MEMOIRS of the American Mathematical Society Number 1006 On the Algebraic Foundations of Bounded Cohomology Theo Bu¨hler November 2011 • Volume 214 • Number 1006 (second of 5 numbers) • ISSN 0065-9266 American Mathematical Society Number 1006 On the Algebraic Foundations of Bounded Cohomology Theo Bu¨hler November2011 • Volume214 • Number1006(secondof5numbers) • ISSN0065-9266 Library of Congress Cataloging-in-Publication Data Bu¨hler,Theo,1978- Onthealgebraicfoundationsofboundedcohomology/TheoBu¨hler. p.cm. —(MemoirsoftheAmericanMathematicalSociety,ISSN0065-9266;no. 1006) “November2011,volume214,number1006(secondof5numbers).” Includesbibliographicalreferences. ISBN978-0-8218-5311-5(alk. paper) 1. Derived categories (Mathematics). 2. Homology theory. 3. Functions of bounded varia- tion. 4.Topologicalspaces. I.Title. QA169.B84 2011 514(cid:2).23—dc23 2011030199 Memoirs of the American Mathematical Society Thisjournalisdevotedentirelytoresearchinpureandappliedmathematics. Publisher Item Identifier. The Publisher Item Identifier (PII) appears as a footnote on theAbstractpageofeacharticle. Thisalphanumericstringofcharactersuniquelyidentifieseach articleandcanbeusedforfuturecataloguing,searching,andelectronicretrieval. Subscription information. Beginning with the January 2010 issue, Memoirs is accessi- ble from www.ams.org/journals. The 2011 subscription begins with volume 209 and consists of six mailings, each containing one or more numbers. Subscription prices are as follows: for pa- per delivery, US$741 list, US$592.80 institutional member; for electronic delivery, US$667 list, US$533.60institutional member. Uponrequest, subscribers topaper delivery ofthis journalare also entitled to receive electronic delivery. If ordering the paper version, subscribers outside the United States and India must pay a postage surcharge of US$69; subscribers in India must pay apostagesurchargeofUS$95. ExpediteddeliverytodestinationsinNorthAmericaUS$58;else- whereUS$167. Subscriptionrenewalsaresubjecttolatefees. Seewww.ams.org/help-faqformore journalsubscriptioninformation. Eachnumbermaybeorderedseparately;pleasespecifynumber whenorderinganindividualnumber. Back number information. Forbackissuesseewww.ams.org/bookstore. Subscriptions and orders should be addressed to the American Mathematical Society, P.O. Box 845904, Boston, MA 02284-5904USA. All orders must be accompanied by payment. Other correspondenceshouldbeaddressedto201CharlesStreet,Providence,RI02904-2294USA. Copying and reprinting. Individual readers of this publication, and nonprofit libraries actingforthem,arepermittedtomakefairuseofthematerial,suchastocopyachapterforuse in teaching or research. Permission is granted to quote brief passages from this publication in reviews,providedthecustomaryacknowledgmentofthesourceisgiven. Republication,systematiccopying,ormultiplereproductionofanymaterialinthispublication is permitted only under license from the American Mathematical Society. Requests for such permissionshouldbeaddressedtotheAcquisitionsDepartment,AmericanMathematicalSociety, 201 Charles Street, Providence, Rhode Island 02904-2294 USA. Requests can also be made by [email protected]. MemoirsoftheAmericanMathematicalSociety(ISSN0065-9266)ispublishedbimonthly(each volume consisting usually of more than one number) by the American Mathematical Society at 201CharlesStreet,Providence,RI02904-2294USA.PeriodicalspostagepaidatProvidence,RI. Postmaster: Send address changes to Memoirs, American Mathematical Society, 201 Charles Street,Providence,RI02904-2294USA. (cid:2)c 2011bytheAmericanMathematicalSociety. Allrightsreserved. Copyrightofindividualarticlesmayreverttothepublicdomain28years afterpublication. ContacttheAMSforcopyrightstatusofindividualarticles. (cid:2) (cid:2) (cid:2) ThispublicationisindexedinScienceCitation IndexR,SciSearchR,ResearchAlertR, (cid:2) (cid:2) CompuMath Citation IndexR,Current ContentsR/Physical,Chemical& Earth Sciences. PrintedintheUnitedStatesofAmerica. (cid:2)∞ Thepaperusedinthisbookisacid-freeandfallswithintheguidelines establishedtoensurepermanenceanddurability. VisittheAMShomepageathttp://www.ams.org/ 10987654321 161514131211 Contents Introduction and Main Results . . . . . . . . . . . . . . . . . . . . . . . . . . vii 1. Reconstructing the Burger-Monod Theory . . . . . . . . . . . . . . . . viii 2. More Functoriality . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xiii 3. Discrete Groups . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xiv 4. Amenable Groups. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xvi 5. (cid:2)1-Homology and Bounded Cohomology for Spaces . . . . . . . . . . .xviii 6. The Canonical Semi-Norm . . . . . . . . . . . . . . . . . . . . . . . . . xx Part 1. Triangulated Categories . . . . . . . . . . . . . . . . . . . . 1 Chapter I. Triangulated Categories . . . . . . . . . . . . . . . . . . . . . . . 3 1. Definition and Elementary Properties. . . . . . . . . . . . . . . . . . . 3 2. Triangle Functors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14 Chapter II. The Derived Category of an Exact Category . . . . . . . . . . . 17 1. Verdier Localization . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17 2. The Derived Category of an Exact Category . . . . . . . . . . . . . . . 24 3. Derived Functors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26 Chapter III. Abstract Truncation: t-Structures and Hearts . . . . . . . . . . 31 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31 2. Abstract Truncation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36 Part 2. Homological Algebra for Bounded Cohomology . . . . . . 45 Chapter IV. Categories of Banach Spaces . . . . . . . . . . . . . . . . . . . 47 1. The Bicomplete Category of Banach Spaces . . . . . . . . . . . . . . . 47 2. The Additive Category of Banach Spaces . . . . . . . . . . . . . . . . 54 Chapter V. Derived Categories of Banach G-Modules . . . . . . . . . . . . . 63 1. Discrete Groups . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63 2. Some Remarks on Topological Groups . . . . . . . . . . . . . . . . . . 70 Part 3. Appendices . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73 Appendix A. Mapping Cones, Homotopy Push-Outs, Mapping Cylinders . . 75 1. The Mapping Cone . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75 Appendix B. Pull-Back of Exact Structures . . . . . . . . . . . . . . . . . . 79 Appendix C. Model Categories . . . . . . . . . . . . . . . . . . . . . . . . . 81 1. Statement of the Result . . . . . . . . . . . . . . . . . . . . . . . . . . 81 iii iv CONTENTS 2. Proof of Theorem 1.2.2 . . . . . . . . . . . . . . . . . . . . . . . . . . . 83 3. Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86 Appendix D. Standard Borel G-Spaces are Regular . . . . . . . . . . . . . . 87 1. Representations Associated to a Group Action . . . . . . . . . . . . . 87 Appendix E. The Existence of Bruhat Functions . . . . . . . . . . . . . . . 91 1. Proper Group Actions . . . . . . . . . . . . . . . . . . . . . . . . . . . 91 Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95 Abstract Bounded cohomology for topological spaces was introduced by Gromov in the late seventies, mainly to describe the simplicial volume invariant. It is an exotic cohomologytheoryforspacesinthatitfailsexcisionandthuscannotberepresented by spectra. Gromov’s basic vanishing result of bounded cohomology for simply connectedspacesimpliesthatboundedcohomologyforspacesisaninvariantofthe fundamental group. To prove this, one is led to introduce a cohomology theory for groups and the present work is concerned with the latter, which has been studied by Gromov, Brooks, Ivanov and Noskov to name but the most important initial contributors. Generalizing these ideas from discrete groups to topological groups, Burger and Monod have developed continuous bounded cohomology in the late nineties. It is a widespread opinion among experts that (continuous) bounded cohomo- logy cannot be interpreted as a derived functor and that triangulated methods break down. We prove that this is wrong. We use the formalism of exact categories and their derived categories in order to construct a classical derived functor on the category of Banach G-modules with valuesinWaelbroeck’sabeliancategory. Thisgivesusanaxiomaticcharacterization of this theory for free and it is a simple matter to reconstruct the classical semi- normed cohomology spaces out of Waelbroeck’s category. Weprovethatthederivedcategoriesofrightboundedandofleftboundedcom- plexes of Banach G-modules are equivalent to the derived category of two abelian categories (one for each boundedness condition), a consequence of the theory of abstracttruncationandheartsoft-structures. Moreover,weprovethatthederived categories of Banach G-modules can be constructed as the homotopy categories of model structures on the categories of chain complexes of Banach G-modules thus proving that the theory fits into yet another standard framework of homological and homotopical algebra. ReceivedbytheeditorFebruary8,2008. ArticleelectronicallypublishedonMarch11,2011;S0065-9266(2011)00618-0. 2000MathematicsSubjectClassification. Primary18E30,Secondary18E10,18G60,46M18, 20J05,57T. Key wordsand phrases. BoundedCohomology,DerivedCategories,DerivedFunctors. Affiliation at time of publication: Departement Mathematik der ETH Zu¨rich, R¨amistrasse101,CH-8092Zu¨rich;email: [email protected]. (cid:2)c2011 American Mathematical Society v Introduction and Main Results This work is based on the remarkable thesis of N. Monod [Mon01] which laid out the foundations for the theory of continuous bounded cohomology with a view towardsapplicationsinrigiditytheory. Itisthedetailedpresentationofjointresults with his advisor M. Burger. The power and the merits of the machinery are well- known and the reader should consult the recent works of Burger, Monod and their collaboratorsin order toget anidea of the state of the art. Since Monod’s thesis is quite heavy-going and the present work is probably even worse in this respect, we recommend the novice to (cid:2)1-homology and bounded cohomology to consult Lo¨h’s beautiful thesis [L¨oh07], the declared—and happily also achieved—goal of which is to provide a lightweight approach to the theory. Weshallfocusonthealgebraicaspectsofthetheoryinordertoprovideagood reasonforthecloseresemblanceofthefundamentalhomological resultsinbounded cohomology to their classical algebraic counterparts. Burger and Monod point out that the definitions of the central notions are inspired by relative homological algebra—cf. also N. V. Ivanov [Iva85]. Our point is that they are nothing but instances of the categorical notions up to a slight detail which stems from the interest in semi-norms. Keeping track of semi-norms in homology is certainly of utmost importance to most applications up to date. Many technicalities are due to them and it seems that they are nothing but a nuisance for theoretical considerations. A clear-cut approach is probably best achieved by dropping the semi-norms and by recovering them only after the cohomology theory has been developed in full. The main upshot of the present work is as follows: (1) Underlyingcontinuous bounded cohomology there is an exactcategory in the sense of Quillen with enough injectives. (2) Continuous bounded cohomology arises from the total derived functor of the maximal invariant submodule functor. (3) The heart of the canonical left t-structure on the derived category of BanachspacesisWaelbroeck’sabeliancategoryofquotientBanachspaces. Continuous bounded cohomology factors over this abelian category and canthusbeviewedasaclassicalderivedfunctor—modulothefactthatits domain of definition is not an abelian category. An ad hoc construction gives back the semi-normed space-valued functor of Burger and Monod. Continuousboundedcohomologyisafunctionalanalyticvariantofgroupcohomol- ogy. Objects of study are Hausdorff topological groups and their strongly continu- ous and isometric representations on Banach spaces. Since the category of Banach spaces is not abelian, the classical language does not apply. However, the category ofBanachG-modulescanbeequippedwithvariousexactstructuresinthesenseof vii viii INTRODUCTION AND MAIN RESULTS Quillen, for which the theory of the derived category allows us to speak of derived functors. At the heart of the matter is Burger-Monod’s “functorial characterization” which tells us that continuous bounded cohomology may be computed as if it were the classical derived functor of the G-invariants. By the work of Burger-Monod we know that cohomology in degree zero coincides with the G-invariants, cohomology vanishes on injectives in higher degrees and that there is a natural long exact sequence for every short exactsequence of G-modules. Thus, bounded cohomology shouldbethederivedfunctoroftheG-invariantsandthemainpurposeofthiswork is to make this precise and to show that bounded cohomology for groups is not as exotic as is often claimed. 1. Reconstructing the Burger-Monod Theory Our construction of bounded cohomology involves three essential choices and one ad hoc construction. These choices are justified by the original work of Burger and Monod, see [Mon01] and [BM02]. Fix as ground field k either the real or complex numbers. Let Ban be the category of Banach spaces over k and bounded k-linear maps. Let G be a Hausdorff topological group and define the category G−Ban of Banach G-modules to be the strongly continuous and isometric representations of G on Banach spaces over k together with the bounded linear G-equivariant maps. This is an additive category which is finitely bicomplete and whose class of kernel-cokernel pairs coincides with the short sequences of Banach G-modules whose underlying sequence of k-vector spaces is short exact. Notice that there is a functor (−)G :G−Ban→Ban associating to a Banach G-module M the closed subspace MG of G-invariant vectors. It is right adjoint to thetrivialmodulefunctorBan→G−BanwhicharisesfromconsideringaBanach space as a Banach G-module with the trivial G-action. The three choices are: (i) On G−Ban choose the exact structure given by the class EG of short se- rel quences whose underlying sequence of Banach spaces is split exact. Burger andMonodhaveessentially shownthat(G−Ban,EG )hasenoughinjectives. rel Therefore the derived category D+(G−Ban,EG ) of left bounded complexes rel is equivalent to the homotopy category K+(I (EG )) of left bounded com- rel plexes of EG -injective Banach G-modules. In particular D+(G−Ban,EG ) rel rel has small Hom-sets and right derived functors exist. (ii) On Ban choose the exact structure given by the class E consisting of all max kernel-cokernel pairs. It is well-known that (Ban,E ) has enough of both max projective and injective objects. (iii) Choose the left truncation functor τ≤0 on Ch(Ban) given on objects by (cid:2) τ≤0E =(···→E−2 →E−1 →Kerd0 →0→···). (cid:2) E It induces the left t-structure (D≤0,D≥0) on D+(Ban,E ) whose heart is (cid:2) (cid:2) max equivalent to Waelbroeck’s abelian category qBan. Let H0 :=τ≥0τ≤0 ∼=τ≤0τ≥0 (cid:2) (cid:2) (cid:2) (cid:2) (cid:2) be the associated homological functor H0 : D+(Ban,E ) → qBan and (cid:2) max write as usual Hn(E)=H0(E[n]) for n∈Z. (cid:2) (cid:2) 1. RECONSTRUCTING THE BURGER-MONOD THEORY ix Remark. For the moment it suffices to know about qBan that it is abelian and that its objects are represented by pairs (E−1 (cid:4)→ E0) of Banach spaces and an injective bounded linear map between them. Moreover, there is an embedding (a fully faithful functor) ι : Ban → qBan given on objects by ι(E) = (0 (cid:4)→ E) and that ι admits a left adjoint because Ban has cokernels, see Lemma 2.2.6 of Chapter III. Letusgiveasimpleproofoftheexistenceofenoughinjectivesin(G−Ban,EG ). rel Underlyingtheargument istheconceptofamonad(triple)arisingfromanadjoint pair of functors, see Weibel [Wei94, Section 8.6]. Recall that a function f : G → E from a Hausdorff topological group G into a Banach space E is called left uniformly continuous if for every net g → 1 in G λ one has sup (cid:5)f(g x)−f(x)(cid:5) −λ−→−−∞→0. x∈G λ E Lemma. The forgetful functor ↓: G−Ban → Ban has a right adjoint given on objects E ∈Ban by the Banach space Clu(G,E)={f :G→E : f is bounded and left uniformly continuous} b equipped with the supremum norm and the G-action (gf)(x)=f(g−1x). Proof. Let L be the forgetful functor and let R = Clu(G,−). Let M ∈ G−Ban b and let E ∈ Ban be any two objects. The unit and counit of the adjunction are the natural transformations η :idG−Ban ⇒RL and ε:LR⇒idBan given by η (m)=(x(cid:8)→x−1m)∈RL(M)=Clu(G,↓M) and ε (f)=f(1)∈E M b E for m∈M and for f∈LR(E)=↓Clu(G,E). Since (cid:5)m(cid:5) =sup (cid:5)η (m)(x)(cid:5) , b M x∈G M M weseethatη (m)isaboundedfunction. Letuscheckthatη (m)isleftuniformly M M continuous, so let g →1 be a net and compute for m∈M λ sup(cid:5)η (m)(g x)−η (m)(x)(cid:5) = sup(cid:5)x−1g−1m−x−1m(cid:5) =(cid:5)g−1m−m(cid:5) M λ M M λ M λ M x∈G x∈G thelasttermconvergestozerobecausetheactionofGonM isstronglycontinuous. Moreover, η :M →RL(M) is G-equivariant because M η (gm)(x)=x−1gm=(g−1x)−1m=η (m)(g−1x)=[gη (m)](x). M M M Thefactthatη andεarenaturaltransformationsisobviousaswellastheonethat the compositions L(M)−L−(−ηM−→) LRL(M)−ε−L−(M−→) L(M) and R(E)−η−R−(−E→) RLR(E)−R−(−ε−E→) R(E) are equal to id and to id . Thus we have L(cid:9)R by [ML98, Ch. IV.1]. (cid:2) L(M) R(E) Corollary. There are enough injectives in (G−Ban,EG ). rel Proof. Choosethesplit exactstructureE onBansothateveryBanachspace min is (projective and) injective. Observe that the forgetful functor ↓:(G−Ban,EG )−→(Ban,E ) rel min
Description: