ( i ) that T = T ( k ) ats b y h yp erb oli isometries, and th us there is a T -in v arian t losed on v ex subset Z ⊆ X of the form Z = Z 1 × R d su h that the T -ation is trivial on the Z 1 fator; this follo ws from Theorem I I.6.8 in [ BH99 ℄ and the prop erness of the ation. Moreo v er, L preserv es Z and its deomp osition Z = Z 1 × R d , ating b y translations on the R d fator ( lo . it. ). Sine L is semi-simple, this translation ation is trivial and th us L preserv es an y Z 1 bre, sa y for instane Z 0 := Z 1 × { 0 } ⊆ Z . F or b oth the existene of a minimal set Y and the ondition ( ∂ Y ) L = ∅ , it sues to sho w that L has no xed p oin t in ∂ Z 0 (Prop osition 4.1 ). W e laim that ∂ Z 0 is Tits-isometri to the spherial building of the Lévi subgroup Z G ( T ) . Indeed, w e kno w from ( ii ) that ∂ X is equiv arian tly isometri to ∂ X model and the building of Z G ( T ) is  haraterised as the p oin ts at distane π / 2 from the b oundary of the T -in v arian t at in ∂ X model . On the other hand, L has maximal semi-simple rank in Z G ( T ) b y the  hoie of P and therefore annot b e on tained in a prop er parab oli subgroup of Z G ( T ) . This sho ws that L has no xed p oin t in ∂ Z 0 and ompletes the pro of.  7.C. No bran hing geo desis. Reall that in a geo desi metri spae X , the spae of diretions Σ x at a p oin t x is the ompletion of the spae e Σ x of geo desi germs equipp ed with the Alexandro v angle metri at x . If X has uniquely extensible geo desis, then e Σ x = Σ x . The follo wing is a result of V. Beresto vskii [ Ber02 ℄ (w e read it in [ Ber ,  3℄; it also follo ws from A. Lyt hak's argumen ts in [ Lyt05 ,  4℄). Theorem 7.8. L et X b e a pr op er CA T(0) sp a e with uniquely extensible ge o desis and x ∈ X . Then e Σ x = Σ x is isometri to a Eulide an spher e.  W e use this result to establish the follo wing. Prop osition 7.9. L et X b e a pr op er CA T(0) sp a e with uniquely extensible ge o desis. Then any total ly dis onne te d lose d sub gr oup D < Is( X ) is disr ete. Pr o of. There is some ompat op en subgroup Q < D , see [ Bou71 , I I I  4 No 6℄. Let x b e a Q -xed p oin t. The isometry group of Σ x is a ompat Lie group b y Theorem 7.8 and th us the image of the pronite group Q in it is nite. Let th us K < Q b e the k ernel of this represen tation, whi h is op en. Denote b y S ( x, r ) the r -sphere around x . The Q - equiv arian t visual map S ( x, r ) → Σ x is a bijetion b y unique extensibilit y . It follo ws that K is trivial.  W e are no w ready for: End of pr o of of The or em 1.1 . Sine the ation is o ompat, it is minimal b y Lemma 3.13 . The fat that extensibilit y of geo desis is inherited b y diret fators of the spae follo ws from the  haraterisation of geo desis in pro duts, see [ BH99 , I.5.3(3)℄. Ea h fator X i is th us a symmetri spae in view of Theorem 7.4 ( iii ). By virtue of Corollary 6.3 ( iii ), the totally disonneted fators D j at b y semi-simple isometries. 32 PIERRE-EMMANUEL CAPRA CE AND NICOLAS MONOD Assume no w that X has uniquely extensible geo desis. F or the same reason as b efore, this prop ert y is inherited b y ea h diret fator of the spae. Th us ea h D j is disrete b y Prop osition 7.9 .  Theorem 7.10. L et X b e a pr op er irr e duible CA T(0) sp a e with uniquely extensible ge o desis. If X admits a non-disr ete gr oup of isometries with ful l limit set but no glob al xe d p oint at innity, then X is a symmetri sp a e. The ondition on xed p oin ts at innit y is neessary in view of E. Hein tze's examples [ Hei74 ℄ of negativ ely urv ed homogeneous manifolds whi h are not symmetri spaes. In fat these spaes onsist of ertain simply onneted soluble Lie groups endo w ed with a left-in v arian t negativ ely urv ed Riemannian metri. Prop osition 7.11. L et X b e a pr op er CA T(0) sp a e with uniquely extensible ge o desis. Then ∂ X has nite dimension. Pr o of. Let x ∈ X and reall that b y Beresto vskii's result quoted in Theorem 7.8 ab o v e, Σ x is isometri to a Eulidean sphere. By denition of the Tits angle, the visual map ∂ X → Σ x asso iating to a geo desi ra y its germ at x is Tits-on tin uous (in fat, 1 -Lips hitz). It is furthermore injetiv e (atually , bijetiv e) b y unique extensibilit y . Therefore, the top ologial dimension of an y  omp at subset of ∂ X is b ounded b y the dimension of the sphere Σ x . The laim follo ws no w from Kleiner's  haraterisation of the dimension of spaes with urv a- ture b ounded ab o v e in terms of the top ologial dimension of ompat subsets (Theorem A in [ Kle99 ℄).  Pr o of of The or em 7.10 . By Lemma 3.13 , the ation of G := Is( X ) is minimal. In view of Prop osition 7.11 , the b oundary ∂ X is nite-dimensional. Th us w e an apply Theorem 1.6 and A ddendum 1.8 . Sine X is irreduible and non-disrete, Prop osition 7.9 implies that G is an almost onneted simple Lie group (unless X = R , in whi h ase X is indeed a symmetri spae). W e onlude b y Theorem 7.4 .  7.D. No op en stabiliser at innit y . The follo wing statemen t sums up some of the pre- eding onsiderations: Corollary 7.12. L et X b e a pr op er ge o desi al ly  omplete CA T(0) sp a e without Eulide an fator suh that some lose d sub gr oup G < Is( X ) ats  o  omp atly. Supp ose that no op en sub gr oup of G xes a p oint at innity. Then we have the fol lowing: (i) X admits a  anoni al e quivariant splitting X ∼ = X 1 × · · · × X p × Y 1 × · · · × Y q wher e e ah X i is a symmetri sp a e and e ah Y j p ossesses a G -e quivariant lo  al ly nite de  omp osition into  omp at  onvex  el ls. (ii) G p ossesses hyp erb oli elements. (iii) Every  omp at sub gr oup of G is  ontaine d in a maximal one; the maximal  omp at sub gr oups fal l into nitely many  onjugay lasses. (iv) Q Z ( G ) = 1 ; in p artiular G has no non-trivial disr ete normal sub gr oup. (v) so c( G ∗ ) is a dir e t pr o dut of p + q non-disr ete har ateristi al ly simple gr oups. Pr o of. (i) F ollo ws from Theorem 1.1 and Corollary 6.7 . (ii) Clear from Corollary 6.10 . (iii) and (iv) Immediate from (i) and Theorem 6.14 (i) and (ii). (v) F ollo ws from (i), (iv) and Prop osition 6.11 .  ISOMETR Y GR OUPS OF NON-POSITIVEL Y CUR VED SP A CES: STR UCTURE THEOR Y 33 7.E. Co ompat stabilisers at innit y . W e undertak e the pro of of Theorem 1.3 whi h desrib es isometrially an y geo desially omplete prop er CA T(0) spae su h that the sta- biliser of ev ery p oin t at innit y ats o ompatly . Remark 7.13. (i) The form ulation of Theorem 1.3 allo ws for symmetri spaes of Eulidean t yp e. (ii) A BassSerre tree is a tree admitting an edge-transitiv e automorphism group; in partiular, it is regular or bi-regular (the regular ase b eing a sp eial ase of Eulidean buildings). Lemma 7.14. L et X b e a pr op er CA T(0) sp a e suh that the stabiliser of every p oint at innity ats  o  omp atly on X . F or any ξ ∈ ∂ X , the set of η ∈ ∂ X with ∠ T ( ξ , η ) = π is  ontaine d in a single orbit under Is( X ) . Pr o of. W rite G = Is( X ) . In view of Prop osition 7.1 applied to G ξ , it sues to pro v e that the G -orbit of an y su h η on tains a p oin t opp osite to ξ . By denition of the Tits angle, there is a sequene { x n } in X su h that ∠ x n ( ξ , η ) tends to π . Sine G ξ ats o ompatly , it on tains a sequene { g n } su h that, up on extrating, g n x n on v erges to some x ∈ X and g n η to some η ′ ∈ ∂ X . The angle semi-on tin uit y argumen ts giv en in the pro of of Prop osition I I.9.5(3) in [ BH99 ℄ sho w that ∠ x ( ξ , η ′ ) = π , realling that all g n x ξ . This means that there is a geo desi σ : R → X through x with σ ( −∞ ) = ξ and σ ( ∞ ) = η ′ . On the other hand, sine G η is o ompat in G , the G -orbit of η is losed in the ne top ology . This means that there is g ∈ G with η ′ = g η , as w as to b e sho wn.  W e shall need another form of angle rigidit y (ompare Prop osition 6.8 ), this time for Tits angles. Prop osition 7.15. L et X b e a ge o desi al ly  omplete pr op er CA T(0) sp a e, G < Is( X ) a lose d total ly dis onne te d sub gr oup and ξ ∈ ∂ X . If the stabiliser G ξ ats  o  omp atly on X , then the G -orbit of ξ is disr ete in the Tits top olo gy. Pr o of. Supp ose for a on tradition that there is a sequene { g n } su h that g n ξ 6 = ξ for all n but ∠ T ( g n ξ , ξ ) tends to zero. Sine G ξ is o ompat, w e an assume that g n on v erges in G ; sine the Tits top ology is ner than the ne top ology for whi h the G -ation is on tin uous, the limit of g n m ust x ξ and w e an therefore assume g n → 1 . Let B ⊆ X b e an op en ball large enough so that G ξ .B = X . Sine b y Lemma 3.13 w e an apply Theorem 6.1 , there is no loss of generalit y in assuming that ea h g n xes B p oin t wise. Let c : R + → X b e a geo desi ra y p oin ting to w ards ξ with c (0) ∈ B . F or ea h n there is r n > 0 su h that c and g n c bran h at the p oin t c ( r n ) . In partiular, g n xes c ( r n ) but not c ( r n + ε ) no matter ho w small ε > 0 . W e no w  ho ose h n ∈ G ξ su h that x n := h n c ( r n ) ∈ B and notie that the sequene k n := h n g n h − 1 n is b ounded sine k n xes x n . W e an therefore assume up on extrating that it on v erges to some k ∈ G ; in view of Theorem 6.1 , w e an further assume that all k n oinide with k on B and in partiular k xes all x n . Sine ∠ T ( k n ξ , ξ ) = ∠ T ( g n ξ , ξ ) , w e also ha v e k ∈ G ξ . Considering an y giv en n , it follo ws no w that k xes the ra y from x n to ξ . Th us k n xes an initial segmen t of this ra y at x n . This is equiv alen t to g n xing an initial segmen t at c ( r n ) of the ra y from c ( r n ) to ξ , on trary to our onstrution.  Here is a rst indiation that our spaes migh t resem ble symmetri spaes or Eulidean buildings: Prop osition 7.16. L et X b e a pr op er CA T(0) sp a e suh that that the stabiliser of every p oint at innity ats  o  omp atly on X . Then any p oint at innity is  ontaine d in an isometri al ly emb e dde d standar d n -spher e in ∂ X , wher e n = dim ∂ X . 34 PIERRE-EMMANUEL CAPRA CE AND NICOLAS MONOD Pr o of. Let η ∈ ∂ X . There is some standard n -sphere S isometrially em b edded in ∂ X b eause X is o ompat (Theorem C in [ Kle99 ℄). By Lemma 3.1 in [ BL05 ℄, there is ξ ∈ S with ∠ T ( ξ , η ) = π . Let ϑ ∈ S b e the an tip o de in S of ξ . In view of Lemma 7.14 , there is an isometry sending ϑ to η . The image of S on tains η .  W e need one more fat for Theorem 1.3 . The b oundary of a CA T(0) spae need not b e omplete, regardless of the geo desi ompleteness of the spae itself; ho w ev er, this is the ase in our situation in view of Prop osition 7.16 : Corollary 7.17. L et X b e a pr op er CA T(0) sp a e suh that that the stabiliser of every p oint at innity ats  o  omp atly on X . Then ∂ X is ge o desi al ly  omplete. Pr o of. Supp ose for a on tradition that some Tits-geo desi ends at ξ ∈ ∂ X and let B ⊆ ∂ X b e a small on v ex Tits-neigh b ourho o d of ξ ; in partiular, B is on tratible. Sine b y Prop osition 7.16 there is an n -sphere through ξ for n = dim ∂ X , the relativ e homology H n ( B , B \ { ξ } ) is non-trivial. Our assumption implies that B \ { ξ } is on tratible b y using the geo desi on tration to some p oin t η ∈ B \ { ξ } on the giv en geo desi ending at ξ . This implies H n ( B , B \ { ξ } ) = 0 , a on tradition. (This argumen t is adapted from [ BH99 , I I.5.12℄.)  End of pr o of of The or em 1.3 . W e shall use b elo w that pro dut deomp ositions preserv e ge- o desi ompleteness (this follo ws e.g. from [ BH99 , I.5.3(3)℄). W e an redue to the ase where X has no Eulidean fator. By Lemma 3.13 , the group G = Is( X ) as w ell as all sta- bilisers of p oin ts at innit y at minimally . In partiular, Prop osition 7.3 ensures that G has no xed p oin t at innit y and w e an apply Theorem 1.6 and A ddendum 1.8 . Therefore, w e an from no w on assume that X is irreduible. If the iden tit y omp onen t G ◦ is non-trivial, then Theorem 1.1 (see also Theorem 7.4 (iii)) ensures that X is a symmetri spae, and w e are done. W e assume heneforth that G is totally disonneted. F or an y ξ ∈ ∂ X , the olletion An t( ξ ) = { η : ∠ T ( ξ , η ) = π } of an tip o des is on tained in a G -orbit b y Lemma 7.14 and hene is Tits-disrete b y Prop osition 7.15 . This disreteness and the geo desi ompleteness of the b oundary (Corollary 7.17 ) are the assumptions needed for Prop osition 4.5 in [ Lyt05 ℄, whi h states that ∂ X is a building. Sine X is irreduible, ∂ X is not a (non-trivial) spherial join, see Theorem I I.9.24 in [ BH99 ℄. Th us, if this building has non-zero dimension, w e onlude from the main result of [ Lee00 ℄ that X is a Eulidean building of higher rank. If on the other hand ∂ X is zero-dimensional, then w e laim that it is homogeneous under G . Indeed, w e kno w already that for an y giv en ξ ∈ ∂ X , the set An t( ξ ) lies in a single orbit. Sine in the presen t ase An t( ξ ) is simply ∂ X \ { ξ } , the laim follo ws from the fat that G has no xed p oin t at innit y . W e ha v e to sho w that X is an edge-transitiv e tree. T o this end, onsider an y p oin t x ∈ X . The isotrop y group G x is op en b y Theorem 6.1 . In partiular, sine G ats transitiv ely on ∂ X and sine G ξ is o ompat, it follo ws that G x has nitely man y orbits in ∂ X . Let ρ 1 , . . . , ρ k b e geo desi ra ys emanating from x and p oin ting to w ards b oundary p oin ts whi h pro vide a omplete set of represen tativ es for the G x -orbits. F or r > 0 suien tly large, the v arious in tersetions of the ra ys ρ 1 , . . . , ρ k with the r -sphere S r ( x ) en tred at x forms a set of k distint p oin ts. This set is a fundamen tal domain for the G x -ation on S r ( x ) . Sine G x has disrete orbits on S r ( x ) , w e dedue from Theorem 6.1 that the sphere S r ( x ) is nite. Sine this holds for an y r > 0 suien tly large and an y x ∈ X , it follo ws that ev ery sphere in X is nite. This implies that X is 1 -dimensional (see [ Kle99 ℄). In other w ords X is a ISOMETR Y GR OUPS OF NON-POSITIVEL Y CUR VED SP A CES: STR UCTURE THEOR Y 35 metri tree. W e denote b y V the set of bran h p oin ts whi h w e shall all the v erties. It remains to sho w that G has at most t w o orbits on V . Giv en ξ ′ ∈ ∂ X , let β ξ ′ : G ξ ′ → R denote the Busemann  harater en tred at ξ ′ (see  2 ). Sine X is a o ompat tree, it follo ws that β ξ ′ has disrete image. Let g ∈ G ξ ′ b e an elemen t su h that β ξ ′ ( g ) is p ositiv e and minimal. Then g is h yp erb oli and translates a geo desi line L . Let ξ ′′ denote the endp oin t of L distint from ξ . Let v ∈ L b e an y v ertex. W e denote b y e ′ and e ′′ the edges of L on taining v and p oin ting resp etiv ely to ξ ′ and ξ ′′ . Giv e an y edge e on taining v with e ′ 6 = e 6 = e ′′ , w e prolong e to a geo desi ra y ρ whose in tersetion with L is redued to { v } . Sine G ξ ′ is transitiv e on ∂ X \ { ξ ′ } there exists g ′ ∈ G ξ ′ su h that g ′ .ξ ′′ = ρ ( ∞ ) . Up on pre-omp osing g ′ with a suitable p o w er of g , w e ma y assume that β ξ ′ ( g ′ ) = 0 . In other w ords g ′ xes v . This sho ws that G ξ ′ ,v is transitiv e on the edges on taining v and dieren t from e ′ . The same argumen t with ξ ′ and ξ ′′ in ter hanged sho ws that G ξ ′′ ,v is transitiv e on the edges on taining v and dieren t from e ′′ . In partiular G v is transitiv e on the edges on taining v . A straigh tforw ard indution on the distane to v no w sho ws that for an y v ertex w ∈ V , the isotrop y group G w is transitiv e on the edges on taining w . This implies that G is indeed edge-transitiv e.  8. A few ases of CA T(0) superrigidity This Setion demonstrates that ertain forms of sup errigidit y an b e obtained b y om- bining the struture results of this pap er with kno wn sup errigidit y te hniques. Mu h more general results will b e established in the ompanion pap er [ CM08b ℄. 8.A. CA T(0) sup errigidit y for some lassial non-uniform latties. Let Γ b e a non- uniform lattie in a simple (real) Lie group G of rank at least 2 . By [ LMR00 , Theorem 2.15℄, unip oten t elemen ts of Γ are exp onen tially distorted. This means that, with resp et to an y nitely generating set of Γ , the w ord length of | u n | is an O (lo g n ) when u is a unip oten t. More generally an elemen t u is alled distorte d if | u n | is sublinear. If Γ is virtually b oundedly generated b y unip oten t elemen ts, one an therefore apply the follo wing xed p oin t priniple: Lemma 8.1. L et Γ b e a gr oup whih is virtual ly b ounde d ly gener ate d by distorte d elements. Then any isometri Γ -ation on a  omplete CA T(0) sp a e suh that elements of zer o tr ans- lation length ar e el lipti has a glob al xe d p oint. Pr o of. F or an y Γ -ation on a CA T(0) spae, the translation length of a distorted elemen t is zero. Th us ev ery su h elemen t has a xed p oin t; the assumption on Γ no w implies that all orbits are b ounded, th us pro viding a xed p oin t [ BH99 , I I.2.8(1)℄.  Bounded generation is a strong prop ert y , whi h onjeturally holds for all (non-uniform) latties of a higher rank semi-simple Lie group. It is kno wn to hold for arithmeti groups in split or quasi-split algebrai groups of a n um b er eld K of K -rank ≥ 2 b y [ T a v90 ℄, as w ell as in a few ases of isotropi but non-quasi-split groups [ ER06 ℄. As notied in a on v ersation with Sh. Mozes, Lemma 8.1 yields the follo wing elemen tary sup errigidit y statemen t. Prop osition 8.2. L et Λ = SL n ( Z [ 1 p 1 ··· p k ]) with n ≥ 3 and p i distint primes and set H = SL n ( Q p 1 ) × · · · × SL n ( Q p k ) . Given any isometri Λ -ation on any  omplete CA T(0) sp a e suh that every element of zer o tr anslation length is el lipti, ther e exists a Λ -invariant lose d  onvex subsp a e on whih the given ation extends uniquely to a  ontinuous H -ation by isometries. 36 PIERRE-EMMANUEL CAPRA CE AND NICOLAS MONOD Pr o of. Let X b e a omplete CA T(0) spae endo w ed with a Λ -ation as in the statemen t. The subgroup Γ = SL n ( Z ) < Λ xes a p oin t b y Lemma 8.1 . The statemen t no w follo ws b eause Γ is the in tersetion of Λ with the op en subgroup SL n ( Z p 1 ) × · · · × SL n ( Z p k ) of H ; for later use, w e isolate this elemen tary fat as Lemma 8.3 b elo w.  Lemma 8.3. L et H b e a top olo gi al gr oup, U < H an op en sub gr oup, Λ < H a dense sub gr oup and Γ = Λ ∩ U . A ny Λ -ation by isometries on a  omplete CA T(0) sp a e with a Γ -xe d p oint admits a Λ -invariant lose d  onvex subsp a e on whih the ation extends  ontinuously to H . Pr o of. Let X b e the CA T(0) spae and x 0 ∈ X a Γ -xed p oin t. F or an y nite subset F ⊆ Λ , let Y F ⊆ X b e the losed on v ex h ull of F x 0 . The losed on v ex h ull Y of Λ x 0 is the losure of the union Y ∞ of the direted family { Y F } . Therefore, sine the ation is isometri and Y is omplete, it sues to sho w that the Λ -ation on Y ∞ is on tin uous for the top ology indued on Λ b y H . Equiv alen tly , it sues to pro v e that all orbital maps Λ → Y ∞ are on tin uous at 1 ∈ Λ . This is the ase ev en for the disrete top ology on Y ∞ b eause the p oin t wise xator of ea h Y F is an in tersetion of nitely man y onjugates of Γ , the latter b eing op en b y denition.  The same argumen ts as b elo w sho w that Theorem 1.14 holds for an y lattie of a higher- rank semi-simple Lie group whi h is b oundedly generated b y distorted elemen ts (and a- ordingly Theorem 1.15 generalises to suitable (S-)arithmeti groups). Pr o of of The or ems 1.14 and 1.15 . W e start with the ase Γ = SL n ( Z ) . By Theorem 1.1 , w e obtain a losed on v ex subspae X ′ whi h splits as a diret pro dut X ′ ∼ = X 1 × · · · × X p × Y 0 × Y 1 × · · · × Y q in an Is( X ′ ) -equiv arian t w a y , where Y 0 ∼ = R n is the Eulidean fator. Ea h totally dison- neted fator D i of Is( X ′ ) ∗ ats b y semi-simple isometries on the orresp onding fator Y i of X ′ b y Corollary 6.3 . Therefore, b y Lemma 8.1 for ea h i = 0 , . . . , q , the indued Γ -ation on Y i has a global xed p oin t, sa y y i . In other w ords Γ stabilises the losed on v ex subset Z := X 1 × · · · × X p × { y 0 } × · · · × { y q } ⊆ X . Note that the isometry group of Z is an almost onneted semi-simple real Lie group L . Com bining Lemma VI I.5.1 and Theorems VI I.5.15 and VI I.6.16 from [ Mar91 ℄, it follo ws that the Zariski losure of the image of Γ in L is a omm uting pro dut L 1 .L 2 , where L 1 is ompat, su h that the orresp onding homomorphism Γ → L 2 extends to a on tin uous homomorphism G → L 2 . W e dene Y ⊆ Z as the xed p oin t set of L 1 . No w L 2 , and hene Γ , stabilises Y . Therefore the on tin uous homomorphism G → L 2 yields a G -ation on Y whi h extends the giv en Γ -ation, as desired. Applying Theorem 7.4 p oin t ( iv ) to the pair L 2 < L ating on Z , w e nd in partiular that L 2 has no xed p oin t at innit y in Y . Th us, up on replaing Y b y a subspae, it is L 2 -minimal. No w Theorem 2.4 in [ CM08b ℄ (whi h is ompletely indep enden t of the presen t onsiderations) implies that the Γ - and G -ations on Y are minimal and without xed p oin t in ∂ Y (although there migh t b e xed p oin ts in ∂ X ). T urning to Theorem 1.15 , the only  hange is that one replaes Lemma 8.1 b y Prop osi- tion 8.2 .  ISOMETR Y GR OUPS OF NON-POSITIVEL Y CUR VED SP A CES: STR UCTURE THEOR Y 37 8.B. CA T(0) sup errigidit y for irreduible latties in pro duts. The aim of this se- tion is to state a v ersion of the sup errigidit y theorem [ Mon06 , Theorem 6℄ with CA T(0) targets. The original statemen t from lo . it. onerns ations of latties on arbitr ary CA T(0) spaes, with redued un b ounded image. The follo wing statemen t sho ws that, when the underlying CA T(0) spae is nie enough, the assumption on the ation an b e onsid- erably w eak ened. W e reall for the statemen t that an y isometri ation on a prop er CA T(0) spae without glob el xed p oin t at innit y admits a  anoni al minimal non-empt y losed on v ex in v arian t subspae, see Remarks 39 in [ Mon06 ℄. Theorem 8.4. L et Γ b e an irr e duible uniform (or squar e-inte gr able we akly  o  omp at) latti e in a pr o dut G = G 1 × · · · × G n of n ≥ 2 lo  al ly  omp at σ - omp at gr oups. L et X b e a pr op er CA T(0) sp a e with nite-dimensional b oundary. Given any Γ -ation on X without xe d p oint at innity, if the  anoni al Γ -minimal subset Y ⊆ X has no Eulide an fator, then the Γ -ation on Y extends to a  ontinuous G -ation by isometries. Remark 8.5. Although the ab o v e ondition on the Eulidean fator in the Γ -minimal sub- spae Y migh t seem a wkw ard, it annot b e a v oided, as illustrated b y Example 64 in [ Mon06 ℄. Notie ho w ev er that if Γ has the prop ert y that an y isometri ation on a nite-dimensional Eulidean spae has a global xed (for example if Γ has Kazhdan's prop ert y (T)), then an y minimal Γ -in v arian t subspae has no Eulidean fator. Pr o of of The or em 8.4 . Let Y ⊆ X b e the anonial subspae realled ab o v e. Then Is( Y ) ats minimally on Y , without xed p oin t at innit y . In partiular w e ma y apply Theorem 1.6 and A ddendum 1.8 . In order to sho w that the Γ -ation on Y extends to a on tin uous G - ation, it is suien t to sho w that the indued Γ -ation on ea h irreduible fator of Y extends to a on tin uous G -ation, fatoring through some G i . But the indued Γ -ation on ea h irreduible fator of Y is redued b y Corollary 3.8 . Th us the result follo ws from [ Mon06 , Theorem 6℄.  Referenes [AB98a℄ Sot A dams and W erner Ballmann, A menable isometry gr oups of Hadamar d sp a es , Math. Ann. 312 (1998), no. 1, 183195. [AB98b℄ Stephanie B. Alexander and Ri hard L. Bishop, W arp e d pr o duts of Hadamar d sp a es , Man usripta Math. 96 (1998), no. 4, 487505. [AB04℄ , Curvatur e b ounds for warp e d pr o duts of metri sp a es , Geom. F unt. Anal. 14 (2004), no. 6, 11431181. [Ale57℄ Alexander D. Alexandro w, Üb er eine V er al lgemeinerung der R iemannshen Ge ometrie , S hr. F ors h ungsinst. Math. 1 (1957), 3384. [Bal95℄ W erner Ballmann, L e tur es on sp a es of nonp ositive urvatur e , DMV Seminar, v ol. 25, Birkhäuser V erlag, Basel, 1995, With an app endix b y Misha Brin. [Ber℄ V aleri   N. Beresto vski  , Busemann sp a es of Aleksandr ov urvatur e b ounde d ab ove , Max-Plan k preprin t Nr. 74/2001. [Ber02℄ , Busemann sp a es with upp er-b ounde d Aleksandr ov urvatur e , Algebra i Analiz 14 (2002), no. 5, 318. [BGP92℄ Y uri Burago, Mikhaïl Gromo v, and Gregory P erel ′ man, A. D. Aleksandr ov sp a es with urvatur es b ounde d b elow , Usp ekhi Mat. Nauk 47 (1992), no. 2(284), 351, 222. [BH99℄ Martin R. Bridson and André Haeiger, Metri sp a es of non-p ositive urvatur e , Grundlehren der Mathematis hen Wissens haften 319, Springer, Berlin, 1999. [BL05℄ Andreas Balser and Alexander Lyt hak, Centers of  onvex subsets of buildings , Ann. Global Anal. Geom. 28 (2005), no. 2, 201209. 38 PIERRE-EMMANUEL CAPRA CE AND NICOLAS MONOD [BM96℄ Mar Burger and Shahar Mozes, CA T(-1) -sp a es, diver gen e gr oups and their  ommensur ators , J. Amer. Math. So . 9 (1996), 5793. [BM00℄ , Gr oups ating on tr e es: fr om lo  al to glob al strutur e , Inst. Hautes Études Si. Publ. Math. (2000), no. 92, 113150 (2001). [BO69℄ Ri hard L. Bishop and Barrett O'Neill, Manifolds of ne gative urvatur e , T rans. Amer. Math. So . 145 (1969), 149. [Bou68℄ Niolas Bourbaki, Éléments de mathématique. Fas. XXXIV. Gr oup es et algèbr es de Lie. Chapitr e IV: Gr oup es de Coxeter et systèmes de Tits. Chapitr e V: Gr oup es engendr és p ar des r éexions. Chapitr e VI: systèmes de r aines , A tualités Sien tiques et Industrielles, No. 1337, Hermann, P aris, 1968. [Bou71℄ , Éléments de mathématique. Top olo gie génér ale. Chapitr es 1 à 4 , Hermann, P aris, 1971. [Cap07℄ Pierre-Emman uel Caprae, A menable gr oups and hadamar d sp a es with a total ly dis onne te d isometry gr oup , Preprin t, a v ailable at , 2007. [Che99℄ Chien-Hsiung Chen, W arp e d pr o duts of metri sp a es of urvatur e b ounde d fr om ab ove , T rans. Amer. Math. So . 351 (1999), no. 12, 47274740. [CK00℄ Christopher B. Crok e and Brue Kleiner, Sp a es with nonp ositive urvatur e and their ide al b ound- aries , T op ology 39 (2000), no. 3, 549556. [CL08℄ Pierre-Emman uel Caprae and Alexander Lyt hak, A t innity of nite-dimensional CA T(0) sp a es , Preprin t, 2008. [CM08a℄ Pierre-Emman uel Caprae and Niolas Mono d, De  omp osing lo  al ly  omp at gr oups into simple pie  es , Preprin t, arXiv:math/0811.4101, 2008. [CM08b℄ , Isometry gr oups of non-p ositively urve d sp a es: Disr ete sub gr oups , Preprin t, 2008. [dR52℄ Georges de Rham, Sur la r e dutibilité d'un esp a e de Riemann , Commen t. Math. Helv. 26 (1952), 328344. [ER06℄ Igor V. Ero v enk o and Andrei S. Rapin h uk, Bounde d gener ation of S -arithmeti sub gr oups of isotr opi ortho gonal gr oups over numb er elds , J. Num b er Theory 119 (2006), no. 1, 2848. [F ar08℄ Benson F arb, Gr oup ations and Hel ly's the or em , Preprin t, 2008. [FL06℄ Thomas F o erts h and Alexander Lyt hak, The de Rham de  omp osition the or em for metri sp a es , GAF A (to app ear), arXiv:math/0605419v1, 2006. [FNS06℄ K o ji F ujiw ara, K oi hi Nagano, and T ak ashi Shio y a, Fixe d p oint sets of p ar ab oli isometries of CA T(0)-sp a es , Commen t. Math. Helv. 81 (2006), no. 2, 305335. [GW71℄ Detlef Gromoll and Joseph A. W olf, Some r elations b etwe en the metri strutur e and the algebr ai strutur e of the fundamental gr oup in manifolds of nonp ositive urvatur e , Bull. Amer. Math. So . 77 (1971), 545552. [Hei74℄ Ernst Hein tze, On homo gene ous manifolds of ne gative urvatur e , Math. Ann. 211 (1974), 2334. [HLS00℄ Christoph Hummel, Urs Lang, and Viktor S hro eder, Convex hul ls in singular sp a es of ne gative urvatur e , Ann. Global Anal. Geom. 18 (2000), no. 2, 191204. [Kle99℄ Brue Kleiner, The lo  al strutur e of length sp a es with urvatur e b ounde d ab ove , Math. Z. 231 (1999), no. 3, 409456. [KM99℄ Anders Karlsson and Gregory A. Margulis, A multipli ative er go di the or em and nonp ositively urve d sp a es , Comm. Math. Ph ys. 208 (1999), no. 1, 107123. [Lee00℄ Bernhard Leeb, A har aterization of irr e duible symmetri sp a es and Eulide an buildings of higher r ank by their asymptoti ge ometry , Bonner Mathematis he S hriften, 326, Univ ersität Bonn Mathematis hes Institut, Bonn, 2000. [LMR00℄ Alexander Lub otzky , Shahar Mozes, and Madabusi San tanam Ragh unathan, The wor d and R ie- mannian metris on latti es of semisimple gr oups , Inst. Hautes Études Si. Publ. Math. (2000), no. 91, 553 (2001). [L Y72℄ H. Blaine La wson, Jr. and Shing-T ung Y au, Comp at manifolds of nonp ositive urvatur e , J. Dif- feren tial Geometry 7 (1972), 211228. [Lyt05℄ Alexander Lyt hak, R igidity of spheri al buildings and joins , Geom. F unt. Anal. 15 (2005), no. 3, 720752. [Mar91℄ Gregory A. Margulis, Disr ete sub gr oups of semisimple Lie gr oups , Ergebnisse der Mathematik und ihrer Grenzgebiete (3) [Results in Mathematis and Related Areas (3)℄, v ol. 17, Springer-V erlag, Berlin, 1991. [Mon01℄ Niolas Mono d, Continuous b ounde d  ohomolo gy of lo  al ly  omp at gr oups , Leture Notes in Math- ematis, v ol. 1758, Springer-V erlag, Berlin, 2001. ISOMETR Y GR OUPS OF NON-POSITIVEL Y CUR VED SP A CES: STR UCTURE THEOR Y 39 [Mon06℄ , Sup errigidity for irr e duible latti es and ge ometri splitting , J. Amer. Math. So . 19 (2006), no. 4, 781814. [Pra77℄ Gopal Prasad, Str ong appr oximation for semi-simple gr oups over funtion elds , Ann. of Math. (2) 105 (1977), no. 3, 553572. [Sw e99℄ Eri L. Sw enson, A ut p oint the or em for CA T(0) gr oups , J. Dieren tial Geom. 53 (1999), no. 2, 327358. [T a v90℄ Oleg I. T a vgen ′ , Bounde d gener ability of Cheval ley gr oups over rings of S -inte ger algebr ai num- b ers , Izv. Ak ad. Nauk SSSR Ser. Mat. 54 (1990), no. 1, 97122, 221222. [WM07℄ Da v e Witte Morris, Bounde d gener ation of SL ( n, A ) (after D. Carter, G. Kel ler, and E. Paige). , New Y ork J. Math. 13 (2007), 383421. UCL, 1348 Louv ain-la-Neuve, Belgium E-mail addr ess : pe.apraeulouvain.be EPFL, 1015 La usanne, Switzerland E-mail addr ess : niolas.monodepfl.h

Original Paper

Loading high-quality paper...

Comments & Academic Discussion

Loading comments...

Leave a Comment