1: \documentclass[american,a4paper,10pt,leqno]{amsart}

2: 

3: 

4: 

5: \listfiles

6: \usepackage{babel,enumerate,amssymb,amsthm}

7: \usepackage[arrow,curve,matrix,tips,2cell]{xy}

8:   \SelectTips{eu}{10} \UseTips

9:   \UseAllTwocells

10: 

11: \DeclareMathAlphabet{\matheurm}{U}{eur}{m}{n}

12:   % AMS Fonts' Euler roman (cursive) medium weight

13:   % used for categories

14: 

15: 

16: \newcommand{\cell}{\operatorname{cell}}

17: \newcommand{\cyclic}{\operatorname{cyclic}}

18: \newcommand{\et}{\operatorname{\'et}}

19: \newcommand{\iso}{\operatorname{iso}}

20: \newcommand{\op}{\operatorname{op}}

21: \newcommand{\pt}{\operatorname{pt}}

22: \newcommand{\per}{\operatorname{per}}

23: \newcommand{\sing}{\operatorname{sing}}

24: 

25: 

26: \newcommand{\Jdtr}{\mathbf{dtr}}

27: \newcommand{\Jntr}{\mathbf{ntr}}

28: \newcommand{\K}{\mathbf{K}}

29: \newcommand{\THH}{\mathbf{T\!H\!H}}

30: \newcommand{\TC}{\mathbf{T\!C}}

31: \newcommand{\trc}{\mathbf{trc}}

32: 

33: 

34: \newcommand{\Ab}{\matheurm{Ab}}

35: \newcommand{\Addcat}{\matheurm{Add~cat}}

36: \newcommand{\Cat}{\matheurm{Cat}}

37: \newcommand{\FGINJ}{\matheurm{FGINJ}}

38: \newcommand{\Groupoids}{{\matheurm{Groupoids}}}

39: \newcommand{\Or}{\matheurm{Or}\hspace*{.1em}}   %% I HAVE MODIFIED THIS MACRO %%

40: \newcommand{\OrG}{\matheurm{Or}G}

41: \newcommand{\Sets}{\matheurm{Sets}}

42: \newcommand{\Sp}{\matheurm{Sp}}

43: 

44: \newcommand{\cat}{\text{-}\matheurm{cat}}

45: 

46: \DeclareMathOperator{\Aut}{Aut}

47: \DeclareMathOperator{\aut}{aut}

48: \DeclareMathOperator{\ch}{ch}

49: \DeclareMathOperator{\coequ}{coequ}

50: \DeclareMathOperator{\coker}{coker}

51: \DeclareMathOperator{\con}{con}

52: \DeclareMathOperator{\diag}{diag}

53: \DeclareMathOperator{\dtr}{dtr}

54: \DeclareMathOperator{\ntr}{ntr}

55: \DeclareMathOperator{\id}{id}

56: \DeclareMathOperator{\ind}{ind}

57: \DeclareMathOperator{\Idem}{Idem}

58: \DeclareMathOperator{\map}{map}

59: \DeclareMathOperator{\mor}{mor}

60: \DeclareMathOperator{\obj}{obj}

61: \DeclareMathOperator{\pr}{pr}

62: \DeclareMathOperator{\res}{res}

63: \DeclareMathOperator{\sub}{sub}

64: \DeclareMathOperator{\tr}{tr}

65: \DeclareMathOperator{\Seg}{Seg}

66: \DeclareMathOperator{\Sw}{Sw}

67: \DeclareMathOperator{\Tor}{Tor}

68: \DeclareMathOperator{\Tot}{Tot}

69: 

70: 

71: \newcommand{\negspace}{\hspace*{-.03em}} %% THIS IS A NEW MACRO %%

72: \DeclareMathOperator{\HH}{H\negspace H}  %% THIS IS A NEW MACRO %%

73: \DeclareMathOperator{\HC}{H\negspace C}  %% THIS IS A NEW MACRO %%

74: \DeclareMathOperator{\HN}{H\negspace N}  %% THIS IS A NEW MACRO %%

75: \DeclareMathOperator{\HP}{H\negspace P}  %% THIS IS A NEW MACRO %%

76: \DeclareMathOperator{\HX}{H\negspace X}  %% THIS IS A NEW MACRO %%

77: \DeclareMathOperator{\CN}{C\negspace N}  %% THIS MACRO HAS BEEN MODIFIED %%

78: \DeclareMathOperator{\DK}{D\negspace\negspace K}  %% THIS MACRO HAS BEEN MODIFIED %%

79: 

80: 

81: \DeclareMathOperator*{\colim}{colim}

82: \DeclareMathOperator*{\sma}{\wedge}  % \smash already defined

83: \DeclareMathOperator*{\bigsma}{\bigwedge}

84: \DeclareMathOperator*{\tensor}{\otimes}

85: 

86: \newcommand{\Fin}  {\mathcal{F}\text{in}}

87: \newcommand{\FCyc}{\mathcal{FC}\text{yc}}

88: \newcommand{\VCyc}{\mathcal{VC}\text{yc}}

89: 

90: 

91:   \newcommand{\IA}{\mathbb{A}}

92:   \newcommand{\IB}{\mathbb{B}}

93:   \newcommand{\IC}{\mathbb{C}}

94:   \newcommand{\ID}{\mathbb{D}}

95:   \newcommand{\IE}{\mathbb{E}}

96:   \newcommand{\IG}{\mathbb{G}}

97:   \newcommand{\IH}{\mathbb{H}}

98:   \newcommand{\II}{\mathbb{I}}

99: \renewcommand{\IJ}{\mathbb{J}}

100:   \newcommand{\IK}{\mathbb{K}}

101:   \newcommand{\IL}{\mathbb{L}}

102:   \newcommand{\IM}{\mathbb{M}}

103:   \newcommand{\IN}{\mathbb{N}}

104:   \newcommand{\IO}{\mathbb{O}}

105:   \newcommand{\IP}{\mathbb{P}}

106:   \newcommand{\IQ}{\mathbb{Q}}

107:   \newcommand{\IR}{\mathbb{R}}

108:   \newcommand{\IS}{\mathbb{S}}

109:   \newcommand{\IT}{\mathbb{T}}

110:   \newcommand{\IU}{\mathbb{U}}

111:   \newcommand{\IV}{\mathbb{V}}

112:   \newcommand{\IW}{\mathbb{W}}

113:   \newcommand{\IX}{\mathbb{X}}

114:   \newcommand{\IY}{\mathbb{Y}}

115:   \newcommand{\IZ}{\mathbb{Z}}

116: 

117:   \newcommand{\JA}{\mathbf{A}}    \newcommand{\Ja}{\mathbf{a}}

118:   \newcommand{\JB}{\mathbf{B}}    \newcommand{\Jb}{\mathbf{b}}

119:   \newcommand{\JC}{\mathbf{C}}    \newcommand{\Jc}{\mathbf{c}}

120:   \newcommand{\JD}{\mathbf{D}}    \newcommand{\Jd}{\mathbf{d}}

121:   \newcommand{\JE}{\mathbf{E}}    \newcommand{\Je}{\mathbf{e}}

122:   \newcommand{\JF}{\mathbf{F}}    \newcommand{\Jf}{\mathbf{f}}

123:   \newcommand{\JG}{\mathbf{G}}    \newcommand{\Jg}{\mathbf{g}}

124:   \newcommand{\JH}{\mathbf{H}}    \newcommand{\Jh}{\mathbf{h}}

125:   \newcommand{\JI}{\mathbf{I}}    \newcommand{\Ji}{\mathbf{i}}

126:   \newcommand{\JJ}{\mathbf{J}}    \newcommand{\Jj}{\mathbf{j}}

127:   \newcommand{\JK}{\mathbf{K}}    \newcommand{\Jk}{\mathbf{k}}

128:   \newcommand{\JL}{\mathbf{L}}    \newcommand{\Jl}{\mathbf{l}}

129:   \newcommand{\JM}{\mathbf{M}}    \newcommand{\Jm}{\mathbf{m}}

130:   \newcommand{\JN}{\mathbf{N}}    \newcommand{\Jn}{\mathbf{n}}

131:   \newcommand{\JO}{\mathbf{O}}    \newcommand{\Jo}{\mathbf{o}}

132:   \newcommand{\JP}{\mathbf{P}}    \newcommand{\Jp}{\mathbf{p}}

133:   \newcommand{\JQ}{\mathbf{Q}}    \newcommand{\Jq}{\mathbf{q}}

134:   \newcommand{\JR}{\mathbf{R}}    \newcommand{\Jr}{\mathbf{r}}

135:   \newcommand{\JS}{\mathbf{S}}    \newcommand{\Js}{\mathbf{s}}

136:   \newcommand{\JT}{\mathbf{T}}    \newcommand{\Jt}{\mathbf{t}}

137:   \newcommand{\JU}{\mathbf{U}}    \newcommand{\Ju}{\mathbf{u}}

138:   \newcommand{\JV}{\mathbf{V}}    \newcommand{\Jv}{\mathbf{v}}

139:   \newcommand{\JW}{\mathbf{W}}    \newcommand{\Jw}{\mathbf{w}}

140:   \newcommand{\JX}{\mathbf{X}}    \newcommand{\Jx}{\mathbf{x}}

141:   \newcommand{\JY}{\mathbf{Y}}    \newcommand{\Jy}{\mathbf{y}}

142:   \newcommand{\JZ}{\mathbf{Z}}    \newcommand{\Jz}{\mathbf{z}}

143: 

144:   \newcommand{\EA}{\matheurm{A}}  \newcommand{\Ea}{\matheurm{a}}

145:   \newcommand{\EB}{\matheurm{B}}  \newcommand{\Eb}{\matheurm{b}}

146:   \newcommand{\EC}{\matheurm{C}}  \newcommand{\Ec}{\matheurm{c}}

147:   \newcommand{\ED}{\matheurm{D}}  \newcommand{\Ed}{\matheurm{d}}

148:   \newcommand{\EE}{\matheurm{E}}  \newcommand{\Ee}{\matheurm{e}}

149:   \newcommand{\EF}{\matheurm{F}}  \newcommand{\Ef}{\matheurm{f}}

150:   \newcommand{\EG}{\matheurm{G}}  \newcommand{\Eg}{\matheurm{g}}

151:   \newcommand{\EH}{\matheurm{H}}  \newcommand{\Eh}{\matheurm{h}}

152:   \newcommand{\EI}{\matheurm{I}}  \newcommand{\Ei}{\matheurm{i}}

153:   \newcommand{\EJ}{\matheurm{J}}  \newcommand{\Ej}{\matheurm{j}}

154:   \newcommand{\EK}{\matheurm{K}}  \newcommand{\Ek}{\matheurm{k}}

155:   \newcommand{\EL}{\matheurm{L}}  \newcommand{\El}{\matheurm{l}}

156:   \newcommand{\EM}{\matheurm{M}}  \newcommand{\Em}{\matheurm{m}}

157:   \newcommand{\EN}{\matheurm{N}}  \newcommand{\En}{\matheurm{n}}

158:   \newcommand{\EO}{\matheurm{O}}  \newcommand{\Eo}{\matheurm{o}}

159:   \newcommand{\EP}{\matheurm{P}}  \newcommand{\Ep}{\matheurm{p}}

160:   \newcommand{\EQ}{\matheurm{Q}}  \newcommand{\Eq}{\matheurm{q}}

161:   \newcommand{\ER}{\matheurm{R}}  \newcommand{\Er}{\matheurm{r}}

162:   \newcommand{\ES}{\matheurm{S}}  \newcommand{\Es}{\matheurm{s}}

163:   \newcommand{\ET}{\matheurm{T}}  \newcommand{\Et}{\matheurm{t}}

164:   \newcommand{\EU}{\matheurm{U}}  \newcommand{\Eu}{\matheurm{u}}

165:   \newcommand{\EV}{\matheurm{V}}  \newcommand{\Ev}{\matheurm{v}}

166:   \newcommand{\EW}{\matheurm{W}}  \newcommand{\Ew}{\matheurm{w}}

167:   \newcommand{\EX}{\matheurm{X}}  \newcommand{\Ex}{\matheurm{x}}

168:   \newcommand{\EY}{\matheurm{Y}}  \newcommand{\Ey}{\matheurm{y}}

169:   \newcommand{\EZ}{\matheurm{Z}}  \newcommand{\Ez}{\matheurm{z}}

170: 

171:   \newcommand{\cala}{\mathcal{A}}

172:   \newcommand{\calb}{\mathcal{B}}

173:   \newcommand{\calc}{\mathcal{C}}

174:   \newcommand{\cald}{\mathcal{D}}

175:   \newcommand{\cale}{\mathcal{E}}

176:   \newcommand{\calf}{\mathcal{F}}

177:   \newcommand{\calg}{\mathcal{G}}

178:   \newcommand{\calh}{\mathcal{H}}

179:   \newcommand{\cali}{\mathcal{I}}

180:   \newcommand{\calj}{\mathcal{J}}

181:   \newcommand{\calk}{\mathcal{K}}

182:   \newcommand{\call}{\mathcal{L}}

183:   \newcommand{\calm}{\mathcal{M}}

184:   \newcommand{\caln}{\mathcal{N}}

185:   \newcommand{\calo}{\mathcal{O}}

186:   \newcommand{\calp}{\mathcal{P}}

187:   \newcommand{\calq}{\mathcal{Q}}

188:   \newcommand{\calr}{\mathcal{R}}

189:   \newcommand{\cals}{\mathcal{S}}

190:   \newcommand{\calt}{\mathcal{T}}

191:   \newcommand{\calu}{\mathcal{U}}

192:   \newcommand{\calv}{\mathcal{V}}

193:   \newcommand{\calw}{\mathcal{W}}

194:   \newcommand{\calx}{\mathcal{X}}

195:   \newcommand{\caly}{\mathcal{Y}}

196:   \newcommand{\calz}{\mathcal{Z}}

197: 

198: \newcommand{\EGF}[2]{E_{#2}(#1)}               % klassifizierender Raum einer Familie CW-version

199: 

200: 

201: %%%%%%%%%%%%%%%%%%%%%%%%%%%

202: %%%%%%%%%%%%%%%%%%%%%%%%%%%

203: 

204: % new macros -- begin

205: 

206: \newcommand{\conjclass}[1]{(#1)} % or: \newcommand{\conjclass}[1]{[#1]}

207: \newcommand{\ssconG}{\underline{\con G}} % or: \newcommand{\ssconG}{\con_\bullet G}

208: \DeclareMathOperator{\assembly}{assembly}

209: \newcommand{\into}{\hookrightarrow}

210: \newcommand{\onto}{\twoheadrightarrow}

211: \DeclareMathOperator{\h}{h}

212: \newcommand{\minusinfinity}{{-\infty}}

213: \newcommand{\downdots}{\raisebox{-.55em}{$\ldots$}}

214: \newcommand{\stackover}[2]{\renewcommand{\arraystretch}{0.2}

215: \begin{array}[t]{c}

216: #2\\ {\scriptstyle #1}\\

217: \end{array}

218: \renewcommand{\arraystretch}{1}}

219: 

220: % new macros -- end

221: 

222: %%%%%%%%%%%%%%%%%%%%%%%%%%%

223: %%%%%%%%%%%%%%%%%%%%%%%%%%%

224: 

225: 

226: 

227: 

228: %\newcommand{\comment}[1][???]{\textsf{[\textbf{Comment:} #1]}}

229: %\newcommand{\comment}[1]{\textsf{[\textbf{Comment:} #1]}}

230: 

231: %%%%%%%%%%%%%%%%%%%%%%%%%%%

232: %%%%%%%%%%%%%%%%%%%%%%%%%%%

233: 

234: % to have the `clean' version, replace the macros

235: 

236: \newcommand{\commentskip}{\medskip} %%% new macro

237: \newcounter{commentcounter}

238: \newcommand{\comment}[1]{\stepcounter{commentcounter}{\bf Comment \arabic{commentcounter}}: {\ttfamily #1} }

239: 

240: % by

241: 

242: %\newcommand{\commentskip}{}

243: %\newcommand{\comment}[1]{}

244: 

245: %%%%%%%%%%%%%%%%%%%%%%%%%%%

246: %%%%%%%%%%%%%%%%%%%%%%%%%%%

247: 

248: \theoremstyle{plain}      \newtheorem{theorem}{Theorem}[section]

249:                           \newtheorem{maintheorem}[theorem]{Main Theorem}

250:                           \newtheorem{lemma}[theorem]{Lemma}

251:                           \newtheorem{corollary}[theorem]{Corollary}

252:                           \newtheorem{proposition}[theorem]{Proposition}

253:                           \newtheorem{addendum}[theorem]{Addendum}

254:                           \newtheorem{conjecture}[theorem]{Conjecture}

255: 

256: 

257: \theoremstyle{definition} \newtheorem{definition}[theorem]{Definition}

258:                           \newtheorem{example}[theorem]{Example}

259:                           \newtheorem{question}[theorem]{Question}

260:                           \newtheorem{remark}[theorem]{Remark}

261:                           \newtheorem{notation}[theorem]{Notation}

262:                           \newtheorem{convention}[theorem]{Convention}

263: 

264: \renewcommand{\labelenumi}{(\roman{enumi})}

265: \renewcommand{\theenumi}{(\roman{enumi})}

266: 

267: 

268: \makeatletter\let\c@equation=\c@theorem\makeatother

269: \renewcommand{\theequation}{\thetheorem}

270: 

271: 

272: \hyphenation{equi-va-riant}

273: \hyphenation{Hoch-schild}

274: 

275:   \title[Detecting $K$-theory by cyclic homology]

276:               {Detecting $K$-Theory by cyclic homology}

277:        \author{Wolfgang L\"uck}

278:               \author{Holger Reich}

279:       \address{Westf\"alische Wilhelms-Universit\"at M\"unster\\

280:                Mathematisches Institut\\

281:                Einsteinstr.~62,

282:                D-48149 M\"unster, Germany}

283:         \email{lueck@math.uni-muenster.de}

284:       \urladdr{http://www.math.uni-muenster.de/u/lueck}

285:         \email{reichh@math.uni-muenster.de}

286:       \urladdr{http://www.math.uni-muenster.de/u/reichh}

287:                \date{\today}

288:      \keywords{}

289:     \subjclass[2000]{}

290: 

291: 

292: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

293: 

294: \begin{document}

295: 

296: \maketitle

297: 

298: \begin{abstract}

299: We discuss which part of the rationalized algebraic $K$-theory of a group ring is detected

300: via trace maps to Hochschild homology, cyclic homology, periodic cyclic or negative cyclic homology.

301: 

302: \smallskip

303: \noindent

304: Key words: algebraic K-theory of group rings, Hochschild homology, cyclic homology, trace maps.\\

305: Mathematics Subject Classification 2000: 19D55.

306: \end{abstract}

307: 

308: 

309: Dedicated to memory of Michel Matthey.

310: 

311: 

312: 

313: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

314: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

315: \setcounter{section}{-1}

316: \section{Introduction and statement of results}

317: 

318: 

319: 

320: Fix a commutative ring $k$, referred to as the \emph{ground ring}. Let $R$ be a $k$-algebra, i.e.\ an associative ring $R$ together

321: with a unital ring homomorphism from $k$ to the center of $R$. We

322: denote by $\HH^{\otimes_k}_{\ast} ( R )$ the \emph{Hochschild homology} of $R$ relative to the ground ring $k$,

323: and similarly by $\HC^{\otimes_k}_{\ast}(R)$, $\HP^{\otimes_k}_{\ast}(R)$ and $\HN_{\ast}^{\otimes_k} (R)$

324: the \emph{cyclic}, the \emph{periodic cyclic} and the \emph{negative cyclic homology} of $R$ relative to $k$.

325: Hochschild homology receives a map from the algebraic $K$-theory,

326: which is known as the \emph{Dennis trace map}. There are variants of the Dennis trace taking values in cyclic,

327: periodic cyclic and negative cyclic homology (sometimes called \emph{Chern characters}), as displayed in the

328: following commutative diagram.

329: \begin{eqnarray} \label{trace-maps}

330: \vcenter{\xymatrix@C=3.2em{

331: & \HN_{\ast}^{\otimes_k} ( R ) \ar[r] \ar[d]^{\h} & \HP_{\ast}^{\otimes_k} ( R ) \ar[d] \\

332: K_{\ast} ( R ) \ar[ru]^-{\ntr} \ar[r]^-{\dtr} & \HH_{\ast}^{\otimes_k} ( R ) \ar[r] & \HC_{\ast}^{\otimes_k} ( R ).

333:          }}

334: \end{eqnarray}

335: For the definition of these maps, see \cite[Chapters~8~and~11]{Loday(1992)} and Section~\ref{sec:trace-maps} below.

336: 

337: In the following we will focus on the case of group rings $RG$, where $G$ is a group and we refer to the

338: $k$-algebra $R$ as the \emph{coefficient ring}.

339: We investigate the following question.

340: \begin{question} \label{que:main-question}

341: Which part of $K_{\ast}( R G ) \otimes_{\IZ} \IQ$ can be detected using linear trace invariants

342: like the Dennis trace to Hochschild homology, or its variants with values in cyclic homology,

343: periodic cyclic homology and negative cyclic homology\,?

344: \end{question}

345: 

346: For any group $G$, we prove ``\emph{detection results}'', which state that certain parts of $K_\ast(RG) \otimes_{\IZ} \IQ$

347: can be detected by the trace maps in diagram~\ref{trace-maps}, accompanied by ``\emph{vanishing results}'',

348: which state that a complement of the part which is then known to be detected is mapped to zero.

349: For the detection results, we only make assumptions on the coefficient ring $R$, whereas for

350: the vanishing results we additionally need the \emph{Farrell-Jones Conjecture} for $RG$ as an input, compare

351: Example~\ref{exa:FJ}. Modulo the Farrell-Jones Conjecture, we will give a complete answer to

352: Question~\ref{que:main-question} for instance in the case of Hochschild and  cyclic homology, when the coefficient

353: ring $R$ is an algebraic number field $F$ or its ring of integers $\calo_F$. We will also give

354: partial results for  periodic cyclic and negative cyclic homology.

355: 

356: 

357: 

358: All detection results are obtained by using only the Dennis trace with values in $\HH_{\ast}^{\otimes_{k}} ( RG )$,

359: whereas all vanishing results hold even for the trace with values

360: in $\HN_{\ast}^{\otimes_{\IZ}} ( RG )$, which, in view of diagram~\eqref{trace-maps}, can be viewed as the best among

361: the considered trace invariants.

362: (Note that for a $k$-algebra $R$ every homomorphism $k^{\prime} \to k$ of commutative rings leads to a homomorphism

363: $\HN_{\ast}^{\otimes_{k^{\prime}}} ( R ) \to \HN_{\ast}^{\otimes_{k}} ( R )$. Similar for Hochschild, cyclic and periodic

364: cyclic homology.)

365: We have no example where the extra effort that goes into the construction of the variants with values in

366: cyclic, periodic cyclic or negative cyclic homology yields more information about $K_{\ast}(RG)\otimes_{\IZ} \IQ$ than

367: one can obtain by Hochschild homology, see also Remark~\ref{rem:Chern-characters} and \ref{remark-LRRV} below.

368: 

369: We will now explain our main results. We introduce some notation.

370: 

371: \begin{notation}

372: Let $G$ be a group and $H$ a subgroup. We write $\langle g \rangle$ for the cyclic

373: subgroup generated by $g \in G$. We denote by $\conjclass{g}$ and by $(H)$ the conjugacy class of $g$ respectively

374: of $H$ in $G$. Let $\con G$ be the set of conjugacy classes of elements of $G$.

375: The set of conjugacy classes of finite cyclic subgroups of $G$ will be denoted by $( \FCyc )$.

376: 

377: Let $Z_G H$ and $N_G H$

378: denote the centralizer and the normalizer of  $H$ in $G$, respectively.

379: The \emph{Weyl group}

380: $W_G H$ is defined as the quotient $N_G H \big/ H \cdot Z_G H$ and coincides for an abelian subgroup $H$ with

381: $N_G H / Z_G H$.

382: \end{notation}

383: 

384: Let $C$ be a finite cyclic group. We will define in \eqref{theta_C} an idempotent

385: $\theta_C \in A(C) \otimes_{\IZ} \IQ$ in the rationalization of the \emph{Burnside ring} $A(C)$ of $C$.

386: Since there is a natural action of $A(C)$ on $K_{\ast}(RC)$, we obtain a corresponding direct summand

387: $$\theta_C\big(K_{\ast}(RC) \otimes_{\IZ} \IQ\big) \; \subseteq\; K_{\ast}(RC) \otimes_{\IZ} \IQ\,.$$

388: In Lemma~\ref{lem:module-over-burnside}, we prove that $\theta_C ( K_{\ast} (RC ) \otimes_{\IZ} \IQ )$ is isomorphic to

389: the \emph{Artin defect}

390: \begin{eqnarray*}

391: \coker \left( \bigoplus_{D \lneqq C} \ind_D^C \colon \bigoplus_{D \lneqq C}

392: K_{\ast} ( RD ) \otimes_{\IZ} \IQ \to

393: K_{\ast} ( RC ) \otimes_{\IZ} \IQ

394: \right),

395: \end{eqnarray*}

396: which measures the part of $K_{\ast} ( RC ) \otimes_{\IZ} \IQ$ which is not obtained by induction from proper subgroups

397: of $C$.

398: 

399: The conjugation action of $N_GC$ on $C$ induces an action of the Weyl group $W_GC = N_GC/Z_GC$

400: on $K_{\ast} ( RC ) \otimes_{\IZ} \IQ$ and thus on $\theta_C(K_{\ast}(RG) \otimes_{\IZ} \IQ)$.

401: There is an obvious $W_GC$-action on $BZ_GC = Z_GC\backslash EN_GC$. These actions are understood

402: in the following statement.

403: 

404: 

405: \begin{theorem}[Main Detection Result] \label{the:main-detection-result}

406: \mbox{}\\

407: Let $G$ be a group, $k$ a commutative ring and $R$ a $k$-algebra. Suppose that the underlying ring of $R$

408: is from the following list\,:

409: \begin{enumerate}

410: \item a finite dimensional semi\-simple algebra $R$ over a field $F$ of characteristic zero;

411: \item a commutative complete local domain $R$ of characteristic zero;

412: \item a commutative Dedekind domain $R$ in which the order of every finite cyclic subgroup of $G$ is invertible and whose quotient field is an algebraic number field.

413: \end{enumerate}

414: Then there exists an injective homomorphism

415: \begin{eqnarray} \label{main-injective-map}

416: \hspace{3em}

417: \bigoplus_{(C) \in (\FCyc)}

418: H_{\ast}(BZ_GC;\IQ) \otimes_{\IQ[W_GC]} \theta_C \big(  K_0(RC) \otimes_{\IZ} \IQ \big)

419: \to K_{\ast} ( RG ) \otimes_{\IZ } \IQ

420: \end{eqnarray}

421: whose image is detected by the Dennis trace map

422: \begin{eqnarray} \label{Dennis-trace}

423: \dtr \colon K_{\ast} ( RG ) \otimes_{\IZ} \IQ \to \HH^{\otimes_{k}}_{\ast}( RG ) \otimes_{\IZ } \IQ\,,

424: \end{eqnarray}

425: in the sense that the composition of the map \eqref{main-injective-map} with $\dtr$ is injective.

426: Also the composition with the map to $\HC^{\otimes_{k}}_{\ast}( RG ) \otimes_{\IZ } \IQ$ remains

427: injective.

428: \end{theorem}

429: 

430: Examples of rings $R$ appearing in the list of Theorem~\ref{the:main-detection-result} are\,:

431: \begin{itemize}

432: 

433: \item fields of characteristic zero;

434: 

435: \item the group ring $FH$ of a finite group $H$ over a field $F$ of characteristic zero;

436: 

437: \item the ring $ \IZ\widehat{_p}$ of $p$-adic integers;

438: 

439: \item for the given $G$, the localization $S^{-1}\calo_F$ of the ring of integers $\calo_F$ in an

440: algebraic number field $F$, for instance $S^{-1}\IZ$, where $S$ is the multiplicative set generated by the

441: orders of all finite cyclic subgroups of $G$.

442: 

443: \end{itemize}

444: 

445: Depending on the choice of the coefficient ring $R$, the description of the source of the map \ref{main-injective-map}

446: can be simplified. We mention two examples. Let $\IQ_\infty$ be the

447: field obtained from $\IQ$ by adjoining all roots of unity.

448: 

449: 

450: \begin{theorem}[Detection Result for $\IQ$ and $\IC$ as coefficients]\label{the:positive for IQ and IC}

451: \mbox{}\\

452: For every group $G$, there exist injective homomorphisms

453: \begin{eqnarray*}

454: \bigoplus_{(C) \in ( \FCyc )} H_{\ast}( B N_G C;\IQ) & \!\!\!\to\!\!\! & K_{\ast} ( \IQ G ) \otimes_{\IZ } \IQ,

455: \\[.2em]

456: \bigoplus_{\conjclass{g} \in \con G , |g| < \infty} H_{\ast}( B Z_G \langle g \rangle;\IQ_\infty) & \!\!\!\to\!\!\! &

457: K_{\ast} ( \IC G ) \otimes_{\IZ } \IQ_\infty\,.

458: \end{eqnarray*}

459: The image of these maps is detected by the Dennis trace map with $\IQ$ and $\IC$ as ground ring, respectively.

460: The coefficient field $\IQ$ (resp.\ $\IC$) can be replaced by any field of characteristic zero (resp.\ any

461: field containing $\IQ_\infty$).

462: \end{theorem}

463: 

464: Theorem~\ref{the:positive for IQ and IC} for $\IQ_\infty$ and $\IC$ as

465: coefficient fields is the main result of the paper by Matthey~\cite{Matthey(2001)}.

466: The techniques there are based on so-called \emph{delocalization}

467: and the computation of the Hochschild homology and of the cyclic homology of group

468: rings with commutative coefficient rings containing $\IQ$ (see

469: \cite[Section~9.7]{Weibel(1994)} and \cite{Burghelea(1985)}). They

470: are quite different from the ones used in the present paper and are

471: exactly suited for the cases studied there and do not seem to be

472: extendable to the situations considered here. Both maps appearing in

473: Theorem~\ref{the:positive for IQ and IC} are optimal in the

474: sense of Theorem~\ref{the:optimal-result} and of

475: Theorem~\ref{the:Another-optimal-result} below, provided that the

476: Farrell-Jones Conjecture holds rationally for $K_{\ast}(\IQ G)$ and

477: $K_{\ast}(\IC G)$ respectively.

478: 

479: 

480: The Main Detection Theorem~\ref{the:main-detection-result}  is obtained by studying the following commutative diagram\,:

481: \begin{eqnarray} \label{diagram-first-appearance}

482: \vcenter{\xymatrix@C=4.5em{

483: H_{\ast}^G ( \underline{E}G ; \JK R ) \ar[d]_{H_{\ast}^G( \underline{E}G ; \Jdtr )}  \ar[r]^-{\assembly} & K_{\ast} ( RG ) \ar[d]^{\dtr} \\

484: \; H_{\ast}^G ( \underline{E}G ; \JH \JH^{\otimes_{\IZ}} R ) \; \ar[r]^-{\assembly} & \HH_{\ast}^{\otimes_{\IZ}} ( RG )

485: }}

486: \end{eqnarray}

487: Here, the horizontal arrows are \emph{generalized assembly maps} for $K$-theory and Hochschild homology respectively, and

488: the left vertical arrow is a suitable version of the Dennis trace map.

489: The $G$-space $\underline{E}G$ is a model for the so-called \emph{classifying space for proper $G$-actions}.

490: Moreover, $H_{\ast}^G( - ; \JK R)$ and $H_{\ast}^G( - ;  \JH \JH^{\otimes_{\IZ}} R )$ are certain

491: $G$-homology theories. We will explain the diagram in more detail in

492: Section~\ref{sec:Outline-of-the-method}. We will prove that the lower horizontal arrow in \eqref{diagram-first-appearance}

493: is split injective, see Theorem~\ref{the:splitHochschild}.

494: In fact, Theorem~\ref{the:splitHochschild} gives a complete picture of the generalized assembly map for

495: Hochschild and cyclic homology.

496: We will also compute the left-hand vertical arrow after rationalization, compare Theorem~\ref{the:Chern} and

497: Propositions \ref{prop:compute-theta-null}, \ref{prop:comp-Z} and \ref{prop:compute-higher-HH-HC}. According

498: to this computation, the left-hand side in \eqref{main-injective-map} is a direct summand in

499: $H_{\ast}( \underline{E}G ; \JK R ) \otimes_{\IZ} \IQ$ on which, for $R$ as in Theorem~\ref{the:main-detection-result},

500: the map

501: \begin{eqnarray} \label{dtr-homology}

502: H^G_{\ast} ( \underline{E}G ; \JK R ) \otimes_{\IZ} \IQ \to H^G_{\ast} ( \underline{E}G ;

503: \JH \JH^{\otimes_{\IZ}} R ) \otimes_{\IZ} \IQ

504: \end{eqnarray}

505: is injective. This will prove Theorem~\ref{the:main-detection-result}. Now, suppose that $R$ is as in case $(i)$ of

506: Theorem~\ref{the:main-detection-result}, with $F$ a number field. Then, it turns out that the map \eqref{dtr-homology}

507: vanishes on a complementary summand. According to the Farrell-Jones Conjecture for $K_{\ast}(RG)$, the upper

508: horizontal arrow in \eqref{diagram-first-appearance} should be an isomorphism (this uses that  $R$ is a regular

509: ring with $\IQ \subseteq R$). Combining these facts, we will deduce the following result.

510: 

511: \begin{theorem}[Vanishing Result for Hochschild and cyclic homology] \label{the:optimal-result}

512: \mbox{}\\

513: Let $G$ be a group, $F$ an algebraic number field, and $R$ be finite dimensional semi\-simple $F$-algebra.

514: Suppose that for some $n\geq 0$, the Farrell-Jones Conjecture holds rationally for $K_{n}(RG)$, see

515: Example~\ref{exa:FJ} below.

516: 

517: Then Theorem~\ref{the:main-detection-result} is optimal for the Hochschild homology trace invariant, in the sense that

518: the Dennis trace map

519: \begin{eqnarray} \label{dennis-with-z-coefficients}

520: \dtr\colon K_{n} ( RG ) \otimes_{\IZ} \IQ \to \HH^{\otimes_{\IZ}}_{n} ( RG )\otimes_{\IZ} \IQ

521: \end{eqnarray}

522: vanishes on a direct summand that is complementary to the image of the injective map \eqref{main-injective-map}

523: in degree $n$. Consequently, also the trace taking values in rationalized cyclic homology

524: $\HC^{\otimes_{\IZ}}_{n} ( RG ) \otimes_{\IZ} \IQ$ vanishes on this complementary summand.

525: \end{theorem}

526: 

527: One might still hope that the refinements of the Dennis trace map with values in periodic cyclic or negative

528: cyclic homology detect more of the rationalized algebraic $K$-theory of $RG$.

529: But one can show that this is not the case if one additionally assumes

530: a finiteness condition on the classifying space $\underline{E}G$.

531: Recall that the $G$-space $\underline{E}G$ is called \emph{cocompact} if the orbit space

532: $G\backslash \underline{E}G$ is compact, in other words, if it consists of finitely many $G$-equivariant cells.

533: Cocompact models for $\underline{E}G$ exist for many interesting groups

534: $G$ such as discrete cocompact subgroups of virtually connected Lie groups, word-hyperbolic groups, arithmetic subgroups of

535: a semi\-simple connected $\IQ$-algebraic group, and

536: mapping class groups (see for instance \cite{Lueck(2004h)}).

537: 

538: 

539: \begin{theorem}[Vanishing Result for periodic and negative cyclic homology] \label{the:Another-optimal-result}

540: \mbox{}\\

541: Let $F$ be an algebraic number field, and $R$ a finite dimensional semi\-simple $F$-algebra.

542: Suppose that for some $n\geq 0$, the Farrell-Jones Conjecture holds rationally for $K_{n} ( RG )$. Suppose

543: further that there exists a cocompact model for the classifying space for proper $G$-actions \underline{E}G.

544: 

545: Then also the refinements of the Dennis trace with values in $\HP^{\otimes_{\IZ}}_{n} ( RG ) \otimes_{\IZ} \IQ$ and

546: in $\HN^{\otimes_{\IZ}}_{n} ( RG ) \otimes_{\IZ} \IQ$ vanish on a direct summand which is complementary to the image

547: of the injective map \eqref{main-injective-map} in degree $n$.

548: \end{theorem}

549: 

550: 

551: The next result is well-known. It shows in particular that the rational group homology

552: $H_{\ast}(BG;\IQ)$ is contained in $K_{\ast}(RG) \otimes_{\IZ} \IQ$ for all commutative rings $R$ of characteristic

553: zero.

554: 

555: \begin{theorem}[Detection Result for commutative rings of characteristic zero]

556: \label{the:detection-result-for-comm.-rings-of-char.-zero}

557: \mbox{}\\

558: Let $R$ be a ring such that the canonical ring homomorphism $\IZ \to R$ induces an

559: injection $\HH_0^{\otimes_{\IZ}}(\IZ) = \IZ \into \HH_0^{\otimes_{\IZ}}(R) = R\big/[R,R] $, for instance a

560: commutative ring of characteristic zero.

561: 

562: Then, for any group $G$, there exists an injective homomorphism

563: \begin{eqnarray} \label{injective-for-z-coefficients}

564: H_{\ast} ( BG ; \IQ ) \to K_{\ast} (RG) \otimes_{\IZ } \IQ

565: \end{eqnarray}

566: whose composition with the Dennis trace map \eqref{Dennis-trace}

567: is injective for every choice of a ground ring $k$ such that $R$ is a $k$-algebra.

568: The corresponding statement holds with Hochschild homology replaced by cyclic homology.

569: \end{theorem}

570: 

571: Special cases of this result are treated for example in \cite[Proposition~6.3.24 on page~366]{Rosenberg(1994)}.

572: 

573: 

574: According to the Farrell-Jones Conjecture, the image of \eqref{injective-for-z-coefficients} should only be a

575: very small part of the rationalized $K$-theory of $RG$.

576: The following result illustrates that, for certain coefficient rings, including $\IZ$, one cannot expect to detect

577: more by linear traces than achieved in Theorem~\ref{the:detection-result-for-comm.-rings-of-char.-zero}.

578: 

579: \begin{theorem}[Vanishing Result for integral coefficients]  \label{the:optimal-Z-coeff}

580: \mbox{}\\

581: Let $S^{-1}\calo_F$ be a localization of a ring of integers $\calo_F$  in an algebraic number field $F$

582: with respect to a (possibly empty) multiplicatively closed subset $S$.

583: Assume that no prime divisor of the order $|H|$ of a nontrivial finite subgroup $H$ of $G$ is invertible in $S^{-1} \calo_F$.

584: Suppose that for some $n\geq 0$, the Farrell-Jones Conjecture holds rationally for

585: $K_n(S^{-1} \calo_F[G])$.

586: 

587: Then the Dennis trace \eqref{dennis-with-z-coefficients} vanishes on a summand in

588: $K_n (S^{-1} \calo_F[G])\otimes_{\IZ} \IQ$

589: which is complementary to the image of

590: the map \eqref{injective-for-z-coefficients} in degree $n$.

591: Consequently, the analogous statement holds for the trace with values in

592: $\HC^{\otimes_{\IZ}}_{n} (S^{-1} \calo_F[G]) \otimes_{\IZ} \IQ$.

593: \end{theorem}

594: 

595: The most interesting case in Theorem~\ref{the:optimal-Z-coeff} is $R = \IZ$.

596: We remark that rationally, the Farrell-Jones Conjecture for $K_{\ast}( \IZ G )$

597: is known in many cases, for example for every subgroup $G$ of a discrete cocompact subgroup of a

598: virtually connected Lie group \cite{Farrell-Jones(1993a)}. For a survey of known results about the

599: Farrell-Jones Conjecture, we refer the reader to \cite{Lueck-Reich(2003b)}.

600: 

601: \begin{remark} \label{rem:Chern-characters}

602: There are further trace invariants (or Chern characters) given by maps $\ch_{n,r}\colon K_{n}

603: ( RG ) \to \HC^{\otimes_{k}}_{n+2r} ( RG )$, for fixed $n$, $r \geq 0$, see~\cite[8.4.6 on page 272 and 11.4.3 on page 371]{Loday(1992)}. This

604: will however produce no new detection results in the spirit of the above statements, since there is a commutative

605: diagram

606: \[

607: \xymatrix@R=3em@C=.5em{

608: & & & & K_{n} ( RG ) \ar[dllll]_{\ntr} \ar[dl] \ar[dr]_(.4){\ch_{n,r}\!\!\!\!\!} \ar[drrrr] & & & & \\

609: \HN_{n}^{\otimes_k} ( RG ) \ar[rrr] & & & \HP_{n}^{\otimes_k} ( RG ) \ar[rr]^-{\overline{S}} & & \HC_{n+2r}^{\otimes_k} ( RG )

610: \ar[rrr]^-{S^{r}} & & & \HC_{n}^{\otimes_k} ( RG ).

611:          }

612: \]

613: \end{remark}

614: 

615: \begin{remark} \label{remark-LRRV}

616: In \cite{Boekstedt-Hsiang-Madsen(1993)}, B{\"o}kstedt, Hsiang and Madsen

617: define the \emph{cyclotomic trace}, a map out of $K$-theory which takes values in

618: \emph{topological cyclic homology}. The cyclotomic trace map can be thought of as an even more elaborate refinement of the

619: Dennis trace map. In contrast to the Dennis trace, it seems that the cyclotomic trace has the potential to detect

620: almost all of the rationalized $K$-theory of an integral group ring. This question is investigated in detail in

621: \cite{Lueck-Reich-Rognes-Varisco(2005)}.

622: \end{remark}

623: 

624: 

625: %\begin{remark}  \label{remark-about-R-not-free-over-k}

626: %%By $\HH^{\otimes_k}( R )$ we denote the Hochschild homology defined via the standard

627: %Hochschild complex using tensor products over $k$ (compare \cite[Chapter~1]{Loday(1992)}).

628: %We do not assume that $R$ (or $RG$) is flat

629: %over $k$. Note that if $R$ is not flat over $k$ then depending on the point of view this may be a

630: %too ``naive'' definition of Hochschild homology, because it does not have the more conceptional interpretation

631: %as $\Tor_k^{R\otimes_k R^{\op}}( R ,R )$, compare \cite[Proposition~1.1.13 on page~12]{Loday(1992)}.

632: %This defect can however be repaired using the framework of relative homological algebra, compare

633: %\cite[Lemma~9.1.3 on page~302]{Weibel(1994)}.

634: %\end{remark}

635: 

636: 

637: \medskip

638: 

639: The paper is organized as follows\,:

640: 

641: \smallskip

642: 

643: \begin{tabular}{ll}

644: \ref{sec:Outline-of-the-method}. & Outline of the method

645: \\

646: \ref{sec:Proofs}. & Proofs

647: \\

648: \ref{sec:trace-for-finite-cyclic}. & The trace maps for finite cyclic groups

649: \\

650: \ref{sec:Notation}. & Notation and general machinery

651: \\

652: \ref{sec:trace-maps}. & The trace maps

653: \\

654: \ref{sec:mackey}. & Equivariant homology theories, induction and Mackey structures

655: \\

656: \ref{sec:eval-chern}. & Evaluating the equivariant Chern character

657: \\

658: \ref{sec:comparing-different-models}. & Comparing different models

659: \\

660: \ref{sec:Splitting-assembly-maps}. & Splitting assembly maps

661: \\

662:  & References

663: \end{tabular}

664: 

665: 

666: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

667: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

668: \section{Outline of the method}\label{sec:Outline-of-the-method}

669: 

670: This paper is concerned with comparing generalized assembly maps for $K$-theory, via the Dennis trace or its refinements,

671: with generalized assembly maps for Hochschild homology, for cyclic, periodic cyclic or negative cyclic homology.

672: Before we explain the general strategy behind our results we briefly explain the concept of a generalized assembly map; for more details the reader is referred

673: to \cite{Davis-Lueck(1998)} and \cite[Section~2 and 6]{Lueck-Reich(2003b)}.

674: 

675: 

676: A \emph{family of subgroups} of a given group $G$ is a non-empty collection of subgroups which is closed under

677: conjugation and finite intersections. Given a family $\calf$ of subgroups, there always exists a $G$-$CW$-complex

678: $\EGF{G}{\calf}$ all of whose isotropy groups lie in $\calf$ and which has the property that for all $H \in \calf$, the

679: fixed subspace $\EGF{G}{\calf}^H$ is a contractible space. A $G$-$CW$-complex with these properties is unique up to

680: $G$-homotopy because it receives a $G$-map from every $G$-$CW$-complex all whose isotropy groups lie in $\calf$ and this $G$-map

681: is unique up to $G$-homotopy. If $\calf= \Fin$ is the \emph{family of finite subgroups}, then one often writes $\underline{E}G$

682: for $\EGF{G}{\Fin}$. For a survey on these spaces, see for instance \cite{Lueck(2004h)}.

683: 

684: %We also write $(\calf)$ for the

685: %\emph{set of conjugacy classes $(H)$ of subgroups $H\in\calf$}.

686: 

687: Let $\Or G$ denote the \emph{orbit category} of $G$. Objects are the homogenous spaces $G/H$ considered as left $G$-spaces, and

688: morphisms are $G$-maps. A functor $\JE$, from the orbit category $\Or G$ to the category of spectra, is called an

689: \emph{$\Or G$-spectrum}. Each $\Or G$-spectrum $\JE$ gives rise to a

690: \emph{$G$-homology theory} $H^G_{\ast}( - ; \JE )$, compare~\cite[Section~6]{Lueck-Reich(2003b)} and the beginning of

691: Section~\ref{sec:mackey} below. Given $\JE$ and a family $\calf$ of subgroups of $G$, the so-called  \emph{generalized assembly map}

692: \begin{eqnarray} \label{assembly-map}

693: \xymatrix@C=4em{

694: H^G_{\ast} \big(\EGF{G}{\calf}; \JE \big) \ar[r]^-{\assembly} & H^G_{\ast}( \pt ; \JE )

695:                }

696: \end{eqnarray}

697: is merely the homomorphism induced by the map $\EGF{G}{\calf} \to \pt$. The group

698: $H_{\ast}^G( \pt ; \JE)$ can be canonically identified with $\pi_{\ast}( \JE( G/G ))$.

699: 

700: \begin{example}[The Farrell-Jones Conjecture] \label{exa:FJ}

701: \mbox{}\\

702: Given an arbitrary ring $R$ and an arbitrary group $G$, there exists a \emph{non-connective $K$-theory $\Or G$-spectrum},

703: denoted by $\JK^{\minusinfinity}R(?)$, such that there is a natural isomorphism

704: $$

705: \pi_n \big(\JK^{\minusinfinity}R ( G/H ) \big) \cong K_n ( RH )

706: $$

707: for all $H\leq G$ and all $n \in \IZ$, compare \cite[Theorem~6.9]{Lueck-Reich(2003b)}.

708: The \emph{Farrell-Jones Conjecture} for $K_{n}(RG)$, \cite[1.6 on page~257]{Farrell-Jones(1993a)},

709: predicts that the generalized assembly map

710: \[

711: \xymatrix@C=4em{

712: H_n^G \big(\EGF{G}{\VCyc};\JK^{\minusinfinity} R \big) \ar[r]^-{\assembly} &

713: H_n^G(\pt;\JK^{\minusinfinity} R) \cong K_n(RG)

714:                  }

715: \]

716: is an isomorphism. Here $\VCyc$ stands for the \emph{family of all virtually cyclic subgroups} of~$G$.

717: A group is called virtually cyclic if it contains a cyclic subgroup of finite index.

718: 

719: %We also point out that there is a map of $\Or G$-spectra $\JK R(?) \to \JK^{\minusinfinity} R(?)$

720: %(the $(-1)$-connective covering map), which is natural in $R$, and which induces an isomorphism

721: %\[

722: %\pi_{n}\big(\JK R ( G/H ) \big) \xrightarrow{\cong} \pi_{n}\big(\JK^{\minusinfinity} R ( G/H ) \big)\,,

723: %\]

724: %that corresponds to the identity of $K_n(RH)$, for every $H\leq G$ and every $n\geq 0$.

725: \end{example}

726: 

727: In Section~\ref{sec:trace-maps}, we will construct the following commutative diagram of connective

728: $\Or G$-spectra and maps (alias natural transformations) between them\,:

729: \begin{eqnarray} \label{spec-level-trace-maps}

730: \vcenter{\xymatrix@C=3.2em{

731: & \JH \JN^{\otimes_k} R \ar[r] \ar[d]^{\Jh} & \JH \JP^{\otimes_k} R \ar[d]  \\

732: \JK R \ar[r]^-{\Jdtr} \ar[ur]^-{\Jntr} & \JH \JH^{\otimes_k} R \ar[r] & \JH \JC^{\otimes_k} R.

733:         }}

734: \end{eqnarray}

735: Decisive properties of these constructions are that for all $n \geq 0$, we have natural isomorphisms

736: \begin{eqnarray} \label{relate-to-classical}

737: \renewcommand{\arraystretch}{1.4}

738: \begin{array}{lcl}

739: \pi_n\big(\JK R ( G/H )\big) & \cong  & K_n ( RH )

740: \\

741: \pi_n\big( \JH \JH^{\otimes_k } R ( G/H ) \big) & \cong & \HH_n^{\otimes_k} ( RH )

742: \\

743:  \pi_n\big( \JH \JC^{\otimes_k } R ( G/H ) \big) & \cong & \HC_n^{\otimes_k} ( RH )

744: \\

745: \pi_n\big( \JH \JP^{\otimes_k } R ( G/H ) \big)  & \cong  &

746: \HP_n^{\otimes_k} ( RH )

747: \\

748: \pi_n\big( \JH \JN^{\otimes_k } R ( G/H ) \big)  & \cong  &

749: \HN_n^{\otimes_k} ( RH )

750: \end{array}

751: \end{eqnarray}

752: and all negative homotopy groups vanish. Note that we need to distinguish between the non-connective version

753: $\JK^{-\infty}R$ and the connective version $\JK R$.

754: Under the identifications above, the maps of $\Or G$-spectra in \eqref{spec-level-trace-maps}

755: evaluated at an orbit $G/H$ induce, on the level of homotopy groups, the maps in \eqref{trace-maps}

756: with $R$ replaced by the corresponding group ring $RH$.

757: 

758: \begin{remark} We found it technically convenient to work, at the level of spectra, with the connective versions

759: of periodic cyclic and negative cyclic homology. Since we are mainly interested in the trace maps (whose source

760: will be the connective $K$-theory spectrum), we do not lose any information.

761: \end{remark}

762: 

763: Since the assembly map \eqref{assembly-map} is natural in the functor $\JE$, we obtain, for each family of

764: subgroups $\calf$ of a group $G$ and for each $n\geq 0$, the commutative diagram

765: \begin{eqnarray} \label{diagram-assembly-dtr}

766: \mbox{\phantom{xxxxxx}}\vcenter{\xymatrix@C=4.5em{

767: H_n^G \big( \EGF{G}{\calf};\JK R \big) \ar[r]^-{\assembly} \ar[d]_{H_n^G( \EGF{G}{\calf} ; \Jntr)} &

768: H_n^G (\pt;\JK R) \cong K_n(RG) \ar@<-2.5em>[d]^{\ntr} \\

769: H_n^G \big( \EGF{G}{\calf}; \JH \JN ^{\otimes_k} R \big) \ar[r]^-{\assembly} \ar[d]_{H_n^G( \EGF{G}{\calf} ; \Jh)} &

770: H_n^G( \pt ; \JH \JN^{\otimes_k} R ) \cong \HN_n^{\otimes_k} ( RG ) \ar@<-2.5em>[d]^{\h} \\

771: H_n^G \big( \EGF{G}{\calf};\JH \JH^{\otimes_k}R \big) \ar[r]^-{\assembly} &

772: H_n^G(\pt;\JH \JH^{\otimes_k}R) \cong \HH_n^{\otimes_k}(RG).

773:          }}

774: \end{eqnarray}

775: The vertical compositions are the corresponding versions of the Dennis trace map.

776: 

777: \smallskip

778: 

779: Our investigation relies on two main ingredients. The first ingredient are splitting and isomorphism

780: results for the assembly maps of Hochschild and cyclic type.

781: 

782: \begin{theorem} [The Isomorphism Conjecture for $\HH$ and $\HC$] \label{the:splitHochschild}

783: \mbox{}\\

784: Let $k$ be a commutative ring, $R$ a $k$-algebra, and $G$ a group. Then the generalized Hochschild homology

785: assembly map

786: \[

787: \xymatrix@C=5em{

788: \;H_{\ast} \big( \EGF{G}{\calf}; \JH \JH^{\otimes_k} R \big)\; \ar[r]^-{\assembly} &

789: H_{\ast} ( \pt ; \JH \JH^{\otimes_k} R )\cong \HH_{\ast}^{\otimes_k}(RG)

790:                }

791: \]

792: is split injective for every family $\calf$.

793: If $\calf$ contains the family of all (finite and infinite) cyclic subgroups, then

794: the map is an isomorphism. The analogous statement holds for $\JH \JC$ in

795: place of $\JH \JH$.

796: \end{theorem}

797: 

798: The fact that the definition of periodic cyclic and of negative cyclic homology involves certain inverse limit processes

799: prevents us from proving the analogous result in

800: these cases without assumptions on the group $G$. But we still have the following statement.

801: 

802: 

803: \begin{addendum} [Splitting Results for the $\HP$ and $\HN$-assembly maps] \label{add:split-HN}

804: \mbox{}\\

805: Suppose that there exists a cocompact  model for the classifying space $\EGF{G}{\calf}$. Then the statement of

806: Theorem~\ref{the:splitHochschild} also holds for $\JH \JP$ and $\JH \JN$ in place of $\JH \JH$.

807: \end{addendum}

808: 

809: The proofs of Theorem~\ref{the:splitHochschild} and Addendum~\ref{add:split-HN} are presented in

810: Section~\ref{sec:Splitting-assembly-maps}.

811: 

812: \begin{remark}

813: We do not know any non-trivial example where the isomorphism statement in Addendum~\ref{add:split-HN}

814: applies, i.e.\ where $\calf$ contains all (finite and infinite) cyclic groups and where, at the same

815: time, $\EGF{G}{\calf}$ has a cocompact model.

816: \end{remark}

817: 

818: The second main ingredient of our investigation is the rational computation

819: of equivariant homology theories from \cite{Lueck(2002b)}.

820: For varying $G$, our $G$-homology theories

821: like $\calh^G_\ast(-)=H_\ast^G( - ; \JK R )$ or $\calh^G_\ast(-)=H_\ast^G(- ; \JH \JH^{\otimes_k} R)$

822: are linked by a so-called \emph{induction structure} and form an \emph{equivariant homology theory} in the sense of

823: \cite{Lueck(2002b)}. Moreover, these homology theories admit a \emph{Mackey structure}.

824: In Section~\ref{sec:mackey}, we review these notions and explain some general principles which allow us to

825: verify that $G$-homology theories like the ones we are interested in indeed admit induction and Mackey structures.

826: In particular, Theorems~0.1 and 0.2 in \cite{Lueck(2002b)} apply and yield an explicit computation of

827: $\calh^G_\ast ( \underline{E}G ) \otimes_{\IZ} \IQ$. In Section~\ref{sec:eval-chern}, we review this

828: computation and discuss a simplification which occurs in the case of $K$-theory, Hochschild,

829: cyclic, periodic cyclic and negative cyclic homology, due to the fact that in all these special cases,

830: we have additionally a module structure over the Swan ring.

831: 

832: 

833: In order to state the result of this computation, we introduce some more notation. For a finite group $G$, we

834: denote by $A(G)$ the \emph{Burnside ring} which is additively generated by isomorphism classes of finite transitive

835: $G$-sets.

836: Let $( \sub G )$ denote the set of conjugacy classes of subgroups of $G$.

837: 

838: The counting fixpoints ring homomorphism

839: \begin{eqnarray} \label{def-chi-G}

840: \chi_G \colon A(G) \to \prod_{(\sub G)} \IZ\,

841: \end{eqnarray}

842: which is induced by sending  a $G$-set $S$ to $( |S^H| )_{(H)}$ becomes

843: an isomorphism after rationalization,

844: compare \cite[page~19]{Dieck(1987)}. For a finite cyclic group $C$, we consider the idempotent

845: \begin{eqnarray} \label{theta_C}

846: \theta_C \; = \; (\chi_C \otimes_{\IZ} \IQ)^{-1} \big((\delta_{CD})_{D}\big) & \in & A(C) \otimes_{\IZ} \IQ\,,

847: \end{eqnarray}

848: where $(\delta_{CD})_D\in \prod_{\sub C} \IQ$ is given by $\delta_{CC}=1$ and $\delta_{CD}= 0$ if $D \neq C$.

849: 

850: \smallskip

851: 

852: Recall that $K_{\ast} ( RC )$ and similarly Hochschild, cyclic, periodic cyclic and negative cyclic homology

853: of $RC$ are modules over the Burnside ring $A(C)$. The action of a $C$-set $S$ is in all cases induced from taking

854: the tensor product over $\IZ$ with the corresponding permutation module $\IZ S$.

855: In Lemma~\ref{lem:module-over-burnside} below, we prove that $\theta_C ( K_{\ast} ( RC ) \otimes_{\IZ} \IQ )$

856: is isomorphic to the $\IQ$-vector space

857: \begin{eqnarray} \label{eq:SC-ist-Artin-defekt}

858: \hspace*{3em}

859: \coker \left( \bigoplus_{D \lneqq C} \ind_D^C \colon \bigoplus_{D \lneqq C}

860: K_{\ast} ( RD ) \otimes_{\IZ} \IQ \to

861: K_{\ast} ( RC ) \otimes_{\IZ} \IQ

862: \right),

863: \end{eqnarray}

864: which is known as the \emph{Artin defect} of $K_{\ast} ( RC ) \otimes_{\IZ} \IQ$.

865: 

866: In Section~\ref{sec:eval-chern} we establish the following result.

867: 

868: \begin{theorem} \label{the:Chern}

869: For each $n\geq 0$, the following

870: diagram commutes and the arrows labelled $\ch^G$ are isomorphisms\,:

871: \begin{eqnarray*}

872: \xymatrix@C=-10mm{

873: \smash[b]{{\displaystyle {\bigoplus_{\substack{p,q \geq 0\\p+q = n}}}\; \displaystyle

874: \bigoplus_{(C) \in (\FCyc)}}}

875: H_p(BZ_GC;\IQ) \otimes_{\IQ[W_GC]} \theta_C \big(  K_q (RC) \otimes_{\IZ} \IQ \big)

876: \ar[dr]^-{\ch^G}_-{\cong}

877: \ar[dd]_{\dtr_\ast}%_-{\id \otimes_{\IQ[W_GC]} \theta_C (\id \otimes_{\IZ} \dtr_q)}

878: & \\

879: & H_n^G(\underline{E}G;\JK R ) \otimes_{\IZ} \IQ

880:  \ar[dd]^{H_n^G( \underline{E}G ; \Jdtr) \otimes_{\IZ} \IQ} %_-{\id \otimes_{\IZ} H_n^G\underline{E}G; \dtr)}

881: \\

882: \smash[b]{{\displaystyle

883: {\bigoplus_{\substack{p,q \geq 0\\p+q = n}}}\; \displaystyle \bigoplus_{(C) \in (\FCyc )}}}

884: H_p(BZ_GC;\IQ) \otimes_{\IQ[W_GC]} \theta_C \big(  \HH_q^{\otimes_k}(RC) \otimes_{\IZ} \IQ \big)

885: \ar[dr]^-{\ch^G}_-{\cong}

886: & \\

887: & H_n^G(\underline{E}G;\JH \JH^{\otimes_k} R) \otimes_{\IZ} \IQ.

888: }

889: \end{eqnarray*}

890: The left-hand vertical arrow is induced by the Dennis trace maps for finite cyclic groups and

891: respects the double direct sum decompositions. The right-hand vertical arrow is induced by

892: the $\Or G$-spectrum Dennis trace $\Jdtr$, compare \eqref{spec-level-trace-maps}.

893: There are similar diagrams and isomorphisms

894: corresponding to each of the other maps in diagram \eqref{spec-level-trace-maps}.

895: \end{theorem}

896: 

897: 

898: \begin{remark} \label{remark-nonconn-conn}

899: The $(-1)$-connected covering map of $\Or G$-spectra $\JK R \to \JK^{-\infty} R$

900: induces for every orbit $G/H$ an isomorphism

901: \[

902: \pi_n ( \JK R ( G/H) ) \to \pi_n ( \JK^{-\infty} R ( G/H) )

903: \]

904: if $n \geq 0$. The source is trivial for $n < 0$. This map induces

905: the following commutative diagram.

906: \begin{eqnarray*}

907: \xymatrix@C=-10mm{

908: \smash[b]{{\displaystyle {\bigoplus_{\substack{p,q \geq 0\\p+q = n}}}\; \displaystyle

909: \bigoplus_{(C) \in (\FCyc)}}}

910: H_p(BZ_GC;\IQ) \otimes_{\IQ[W_GC]} \theta_C \big(  K_q (RC) \otimes_{\IZ} \IQ \big)

911: \ar[dr]^-{\ch^G}_-{\cong}

912: \ar[dd]%_-{\id \otimes_{\IQ[W_GC]} \theta_C (\id \otimes_{\IZ} \dtr_q)}

913: & \\

914: & H_n^G(\underline{E}G;\JK R ) \otimes_{\IZ} \IQ

915:  \ar[dd]%_-{\id \otimes_{\IZ} H_n^G\underline{E}G; \dtr)}

916: \\

917: \smash[b]{{\displaystyle

918: {\bigoplus_{\substack{p,q \in \IZ \\p+q = n}}}\; \displaystyle \bigoplus_{(C) \in (\FCyc )}}}

919: H_p(BZ_GC;\IQ) \otimes_{\IQ[W_GC]} \theta_C \big(  K_q (RC) \otimes_{\IZ} \IQ \big)

920: \ar[dr]^-{\ch^G}_-{\cong}

921: & \\

922: & H_n^G(\underline{E}G;\JK^{\minusinfinity} R) \otimes_{\IZ} \IQ.

923: }

924: \end{eqnarray*}

925: Here the arrows

926: labelled $\ch^G$ are isomorphisms.

927: Note the restriction $p$, $q \geq 0$ for the sum in the upper left hand corner.

928: \end{remark}

929: 

930: 

931: 

932: 

933: \subsection{} \label{subsec:general-strategy}

934: {\bf General strategy.}

935: We now explain the strategy behind all the results that appeared in the introduction.

936: If we combine the diagram appearing in Theorem~\ref{the:Chern} with diagram \eqref{diagram-assembly-dtr},

937: for each $n\geq 0$, we get a commutative diagram

938: \begin{eqnarray*} \label{diagram3}

939: \xymatrix@C=-1mm{

940: \smash[b]{{\displaystyle {\bigoplus_{\substack{p,q \geq 0\\p+q = n}}}\; \displaystyle

941: \bigoplus_{(C) \in (\FCyc)}}}

942: H_p(BZ_GC;\IQ) \otimes_{\IQ[W_GC]} \theta_C \big(  K_q (RC) \otimes_{\IZ} \IQ \big)

943: \ar[dr]%^{\assembly\circ\ch^{G}}

944: \ar[dd]_{\dtr_\ast}%_-{\id \otimes_{\IQ[W_GC]} \theta_C (\id \otimes_{\IZ} \dtr_q)}

945: & \\

946: & K_n(RG) \otimes_{\IZ} \IQ

947:  \ar[dd]^{\dtr}

948: \\

949: \smash[b]{{\displaystyle

950: {\bigoplus_{\substack{p,q \geq 0\\p+q = n}}}\; \displaystyle \bigoplus_{(C) \in (\FCyc )}}}

951: H_p(BZ_GC;\IQ) \otimes_{\IQ[W_GC]} \theta_C \big(  \HH_q^{\otimes_\IZ}(RC) \otimes_{\IZ} \IQ \big)

952: \ar[dr]

953: & \\

954: & \;\HH_n^{\otimes_{\IZ}}(RG) \otimes_{\IZ} \IQ.\;

955: }

956: \end{eqnarray*}

957: Because of Theorem~\ref{the:splitHochschild} and the isomorphism statement in Theorem~\ref{the:Chern} the lower horizontal map

958: is injective.

959: There is an analogue of the commutative diagram above, where

960: the upper row is the same and $\HH$ is replaced by $\HC$ in the bottom row. Also in this case we know that the lower horizontal map

961: is injective because of

962: Theorem~\ref{the:splitHochschild} and \ref{the:Chern}.

963: 

964: \smallskip

965: 

966: Observe that $W_G C$ is always a finite group, hence $\IQ[W_G C]$ is a semi\-simple ring, so that every module

967: over $\IQ W_G C$ is flat and the functor $H_p( B Z_G C ; \IQ) \otimes_{\IQ W_G C} ( - )$

968: preserves injectivity.

969: 

970: For $q\geq 0$ given, we see that suitable injectivity results

971: about the maps

972: \begin{eqnarray} \label{eq:the-map-theta-C}

973: \theta_C \big( K_q( RC ) \otimes_{\IZ} \IQ \big)  \to \theta_C \big( \HH_q^{\otimes_k} ( RC ) \otimes_{\IZ} \IQ \big)

974: \end{eqnarray}

975: for the finite cyclic subgroups $C\leq G$ lead to the proof

976: of detection results in degree $n$. These maps \eqref{eq:the-map-theta-C} will be studied in

977: Section~\ref{sec:trace-for-finite-cyclic}.

978: 

979: If $R$ is a regular ring containing $\IQ$, then the family $\VCyc$ of virtually cyclic subgroups

980: can be replaced by the family $\Fin$ of finite subgroups and the non-connective $K$-theory $\Or G$-spectrum

981: $\JK^{\minusinfinity} R(?)$ by its connective version $\JK R(?)$ in the statement of the Farrell-Jones Conjecture,

982: i.e., in this case, the Farrell-Jones Conjecture for $K_n(RG)$, for some $n\in\IZ$, is equivalent to the statement

983: that the assembly map

984: $$

985: \xymatrix@C=4em{

986: H_n^G(\underline{E}G; \JK R) \ar[r]^-{\assembly} & K_n(RG)

987:                }

988: $$

989: is an isomorphism if $n \geq 0$ and to the statement that $K_n(RG) = 0$ if $n \leq -1$

990: (see \cite[Proposition~2.14]{Lueck-Reich(2003b)}). As a consequence, the upper horizontal

991: arrow in the diagram above (where $n\geq 0$) is bijective if the Farrell-Jones Conjecture

992: is true rationally for $K_n(RG)$.

993: 

994: So for $q\geq 0$ given, we see that suitable vanishing results

995: about the maps \eqref{eq:the-map-theta-C}

996: (and about their analogues involving cyclic homology)

997: combined with the assumption that the Farrell-Jones conjecture holds rationally for $K_n(RG)$ lead to the proof of

998: vanishing results in degree $n$.

999: 

1000: \section{Proofs} \label{sec:Proofs}

1001: 

1002: 

1003: Based on the strategy explained in the previous paragraphs we now give the proofs

1004: of the theorems stated in the introduction,

1005: modulo the following results\,: Theorem~\ref{the:splitHochschild} and Addendum~\ref{add:split-HN}

1006: (proved in Section~\ref{sec:Splitting-assembly-maps}); Theorems \ref{the:Chern} (proved in Section~\ref{sec:eval-chern}, using Sections

1007: \ref{sec:Notation}--\ref{sec:mackey});

1008: and the results of

1009: Section~\ref{sec:trace-for-finite-cyclic} (which is self-contained, except for Lemma~\ref{lem:module-over-burnside}

1010: whose proof is independent of the rest of the paper).

1011: 

1012: 

1013: \subsection{}

1014: {\bf Proof of Theorem~\ref{the:main-detection-result}.}

1015: After the general strategy~\ref{subsec:general-strategy}, the necessary injectivity result to complete the proof

1016: appears in Proposition~\ref{prop:compute-theta-null} below.

1017: \qed

1018: 

1019: 

1020: \subsection{}

1021: {\bf Proof of Theorem~\ref{the:optimal-result}.}

1022: The result follows directly from the general strategy~\ref{subsec:general-strategy} and the vanishing result

1023: stated as Proposition~\ref{prop:compute-higher-HH-HC} below.

1024: \qed

1025: 

1026: \subsection{}

1027: {\bf Proof of Theorem~\ref{the:Another-optimal-result}.}

1028: The proof is completely analogous to that of Theorem~\ref{the:optimal-result}. The extra condition that there

1029: is a cocompact model for $\underline{E}G$ is only needed to apply Addendum~\ref{add:split-HN} in place of

1030: Theorem~\ref{the:splitHochschild}.

1031: \qed

1032: 

1033: \subsection{}

1034: {\bf Proof of Theorem~\ref{the:positive for IQ and IC}.}

1035: The next lemma explains why Theorem~\ref{the:positive for IQ and IC} for $\IQ$ as coefficients

1036: follows from Theorem~\ref{the:main-detection-result}. The case of $\IC$ as coefficients is proven

1037: similarly, compare \cite[Example~8.11]{Lueck(2002b)}.

1038: \qed

1039: 

1040: \begin{lemma} \label{lem:compute-theta}

1041: \begin{enumerate}

1042: \item \label{lem:compute-theta-i} Let $C$ be a finite cyclic group. Then one has

1043: \[

1044: \theta_C \big( K_0 ( \IQ C ) \otimes_{\IZ} \IQ \big) \; \cong \; \IQ

1045: \]

1046: and every group automorphism of $C$ induces the identity on $\IQ$.

1047: 

1048: \item  \label{lem:compute-theta-ii} For any group $G$ and finite cyclic subgroup $C \leq G$, the map

1049: $$H_{\ast}(BZ_GC;\IQ) \otimes_{\IQ[W_GC]} \IQ \xrightarrow{\cong} H_{\ast}(BN_GC;\IQ)$$

1050: induced by the inclusion $Z_GC \into N_GC$ is an isomorphism. Here $\IQ$ carries the trivial $W_GC$-action.

1051: \end{enumerate}

1052: \end{lemma}

1053: \begin{proof}

1054: \ref{lem:compute-theta-i}

1055: There is a commutative diagram

1056: \[

1057: \xymatrix{

1058: A(C) \otimes_{\IZ} \IQ \ar[d]_{\chi_C \otimes_{\IZ} \IQ}^-{\cong} \ar[r]^-{\cong} & K_0 ( \IQ C )

1059: \otimes_{\IZ} \IQ \ar[d]^-{\cong} \\

1060: \prod_{D \in \sub C} \IQ \ar[r]^-{\cong} & \map( \sub C , \IQ )

1061:          }

1062: \]

1063: Here, the upper horizontal map sends a $C$-set to the corresponding permutation module.

1064: The product in the lower left corner is taken

1065: over the set $\sub C$ of all subgroups of $C$ and the left-hand vertical arrow is given by

1066: sending the class of a $C$-set $S$ to $(|S^D|)_D$ and is an isomorphism, as already mentioned after \eqref{def-chi-G}.

1067: The right-hand vertical map is given by sending a rational representation $V$ to its character, i.e.\ if $d$ generates the subgroup $\langle d \rangle$, then

1068: $\langle d \rangle \mapsto \tr_{\IQ} (d \colon V \to V)$. This map is also an isomorphism, compare

1069: \cite[II.\,\S\,12]{Serre(1977)}.

1070: The lower horizontal map is the isomorphism given by sending $(x_D)_{D \in \sub C}$ to $(D \mapsto x_D)$.

1071: The diagram is natural with respect to automorphisms of

1072: $C$. By definition, $\theta_C \in A(C) \otimes_{\IZ} \IQ$  corresponds to the idempotent

1073: $(\delta_{CD})_D$ in the lower left-hand corner. Now, the result follows.

1074: 

1075: \smallskip

1076: 

1077: \ref{lem:compute-theta-ii} This follows from the Lyndon-Hochschild-Serre spectral

1078: sequence of the fibration $BZ_GC \to BN_GC \to BW_GC$ and from the fact that,

1079: the group $W_GC$ being finite, for any $\IQ[W_GC]$-module $M$, the

1080: $\IQ$-vector space $H_p(C_\ast(EW_GC) \otimes_{\IZ[W_GC]} M)$ is isomorphic to

1081: $M \otimes_{\IQ[W_GC]} \IQ$ for $p = 0$ and trivial for $p \geq 1$.

1082: \end{proof}

1083: 

1084: %This completes the proof of Theorem~\ref{the:positive for IQ and IC}.

1085: %\qed

1086: 

1087: 

1088: \subsection{}

1089: {\bf Proof of Theorem~\ref{the:detection-result-for-comm.-rings-of-char.-zero}.}

1090: The proof is analogous to that of Theorem~\ref{the:main-detection-result}, with the exception that

1091: we do not use Proposition~\ref{prop:compute-theta-null} but the following consequences of the hypothesis

1092: on $R$ made in the statement\,: the diagram

1093: \[

1094: \xymatrix{

1095: K_0(\IZ) \ar[r]^-{\dtr}_-{\cong} \ar[d] & \HH_0^{\otimes_{\IZ}} ( \IZ) = \IZ \ar[d]

1096: \\

1097: K_0(R) \ar[r]^-{\dtr}  & \HH_0^{\otimes_{\IZ}}(R)

1098: }

1099: \]

1100: commutes, the upper horizontal map is an isomorphism and both vertical arrows are injective.

1101: The map \eqref{injective-for-z-coefficients} is now defined as the restriction of the upper horizontal arrow

1102: of the diagram appearing in \ref{subsec:general-strategy}, in degree $n$, to the summand for $q = 0$ and

1103: $C = \{e\}$ and then further to the $\IQ$-submodule

1104: $$H_p(BG;\IQ) \cong H_p(BG;\IQ) \otimes_{\IQ} \big( K_0(\IZ) \otimes_{\IZ} \IQ \big) \; \subseteq \;

1105: H_p(BG;\IQ) \otimes_{\IQ} \big( K_0(R) \otimes_{\IZ} \IQ \big)$$

1106: (here, $p=n$). Injectivity of \eqref{injective-for-z-coefficients} is now clear from the general

1107: strategy~\ref{subsec:general-strategy}.

1108: \qed

1109: 

1110: 

1111: \subsection{}

1112: {\bf Proof of Theorem~\ref{the:optimal-Z-coeff}.}

1113: For the given $n\geq 0$, the diagram

1114: \begin{eqnarray*}

1115: \qquad\qquad\xymatrix{

1116: H_n^G \big( \EGF{G}{\Fin};\JK^{\minusinfinity} S^{-1} \calo_F \big) \ar[r] \ar[d]&

1117: H_n^G \big( \EGF{G}{\Fin};\JK^{\minusinfinity} F \big) \ar[d]^-{\cong}

1118: \\

1119: H_n^G \big( \EGF{G}{\VCyc};\JK^{\minusinfinity} S^{-1} \calo_F \big) \ar[r] \ar[d]_-{\assembly}^-{\cong_{\IQ}}  &

1120: H_n^G \big( \EGF{G}{\VCyc};\JK^{\minusinfinity} F \big) \ar[d]^-{\assembly}

1121: \\

1122: \hspace*{-7.5em}K_n \big( S^{-1} \calo_F[G] \big) \cong H_n^G(\pt;\JK^{\minusinfinity} S^{-1} \calo_F) \ar[r]

1123: \ar[d]_{\dtr} & H_n^G(\pt;\JK^{\minusinfinity} F) \cong K_n(FG)\hspace*{-4.5em} \ar[d]^{\dtr}

1124: \\

1125: \HH^{\otimes_{\IZ}}_n \big( S^{-1} \calo_F[G] \big) \ar[r]^-{\cong_{\IQ}} &

1126: \HH^{\otimes_{\IZ}}_n(FG)

1127: }

1128: \end{eqnarray*}

1129: commutes, where the vertical maps are induced by the, up to $G$-homotopy, unique $G$-maps or are

1130: given by the trace maps; the horizontal arrows are induced by the inclusion of rings $S^{-1} \calo_F \subseteq F$.

1131: Some explanations are in order for the indicated integral respectively rational isomorphisms.

1132: 

1133: \smallskip

1134: 

1135: For every ring $R$, there are isomorphisms

1136: $\HH^{\otimes_{\IZ}}_{\ast} ( R ) \otimes_{\IZ} \IQ \cong

1137: \HH^{\otimes_{\IZ}}_{\ast} ( R \otimes_{\IZ} \IQ )$ and

1138: $\HC^{\otimes_{\IZ}}_{\ast} ( R ) \otimes_{\IZ} \IQ \cong

1139: \HC^{\otimes_{\IZ}}_{\ast} ( R \otimes_{\IZ} \IQ )$, because

1140: $\CN_{\bullet}^{\otimes_{\IZ}} ( R \otimes_{\IZ} \IQ ) \cong

1141: \CN_{\bullet}^{\otimes_{\IZ}} (R) \otimes_{\IZ} \IQ$ and because

1142: the functor $(-)\otimes_{\IZ} \IQ$ commutes with homology and with

1143: $\Tot^{\oplus}$. (For the notation, see

1144: Subsections \ref{subsec:nerves} and \ref{subsec:simpl-dold-kan} below.)

1145: Here, we use that for the total complex occurring in the

1146: definition of cyclic homology it does not matter whether one takes

1147: $\Tot^{\oplus}$ or $\Tot^{\prod}$. Note that a corresponding

1148: statement is false for $\HP$ and $\HN$. Hence the bottom horizontal arrow in the diagram above

1149: is rationally bijective since $S^{-1} \calo_F \otimes_{\IZ} \IQ \cong F$.

1150: 

1151: \smallskip

1152: 

1153: The  middle left vertical arrow is rationally  bijective,

1154: since we assume that the Farrell-Jones Conjecture holds rationally for $K_n(S^{-1} \calo_F[G])$.

1155: 

1156: \smallskip

1157: 

1158: Since $F$ is a regular ring and contains $\IQ$, the top right vertical arrow is an isomorphism

1159: by \cite[Proposition~2.14]{Lueck-Reich(2003b)}, see also Subsection~\ref{subsec:general-strategy}.

1160: 

1161: \smallskip

1162: 

1163: Bartels \cite{Bartels(2003d)} has constructed, for every ring $R$ and every $m\in\IZ$, a retraction

1164: $$r(R)_m \colon H_m^G \big( \EGF{G}{\VCyc};\JK^{\minusinfinity} R \big) \; \to \; H_m^G \big( \EGF{G}{\Fin};

1165: \JK^{\minusinfinity} R \big)$$

1166: of the canonical map $H_m^G(\EGF{G}{\Fin};\JK^{\minusinfinity} R) \to H_m^G(\EGF{G}{\VCyc};\JK^{\minusinfinity} R)$,

1167: which is natural in~$R$. We obtain a decomposition, natural in $R$,

1168: $$H_m^G \big( \EGF{G}{\VCyc};\JK^{\minusinfinity}R \big) \; \cong \;

1169: H_m^G \big( \EGF{G}{\Fin};\JK^{\minusinfinity}R \big) \oplus \ker \big( r(R)_m \big)\,.$$

1170: Therefore, we conclude from the commutative diagram above that the composition

1171: \[ %\label{composition-vcyc-to-HC}

1172: H_n^G \big( \EGF{G}{\VCyc};\JK^{\minusinfinity} S^{-1} \calo_F \big) \xrightarrow{\cong_{\IQ}}

1173: H_n^G ( \pt;\JK^{\minusinfinity} S^{-1} \calo_F ) \xrightarrow{\dtr} \HH^{\otimes_{\IZ}}_n \big( S^{-1} \calo_F[G] \big)\,,

1174: \]

1175: after tensoring with $\IQ$, contains $\ker(r(S^{-1} \calo_F)_n) \otimes_\IZ \IQ$ in its kernel,

1176: because $\ker(r(F)_n)=0$. So, to study injectivity properties of the Dennis trace map

1177: we can focus attention on the composition

1178: \begin{multline*}

1179: H_n^G \big( \EGF{G}{\Fin};\JK^{\minusinfinity} S^{-1} \calo_F \big) \otimes_\IZ \IQ \into

1180: H_n^G \big( \EGF{G}{\VCyc};\JK^{\minusinfinity} S^{-1} \calo_F \big) \otimes_\IZ \IQ

1181: \\

1182: \xrightarrow{\cong} H_n^G ( \pt;\JK^{\minusinfinity} S^{-1} \calo_F ) \otimes_\IZ \IQ

1183: \xrightarrow{\dtr} \HH^{\otimes_{\IZ}}_n \big( S^{-1} \calo_F[G] \big) \otimes_\IZ \IQ\,.

1184: \end{multline*}

1185: 

1186: \smallskip

1187: 

1188: By naturality of the bottom isomorphism in Remark~\ref{remark-nonconn-conn}, there is

1189: a commutative diagram

1190: \begin{eqnarray*}

1191: \xymatrix@C=-30mm{

1192: \smash[b]{{\displaystyle {\bigoplus_{\substack{p,q \in \IZ \\p+q = n}}}\; \displaystyle

1193: \bigoplus_{(C) \in (\FCyc)}}}

1194: H_p(BZ_GC;\IQ) \otimes_{\IQ[W_GC]} \theta_C \big(  K_q (S^{-1} \calo_F [C]) \otimes_{\IZ} \IQ \big)

1195: \ar[dr]^-{\ch^G}_-{\cong}

1196: \ar[dd]

1197: & \\

1198: & H_n^G \big( \EGF{G}{\Fin};\JK^{\minusinfinity} S^{-1} \calo_F \big) \otimes_{\IZ} \IQ

1199:  \ar[dd]

1200: \\

1201: \smash[b]{{\displaystyle

1202: {\bigoplus_{\substack{p,q \in \IZ \\p+q = n}}}\; \displaystyle \bigoplus_{(C) \in (\FCyc )}}}

1203: H_p(BZ_GC;\IQ) \otimes_{\IQ[W_GC]} \theta_C \big(  K_q (FC)  \otimes_{\IZ} \IQ \big)

1204: \ar[dr]^-{\ch^G}_-{\cong}

1205: & \\

1206: & H_n^G \big( \EGF{G}{\Fin};\JK^{\minusinfinity} F \big) \otimes_{\IZ} \IQ.

1207: }

1208: \end{eqnarray*}

1209: Now, consider the composition

1210: \begin{multline} \label{composition-oplus-to-HC}

1211: \bigoplus_{\substack{p,q \in \IZ \\p+q = n}}\;

1212: \bigoplus_{(C) \in (\FCyc)}

1213: H_p(BZ_GC;\IQ) \otimes_{\IQ[W_GC]} \theta_C \big(  K_q (S^{-1} \calo_F [C]) \otimes_{\IZ} \IQ \big)

1214: \\[.6em]

1215: \qquad \stackover{\cong\,}{\xrightarrow{\ch^G}}

1216: H_n^G \big( \EGF{G}{\Fin};\JK^{\minusinfinity} S^{-1} \calo_F \big) \otimes_\IZ \IQ

1217: \into H_n^G \big( \EGF{G}{\VCyc};\JK^{\minusinfinity} S^{-1} \calo_F \big) \otimes_\IZ \IQ

1218: \\[1.4em]

1219: \xrightarrow{\cong}

1220: H_n^G ( \pt;\JK^{\minusinfinity} S^{-1} \calo_F ) \otimes_\IZ \IQ

1221: \xrightarrow{\dtr} \HH^{\otimes_{\IZ}}_n \big( S^{-1} \calo_F[G] \big) \otimes_\IZ \IQ\,.

1222: \end{multline}

1223: By the previous two diagrams, the composition \eqref{composition-oplus-to-HC} takes each of the direct summands

1224: for $q \leq -1$ to zero, since $K_q(FC) = 0$ for $q \leq -1$ (the ring $FC$ being regular).

1225: 

1226: Combining the commutativity of the diagrams occurring in Theorems \ref{the:Chern} and Remark~\ref{remark-nonconn-conn}

1227: (for $R=S^{-1} \calo_F$), we deduce that the composition \eqref{composition-oplus-to-HC} restricted to a

1228: direct summand with $p,q\geq 0$ and with $C$ arbitrary factorizes through the $\IQ$-vector space

1229: \[

1230: H_p(BZ_GC;\IQ) \otimes_{\IQ[W_GC]} \theta_C \big(  \HH^{\otimes_\IZ}_q (S^{-1} \calo_F [C]) \otimes_{\IZ} \IQ \big)\,.

1231: \]

1232: Using the isomorphism $\HH^{\otimes_{\IZ}}_\ast ( S^{-1} \calo_F[G] ) \otimes_\IZ \IQ \cong \HH^{\otimes_{\IZ}}_\ast (FG)

1233: \otimes_\IZ \IQ$, from the vanishing result stated as Proposition~\ref{prop:compute-higher-HH-HC} below, we conclude that

1234: the composition \eqref{composition-oplus-to-HC} vanishes on all summands with $q \geq 1$.

1235: 

1236: \smallskip

1237: 

1238: Finally, Proposition~\ref{prop:comp-Z} below implies

1239: that the composition \eqref{composition-oplus-to-HC} vanishes on all summands with $q = 0$ and $C \neq \{e\}$, and is

1240: injective on the summand for $q = 0$ and $C = \{e\}$. But the restriction of the composition

1241: \eqref{composition-oplus-to-HC} to the summand with $q = 0$ and $C = \{e\}$ is precisely the

1242: composition of the injective map \eqref{injective-for-z-coefficients} with the Dennis trace, simply

1243: by Remark~\ref{remark-nonconn-conn} and by construction of the map \eqref{injective-for-z-coefficients}

1244: (see the proof of Theorem~\ref{the:detection-result-for-comm.-rings-of-char.-zero} above). This finishes

1245: the proof of Theorem~\ref{the:optimal-Z-coeff}.

1246: \qed

1247: 

1248: 

1249: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1250: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1251: \section{The trace maps for finite cyclic groups} \label{sec:trace-for-finite-cyclic}

1252: 

1253: In this section, for a finite cyclic group $C$, a coefficient $k$-algebra $R$, and $q\geq 0$,

1254: we investigate the trace map

1255: \begin{eqnarray} \label{eq:trace-map-with-thetaC}

1256: \theta_C \big( K_q( RC ) \otimes_{\IZ} \IQ \big)  \to \theta_C \big( \HH_q^{\otimes_k} ( RC ) \otimes_{\IZ} \IQ \big)

1257: \end{eqnarray}

1258: and its variants using cyclic, periodic cyclic and negative cyclic homology.

1259: All results concerning the map \eqref{eq:trace-map-with-thetaC} with $q > 0$ will in fact be vanishing results

1260: stating that the map is the zero map.

1261: 

1262: 

1263: \begin{remark}

1264: Note that for a commutative ring $k$ and every $k$-algebra $R$,

1265: the canonical maps

1266: \[

1267: \xymatrix{

1268: \HH_0^{\otimes_{\IZ}}(R) \ar[d]_{\cong} \ar[r]^-{\cong} &  \HH_0^{\otimes_k}(R) \ar[d]^{\cong} \\

1269: \HC_0^{\otimes_{\IZ}} (R) \ar[r]^-{\cong} & \HC_0^{\otimes_k} (R)

1270:          }

1271: \]

1272: are all isomorphisms, because all four groups can be identified with $R / [R , R]$.

1273: The following results about $HH_0^{\otimes_{\IZ}}$ hence also apply to other ground rings and to cyclic homology.

1274: \end{remark}

1275: 

1276: 

1277: \begin{proposition} \label{prop:compute-theta-null}

1278: Let $G$ be a finite group. Suppose that the ring $R$ is from the following list\,:

1279: \begin{enumerate}

1280: \item \label{injective-i}

1281: a finite dimensional semi\-simple algebra $R$ over a field $F$ of characteristic zero;

1282: \item \label{injective-ii}

1283: a commutative complete local domain $R$ of characteristic zero;

1284: \item \label{injective-iii}

1285: a commutative Dedekind domain $R$ whose quotient field $F$ is an algebraic number field and for

1286: which $|G| \in R$ is invertible.

1287: \end{enumerate}

1288: Then the trace map

1289: $K_0 (R G) \to \HH_0^{\otimes_{\IZ}} ( RG )$ is injective in cases \ref{injective-i} and \ref{injective-ii}  and

1290: is rationally injective in case \ref{injective-iii}. This implies in all cases that for a finite cyclic group $C$,

1291: the induced  map,

1292: \[

1293: \theta_C \big( K_0( RC ) \otimes_{\IZ} \IQ \big)  \to

1294: \theta_C \big( \HH_0^{\otimes_{\IZ}}( RC ) \otimes_{\IZ} \IQ \big)

1295: \]

1296: is injective. Moreover, in all cases, except possibly in case \ref{injective-ii}, the $\IQ$-vector space

1297: $\theta_C ( K_0 (RC) \otimes_{\IZ} \IQ )$ is non-trivial.

1298: \end{proposition}

1299: 

1300: \begin{proof}

1301: \ref{injective-i}

1302: We first prove injectivity of the trace $K_0( RG ) \to \HH_0^{\otimes_{\IZ}}( RG )$.

1303: Since $R$ is semi\-simple and the order of $G$ is invertible in $R$, the ring $RG$ is

1304: semi\-simple as well, see for example Theorem~6.1 in \cite{Lam(1991)}.

1305: Using the Wedderburn-Artin Theorem \cite[Theorem~3.5]{Lam(1991)} and the fact that the trace map is compatible with

1306: finite products of rings and with  Morita isomorphisms

1307: \cite[Theorem~1.2.4 on page~17 and  Theorem~1.2.15 on page~21]{Loday(1992)},

1308: it suffices to show that the trace map

1309: \[

1310: \dtr\colon K_0 ( D ) \to \HH_0^{\otimes_{\IZ}} ( D )

1311: \]

1312: is injective in the case where $D$ is a skew-field which is a finite dimensional algebra over a field $F$ of

1313: characteristic zero. The following diagram commutes, where the vertical maps are given by restriction to $F$\,:

1314: \[

1315: \xymatrix{

1316: K_0 ( D ) \ar[d]_{\res} \ar[r]^-{\dtr} & \HH_0^{\otimes_{\IZ}} ( D ) \ar[d]^{\res} \\

1317: \; K_0 ( F ) \; \ar[r]^-{\dtr} & \HH_0^{\otimes_{\IZ}} ( F )

1318:          }

1319: \]

1320: The left vertical map can be identified with the map $\dim_F(D) \cdot \id \colon \IZ \to \IZ$

1321: and is hence injective. The trace map $K_0(F) \to \HH_0^{\otimes_{\IZ}}(F)$ can be identified with the inclusion

1322: $\IZ \to F$. This proves injectivity of the Dennis trace $K_0( RG )\to \HH_0^{\otimes_{\IZ}}( RG )$.

1323: 

1324: \smallskip

1325: 

1326: Let $R$ be a finite dimensional $F$-algebra. Then induction and restriction with respect to the inclusion

1327: $FG \to RG$ induces maps $\ind \colon K_0(FG) \to K_0(RG)$ and

1328: $\res \colon K_0(RG) \to K_0(FG)$ such that $\res \circ \ind = \dim_F(R) \cdot \id$.

1329: Hence the map $\ind \colon K_0(FG)\otimes_{\IZ} \IQ \to K_0(RG)\otimes_{\IZ} \IQ$ is injective. For $G=C$

1330: a finite cyclic group, this restricts to an injective map

1331: $$ \theta_C \big(K_0(FC)\otimes_{\IZ} \IQ\big) \to \theta_C \big(K_0(RC)\otimes_{\IZ} \IQ\big)\,.$$

1332: Since $F$ is a field of characteristic zero there exists a commutative diagram of ring homomorphisms

1333: \[

1334: \xymatrix{

1335: K_0 ( \IQ C )  \otimes_{\IZ} \IQ \ar[d] \ar[r]^-{\cong} & \map( \sub C , \IQ ) \ar[d] \\

1336: K_0 ( F C ) \otimes_{\IZ} F \ar[r]^-{\cong} & \map (  \Gamma_{F,C}\backslash \con C , F ).

1337:          }

1338: \]

1339: Here, the set $\con C$ of conjugacy classes of elements of $C$ identifies with $C$. Set $m=|C|$ and let

1340: $\mu_m\cong\IZ/m\IZ$ be the group of $m$th roots of $1$ in an algebraic closure of $F$. The action of

1341: the Galois group $G( F(\mu_m) | F )$ on $\mu_m$ determines a subgroup $\Gamma_{F,C}$ of $(\IZ/m \IZ)^{\times}

1342: \cong\Aut (\mu_m)$. An element $t \in \Gamma_{F,C}$ operates on $\con C$ by sending

1343: (the conjugacy class of) the element $c$ to $c^t$.

1344: The set of orbits under this action is $ \Gamma_{F,C}\backslash \con C$.

1345: Note that for $F=\IQ$, the group $\Gamma_{\IQ , C}$ is the whole group

1346: $(\IZ / m \IZ)^{\times}$ and $\Gamma_{\IQ , C}\backslash \con C$ can be

1347: identified with $\sub C$, the set of subgroups of $C$.

1348: So, the first line in the diagram is a special case of the second.

1349: The right-hand vertical map is contravariantly induced from the quotient map

1350: $\Gamma_{F,C}\backslash \con C \to \sub C$ and is in particular injective.

1351: The horizontal maps are given by sending a representation to its character. They are isomorphisms by

1352: \cite[II.~\S~12]{Serre(1977)}.

1353: Hence $\theta_C ( K_0 ( \IQ C ) \otimes_{\IZ} \IQ )$ injects in

1354: $\theta_C ( K_0 ( F C ) \otimes_{\IZ} F )$. We have shown in

1355: Lemma~\ref{lem:compute-theta} that

1356: $\theta_C ( K_0 ( \IQ C ) \otimes_{\IZ} \IQ )$ is non-trivial. Hence

1357: $\theta_C (K_0(RC)\otimes_{\IZ} \IQ)$ is non-trivial as well.

1358: 

1359: \smallskip

1360: 

1361: \ref{injective-ii}

1362:  According to Theorem~6.1 in \cite{Swan(1960a)}, the left-hand vertical map in the commutative diagram

1363: \[

1364: \xymatrix{

1365: K_0 ( R G ) \ar[r]^-{\dtr} \ar[d] &  \HH_0^{\otimes_{\IZ}} (RG ) \ar[d] \\

1366: \; K_0 ( FG ) \;\ar[r]^-{\dtr} & \HH_0^{\otimes_{\IZ}} ( FG )

1367:          }

1368: \]

1369: is injective. Here $F$ is the quotient field of $R$. The bottom map is injective by \ref{injective-i}.

1370: 

1371: \smallskip

1372: 

1373: \ref{injective-iii} Since any Dedekind ring is regular, the ring

1374: $R$  is a regular domain in which the order of $G$ is invertible.

1375: Hence $RG$ and $FG$ are regular, compare \cite[Proof of

1376: Proposition~2.14]{Lueck-Reich(2003b)}. For any regular ring $S$, the

1377: obvious map $K_0(S) \to G_0(S)$, with $G_0(S)$ the Grothendieck group of

1378: finitely generated $S$-modules, is bijective \cite[Corollary~38.51 on page~29]{Curtis-Reiner(1987)}.

1379: Therefore, the map $K_0 (RG) \to K_0 (FG)$ can be identified with the map

1380: \[

1381: G_0 (R G) \to G_0 ( FG )\,.

1382: \]

1383: This map has a finite kernel and is surjective under our assumptions on

1384: $R$ and $F$ \cite[Theorem~38.42 on page~22 and Theorem~39.14 on page~51]{Curtis-Reiner(1987)}.

1385: We infer that $K_0 ( RG ) \to K_0 ( FG )$ is rationally bijective.

1386: Using the corresponding commutative square involving the trace maps we have reduced our claim to the case \ref{injective-i}.

1387: \end{proof}

1388: 

1389: \begin{proposition} \label{prop:comp-Z}

1390: Let $S^{-1}\calo_F$ be a localization of the ring of integers $\calo_F$ in an algebraic number field $F$.

1391: Then the canonical map

1392: \[

1393: K_0 ( \IZ ) \otimes_{\IZ} {\IQ} \xrightarrow{\cong} K_0( S^{-1} \calo_F)

1394: \otimes_{\IZ} \IQ

1395: \]

1396: is an isomorphism and the trace map

1397: $$\dtr\colon K_0 ( S^{-1} \calo_F )\otimes_{\IZ} \IQ \to \HH_0^{\otimes_{\IZ}} ( S^{-1} \calo_F )\otimes_{\IZ} \IQ$$

1398: is injective. If $C$ is a non-trivial finite cyclic group and no prime divisor of its order $|C|$

1399: is invertible in $S^{-1}\calo_F$, then

1400: \[

1401: \theta_C ( K_0 ( S^{-1} \calo_F C ) \otimes_{\IZ} \IQ )=0.

1402: \]

1403: \end{proposition}

1404: 

1405: \begin{proof}

1406: According to a result of Swan \cite[Proposition~9.1]{Swan(1960a)}, the canonical map $K_0 ( \IZ ) \otimes_{\IZ} {\IQ}

1407: \to K_0( S^{-1} \calo_F [G]) \otimes_{\IZ} \IQ$ is an isomorphism for a finite group $G$ if no prime divisor of $|G|

1408: \in\calo_F$ occurs in $S$. As a consequence, the Artin defect \eqref{eq:SC-ist-Artin-defekt} of $K_0 ( S^{-1} \calo_F C )

1409: \otimes_{\IZ} \IQ$ (i.e.\ in degree $0$) vanishes. The result now follows from the identification, which will be proved

1410: in Lemma~\ref{lem:module-over-burnside} below, of $\theta_C ( K_0 ( S^{-1} \calo_F C ) \otimes_{\IZ} \IQ )$

1411: with the Artin defect.

1412: \end{proof}

1413: 

1414: We next collect the results which state that the trace map is the zero map in higher degrees.

1415: Note that all linear trace maps factorize through $\HN^{\otimes_{\IZ}}_\ast$. The following

1416: result implies that they all vanish in positive degrees for suitable rings $R$.

1417: 

1418: \begin{proposition} \label{prop:compute-higher-HH-HC}

1419: Let $F$ be an algebraic number field and $R$ a finite dimensional semi\-simple $F$-algebra.

1420: Then, for every finite cyclic group $C$ and for every $n\geq 1$, we have

1421: \[

1422: \HH_n^{\otimes_{\IZ}} ( RC) \otimes_{\IZ} \IQ= 0 \quad \mbox{and} \quad

1423: \HN^{\otimes_{\IZ}}_n ( RC ) \otimes_{\IZ} \IQ  = 0\,.

1424: \]

1425: \end{proposition}

1426: 

1427: \begin{proof}

1428: Analogously to the proof of

1429: Proposition~\ref{prop:compute-theta-null} \ref{injective-i}, one reduces the claim to the case

1430: where the ring $RC$ is replaced by a skew-field $D$ which is a finite dimensional algebra over

1431: an algebraic number field $F$. Let $\overline{F}$ be a \emph{splitting

1432: field} for $D$, i.e.\ a finite field extension $\overline{F}$ of $F$ such

1433: that $\overline{F} \otimes_F D \cong M_n(\overline{F})$, for some $n\geq 1$, see

1434: \cite[Corollary~7.22 on page~155]{Curtis-Reiner(1981)}. Induction

1435: and restriction for $D \subseteq \overline{F} \otimes_{F} D$ yield maps $\ind

1436: \colon K_\ast(D) \to K_\ast(\overline{F} \otimes_{F} D)$ and $\res \colon

1437: K_\ast(\overline{F} \otimes_F D ) \to K_\ast(D)$ such that $\res \circ

1438: \ind = \dim_F(\overline{F}) \cdot \id$. Hence $\ind \colon K_\ast(D)

1439: \to K_\ast(\overline{F} \otimes_F D)$ is rationally injective. The same

1440: procedure applies to Hochschild homology, cyclic, periodic cyclic and negative

1441: cyclic homology, and all these induction and restriction maps are compatible

1442: with the various trace maps. Applying Morita invariance, it thus suffices to prove that

1443: \[

1444: \HH_n^{\otimes_{\IZ}} (\overline{F}) \otimes_{\IZ} \IQ= 0 \quad \mbox{and} \quad

1445: \HN^{\otimes_{\IZ}}_n (\overline{F}) \otimes_{\IZ} \IQ  = 0\,,

1446: \]

1447: for every $n\geq 1$.

1448: For every $\IQ$-algebra $A$, there is obviously an isomorphism $\CN_{\bullet}^{\otimes_{\IZ}} ( A )

1449: \cong \CN_{\bullet}^{\otimes_{\IQ}}( A )$ of cyclic nerves, see Subsection~\ref{subsec:nerves} for

1450: the notation; hence an isomorphism $\HX_{\ast}^{\otimes_{\IZ}} ( A ) \cong \HX_{\ast}^{\otimes_{\IQ}} ( A )$, where

1451: $\HX$ stands for $\HH$, $\HC$, $\HP$ or $\HN$.

1452: So, we may consider $\HX_\ast^{\otimes_{\IQ}}$ in place of $\HX_\ast^{\otimes_{\IZ}}$ in the sequel.

1453: 

1454: 

1455: By the Hochschild-Kostant-Rosenberg Theorem, one has

1456: $\HH_{\ast}^{\otimes_{\IQ}} ( \overline{F} ) \cong \Lambda_{ \overline{F}}^{\ast} \Omega^1_{ \overline{F}| \IQ}$, compare

1457: \cite[Theorem~3.4.4 on page~103]{Loday(1992)}. But $\Omega^1_{ \overline{F} | \IQ} = 0$

1458: because $ \overline{F}$ is a finite separable extension of $\IQ$ (\cite[Corollary~16.16]{Eisenbud(1995)});

1459: therefore $\HH_{\ast}^{\otimes_{\IQ}} ( \overline{F} ) \cong  \overline{F}$ and is concentrated in degree $0$.

1460: From the long exact sequence

1461: \[

1462: \ldots \to \HH_n^{\otimes_{\IQ}}( \overline{F} )\to  \HC_n^{\otimes_{\IQ}}( \overline{F} )\xrightarrow{S}  \HC_{n-2}^{\otimes_{\IQ}}

1463: ( \overline{F} )

1464: \to \HH_{n-1}^{\otimes_{\IQ}}( \overline{F} )\to \ldots

1465: \]

1466: it follows that $\HC_{\ast}^{\otimes_{\IQ}} (  \overline{F} )$ is isomorphic to $ \overline{F}$ in each even non-negative degree,

1467: and is zero otherwise. Since the periodicity map $S$ is an isomorphism as soon as its target is non-trivial, the periodic cyclic

1468: homology is the inverse limit $\HP^{\otimes_{\IQ}}_n ( \overline{F} )= \lim_k \HC^{\otimes_{\IQ}}_{n + 2k} (  \overline{F} )$

1469: and hence is concentrated in (all) even degrees, with a copy of $\overline{F}$ in each such degree, compare

1470: \cite[5.1.10 on page~163]{Loday(1992)} and also Remark~\ref{rem:HP-and-Z} below. In the long exact sequence

1471: \[

1472: \ldots \to \HN_{n}^{\otimes_{\IQ}} ( \overline{F} )\to \HP_n^{\otimes_{\IQ}} ( \overline{F} ) \xrightarrow{\overline{S}}

1473: \HC_{n-2}^{\otimes_{\IQ}} ( \overline{F} )\to

1474: \HN_{n-1}^{\otimes_{\IQ}} (  \overline{F} ) \to \ldots

1475: \]

1476: compare~\cite[Proposition~5.1.5 on page~160]{Loday(1992)}, the map $\overline{S}$ is

1477: then an isomorphism whenever its target is non-trivial. It follows

1478: that $\HN^{\otimes_{\IQ}}_\ast ( \overline{F} )$ is concentrated in

1479: non-positive even degrees (with a copy of $\overline{F}$ in each such degree).

1480: \end{proof}

1481: 

1482: \begin{remark} \label{rem:HP-and-Z}

1483: We could not decide the question whether for an odd $n \geq 1$ and a finite cyclic group $C$, the map

1484: $\theta_C ( K_n ( \IZ C ) \otimes_{\IZ} \IQ ) \to \theta_C ( \HN_n^{\otimes_{\IZ}} ( \IZ C ) \otimes_{\IZ} \IQ)$,

1485: or the corresponding map to periodic cyclic homology, is non-trivial. The calculations in \cite{Kassel(1989)}

1486: and \cite{Cortinas-Guccione-Villamayor(1989)} show that a finer analysis of the trace map is needed in order

1487: to settle the problem. The difficulty is that the $\lim^1$-terms in the computation of $\HP$ out of $\HC$

1488: might contribute to non-torsion elements in odd positive degrees.

1489: \end{remark}

1490: 

1491: 

1492: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1493: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1494: \section{Notation and generalities}\label{sec:Notation}

1495: 

1496: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1497: \subsection{Categories and $k$-linear categories} \label{subsec:cat-and-linear-cat}

1498: 

1499: Let $k$ be a commutative ring. A \emph{$k$-linear category} is a small category which is enriched over

1500: $k$-modules, i.e.\ each morphism set $\hom_{\cala}(c,d)$, with $c,d\in \obj \cala$, has the structure

1501: of a $k$-module, composition of morphisms is bilinear and satisfies the usual associativity axiom; moreover,

1502: there are unit maps $k \to \hom_{\cala} (c , c)$, for every object $c$, satisfying a unit axiom.

1503: Compare \cite[I.8 on page~27, VII.7 on page~181]{MacLane(1971)}. Let $R$ be a $k$-algebra.

1504: For any small category~$\calc$, we can form the associated $k$-linear category $R\calc$.

1505: It has the same objects as $\calc$ and the morphism $k$-modules are obtained as

1506: the free $R$-module over the morphism sets of~$\calc$, i.e.\

1507: \[

1508: \hom_{R \calc} (c,d) = R \mor_{\calc} (c,d)\,.

1509: \]

1510: In fact, this yields a functor $R(-)$  from small categories to $k$-linear categories. Given a $k$-linear category $\cala$,

1511: we denote by $\cala_{\oplus}$ the $k$-linear category whose objects are finite sequences of objects of $\cala$, and

1512: whose morphisms are ``matrices'' of morphisms in $\cala$ with the obvious ``matrix product'' as composition.

1513: Concatenation of sequences yields a sum denoted by ``\,$\oplus$\,'' and we hence obtain, functorially, a

1514: \emph{$k$-linear category with finite sums}, compare \cite[VIII.2 Exercise~6 on page~194]{MacLane(1971)}.

1515: If we consider a $k$-algebra $R$ as a $k$-linear category with one object

1516: then $R_{\oplus}$ is a small model

1517: for the category of finitely generated free left $R$-modules.

1518: 

1519: 

1520: 

1521: 

1522: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1523: \subsection{Nerves and cyclic nerves} \label{subsec:nerves}

1524: 

1525: Let $\calc$ be a small category and let $\cala$ be a $k$-linear category. The \emph{cyclic nerve} of $\calc$

1526: and the \emph{$k$-linear cyclic nerve} of $\cala$ are respectively denoted by

1527: \[

1528: \CN_{\bullet} \calc \quad \mbox{and} \quad \CN_{\bullet}^{\otimes_k} \cala\,.

1529: \]

1530: Depending on the context, they are considered as a cyclic set or as a simplicial set, respectively as a cyclic $k$-module

1531: or as simplicial $k$-module. Recall that by definition, we have

1532: \begin{eqnarray*}

1533: \CN_{[q]} \calc & = &

1534: \coprod_{c_0 , c_1 , \ldots ,c_q \in  \obj \calc} \mor_{\calc} (c_1, c_0) \times \dots \times \mor_{\calc} ( c_q , c_{q-1} )

1535: \times \mor_{\calc}( c_0 , c_q ) ,  \\

1536: \CN_{[q]}^{\otimes_k} \cala & = &

1537: \bigoplus_{c_0 , c_1 , \ldots ,c_q \in  \obj \cala} \hom_{\cala} (c_1, c_0) \otimes_k \dots \otimes_k \hom_{\cala} ( c_q , c_{q-1} )

1538: \otimes_k  \hom_{\cala}( c_0 , c_q ) .  \\

1539: \end{eqnarray*}

1540: The simplicial and cyclic structure maps are induced by composition, insertion of identities and cyclic permutations of

1541: morphisms. For more details, see \cite[2.3]{Waldhausen(1979)}, \cite{Goodwillie(1985)} and \cite{Dundas-McCarthy(1994)}.

1542: The (ordinary) nerve of a small category $\calc$ will always  be considered as a simplicial category and denoted by

1543: $\caln_{\bullet}\calc$. We will write $\obj \caln_{\bullet} \calc$ for the underlying simplicial set of objects.

1544: 

1545: 

1546: 

1547: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1548: \subsection{Simplicial abelian groups and chain complexes} \label{subsec:simpl-dold-kan}

1549: 

1550: If we are given a simplicial abelian group $M_{\bullet}$, we

1551: denote by $\DK_{\ast}(M_{\bullet})$ the associated normalized

1552: chain complex. For a chain complex of abelian groups $C_{\ast}$

1553: which is concentrated in non-negative degrees, we denote by

1554: $\DK_{\bullet}(C_{\ast})$ the simplicial abelian group that is

1555: associated to it under the Dold-Kan correspondence. For details

1556: see \cite[Section~8.4]{Weibel(1994)}. In particular, recall that

1557: there are  natural isomorphism

1558: $\DK_{\bullet}(\DK_{\ast}(M_{\bullet})) \cong M_{\bullet}$ and

1559: $\DK_{\ast}(\DK_{\bullet}(C_{\ast})) \cong C_{\ast}$.

1560: 

1561: The \emph{good truncation} $\tau_{\scriptscriptstyle\geq 0} C_{\ast}$ of a chain complex ¤$C_{\ast}$

1562: is defined as the non-negative chain complex which coincides with $C_{\ast}$ in

1563: strictly positive degrees, has the $0$-cycles $Z_0(C_{\ast})$ in degree $0$, and only trivial

1564: modules in negative degrees. Given a bicomplex $C_{\ast \ast}$, we denote by

1565: \[

1566: \Tot^{\oplus} C_{\ast \ast} \quad  \mbox{and} \quad

1567: \Tot^{\prod} C_{\ast \ast}

1568: \]

1569: the total complexes formed using, respectively, the direct sum or the direct product, compare

1570: \cite[1.2.6 on page~8]{Weibel(1994)}.

1571: 

1572: 

1573: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1574: \subsection{Spectra, $\Gamma$-spaces and Eilenberg-Mc\,Lane spectra} \label{subsec:Gamma-spaces}

1575: 

1576: For us, a \emph{spectrum} consists of a sequence $\JE$ of pointed spaces $E_n$, with $n\geq 0$, together with pointed

1577: maps $s_n \colon S^1 \sma E_n \to E_{n+1}$. We do \emph{not} require that the adjoints  $\sigma_n \colon E_n \to \Omega E_{n+1}$ of these maps are homotopy equivalences.

1578: A map of spectra $\Jf \colon \JE \to \JE^{\prime}$ consists of a sequence of maps $f_n \colon E_n \to E_n^{\prime}$ such that

1579: $f_{n+1} \circ s_n  = s_n^{\prime} \circ \id_{S^1} \sma f_n$. One defines in the usual way the homotopy groups

1580: as $\pi_{n}(\JE) = \colim_k \pi_{n+k}( E_k)$, with $n\in\IZ$. The spectrum $\JE$ is \emph{connective} if $\pi_{n}(\JE) = 0$

1581: for all $n<0$. A map of spectra is called a \emph{stable weak equivalence},

1582: or to be short, an \emph{equivalence}, if it induces an isomorphism on all homotopy groups. A \emph{spectrum of simplicial sets}

1583: is defined similarly, using pointed simplicial sets in place

1584: of pointed spaces. Such spectra can be realized and then yield spectra in the sense above.

1585: We denote by $\JS$ the sphere spectrum (as a spectrum of simplicial sets).

1586: 

1587: 

1588: Let $\Gamma^{\op}$ denote the small model for the category of finite pointed sets whose objects are $k_{+}=

1589: \{ + , 1, \ldots , k \}$, with $k\geq 0$, and whose morphisms are pointed maps.

1590: A \emph{$\Gamma$-space} $\IE$ is a functor from the category $\Gamma^{\op}$ to the category of

1591: pointed simplicial sets which sends $0_+ = \{ + \}$ to the (simplicial) point. Every $\Gamma$-space $\IE$ can be

1592: extended in an essentially unique way to an endofunctor of the category of pointed simplicial sets

1593: which we again denote by $\IE$. By evaluation on the simplicial spheres, a $\Gamma$-space $\IE$

1594: gives rise to a spectrum of simplicial sets denoted by $\IE ( \JS ) $. The realization $| \IE ( \JS) |$ is then a

1595: spectrum in the sense defined above. A $\Gamma$-space $\IE$ is called \emph{special} if the map

1596: $\IE(k_+) \to \IE(1_+ ) \times\dots \times \IE (1_+)$  induced by the projections

1597: $p_i \colon  k_+ \to 1_+$, with $i=1, \ldots , k$, is a weak equivalence for every $k$.

1598: Here, $p_i ( j )$ is $1$ if $j=i$, and is $+$ otherwise. For more information

1599: on spectra and $\Gamma$-spaces, we refer to \cite{Bousfield-Friedlander(1978)} and \cite{Lydakis(1999)}.

1600: 

1601: \smallskip

1602: 

1603: An important example of a $\Gamma$-space is the \emph{Eilenberg-Mc\,Lane $\Gamma$-space} $\IH M_{\bullet}$

1604: associated to a simplicial abelian group $M_{\bullet}$\,. Its value on the finite pointed set $k_+$ is given by

1605: the simplicial abelian group $\IH M_{\bullet} (k_+) = \widetilde{\IZ}[k_+] \otimes_{\IZ} M_{\bullet}$. Here

1606: $\IZ [S]$ denotes the free abelian group generated by the set $S$, and, if the set $S$ is pointed with $s_0$ as

1607: base-point, $\widetilde{\IZ}[S]=\IZ [S]\big/\IZ[s_{0}]$ is the corresponding reduced group. The spectrum

1608: $\JH M_{\bullet} = | \IH M_{\bullet} ( \JS) |$ is a model for the \emph{Eilenberg-MacLane spectrum} associated to

1609: $M_{\bullet}$\,.

1610: The $\Gamma$-space $\IH M_{\bullet}$ is

1611: \emph{very special} in the sense of \cite[page~98]{Bousfield-Friedlander(1978)} and by \cite[Theorem~4.2]{Bousfield-Friedlander(1978)},

1612: the homotopy groups of the spectrum $\JH M_{\bullet}$ coincide with the (unstable) homotopy groups of

1613: (the realization of) $\IH M_{\bullet}( S^0 )$ and hence of $M_{\bullet}$, and consequently with the homology groups of the

1614: associated chain complex $\DK_{\ast}(M_{\bullet})$. So we have natural isomorphisms

1615: \begin{eqnarray} \label{homotopy-HM}

1616: \pi_\ast(\JH M_{\bullet}) \cong \pi_\ast\big(\big|\IH M_{\bullet}( S^0 )\big|\big) \cong \pi_\ast\big(|M_{\bullet}|\big)

1617: \cong H_{*}\big(\DK_{\ast}(M_{\bullet})\big)\,.

1618: \end{eqnarray}

1619: 

1620: 

1621: 

1622: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1623: \subsection{Cyclic, periodic cyclic and negative cyclic homology} \label{define-bicomplexes}

1624: 

1625: If $Z_{\bullet}$ is a cyclic object in the category of abelian groups, then we denote by $B_{\ast\ast}(Z_{\bullet})$,

1626: $B_{\ast\ast}^{\per}( Z_{\bullet} )$ and $B_{\ast\ast}^{-}( Z_{\bullet} )$ the \emph{cyclic}, \emph{periodic cyclic} and

1627: \emph{negative cyclic bicomplexes}, see~\cite[pages 161--162]{Loday(1992)}. For the good truncations

1628: of the associated total complexes, we write

1629: \begin{eqnarray*}

1630: C_{\ast}^{\HC} (Z_{\bullet}) & = & \Tot^{\prod} B_{\ast\ast}(Z_{\bullet}) \\

1631: C_{\ast}^{\HP} (Z_{\bullet}) & = & \tau_{\scriptscriptstyle\geq 0} \Tot^{\prod} B_{\ast\ast}^{\per}(Z_{\bullet}) \\

1632: C_{\ast}^{\HN} (Z_{\bullet}) & = & \tau_{\scriptscriptstyle\geq 0} \Tot^{\prod} B_{\ast\ast}^{-}(Z_{\bullet})\,.

1633: \end{eqnarray*}

1634: In order to have a uniform notation, it is also convenient to write

1635: \begin{eqnarray*}

1636: C_{\ast}^{\HH} (Z_{\bullet}) & = &  \DK_{\ast} ( Z_{\bullet} ).

1637: \end{eqnarray*}

1638: There is a commutative

1639: diagram of chain complexes

1640: \begin{eqnarray} \label{dia-square-nphc}

1641: \vcenter{\xymatrix{

1642: C_{\ast}^{\HN} ( Z_{\bullet} ) \ar[r] \ar[d]_{{\rm h}_\ast} & C_{\ast}^{\HP} ( Z_{\bullet} ) \ar[d] \\

1643: C_{\ast}^{\HH} ( Z_{\bullet} ) \ar[r] & C_{\ast}^{\HC} ( Z_{\bullet} )

1644: }}

1645: \end{eqnarray}

1646: where the horizontal arrows are induced by inclusions of sub-bicomplexes and the vertical arrows by

1647: projections onto quotient bicomplexes. Let $k$ be a commutative ring. If $Z_{\bullet}$ is the $k$-linear

1648: cyclic nerve $\CN_{\bullet}^{\otimes_k} ( \cala )$ of a $k$-linear category $\cala$, we abbreviate

1649: \begin{eqnarray*}

1650: C_{\ast}^{\HX^{\otimes_{\!k}}}( \cala ) & = &  C_{\ast}^{\HX} \big( \CN_{\bullet}^{\otimes_k} ( \cala ) \big)\,.

1651: \end{eqnarray*}

1652: Here $\HX$ stands for $\HH$, $\HC$, $\HP$ or $\HN$.

1653: The corresponding simplicial abelian group and the corresponding Eilenberg-Mc\,Lane spectrum will be denoted

1654: \begin{eqnarray*}

1655: C_{\bullet}^{\HX^{\otimes_{\!k}}} ( \cala ) & = & \DK_{\bullet} \big( C_{\ast}^{\HX^{\otimes_{\!k}}} (\cala) \big) \\

1656: \JH \JX^{\otimes_k} ( \cala ) & = & \JH C_{\bullet}^{\HX^{\otimes_{\!k}}} (\cala)\,.

1657: \end{eqnarray*}

1658: In particular we have the map $\Jh \colon \JH \JN^{\otimes_k} ( \cala ) \to \JH \JH^{\otimes_k} ( \cala ) $

1659: induced from the map $h_{\ast}$ in \eqref{dia-square-nphc}.

1660: If $R$ is a $k$-algebra we can consider it as a $k$-linear category with one object. Then the

1661: homology groups of $C_{\ast}^{\HX^{\otimes_{\!k}}} ( R )$ as defined above coincide in non-negative degrees with the groups

1662: $\HX^{\otimes_k}_\ast(R)$ that appear in the literature, for instance in \cite{Loday(1992)}. Often negative

1663: cyclic homology $\HN_\ast^{\otimes_k}(R)$ is denoted by $\HC_\ast^-(R)$ or $\HC_\ast^-(R|k)$ in the literature.

1664: 

1665: 

1666: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1667: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1668: \section{The trace maps} \label{sec:trace-maps}

1669: 

1670: The aim of this section is to produce the diagram \eqref{spec-level-trace-maps}, i.e.\ the trace maps as maps of

1671: $\Or G$-spectra. We will concentrate on the part of the diagram involving $K$-theory, Hochschild homology and negative

1672: cyclic homology. The remaining arrows are obtained by straightforward modifications.

1673: 

1674: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1675: \subsection{The trace maps for additive categories} \label{subsec-trace-additive-categories}

1676: 

1677: We now review the construction of $K$-theory for additive categories, and of the map $\h$

1678: and the trace maps $\ntr$ and $\dtr$ for $k$-linear categories with finite sums, following the

1679: ideas of \cite{McCarthy(1994)}, \cite{Dundas-McCarthy(1996)} and \cite{Dundas(2000)}.

1680: 

1681: 

1682: The following commutative diagram is natural in the $k$-linear category $\cala$\,:

1683: {\small\begin{eqnarray} \label{babytrace}

1684: \hspace*{-4.5em}\vcenter{\mbox{\phantom{xxxxxxxx}}\xymatrix@C=1.7em{

1685: & \qquad\quad C_{[0]}^{\HN^{\otimes_{\!k}}} ( \cala ) = \DK_{0}( C_{\ast}^{\HN^{\otimes_{\!k}}} ( \cala ))

1686: \ar@<-2.5em>[d]^{{\rm h}_0} \ar[r] &  \DK_{\bullet} ( C_{\ast}^{\HN^{\otimes_{\!k}}}( \cala )) = C_{\bullet}^{\HN^{\otimes_{\!k}}}

1687:  ( \cala) \ar@<-2.8em>[d]^{{\rm h}_\bullet=\DK_\bullet({\rm h}_\ast)} \\

1688: \obj \cala \ar[r]^-{\dtr_0} \ar@(dl,u)[ur]^(.46){\ntr_0} & C_{[0]}^{\HH^{\otimes_{\!k}}}( \cala ) = \CN_{0}^{\otimes_k}

1689: \cala \ar[r] & \DK_{\bullet} ( C^{\HH^{\otimes_{\!k}}}_{\ast}( \cala)) \cong \CN_{\bullet}^{\otimes_k}( \cala ).

1690:          }}

1691: \end{eqnarray}}

1692: \hspace*{-.4em}Here, the lower horizontal map $\dtr_0$ is given by sending an object to the corresponding identity

1693: morphism. The lift $\ntr_0$ of this map is explicitly described on page~286 in \cite{McCarthy(1994)}. The remaining

1694: horizontal maps are just the inclusions of the zero simplices. The vertical maps are induced by the map $\h_\ast$ in

1695: diagram \eqref{dia-square-nphc}. The isomorphism in the bottom right corner is a special case

1696: of the natural isomorphism $\DK_{\bullet} ( \DK_{\ast} (M_{\bullet})) \cong M_{\bullet}$, compare Subsection~\ref{subsec:simpl-dold-kan}.

1697: It will be considered as an identification in the following.

1698: 

1699: 

1700: 

1701: The model for the trace maps, for a given $k$-linear category with finite sums $\cala$

1702: will be obtained by replacing

1703: $\cala$ in the diagram above by a suitable simplicial $k$-linear $\Gamma$-category.

1704: On the $K$-theory side, we will use the fact that $\cala$ has finite sums; on the Hochschild

1705: side, we will use the $k$-linear structure.

1706: 

1707: 

1708: Let $\cala$ be a small category with finite sums.

1709: We can then apply the Segal construction

1710: which yields a $\Gamma$-category $\Seg \cala$, that is, a functor from $\Gamma^{\op}$ to the

1711: category of small categories, compare~\cite[Definition~3.2]{Dundas(2000)}

1712: and \cite[Section~2]{Segal(1974)}.

1713: 

1714: 

1715: Recall that we consider the nerve of a category as a simplicial category.

1716: Let

1717: $\caln_{\bullet}^{\iso}\cala$ be the simplicial subcategory of $\caln_{\bullet} \cala$ for which the objects in

1718: $\caln_{[q]}^{\iso} \cala$ are $q$-tuples of composable isomorphisms in $\cala$, whereas there is no restriction on

1719: the morphisms. Observe that $\obj \caln_{\bullet}^{\iso} \calc = \obj \caln_{\bullet} \iso \calc$, where $\iso \calc$

1720: stands for the subcategory of isomorphisms.

1721: 

1722: 

1723: 

1724: 

1725: 

1726: 

1727: The \emph{connective $K$-theory spectrum}

1728: $\JK( \cala )$ can now be defined as the spectrum associated to the $\Gamma$-space $\obj

1729: \caln_{\bullet}^{\iso} \Seg \cala$, that is,

1730: \begin{eqnarray} \label{def-K-theory}

1731: \JK ( \cala ) & = & \big| ( \obj \caln_{\bullet}^{\iso}  \Seg \cala ) ( \JS ) \big|\,.

1732: \end{eqnarray}

1733: For a comparison to other definitions of $K$-theory, see \cite[Section~1.8]{Waldhausen(1985)}.

1734: 

1735: 

1736: 

1737: We proceed to discuss the trace maps. Let $\cala$ be a $k$-linear category with finite sums. Recall that $\Delta$

1738: is the category of finite ordered sets $[n]=\{0 \leq 1 \leq \ldots \leq  n\}$, with $n\geq 0$, and monotone

1739: maps as morphisms. Observe that $\caln_{\bullet}^{\iso}  \Seg \cala$ is a functor from $\Delta^{\op} \times \Gamma^{\op}$

1740: to $k$-linear categories and it hence makes sense to apply the cyclic nerve constructions.

1741: Since the diagram \eqref{babytrace} is natural in $\cala$

1742: we obtain maps of simplicial $\Gamma$-spaces (alias natural transformations of functors from

1743: $\Delta^{\op} \times \Delta^{\op} \times \Gamma^{\op}$ to the category of pointed sets)

1744: \begin{eqnarray} \label{dia-basic-model-trace-map}

1745: \vcenter{\xymatrix{

1746: & \DK_{\bullet} C_{\ast}^{\HN^{\otimes_{\!k}}} \caln_{\bullet}^{\iso} \Seg \cala \ar[d]^{\h_{\bullet\bullet}} \\

1747: \obj \caln_{\bullet}^{\iso} \Seg \cala

1748: \ar[r]^-{\dtr_{\bullet\bullet}} \ar@(ld,u)[ur]^{\ntr_{\bullet\bullet}} &

1749: \CN_{\bullet}^{\otimes_k} \caln_{\bullet}^{\iso} \Seg \cala.

1750:          }}

1751: \end{eqnarray}

1752: Here $\obj \caln_{\bullet}^{\iso}  \Seg \cala$ is constant in one of the simplicial directions. Taking the

1753: diagonal of the two simplicial directions and passing to the

1754: associated spectra yields the model for the trace maps that we will use. It remains to identify the

1755: objects on the right in \eqref{dia-basic-model-trace-map} with our more standard definitions of Hochschild

1756: and negative cyclic homology.

1757: 

1758: 

1759: \begin{lemma} \label{zig-zag-lemma}

1760: Let $\cala$ be a $k$-linear category with finite sums.

1761: There is a zigzag of stable weak equivalences, natural in $\cala$,  between

1762: \[

1763: \xymatrix@R=.5em{

1764: \JH \JN^{\otimes_k} (\cala ) = \JH \DK_{\bullet} C_{\ast}^{\HN^{\otimes_{\!k}}} \cala \ar@<2.4em>[dd]^{\Jh} & &

1765: \big| ( \DK_{\bullet} C_{\ast}^{\HN^{\otimes_{\!k}}} \caln^{\iso}_{\bullet} \Seg \cala) ( \JS ) \big|

1766: \ar[dd]^{|\h_{\bullet\bullet}(\JS)|} \\

1767: & \mbox{and} & \\

1768: \JH \JH^{\otimes_k} (\cala ) = \JH \CN_{\bullet}^{\otimes_k} \cala & &

1769: \big| ( \CN_{\bullet}^{\otimes_k} \caln^{\iso}_{\bullet} \Seg \cala) ( \JS ) \big|.

1770:          }

1771: \]

1772: \end{lemma}

1773: 

1774: \begin{proof}

1775: Consider, for each $q$, the inclusion of the zero simplices

1776: $i\colon \cala = \caln_{[0]}^{\iso} \cala \to \caln_{[q]}^{\iso} \cala$. There is a

1777: left inverse $p$ (forget everything but the zero-th object) and an obvious natural transformation

1778: between $i \circ p$ and the identity which is objectwise an isomorphism. This induces a special

1779: homotopy equivalence (\cite[Definition~2.3.2]{McCarthy(1994)}) and hence in particular a homotopy

1780: equivalence of cyclic nerves $\CN_{\bullet}^{\otimes_k} \cala \xrightarrow{\simeq} \CN_{\bullet}^{\otimes_k}

1781: \caln_{\bullet}^{\iso} \cala$, which passes to a chain homotopy equivalence

1782: on the negative cyclic construction, compare \cite[Proposition~2.4.1]{McCarthy(1994)}. So

1783: we get rid of the $\caln_{\bullet}^{\iso}$'s in the expressions above. The rest now follows by applying

1784: the following lemma to the map

1785: \[

1786: \IH(h_{\bullet})\colon\IH \DK_{\bullet} C_{\ast}^{\HN^{\otimes_{\!k}}} \Seg \cala \to \IH

1787: \CN_{\bullet}^{\otimes_k}  \Seg \cala

1788: \]

1789: of bi-$\Gamma$-spaces, provided we can prove that the source and the target are both special in both variables

1790: (see~\ref{subsec:Gamma-spaces}). Specialness in the Eilenberg-Mc\,Lane-variable is standard and follows immediately

1791: from the definition of the functor $\IH(-)$. Being special in the Segal-variable means in the case of the first

1792: bi-$\Gamma$-space that for every $l_+$ and $k_+$, the following composition is a weak equivalence\,:

1793: \begin{eqnarray*}

1794: \widetilde{\IZ} [l_+] \otimes_{\IZ} \big(

1795: \CN^{\otimes_k}_{\bullet} \Seg \cala (k_+) \big) & \to &

1796: \widetilde{\IZ} [l_+] \otimes_{\IZ} \big(

1797: \CN^{\otimes_k}_{\bullet} ( \cala \times \dots \times \cala ) \big) \\[.2em]

1798: & \to &

1799: \widetilde{\IZ} [l_+] \otimes_{\IZ} \big(

1800: \CN^{\otimes_k}_{\bullet} \cala \times \dots \times \CN^{\otimes_k}_{\bullet} \cala \big) \\[.2em]

1801: & \to &

1802: \widetilde{\IZ} [l_+] \otimes_{\IZ}

1803: \CN^{\otimes_k}_{\bullet} \cala \times \dots \times

1804: \widetilde{\IZ} [l_+] \otimes_{\IZ}

1805: \CN^{\otimes_k}_{\bullet} \cala\,.

1806: \end{eqnarray*}

1807: This is clearly true for the last map.

1808: The Segal construction is designed in such a way that $\Seg \cala ( k_+ ) \to \cala \times \dots \times \cala$

1809: is an equivalence of categories. By \cite[Proposition~2.4.1]{McCarthy(1994)}, this passes to an equivalence

1810: on the cyclic constructions and yields that the first map is an equivalence.

1811: Proposition~2.4.9 in \cite{McCarthy(1994)} deals with the second map. The argument for the second bi-$\Gamma$-space is

1812: analogous.

1813: \end{proof}

1814: 

1815: \begin{lemma} \label{bi-special-bi-gamma}

1816: Suppose that $(k_+,l_+) \mapsto \IA (k_+ , l_+)$ is a bi-$\Gamma$-space which is special

1817: in both variables, i.e.\ for every fixed $l_+$, the $\Gamma$-space $k_+ \mapsto \IA ( k_+ , l_+ )$

1818: is special, and similarly in the other variable. Then there is a natural zigzag of

1819: stable weak equivalences of spectra of simplicial sets between

1820: \[

1821: \IA ( 1_+ , \JS ) \quad \mbox{and} \quad \IA ( \JS , 1_+ )\,.

1822: \]

1823: \end{lemma}

1824: 

1825: \begin{proof}

1826: There is a naive definition of a bi-spectrum (of simplicial sets) as a collection $\IE$ of pointed simplicial sets $E_{n,m}$,

1827: with $n \geq 0$ and $m \geq 0$, together with horizontal and vertical pointed structure maps

1828: $\sigma_h \colon E_{n,m} \sma S^1 \to E_{n+1,m}$ and $\sigma_v \colon S^1 \sma   E_{n,m} \to E_{n,m+1}$

1829: satisfying

1830: \[

1831: \sigma_h \circ (\sigma_v \sma S^1) = \sigma_v \circ (S^1 \sma \sigma_h)\,.

1832: \]

1833: After some choice of a poset map $\mu \colon \IN_0 \to\IN_0 \times \IN_0$ satisfying a suitable cofinality condition,

1834: for example $\mu(2n)=(n,n)$ and $\mu(2n+1)=(n+1,n)$, one can form the diagonal spectrum

1835: $\diag_{\mu} \IE$, compare \cite[Section~1.3]{Jardine(1997)}.

1836: 

1837: 

1838: 

1839: A \emph{bi-$\Gamma$-space} $\IA$ is a functor from $\Gamma^{\op}$ to the category of $\Gamma$-spaces, denoted

1840: by $k_+\mapsto\IA(k_+,-)$, and such that $\IA(0_+,l_+)$ is the (simplicial) point for each $l_+$.

1841: Every bi-$\Gamma$-space $\IA$ gives rise to a simplicial bi-spectrum $\IA(\JS , \JS^{\prime})$ in the naive

1842: sense above, with $\IA(\JS , \JS^{\prime})_{n,m}=\IA ( S^n , S^m )$. Here $\JS^{\prime}$ is

1843: just a copy of the simplicial sphere spectrum $\JS$ which we want to distinguish in the notation.

1844: 

1845: 

1846: 

1847: 

1848: There are maps of bi-$\Gamma$-spaces

1849: \[

1850: \xymatrix{

1851: \IA ( 1_+ , k_+ \sma l_+ ) & k_+ \sma \IA( 1_+ , l_+ ) \ar[l] \ar[r] & \IA ( k_+ , l_+ )

1852:          }

1853: \]

1854: which extend to maps of bi-spectra. We claim that for every pointed simplicial set $Y$, the corresponding maps of simplicial

1855: spectra

1856: \begin{eqnarray} \label{left-and-right}

1857: \xymatrix{

1858: \IA ( 1_+ , Y \sma \JS ) & Y \sma \IA( 1_+ , \JS) \ar[l] \ar[r] & \IA ( Y , \JS )

1859:          }

1860: \end{eqnarray}

1861: are stable weak equivalences. For the first map, this is \cite[Lemma~4.1]{Bousfield-Friedlander(1978)}

1862: and no specialness assumption is needed. For the second map, one argues as follows.

1863: For a pointed simplicial set $X$, the simplicial set $\IA ( 1_+ , X)$ is at least as connected

1864: as $X$ \cite[Proposition~5.20]{Lydakis(1999)}. Now, the composition

1865: \[

1866: k_+ \sma \IA( 1_+ , X ) \to \IA( k_+ , X ) \to \IA ( 1_+ , X ) \times \dots \times \IA

1867: ( 1_+ , X )

1868: \]

1869: is the inclusion of a $k$-fold wedge into the corresponding $k$-fold product and hence roughly twice as connected as

1870: $\IA (1_+ , X)$. Since the second map is a weak equivalence (by the assumption that $\IA$ is special in the

1871: second variable), we conclude that the connectivity of the first map grows faster than $n$ for $X=S^n$.

1872: The same statement holds for arbitrary pointed simplicial sets $Y$ in place of $k_+$

1873: by a careful version of the Realization Lemma for bisimplicial sets (realization preserves connectivity --- compare

1874: \cite[Lemma~2.1.1]{Waldhausen(1979)}). So, the second map in \eqref{left-and-right} indeed is a stable weak equivalence,

1875: proving the claim above.

1876: 

1877: 

1878: 

1879: If we now apply the elementary Lemma~1.28 from \cite{Jardine(1997)} (this is a Realization Lemma for bi-spectra), we

1880: obtain a zigzag of weak equivalences of spectra of simplicial sets between

1881: $\IA ( 1_+ , \diag_{\mu} \JS \sma \JS^{\prime} )=\diag_{\mu} \IA ( 1_+ , \JS \sma \JS^{\prime} )$

1882: and $\diag_{\mu}\IA ( \JS , \JS^{\prime} )$. The pointed isomorphism between

1883: $S^0$ and the $0$th simplicial set of the spectrum $\diag_{\mu} \JS \sma \JS^{\prime}$ determines

1884: uniquely a map of spectra $\JS \to \diag_{\mu} \JS \sma \JS^{\prime}$ which clearly is an isomorphism. In

1885: total, we have constructed a zigzag of stable weak equivalences between $\IA( 1_+ , \JS)$

1886: and $\diag_{\mu} \IA(\JS , \JS^{\prime})$. The result now follows by symmetry (using specialness in the first variable).

1887: \end{proof}

1888: 

1889: 

1890: Summarizing we have that for a $k$-linear category $\cala$ with finite sums, the model for the

1891: trace maps, at the level of spectra, is given by a  commutative diagram of the following form

1892: {\small\[

1893: \xymatrix@C=1.2em{

1894: & \big| ( \DK_{\bullet} C_{\ast}^{\HN^{\otimes_{\!k}}} \caln^{\iso}_{\bullet} \Seg \cala) ( \JS ) \big|

1895: \ar[d]^{|\h_{\bullet\bullet}(\JS)|}

1896: \ar[r]^-{\simeq} & & \dots \ar[l]_-{\simeq} \ar[r]^-{\simeq} & \JH \JN^{\otimes_k} (\cala ) \ar[d]^{\Jh}  \\

1897: \JK ( \cala ) = \big| ( \obj \caln_{\bullet}^{\iso}  \Seg \cala ) ( \JS ) \big|

1898: \ar[r]^-{\Jdtr} \ar@(ld,u)[ur]^{\Jntr} &

1899: \;\;\; \big| ( \CN_{\bullet}^{\otimes_k} \caln^{\iso}_{\bullet} \Seg \cala) ( \JS ) \big| \;\;\; \ar[r]^-{\simeq} & &

1900: \dots \ar[l]_-{\simeq} \ar[r]^-{\simeq} & \JH \JH^{\otimes_k} (\cala ).

1901:          }

1902: \]}

1903: 

1904: 

1905: 

1906: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1907: \subsection{The trace maps as maps of spectra over the orbit category} \label{subsec:trace-maps}

1908: 

1909: We will now define the $\Or G$-spectra representing $K$-theory, Hochschild homology and other

1910: cyclic homology theories, and the trace maps which appear in \eqref{spec-level-trace-maps}.

1911: 

1912: 

1913: \smallskip

1914: 

1915: Given a $G$-set $S$, let $\calg^G(S)$ denote the associated \emph{transport groupoid}, i.e.\ the category

1916: whose objects are the elements of $S$ and where the set of morphisms from $s \in S$ to $t \in S$ is given

1917: by $\mor(s,t)=\{ g \in G \,|\, gs=t \}$. Given  a $k$-algebra $R$ we can compose the functor $\calg^G(?)$

1918: with the functors $R( -)$ and $( - )_{\oplus}$ (compare Subsection~\ref{subsec:cat-and-linear-cat})

1919: to obtain a functor

1920: \begin{eqnarray*}

1921: R \calg^G ( ? )_{\oplus}\colon\Or G \to  k\mbox{-}\Cat_{\oplus}\,,\quad G/H \mapsto R \calg^G ( G/H )_{\oplus}\,,

1922: \end{eqnarray*}

1923: where $k\mbox{-}\Cat_{\oplus}$ denotes the \emph{category of small $k$-linear categories with finite sums},

1924: whose morphisms are $k$-linear functors (and hence respect the sum, compare \cite[VIII.2~Prop.4 on page 193]{MacLane(1971)}). The \emph{idempotent completion} $\Idem \cala$

1925: of a category $\cala$ has as objects the idempotent endomorphisms in $\cala$, i.e.\ morphisms $p \colon  c \to c$

1926: with $p\circ p=p$; a morphism from $p\colon c \to c$ to $q\colon d \to d$ is given

1927: by a morphism $f\colon c \to d$ with $q\circ f= f \circ p$. The idempotent completion of a $k$-linear category

1928: is again $k$-linear. For a small category $\calc$, the idempotent completion of $R \calc_{\oplus}$ is a $k$-linear

1929: category with finite sums. For an arbitrary ring $S$, the category $\Idem S_{\oplus}$

1930: is a small model for the category of finitely generated projective left $S$-modules.

1931: 

1932: 

1933: 

1934: Let $R$ be a $k$-algebra and $H$ a subgroup of $G$. Consider the commutative diagram of $k$-linear categories

1935: \begin{eqnarray} \label{diagram-k-lin-equivalences}

1936: \hspace*{3em}\vcenter{\xymatrix@C=2em{

1937: RH \ar[r] \ar[d] & RH_{\oplus} \ar[r] \ar[d] & \Idem RH_{\oplus} \ar[d] \\

1938: R\calg^G( G/H )  \ar[r] & R\calg^G (G/H)_{\oplus} \ar[r] & \Idem R\calg^G (G/H)_{\oplus} .

1939:          }}

1940: \end{eqnarray}

1941: The vertical functors are all induced from considering $H$ as the full subcategory of $\calg^G( G/H)$

1942: on the object $eH \in G/H= \obj \calg^G(G/H)$. All vertical functors are $k$-linear equivalences

1943: and the two right hand functors are cofinal inclusions into the corresponding  idempotent

1944: completions. Hence it follows from \cite[Proposition~2.4.1 and 2.4.2]{McCarthy(1994)} that all functors

1945: in the diagram above induce equivalences if one applies Hochschild homology or one of the cyclic homology

1946: theories, i.e.\ $\JH \JX^{\otimes_k} ( - )$. Observe that our $K$-theory functor $\JK( - )$ can only be

1947: applied to the four categories on the right (they have finite sums). The two right-hand vertical maps

1948: induce isomorphisms on all higher $K$-groups, however, $K_0$ may differ for a category with finite sums and

1949: its idempotent completion.

1950: 

1951: \smallskip

1952: 

1953: Finally, define  $\Or G$-spectra $\JK R(?)$ and $\JH \JX^{\otimes_k}R(?)$ by

1954: \begin{eqnarray}

1955: \JK R (G/H) & = & \JK \Idem R \calg^G( G/H )_{\oplus} \label{def-KR-OrG-spectrum} \\

1956: \JH \JX^{\otimes_k} R (G/H) & = & \JH \JX^{\otimes_k} R \calg^G( G/H )\,. \label{def-HX-OrG-spectrum}

1957: \end{eqnarray}

1958: Here again $\JH \JX$ stands for $\JH \JH$, $\JH \JC$ , $\JH \JP$ or $\JH \JN$.

1959: Compare \eqref{def-K-theory} and the notation introduced in Subsection~\ref{define-bicomplexes}.

1960: The discussion above and the one in Subsection~\ref{define-bicomplexes} verify all the isomorphisms

1961: claimed in \eqref{relate-to-classical}.

1962: 

1963: \smallskip

1964: 

1965: Now, apply the construction of \eqref{dia-basic-model-trace-map} in the case where the additive category

1966: $\cala$ is $\Idem R \calg^G( G/H )_{\oplus}$. Using the equivalences (discussed above) induced by the map

1967: $\Idem R \calg^G( G/H )_{\oplus}\leftarrow R\calg^G( G/H )$ and the equivalences appearing in the

1968: diagram at the end of Subsection~\ref{subsec-trace-additive-categories}, we obtain a commutative diagram

1969: of connective $\Or G$-spectra of the shape

1970: \[

1971: \xymatrix@C=1.7em{

1972: & & {}^{\prime}\JH\JN^{\otimes_k}R \ar[d]& \ar[l]_-{\simeq} {}^{\prime\prime}\JH\JN^{\otimes_k}R \ar[d]

1973: \ar[r]^-{\simeq} & & \dots \ar[l]_-{\simeq} \ar[r]^-{\simeq} & \JH \JN^{\otimes_k} R \ar[d]^{\Jh} \\

1974: \JK R \ar[rr]^-{\Jdtr} \ar[urr]^{\Jntr} & & {}^{\prime}\JH\JH^{\otimes_k}R & \ar[l]_-{\simeq} {}^{\prime\prime}\JH

1975: \JH^{\otimes_k}R \ar[r]^-{\simeq} & & \dots \ar[l]_-{\simeq} \ar[r]^-{\simeq} & \JH \JH^{\otimes_k} R,

1976:          }

1977: \]

1978: where all arrows labelled with a ``\,$\simeq$\,'' (in particular all those pointing left) are objectwise

1979: stable weak equivalences.

1980: 

1981: 

1982: 

1983: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1984: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1985: \section{Equivariant homology theories, induction  and Mackey structures} \label{sec:mackey}

1986: 

1987: A \emph{$G$-homology theory} is a collection of functors $\calh_\ast^{G}(-)=\{\calh_n^{G}(-)\}_{n\in\IZ}$

1988: from the category of (pairs of) $G$-$CW$-complexes to the category of abelian groups, which satisfies the $G$-analogues

1989: of the usual axioms for a generalized homology theory, compare \cite[2.1.4]{Lueck-Reich(2003b)}.

1990: 

1991: \smallskip

1992: 

1993: For example, every $\Or G$-spectrum $\JE=\JE(?)$ gives rise to a $G$-homology theory $H_\ast^G ( - ; \JE )$ by setting,

1994: for a $G$-$CW$-complex $X$,

1995: \begin{eqnarray*}

1996: H_\ast^G ( X ; \JE ) & = & \pi_\ast \big( X_+^? \sma_{\Or G} \JE(?) \big)

1997: \end{eqnarray*}

1998: and more generally, for a pair of $G$-$CW$-complexes $(X,A)$,

1999: \begin{eqnarray*}

2000: H_\ast^G ( X , A ; \JE ) & = & \pi_\ast \big( (X_+/A_+)^? \sma_{\Or G} \JE(?) \big)\,.

2001: \end{eqnarray*}

2002: Here, for a $G$-space $Y$, the symbol $Y_+$ denotes the space $Y$ with a disjoint base-point added (viewed as a

2003: $G$-fixpoint), and $Y^?$ stands for the \emph{fixpoint functor} $\map_G( - , Y)$, considered as a contravariant functor

2004: from $\Or G$ to the category of spaces; and $X_+^? \sma_{\Or G} \JE (?)$ is the \emph{balanced smash product} of

2005: a contravariant pointed $\Or G$-space and a covariant $\Or G$-spectrum. It is constructed by applying levelwise the

2006: \emph{balanced smash product}

2007: {\small\begin{multline}  \label{balanced-smash}

2008: Y \sma_{\Or G} Z  =

2009: \xymatrix@C=1.5em{

2010:  \coequ \Big(

2011: \bigvee_{f \in \mor \Or G} Y( t(f) ) \sma Z( s(f) ) \ar@<3pt>[r] \ar@<-3pt>[r] &

2012: \bigvee_{G/H \in \Or G} Y( H ) \sma Z( H ) \Big)

2013: }

2014: \end{multline}}

2015: \hspace*{-.3em}of a contravariant pointed $\Or G$-space $Y(?)$ and a covariant pointed $\Or G$-space $Z(?)$;

2016: here, $s(f)$ stands for the source and $t(f)$ for the target of the morphism $f \in \mor \Or G$,

2017: $\coequ$ is the \emph{coequalizer}, and the two indicated maps are defined by $f^{\ast} \sma \id$

2018: and $\id \sma f_{\ast}$ on the wedge-summand corresponding to $f$. We repeat that $H_{\ast}^G( \pt ; \JE)$ identifies

2019: with $\pi_{\ast}( \JE( G/G ))$. For details, we refer to \cite{Davis-Lueck(1998)} and \cite[Chapter~6]{Lueck-Reich(2003b)}.

2020: 

2021: \smallskip

2022: 

2023: For a group homomorphism $\alpha \colon H \to G$ and an $H$-$CW$-complex $X$, let $\ind_{\alpha} X$ be the quotient

2024: of $G \times X$ by the right action of $H$ given by $(g,x)h = (g \alpha (h) , h^{-1} x)$. An \emph{equivariant homology

2025: theory} $\calh^?_\ast=\calh^?_\ast(-)$ consists of a $G$-homology theory for each group $G$ together with natural induction

2026: isomorphisms

2027: \[

2028: \ind_{\alpha} \colon \calh^H_\ast ( X , A ) \xrightarrow{\cong} \calh^G_\ast ( \ind_{\alpha} X , \ind_{\alpha} A )

2029: \]

2030: for each group homomorphism $\alpha \colon H \to G$ and each $H$-$CW$-pair $(X,A)$ such that $\ker \alpha$ acts freely on $X$.

2031: The induction isomorphisms need to verify certain natural axioms, compare \cite[6.1]{Lueck-Reich(2003b)}. We refer to

2032: the collection of induction isomorphisms as an ``\emph{induction structure}''.

2033: 

2034: \smallskip

2035: 

2036: Suppose that an $\Or G$-spectrum $\JD(?)$ is a composition of functors $\JD = \JE \circ \calg^G (?)$, where

2037: $\JE \colon \Groupoids \to \Sp$ is a functor from the \emph{category of small groupoids} to the \emph{category of spectra}.

2038: If $\JE$ is a \emph{homotopy functor}, i.e.\ sends equivalences of groupoids to stable weak equivalences of spectra,

2039: then, according to \cite[Proposition~6.10]{Lueck-Reich(2003b)} and \cite{SauerJ(2002)}, there is a `naturally' defined

2040: induction structure for the collection of $G$-homology theories, one for each group $G$, given by $H_\ast^G ( - ; \JE

2041: \circ \calg^G )$. Hence each homotopy functor $\JE \colon \Groupoids \to \Sp$ determines an equivariant homology theory

2042: $H_\ast^? ( - ; \JE \circ \calg^G )$.

2043: 

2044: \smallskip

2045: 

2046: Given an equivariant homology theory $\calh_\ast^?(-)$, one can, for each $n\in\IZ$, construct a covariant functor

2047: from $\FGINJ$, i.e.\ \emph{the category of finite groups and injective group homomorphisms}, to $\Ab$, i.e.\ the

2048: \emph{category of abelian groups}, by setting

2049: \begin{eqnarray} \label{mackey-co-a}

2050: M_{\ast} \colon \FGINJ \to \Ab\,, \quad G \mapsto \calh^G_n( \pt )\,;

2051: \end{eqnarray}

2052: for a group monomorphism $\alpha \colon H \into G$, we define $M_{\ast}(\alpha)$ as the composition

2053: \begin{eqnarray} \label{mackey-co-b}

2054: \xymatrix{

2055: M_{\ast} ( H ) = \calh^H_n ( \pt ) \ar[r]^-{\ind_{\alpha}} & \calh_n^G \big( G/ \alpha(H) \big) \ar[r]^-{\calh_n^G (\pr)} &

2056: \calh^G_n ( \pt ) = M_{\ast} (H)\,,

2057:          }

2058: \end{eqnarray}

2059: where $pr$ is the projection onto the point.

2060: 

2061: \smallskip

2062: 

2063: A \emph{Mackey functor} $M$ is a pair $(M_{\ast} , M^{\ast})$ consisting of a co- and a contravariant functor $\FGINJ \to \Ab$

2064: which agree on objects, i.e.\ $M_{\ast}( H ) = M^{\ast} ( H )$ (merely denoted by $M(H)$), and satisfy the following axioms.

2065: \begin{enumerate}

2066: \item For an inner automorphism $c_g \colon G \to G$, $h \mapsto g^{-1} h g$ with $g \in G$ one has

2067: $M_{\ast} ( c_g ) = \id \colon M(G) \to M(G)$.

2068: \item

2069: If $f \colon G \xrightarrow{\cong} H$ is an isomorphism, then one has $M_{\ast} (f) \circ M^{\ast} (f) = \id$ and

2070: $M^{\ast} (f) \circ M_{\ast} (f) = \id$.

2071: \item

2072: There is a double coset formula, i.e., for two subgroups $H,K \leq G$, one has

2073: $$

2074: \renewcommand{\arraystretch}{1.5}

2075: \begin{array}{l}

2076: \qquad\quad\;\; M^{\ast} \big( i \colon K \to G \big) \circ M_{\ast} \big( i \colon H \to G \big) \; = \qquad \\[.2em]

2077: \qquad\qquad\quad\;  \; {\displaystyle\sum_{KgH \in K\backslash G / H}} M_{\ast} \big( c_g \colon H \cap g^{-1} K g \to K \big)

2078: \circ M^{\ast} \big( i \colon H \cap g^{-1} K g \to H \big)\,,

2079: \end{array}

2080: $$

2081: where $c_g( h) = g^{-1}hg$ and $i$ in each case denotes the inclusion.

2082: \end{enumerate}

2083: 

2084: \smallskip

2085: 

2086: If, for every $n\in\IZ$, the covariant functor $M_{\ast}$ we associated in

2087: \eqref{mackey-co-a} and  \eqref{mackey-co-b} to an equivariant homology theory $\calh^?_\ast(-)$

2088: can be extended to a Mackey functor, then we say that the equivariant homology theory admits a

2089: ``\emph{Mackey structure}''.

2090: 

2091: \smallskip

2092: 

2093: Let $R$ be a $k$-algebra. We will consider compositions of functors of the form

2094: \[

2095: \xymatrix@C=3em{

2096: \Or G \ar[r]^-{\calg^G( - )} & \Groupoids \ar[r]^-{R ( - )_{\oplus}} &

2097: k\mbox{-}\Cat_{\oplus} \ar[r]^-{\JF} & \Sp\,.

2098:          }

2099: \]

2100: Recall that $k\mbox{-}\Cat_{\oplus}$ denotes the category of small $k$-linear categories with finite sums.

2101: 

2102: The $\Or G$-spectra we are mainly interested in, namely $\JK R(?)$ and  $\JH \JX^{\otimes_k}R(?)$,

2103: are defined (up to equivalence for the latter) as such a composition with $\JF$ being the composite functor

2104: $\JK\circ\Idem(-)$ for the former, see \eqref{def-K-theory} and \eqref{def-KR-OrG-spectrum}, and being

2105: $\JH \JX^{\otimes_k}(-)$ for the latter, see Subsection~\ref{define-bicomplexes} and  the discussion following

2106: diagram \eqref{diagram-k-lin-equivalences}, and \eqref{def-HX-OrG-spectrum}. It turns out that the non-connective

2107: $K$-theory $\Or G$-spectrum $\JK^{\minusinfinity}R(?)$ of Example~\ref{exa:FJ} is also such a composition.

2108: In that case $\JF$ is the \emph{Pedersen-Weibel functor} (defined on $\Cat_{\oplus}$), compare \cite{Pedersen-Weibel(1985)}.

2109: Up to equivalence a model for the $(-1)$-connective covering map of $\Or G$-spectra $\JK R(?)\to\JK^{\minusinfinity}R(?)$ mentioned in~\ref{exa:FJ} is induced by

2110: a specific natural transformation between the corresponding $\JF$'s.

2111: 

2112: So, consider a functor $\JF\colon k\mbox{-}\Cat_{\oplus} \to \Sp$. We call $\JF$ a \emph{homotopy functor} if it takes

2113: $k$-linear equivalences to stable weak equivalences of spectra. We call $\JF$ \emph{additive} if for every $k$-linear

2114: functors $f,g \colon \cala \to \calb$ between $k$-linear categories with finite sums,

2115: \begin{eqnarray} \label{the-sum-property}

2116: \pi_{\ast} ( \JF ( f \oplus g ) ) = \pi_{\ast} ( \JF (f) ) + \pi_{\ast} ( \JF (g) )

2117: \end{eqnarray}

2118: holds; here, $f \oplus g \colon \cala \to \calb$ is the composition

2119: \[

2120: \xymatrix{

2121: \cala \ar[r]^-{\diag} & \cala \times \cala \ar[r]^-{f \times g} & \calb \times \calb \ar[r]^-{\oplus} & \calb\,,

2122:          }

2123: \]

2124: where $\diag$ denotes the \emph{diagonal embedding} and $\oplus$ is the sum in $\calb$.

2125: 

2126: \begin{proposition} \label{prop:homotopy-functor}

2127: Suppose that $\JF\colon k\mbox{-}\Cat_{\oplus} \to \Sp$ is a homotopy functor. Then, the composite functor

2128: $\JF \circ R (- )_{\oplus}$ is a homotopy functor; in particular, it determines an equivariant homology

2129: theory whose underlying $G$-homology theory, for a group~$G$, is given by the $\Or G$-spectrum $\JF R \calg^G

2130: (?)_{\oplus}$\,, that is, by

2131: \begin{eqnarray*}

2132: H_\ast^G \big( X , A ; \JF R \calg^G(?)_{\oplus} \big) & = & \pi_\ast \big( (X_+/A_+)^? \sma_{\Or G} \JF R\calg^G

2133: ( ? )_{\oplus} \big)\,.

2134: \end{eqnarray*}

2135: If $\JF$ is additive then this equivariant homology theory admits a Mackey structure.

2136: \end{proposition}

2137: 

2138: \begin{proof}

2139: The first part is clear. For the second, we need to define the contravariant half of the Mackey functor and verify the axioms.

2140: For a given ring $S$ let  $\calf(S)$ denote the \emph{category of finitely generated free left $S$-modules},

2141: which is of course not a small category. If we consider a group $H$ as a groupoid with one object, then  $RH_{\oplus}$

2142: is a small model for the category of finitely generated free

2143: left $RH$-modules and there is an inclusion functor $i_H \colon RH_{\oplus} \to \calf(RH)$ which is

2144: an equivalence of categories. We choose a functor $p_H \colon \calf(RH) \to RH_{\oplus}$ such that

2145: $p_H \circ i_H \simeq \id$ and $i_H \circ p_H \simeq \id$. Here, $f \simeq g$ indicates that there

2146: exists a natural transformation through isomorphisms.

2147: Given a homomorphism $\alpha \colon H \to G$ between finite groups, there are the usual induction

2148: and restriction functors $\ind_{\alpha} \colon \calf( RH) \to \calf (RG)$ and $\res_{\alpha} \colon

2149: \calf (RG ) \to \calf ( RH )$. For $n\in\IZ$, we define induction and restriction homomorphisms

2150: \[

2151: \ind_{\alpha} \colon \pi_n \big( \JF RH_{\oplus} \big) \to \pi_n \big( \JF RG_{\oplus} \big) \quad \mbox{and} \quad

2152: \res_{\alpha} \colon \pi_n \big( \JF RG_{\oplus} \big) \to \pi_n \big( \JF RH_{\oplus} \big)

2153: \]

2154: as $\ind_{\alpha} = \pi_n ( \JF ( p_G \circ \ind_{\alpha} \circ i_H ))$ and

2155: $\res_{\alpha} = \pi_n ( \JF ( p_H \circ \res_{\alpha} \circ i_G ))$.

2156: Since $f \simeq g$ implies $\pi_n ( \JF ( f) ) = \pi_n ( \JF (g) )$, this does not depend on the

2157: choice of $p_H$ and $p_G$.

2158: 

2159: \smallskip

2160: 

2161: Unravelling the definitions, one checks that under the identifications

2162: \[

2163: M(H) = \pi_n \big( \pt_+^? \sma_{\Or G} \JF R \calg^G( ? )_{\oplus} \big) \cong \pi_n \big( \JF R \calg^G ( H/H )_{\oplus} \big)

2164: \cong \pi_n \big( \JF RH_{\oplus} \big)\,,

2165: \]

2166: the induction homomorphism $ M_{\ast} ( \alpha )$ from \eqref{mackey-co-b} coincides with the

2167: induction homomorphism we have just constructed. Using the same identifications, we consider the map

2168: $\res_{\alpha}$ constructed above as a map $M(G) \to M(H)$ and denote it by $M^{\ast}( \alpha )$. The

2169: axioms now follow since each of the remaining equalities corresponds to a well-known natural isomorphism

2170: between functors on categories of finitely generated free left modules; for the third axiom, one uses

2171: \eqref{the-sum-property}, i.e.\ additivity of $\JF$.

2172: \end{proof}

2173: 

2174: The functors $\JF$ that are responsible for $\JK^{-\infty}(?)$, $\JK(?)$ and $\JH \JX^{\otimes_k} (?)$ are

2175: homotopy invariant and additive.

2176: We hence obtain the corresponding equivariant homology theories with Mackey structures

2177: given, at a group $G$, by $H_\ast^G ( - ; \JK R )$, $H_\ast^G ( - ; \JK^{\minusinfinity} R )$ and by $H_\ast^G ( - ;\JH \JX^{\otimes_k} R )$.

2178: The maps between these theories that are induced from the maps of $\Or G$-spectra that we have discussed above

2179: are  compatible with the

2180: induction and Mackey structures.

2181: 

2182: 

2183: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

2184: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

2185: \section{Evaluating the equivariant Chern character}

2186: \label{sec:eval-chern}

2187: In this section, we prove Theorem \ref{the:Chern} which is a slight improvement

2188: of results in~\cite{Lueck(2002b)}.

2189: 

2190: 

2191: In the previous section we have verified that the assumptions of Theorem~0.1 and of

2192: Theorem~0.2 in \cite{Lueck(2002b)} are satisfied in the case where the equivariant homology theory $\calh^?_\ast ( - )$

2193: is given, at a group $G$, by $H_\ast^G ( - ; \JK R ) \otimes_{\IZ} \IQ$, by $H_\ast^G ( - ; \JK^{\minusinfinity} R )

2194: \otimes_{\IZ} \IQ$, or by $H_\ast^G ( - ; \JH\JX^{\otimes_k} R ) \otimes_{\IZ} \IQ$. Let $M$ be a Mackey functor,

2195: for instance $H\mapsto\calh^H_n ( \pt )$ for $n\in\IZ$ fixed. For a finite group $H$, recall the notation

2196: \begin{eqnarray} \label{eq:SC-definition}

2197: \hspace*{3em}

2198: S_H \big( M(H) \big) \; = \;

2199: \coker \left( \bigoplus_{K \lneqq H} \ind_K^H \colon \bigoplus_{K \lneqq H}

2200: M(K) \to

2201: M(H)

2202: \right)

2203: \end{eqnarray}

2204: from \cite{Lueck(2002b)}. Observe for example in the case of $K$-theory that this specializes to \eqref{eq:SC-ist-Artin-defekt}.

2205: We obtain from \cite[Theorems~0.1 and 0.2]{Lueck(2002b)}, for every $G$-$CW$-complex $X$ which is \emph{proper} (i.e.\ with all

2206: stabilizers finite) and every $n\in\IZ$, a canonical isomorphism

2207: \[

2208: \calh_n^G ( X ) \; \cong \; \displaystyle \bigoplus_{p + q = n} \; \displaystyle \bigoplus_{(H) \in (\Fin)} H_p

2209: \big( Z_G H \backslash  X^H ; \IQ \big) \otimes_{\IQ [W_G H ] } S_H \big( \calh^H_{q} ( \pt ) \big)\,,

2210: \]

2211: where $(\Fin)$ denotes the set of conjugacy classes of finite subgroups of $G$. This isomorphism

2212: is natural in $X$ and also in the equivariant homology theory with Mackey structure $\calh^?_\ast ( - )$

2213: (i.e.\ for natural transformations of equivariant homology theories respecting the induction and Mackey

2214: structures). Now, take $X = \underline{E}G$. As in Lemma~8.1 in \cite{Lueck-Reich-Varisco(2003)}, one shows

2215: that the projections

2216: \[

2217: Z_G H \backslash \underline{E}G^H  \leftarrow EZ_G H \times_{Z_G H }  \underline{E}G^H \to EZ_G H / Z_G H = BZ_G H

2218: \]

2219: induce isomorphisms on rational homology. Theorems \ref{the:Chern} now follows

2220: from the next two lemmas.

2221: 

2222: \begin{lemma} \label{lemma:vanish-not-cyclic}

2223: Let $R$ be a  ring and let $H$ be a finite group. If $H$ is not cyclic, then

2224: \[

2225: S_H \big( K_n(RH) \tensor_{\IZ} \IQ \big) = 0 \quad \mbox{and} \quad S_H \big( \HX^{\otimes_k}_n ( R H )

2226: \otimes_{\IZ} \IQ \big) = 0\,,

2227: \]

2228: for all $n \in \IZ$.

2229: \end{lemma}

2230: 

2231: 

2232: \begin{proof}

2233: For a group $H$, let $\Sw (H,\IZ)$ be its Swan group, i.e.\ the Grothendieck group of left $\IZ H$-modules

2234: which are finitely generated as abelian groups. Let $\Sw^f(H,\IZ)$ be the Grothendieck

2235: group of left $\IZ H$-modules which are finitely generated free as abelian groups.

2236: The obvious map $\Sw^f(H,\IZ) \to \Sw(H,\IZ)$ is an isomorphism,

2237: see \cite[Proposition~1.1 on page~553]{Swan(1960a)}.

2238: If $H$ is a finite group, then  $\Sw^f(H ,\IZ)$, and hence also $\Sw(H,\IZ)$, has the structure of

2239: a commutative associative ring, where multiplication is induced by the tensor product over $\IZ$ equipped with the

2240: diagonal $H$-action. The tensor product over $\IZ$ equipped with the diagonal action also leads to a $\Sw^f( H , \IZ)$-module

2241: structure, and hence a $\Sw ( H , \IZ)$-module structure,

2242: on $K_n ( RH )$ for each $n \in \IZ$ and each coefficient ring $R$. For an injective group homomorphism

2243: $\alpha \colon H \into K$ between finite groups, we have the

2244: usual induction and restriction homomorphisms $\ind_H^K \colon \Sw( H , \IZ ) \to \Sw( K , \IZ )$ and

2245: $\res^K_H \colon \Sw( K , \IZ ) \to \Sw( H , \IZ )$. It is not difficult to check that with these structures,

2246: $\Sw (- , \IZ)$ is a Green ring functor with values in abelian groups and that, for each $n \in \IZ$,

2247: the functor $K_n ( R (-) )$ is a module over it (compare~\cite[Sections 7 and 8]{Lueck(2002b)}). Now, by a result

2248: of Swan \cite[Corollary~4.2 on page~560]{Swan(1960a)}, for every finite group $H$, the cokernel of the map

2249: \[

2250:     \bigoplus_{\substack{C\lneqq H \\ C \mbox{ \small cyclic}}}

2251:      \ind_C^H\colon \bigoplus_{\substack{C\lneqq H \\ C \mbox{ \small cyclic}}}

2252:      \Sw(C,\IZ) \tensor_{\IZ}\IQ \to \Sw(H , \IZ) \tensor_{\IZ}\IQ

2253: \]

2254: is annihilated by $|H|^2$. With suitable elements $x_C \in \Sw ( C , \IZ )$, we can hence write

2255: \[

2256: |H|^2 \cdot [ \IZ ] = \sum_{\substack{C\lneqq H \\ C \mbox{ \small cyclic}}} \ind_C^H (x_C )\,.

2257: \]

2258: Therefore, up to multiplication by $|H|^2$, every element $y \in K_n ( R H )$ is induced from proper cyclic subgroups, since

2259: \[

2260: |H|^2 \cdot y = |H|^2 \cdot [ \IZ ] \cdot y =

2261: \sum_{\substack{C\lneqq H \\ C \mbox{ \small cyclic}}} \ind_C^H (x_C ) \cdot y  =

2262: \sum_{\substack{C\lneqq H \\ C \mbox{ \small cyclic}}} \ind_C^H (x_C \cdot \res_C^H y)\,.

2263: \]

2264: The argument for Hochschild homology and its cyclic variants is similar. The module structure over the Swan

2265: ring is also in that case induced by the tensor product over $\IZ$.

2266: \end{proof}

2267: 

2268: \begin{remark}

2269: More generally, the proof of Lemma~\ref{lemma:vanish-not-cyclic} works for every module over the rationalized

2270: Swan group $\Sw( - , \IZ)\otimes_{\IZ} \IQ$ considered as a Green ring functor. Note that such a statement does

2271: not hold in general for modules over the rationalized Burnside ring $A( - )\otimes_{\IZ} \IQ$ viewed as a Green

2272: ring functor.

2273: \end{remark}

2274: 

2275: \begin{lemma} \label{lem:module-over-burnside}

2276: Let $C$ be a finite cyclic group and $M$ a Mackey functor with values in $\IQ$-modules.

2277: Keeping notation as in \eqref{theta_C} and \eqref{eq:SC-definition}, there is

2278: a natural isomorphism

2279: \[

2280: \theta_{C} \big( M ( C ) \big) \; \cong \; S_C \big( M ( C ) \big)\,.

2281: \]

2282: \end{lemma}

2283: 

2284: \begin{proof} Let $D$ be a subgroup of $C$. The ring homomorphism $\chi_C$ of \eqref{def-chi-G}

2285: sends $[C /D]$ to $(x_{(E)})_{(E)}$, where $x_{(E)} = |(C/D)^{E}|$ and hence $x_{(E)}=[C \colon D]$ if $E \leq D$ and

2286: is $0$ otherwise. Therefore the maps $i_D^C$ and $r^C_D$ which make the diagrams

2287: \[

2288: \xymatrix{

2289: A(D) \ar[d]_-{\chi_{D}} \ar[r]^-{\ind_D^C} &  A(C) \ar[d]^-{\chi_{C}}  &

2290: A(C) \ar[d]_-{\chi_{C}} \ar[r]^-{\res^C_D} &  A(D) \ar[d]^-{\chi_{D}} \\

2291: \prod_{\sub D} \IQ  \ar[r]^-{i_D^C} & \prod_{\sub C} \IQ  & \prod_{\sub C} \IQ  \ar[r]^-{r^C_D} &  \prod_{\sub D} \IQ

2292:          }

2293: \]

2294: commutative are easily seen to be given as follows. The map $i_D^C$

2295: is multiplication by the index $[C \colon D]$ followed by the inclusion of the factors corresponding to subgroups of

2296: $C$ contained in $D$. The map $r^C_D$ is the projection onto the factors corresponding to subgroups of $C$ contained

2297: in $D$. In particular, $\chi_C(\ind_D^C (\theta_D))$, considered as a function, is supported only on $(D)$ and takes

2298: there the value $[C \colon D]$. As a consequence, in $A(C) \otimes_{\IZ} \IQ$, we have

2299: \[

2300: 1 = [C/C] = \sum_{D \leq C} \frac{1}{[C \colon D]} \ind_D^C ( \theta_D )\,.

2301: \]

2302: Each element in the image of the map $1- \theta_C \colon M(C) \to M(C)$ lies in the image of

2303: $I=\bigoplus_{D \lneqq C} \ind_C^D$, because

2304: \[

2305: (1- \theta_C) x = \left( \sum_{D \lneqq C} \frac{1}{[C \colon D]} \ind_D^C  \theta_D  \right) x

2306: = \sum_{D \lneqq C} \frac{1}{[C \colon D]} \ind_D^C  ( \theta_D  \res^C_D x )\,.

2307: \]

2308: Moreover $\theta_C \colon M(C) \to M(C)$ vanishes on the image of this map $I$; indeed, for $D \lneqq C$,

2309: it follows from the description of $r^C_D$ that $\res^C_D ( \theta_C ) =0$, and therefore

2310: \[

2311: \theta_C  \ind_D^C y  = \ind_D^C ( \res^C_D (\theta_C) y ) = \ind_D^C ( 0 \cdot y ) = 0\,.

2312: \]

2313: So, the cokernel $S_C(M(C))$ of $I$ is isomorphic to the image $\theta_C(M(C))$ of $\theta_C$.

2314: \end{proof}

2315: 

2316: 

2317: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

2318: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

2319: \section{Comparing different models}\label{sec:comparing-different-models}

2320: 

2321: In order to prove splitting results in Section~\ref{sec:Splitting-assembly-maps},

2322: we will work with a chain complex version and  occasionally with a simplicial abelian

2323: group version of the equivariant homology theory that is associated to Hochschild homology.

2324: In the present section, we define these versions and  prove that they all agree.

2325: 

2326: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

2327: 

2328: \smallskip

2329: 

2330: Again, fix a group $G$.

2331: A construction analogous to the balanced smash product \eqref{balanced-smash}, but with smash products ``\,$\wedge$\,''

2332: replaced by tensor products over $\IZ$, and with wedge sums ``\,$\vee$\,'' replaced by direct sums, yields the notion

2333: of \emph{balanced tensor product} $M(?) \otimes_{\IZ \Or G} N(?)$ of a co- and a contravariant $\IZ

2334: \Or G$-module $M(?)$ and $N(?)$. Here by definition a co- or contravariant \emph{$\IZ

2335: \Or G$-module} is a co- respectively  contravariant functor

2336: from $\Or G$ to abelian groups.

2337: Let $C_{\ast} =

2338: C_{\ast} ( ? )$ be a covariant \emph{$\IZ \Or G$-chain complex}, i.e.\ a functor from the orbit category to the category

2339: of chain complexes of abelian groups. We define the \emph{$G$-equivariant Bredon hyperhomology} of a pair of

2340: $G$-$CW$-complexes $(X,A)$ with coefficients in $C_{\ast}$ as

2341: \begin{eqnarray*}

2342: H_\ast^G ( X, A ; C_{\ast}) & = & H_\ast \Big( \Tot^{\oplus} \big( \widetilde{C}_{\ast}^{\sing}( (X_+/A_+)^? )

2343: \otimes_{\IZ \Or G} C_{\ast}(?) \big) \Big)\,.

2344: \end{eqnarray*}

2345: Here, for a pointed $G$-space $Y=(Y,y_0)$ (where $y_0$ is a

2346: $G$-fixpoint), the functor which sends $G/H$ to the reduced

2347: singular chain complex of $Y^H$ is denoted by

2348: $\widetilde{C}_{\ast}^{\sing}(Y^?)$. In this construction, up to

2349: canonical isomorphism, we can replace

2350: $\widetilde{C}_{\ast}^{\sing}(-)$ by the \emph{reduced cellular

2351: chain complex} $\widetilde{C}_{\ast}^{\cell}(-)$ (this will be

2352: needed in Subsection~\ref{subsec:splitting-HN-HP-assembly}).

2353: 

2354: 

2355: For a \emph{simplicial $\IZ\Or G$-module} $M_{\bullet}=M_{\bullet} (?)$, i.e.\ a covariant functor from $\Or G$

2356: to the category of simplicial abelian groups, we define similarly

2357: \begin{eqnarray*}

2358: H_{\ast}^G ( X, A ; M_{\bullet}) & = & \pi_{\ast} \Big( \big|\widetilde{\IZ} \big[ S_{\bullet} ( (X_+/A_+)^? ) \big]

2359: \otimes_{\IZ \Or G} M_{\bullet} ( ? ) \big| \Big)\,.

2360: \end{eqnarray*}

2361: Here, $S_{\bullet}$ stands for the singular simplicial set associated to a topological space. For a pointed

2362: simplicial set $Y_\bullet=(Y_\bullet, y_0)$, we set $\widetilde{\IZ} [Y_\bullet] = \IZ [Y_\bullet] \big/ \IZ [y_0]$

2363: and the tensor

2364: products of simplicial abelian groups are taken degreewise.

2365: 

2366: For an $\Or G$-spectrum $\JE=\JE(?)$, recall that we use the notation

2367: \begin{eqnarray*}

2368: H_{\ast}^G( X,A ; \JE ) & = & \pi_{\ast} \big( (X_+/A_+)^? \sma_{\Or G} \JE(?) \big)\,.

2369: \end{eqnarray*}

2370: 

2371: 

2372: 

2373: 

2374: A simplicial $\IZ\Or G$-module $M_{\bullet}= M_{\bullet}(?)$ gives rise to a $\IZ \Or G$-chain complex

2375: $\DK_{\ast}\!M_{\bullet}$ {\sl via} the Dold-Kan correspondence and determines an $\Or G$-spectrum

2376: $\JH M_{\bullet}$ {\sl via} the Eilenberg-Mc\,Lane functor (see Subsections~\ref{subsec:simpl-dold-kan}

2377: and \ref{subsec:Gamma-spaces}). The following proposition specializes to a well-known fact in the

2378: case where $G$ is the trivial group.

2379: 

2380: \begin{proposition} \label{prop:compare}

2381: Let $M_{\bullet}$ be a functor from $\Or G$ to simplicial abelian groups. Then, there are natural

2382: isomorphisms of $G$-homology theories defined on pairs of $G$-$CW$-complexes

2383: \begin{eqnarray*}

2384: H_\ast^G ( - ; \JH M_{\bullet} )  \xrightarrow{\cong}

2385: H_\ast^G( - ; M_{\bullet} )

2386: \xrightarrow{\cong}

2387: H_\ast^G( - ; \DK_{\ast}\!M_{\bullet} )\,.

2388: \end{eqnarray*}

2389: \end{proposition}

2390: 

2391: 

2392: 

2393: In particular we have natural isomorphisms

2394: \begin{eqnarray} \label{some-isos}

2395: H_\ast^G ( - ; \JH \JX^{\otimes_k}R ) \; \cong \; H_\ast^G( - ; C_{\bullet}^{\HX^{\otimes_{\!k}}\!R}) \; \cong \;

2396: H_\ast^G( - ; C_{\ast}^{\HX^{\otimes_{\!k}}\!R} )\,.

2397: \end{eqnarray}

2398: Here we have used the notation

2399: \begin{eqnarray}

2400: C_{\bullet}^{\HX^{\otimes_{\!k}}\!R}(? ) & = &  C_{\bullet}^{\HX^{\otimes_{\!k}}} R \calg^G(?) \label{def-C-bullet-HXkR} \\

2401: C_{\ast}^{\HX^{\otimes_{\!k}}\!R}(? ) & = & C_{\ast}^{\HX^{\otimes_{\!k}}} R \calg^G(? )\,, \label{def-C-ast-HXkR}

2402: \end{eqnarray}

2403: for the indicated simplicial $\IZ \Or G$-module respectively $\IZ \Or G$-chain complex,

2404: compare Subsection~\ref{define-bicomplexes}, and \eqref{def-HX-OrG-spectrum}.

2405: 

2406: 

2407: \begin{proof}[Proof of Proposition~\ref{prop:compare}]

2408: We discuss the first natural transformation in the absolute case, i.e.\ for $A= \varnothing$ (the general

2409: case is similar).

2410: For a spectrum $\JE$ in simplicial sets we denote by $| \JE |$ the associated (topological) spectrum.

2411: For every $G$-$CW$-complex $X$ and every $\Or G$-spectrum $\JE=\JE(?)$ in simplicial sets,

2412: there is a natural equivalence and, since realization commutes with taking coequalizers, a natural homeomorphism

2413: \[

2414: X_+ \sma_{\Or G} | \JE |

2415: \xleftarrow{\simeq}

2416: |S_{\bullet} X_+ | \sma_{\Or G} | \JE |

2417: \xrightarrow{\cong}

2418: \big|S_{\bullet} X_+ \sma_{\Or G} \JE \big|\,.

2419:  \]

2420: Note that there is an obvious natural isomorphism of spectra

2421: $$

2422: S_{\bullet} X_+ \sma_{\Or G}\IH M_{\bullet}(\JS) \; \cong \; \Big(S_{\bullet} X_+ \sma_{\Or G}\IH M_{\bullet}\Big)(\JS)\,.

2423: $$

2424: Observe also that for an (unpointed) simplicial set $Y_\bullet$, there is an isomorphism of $\Gamma$-spaces

2425: \begin{eqnarray} \label{easy-iso-for-prop-compare}

2426: \IZ[Y_\bullet] \otimes_{\IZ \Or G} \IH M_{\bullet} \; \cong \;

2427: \IH \big( \IZ[Y_\bullet] \otimes_{\IZ \Or G} M_{\bullet} \big)\,.

2428: \end{eqnarray}

2429: By \eqref{homotopy-HM}, the homotopy groups of the spectrum associated to the right hand side are given by the (unstable)

2430: homotopy groups of (the realization of) $\IZ[Y_\bullet] \otimes_{\Or G} M_{\bullet}$\,.

2431: 

2432: So, observing that $\widetilde{\IZ}

2433: (S_\bullet X_+) \cong \IZ[S_\bullet X]$, to produce the first natural transformation of the statement, it will suffice

2434: to define a natural transformation of $\Gamma$-spaces

2435: \begin{eqnarray} \label{nat-trans-for-prop-compare}

2436: S_{\bullet} X_+ \sma_{\Or G} \IH M_{\bullet} \to \widetilde{\IZ}[ S_{\bullet} X_+ ] \otimes_{\IZ \Or G} \IH M_{\bullet}\,.

2437: \end{eqnarray}

2438: More generally, for every contravariant functor $Z_\bullet=Z_\bullet(?)$ from $\Or G$ to pointed simplicial sets and

2439: every covariant functor $N_{\bullet}=N_{\bullet}(?)$ from $\Or G$ to simplicial abelian groups,

2440: we will construct a natural transformation

2441: \[

2442: Z_{\bullet} \sma_{\Or G}  N_{\bullet} \to

2443: \widetilde{\IZ} [ Z_{\bullet} ] \otimes_{\IZ \Or G}  N_{\bullet}\,.

2444: \]

2445: To produce the map we use the following facts. The left-hand side is defined as a coequalizer in the category of pointed

2446: simplicial sets completely analogous to \eqref{balanced-smash}, and the right-hand side similarly as a coequalizer in the category of

2447: simplicial abelian groups.

2448: Let $U$ denote the forgetful functor from simplicial abelian groups to pointed simplicial sets.

2449: For a pointed simplicial set $X$, a simplicial abelian group $A$ and a family $A_i$, $i \in I$ of simplicial abelian groups there are

2450: obvious natural maps $X \sma UA \to U ( \widetilde{\IZ} [ X ] \otimes A )$ and $\bigvee_{i \in I} UA_i \to U(\bigoplus_{i \in I} A_i)$.

2451: Given two maps $f ,g \colon A \to B$ of simplicial abelian groups

2452: there is an obvious natural map $\coequ( Uf , Ug ) \to U \coequ(f,g)$.

2453: Combining these facts one easily constructs the required natural transformation above.

2454: 

2455: 

2456: Now, we show that the first natural transformation of the statement is an isomorphism. If a natural transformation between

2457: $G$-homology theories induces an isomorphism when evaluated on all orbits $G/H$, then it induces an isomorphism for all pairs

2458: of $G$-$CW$-complexes by a well-known argument. Unravelling the construction of the first natural transformation, it suffices

2459: to check that for every orbit $G/H$, the map

2460: $$

2461: %S_{\bullet} (G/H)_+ \sma_{\Or G}\IH M_{\bullet}(\JS) =

2462: \Big(S_{\bullet} (G/H)_+ \sma_{\Or G}\IH M_{\bullet}\Big)(\JS) \to

2463: \IH \big( \widetilde{\IZ}[S_{\bullet} (G/H)_+] \otimes_{\IZ \Or G} M_{\bullet} \big)(\JS)

2464: $$

2465: induced by \eqref{easy-iso-for-prop-compare} and \eqref{nat-trans-for-prop-compare} is a stable weak equivalence.

2466: Before evaluation at $\JS$, both sides are canonically isomorphic to the $\Gamma$-space $\IH M_{\bullet}(G/H)$

2467: by suitable analogues of Lemma~\ref{coyoneda}.

2468: We leave it to the reader to verify that we indeed have $G$-homology

2469: theories here, compare \cite[Lemma~4.2]{Davis-Lueck(1998)}.

2470: 

2471: 

2472: We now construct the second natural transformation of the

2473: statement and prove at the same time that it is an isomorphism.

2474: For a bisimplicial abelian group $A_{\bullet \bullet}$, let

2475: $C_{\ast\ast} ( A_{\bullet \bullet} )$ denote the associated

2476: bicomplex, compare \cite[page~275]{Weibel(1994)}. Note that given

2477: two simplicial abelian groups $C_\bullet$ and $D_\bullet$, there

2478: is a natural isomorphism of bicomplexes $\DK_\ast (C_\bullet)

2479: \otimes_\IZ \DK_\ast(D_\bullet) \cong C_{\ast \ast} (C_\bullet

2480: \otimes_\IZ D_\bullet)$, where $C_\bullet \otimes_\IZ D_\bullet$

2481: is viewed as a bisimplicial abelian group. Note also that

2482: $\DK_\ast \big( \widetilde{\IZ}[ S_\bullet (X) ]\big) =

2483: \widetilde{C}^{\sing}_\ast (X)$, for every space $X$. The

2484: degreewise tensor products of simplicial abelian groups appearing

2485: in the source of the second natural transformation can be thought

2486: of as diagonals of the corresponding bisimplicial sets. Applying

2487: all these observations and using again the definition of the

2488: balanced tensor product in terms of coequalizers, it suffices to

2489: observe that for every pair of maps $f_{\bullet \bullet},

2490: g_{\bullet \bullet} \colon A_{\bullet \bullet} \to B_{\bullet

2491: \bullet}$ of bisimplicial abelian groups, we have the following

2492: chain of isomorphisms

2493: \begin{eqnarray*}

2494: \pi_\ast \big( \big| \coequ ( \diag f_{\bullet \bullet} , \diag g_{\bullet \bullet}) \big| \big)

2495: & \cong & \pi_\ast \big( \big| \diag \coequ ( f_{\bullet \bullet} , g_{\bullet \bullet}) \big| \big) \\

2496: & \cong & H_\ast \big( \Tot^{\oplus} C_{\ast \ast} (\coequ ( f_{\bullet \bullet} , g_{\bullet \bullet}) ) \big) \\

2497: & \cong & H_\ast \big( \coequ ( \Tot^{\oplus} C_{\ast \ast} ( f_{\bullet \bullet}) , \Tot^{\oplus} C_{\ast \ast}( g_{\bullet \bullet}) ) \big)\,,

2498: \end{eqnarray*}

2499: where the second isomorphism is the Eilenberg-Zilber Theorem as formulated in \cite[Theorem~8.5.1 on page~276]{Weibel(1994)}.

2500: \end{proof}

2501: 

2502: 

2503: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

2504: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

2505: \section{Splitting assembly maps}\label{sec:Splitting-assembly-maps}

2506: 

2507: In this section, we prove Theorem~\ref{the:splitHochschild} and Addendum~\ref{add:split-HN},

2508: i.e.\ the splitting and isomorphism results for the assembly maps in Hochschild, cyclic,

2509: periodic cyclic and negative cyclic homology. We begin with the case of Hochschild homology.

2510: 

2511: 

2512: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

2513: \subsection{Splitting the Hochschild homology assembly map} \label{subsec:splitting-HH}

2514: 

2515: Fix a group $G$ and let $S$ be a $G$-set. Recall that $\con G$ denotes the set of conjugacy classes of $G$.

2516: Sending  a $q$-simplex

2517: $(g_0, \ldots , g_q)$ in $\CN_{\bullet} \calg^G(S)$ to the conjugacyclass $\conjclass{g_0 \cdots g_q}$

2518: yields a map of cyclic sets

2519: \[

2520: \CN_{\bullet} \calg^G(S) \to \con G\,.

2521: \]

2522: Here $\con G$ is considered as a  constant cyclic set.

2523: The cyclic nerve decomposes, as a cyclic set, into the disjoint union of the corresponding pre-images, namely

2524: \begin{eqnarray}\label{decomposition}

2525: \CN_{\bullet} \calg^G (S) \; = \; \coprod_{\conjclass{c} \in \con G } \CN_{\bullet \conjclass{c} } \calg^G (S)\,.

2526: \end{eqnarray}

2527: Observe that $\CN_{\bullet \conjclass{c} } \calg^G (G/H) \neq \varnothing$ implies that $\langle c \rangle$ is

2528: subconjugate to $H$.

2529: For every small category $\calc$, we have a natural isomorphism

2530: \[

2531: k\CN_{\bullet} \calc \; \cong \; \CN_{\bullet}^{\otimes_k}   k \calc

2532: \]

2533: and, because of the isomorphism $R \calc \cong R \otimes_k k \calc$, also

2534: \begin{eqnarray} \label{decompose-k-R}

2535: \CN_{\bullet}^{\otimes_k} R \calc \; \cong \; (\CN_{\bullet}^{\otimes_k} k \calc ) \otimes_k ( \CN_{\bullet}^{\otimes_k} R)\,.

2536: \end{eqnarray}

2537: We therefore obtain an induced  decomposition for the $k$-linear cyclic nerve of $R \calg^G (S)$, that we denote by

2538: \begin{eqnarray} \label{linear-splitting}

2539: \CN_{\bullet}^{\otimes_k} R \calg^G (S) & = &

2540: \bigoplus_{\conjclass{c} \in \con G } \CN_{\bullet \conjclass{c}}^{\otimes_k} R \calg^G (S) \\

2541: & = &

2542: \bigoplus_{\conjclass{c} \in \con G } \CN_{\bullet \conjclass{c} }^{\otimes_k}  k\calg^G (S) \otimes_k \CN_{\bullet}^{\otimes} R\,.

2543: \label{decompositionlinear}

2544: \end{eqnarray}

2545: For typographical reasons we introduce the following abbreviation

2546: for the corresponding decomposition of simplicial $\IZ\Or G$-modules\,:

2547: \begin{eqnarray} \label{decomposition of C_bullet^HH}

2548: C_{\bullet}^{\HH^{\otimes_{\!k}}\!R} (?) & = &  \bigoplus_{\conjclass{c} \in \con G }

2549: C_{\bullet \conjclass{c}}^{\HH^{\otimes_{\!k}}\!R} (?)\,;

2550: \end{eqnarray}

2551: see \eqref{def-C-bullet-HXkR} for the notation. Using the identifications \eqref{some-isos}

2552: and the decomposition \eqref{decomposition of C_bullet^HH}, the Hochschild homology generalized assembly map

2553: $$

2554: \xymatrix@C=4em{

2555: H_n^G(\EGF{G}{\calf};\JH \JH^{\otimes_k}R) \ar[r]^-{\assembly} &

2556: H_n^G(\pt;\JH \JH^{\otimes_k}R) \cong \HH_n^{\otimes_k}(RG)

2557:                }

2558: $$

2559: appearing in diagram \eqref{diagram-assembly-dtr} can be identified with the upper horizontal map in the following

2560: commutative diagram

2561: \begin{eqnarray} \label{proof-diagram}

2562: \hspace*{2.3em}

2563: \vcenter{\xymatrix@C=3.8em{

2564: H_{n}^G\big(\EGF{G}{\calf} ; \bigoplus_{\conjclass{c} \in \con G} C_{\bullet \conjclass{c}}^{\HH^{\otimes_{\!k}}\!R}\big)

2565: \ar[d]_{\id \otimes pr_{\calf}} \ar[r]^-{\assembly}

2566: &

2567: H_{n}^G\big( \pt  ; \bigoplus_{\conjclass{c} \in \con G} C_{\bullet \conjclass{c}}^{\HH^{\otimes_{\!k}}\!R}\big)

2568: \ar[d]^{\id \otimes pr_{\calf}}

2569: \\

2570: H_{n}^G\big(\EGF{G}{\calf} ; \smash[b]{\bigoplus_{\stackrel{\conjclass{c} \in \con G}{\langle c \rangle \in \calf}}}

2571: C_{\bullet \conjclass{c}}^{\HH^{\otimes_{\!k}}\!R}\big) \ar[r]

2572: &

2573: H_{n}^G\big( \pt ; \smash[b]{\bigoplus_{\stackrel{\conjclass{c} \in \con G}{\langle c \rangle \in \calf}}}

2574: C_{\bullet \conjclass{c}}^{\HH^{\otimes_{\!k}}\!R}\big).

2575:          }}

2576: \end{eqnarray}

2577: 

2578: \smallskip

2579: 

2580: \noindent

2581: Here, the vertical maps are induced by the projection $pr_{\calf}$ onto the summands for which the cyclic subgroup

2582: $\langle c \rangle$ belongs to the family $\calf$. Note that $pr_{\calf}$ is the identity map if $\calf$ contains

2583: all cyclic subgroups of $G$. The statement about Hochschild homology in Theorem~\ref{the:splitHochschild} now follows

2584: directly from the following two lemmas.

2585: 

2586: \begin{lemma} \label{leftvertical}

2587: For every family $\calf$, the left vertical map in \eqref{proof-diagram} is an isomorphism.

2588: \end{lemma}

2589: 

2590: \begin{lemma} \label{lowerhorizontal}

2591: For every family $\calf$, the bottom map in \eqref{proof-diagram} is an isomorphism.

2592: \end{lemma}

2593: 

2594: The proofs of Lemmas~\ref{leftvertical} and \ref{lowerhorizontal} will occupy the rest of this subsection. They rely on

2595: the following computation of the cyclic nerve of a transport groupoid.

2596: 

2597: \smallskip

2598: 

2599: Let $E_{\bullet} G$ be the simplicial set given by $N_{\bullet} \calg^G( G/1 )$. In words\,: consider $G$ as a category with

2600: $G$ as set of objects and precisely one morphism between any two objects, and then take the nerve of this category.

2601: This is a simplicial model for the universal free $G$-space which is usually denoted by $EG$.

2602: For $c \in G$ let $\langle c \rangle$ be the cyclic subgroup generated by $c$.

2603: For $h \in N_G \langle c \rangle$, let $R_h \in \map_G( G / \langle c \rangle, G/ \langle c \rangle )$

2604: be the map given by $R_h ( g \langle c \rangle ) = gh \langle c \rangle$. For every $G$-set $S$,

2605: precomposing with

2606: $R_h$ yields a left action of $Z_G \langle c \rangle \leq N_G \langle c \rangle$

2607: on $\map_G (G/\langle c \rangle , S)$.

2608: 

2609: \begin{proposition} \label{keyproposition}

2610: For a group $G$, choose a representative $c \in \conjclass{c}$ for each conjugacy class $\conjclass{c} \in \con G$.

2611: Let $\langle c \rangle$ denote the cyclic subgroup it generates.

2612: There is a map of $\Or G$-simplicial sets (depending on the choice)

2613: \[

2614: \coprod_{\conjclass{c} \in \con G}

2615: E_{\bullet} Z_G \langle c \rangle

2616: \times_{Z_G \langle c \rangle}

2617: \map_{G}( G/ \langle c \rangle , ? )

2618: \to \CN_{\bullet} \calg^{G}( ? )\,.

2619: \]

2620: This map is objectwise a simplicial homotopy equivalence, and is compatible with the decomposition

2621: \eqref{decomposition} of the target.

2622: \end{proposition}

2623: 

2624: \begin{remark}

2625: There seems to be no obvious cyclic structure on the source of the map above.

2626: \end{remark}

2627: 

2628: \begin{proof}[Proof of Proposition \ref{keyproposition}]

2629: We first introduce some more notation. Given a groupoid $\calg$, we denote by $\aut \calg$ its

2630: category of automorphisms, i.e.\ the category whose objects are automorphisms $h\colon s \to s$

2631: in $\calg$ and where a morphism from $h\colon s \to s$ to $h'\colon t \to t$

2632: is given by a morphism $g\colon s \to t$ satisfying $h' \circ g = g \circ h$. In the case where

2633: $\calg=\calg^G (S)$, the conjugacy class $\conjclass{h} \in \con G$ associated to an object $h\colon s

2634: \to s$ in $\aut \calg^G ( S )$ does only depend on the isomorphism class of this object.

2635: This yields a well-defined map of simplicial sets

2636: \[

2637: N_{\bullet} \aut \calg^G (S) \; \longrightarrow \; \con G\;,\qquad

2638: \vcenter{\xymatrix@R=1.5em@C=1.9em{

2639: s_0 \ar[d]_{h_0} & s_1  \ar[d]_{h_1} \ar[l]_-{g_0} & \downdots \ar[l]_-{g_1} & s_q \ar[d]_{h_q} \ar[l]_-{\,g_{q-1}} \\

2640: s_0 & s_1 \ar[l]^-{g_0} & \downdots \ar[l]^-{g_1} & s_q \ar[l]^-{\,g_{q-1}}.

2641:          }}

2642: \quad\longmapsto\quad\conjclass{h_0}\,,

2643: \]

2644: where $\con G$ is considered as a constant simplicial set.

2645: Let $\aut_{\conjclass{c}} \calg^G (S)$ denote the full subcategory of $\aut \calg^G (S)$ on the objects $h\colon s

2646: \to  s$ with $h \in \conjclass{c}$.

2647: The decomposition of the nerve into pre-images

2648: under the map to $\con G$ above is given by

2649: \[

2650: N_{\bullet} \aut \calg^G (S) \; = \; \coprod_{\conjclass{c} \in \con G }  N_{\bullet} \aut_{\conjclass{c}} \calg^G (S )\,.

2651: \]

2652: The components of the map in Proposition~\ref{keyproposition} are obtained

2653: as the

2654: composition of the three maps

2655: \[

2656: \xymatrix{

2657: E_{\bullet} Z_G \langle c \rangle

2658: \times_{Z_G \langle c \rangle}

2659: \map_G( G/ \langle c \rangle , ? )

2660: \ar[r] &

2661: N_{\bullet} \calg^{Z_G \langle c \rangle} \big( \map_G ( G/ \langle c \rangle , ? ) \big)  \ar[dl] \\

2662: N_{\bullet} \aut_{\conjclass{c}} \calg^G ( ? )

2663: \ar[r] & \CN_{\bullet \conjclass{c}} \calg^G( ? )

2664:          }

2665: \]

2666: which are constructed in

2667: the following lemma.  Proposition~\ref{keyproposition} is an immediate consequence of that lemma.

2668: \end{proof}

2669: 

2670: \begin{lemma}

2671: Let $G$ be a group and $S$ a $G$-set.

2672: \begin{enumerate}

2673: \item

2674: There is a simplicial isomorphism $E_{\bullet} G \times_G S \to N_{\bullet} \calg^G (S)$.

2675: \item

2676: For $\conjclass{c} \in \con G$, choose a representative $c \in \conjclass{c}$. There is an equivalence of categories

2677: \[

2678: \calg^{Z_G \langle c \rangle} \big( \map_G ( G / \langle c \rangle , S ) \big) \to \aut_{\conjclass{c}}

2679: \calg^G ( S )\,,

2680: \]

2681: which depends on the choice.

2682: \item

2683: For every groupoid $\calg$, there is a simplicial isomorphism

2684: \[

2685: N_{\bullet} \aut \calg  \to \CN_{\bullet} \calg\,.

2686: \]

2687: If $\calg=\calg^G (S)$ then the isomorphism commutes with the maps to $\con G$.

2688: \end{enumerate}

2689: All three constructions are natural with respect to $S$.

2690: \end{lemma}

2691: 

2692: \begin{proof}

2693: (i) The isomorphism $E_\bullet G \times_G S \to N_\bullet \calg^G(S)$ is given, on level $q$, by

2694: \[

2695: \xymatrix@C=2.5em{

2696: \Big[\, g_0  &  g_1 \ar[l]_-{\;g_0 g_1^{-1} } & \downdots \ar[l]_-{\;g_1 g_2^{-1}} & g_q \ar[l]_-{\;g_{q-1} g_q^{-1}}

2697: \, , \, s \,\Big]

2698: \;\;\longmapsto\;\;

2699: \Big(g_0 s &  g_1 s \ar[l]_-{\;g_0 g_1^{-1} } & \downdots \ar[l]_-{\;g_1 g_2^{-1}} & g_q s \ar[l]_-{\;g_{q-1} g_q^{-1}}\Big)\,.

2700:          }

2701: \]

2702: 

2703: \smallskip

2704: 

2705: (ii) The functor sends an object $\phi \in \map_G ( G/\langle c \rangle , S )$ to the automorphism

2706: $c \colon  \phi(e \langle c \rangle ) \to \phi(e \langle c \rangle ) $. Here $e$ is the trivial

2707: element in $G$ and $c$ the chosen representative in $\conjclass{c}$. A morphism $z \colon \phi \to z \phi$, with

2708: $z\in Z_G \langle c \rangle$, is taken to the (iso)morphism $z^{-1} \colon \phi(e \langle c \rangle ) \to z^{-1} \phi

2709: ( e \langle c \rangle )$. The functor is full and faithful and every object

2710: in the target category is isomorphic to an image object.

2711: 

2712: \smallskip

2713: 

2714: (iii) The isomorphism $N_\bullet \aut \calg  \to \CN_\bullet \calg$ is given, on level $q$, by

2715: \[

2716: \;\;\vcenter{\xymatrix@R=1.5em@C=1.9em{

2717: s_0 \ar[d]_{h_0} &

2718: s_1  \ar[d]_{h_1} \ar[l]_-{g_0} &

2719: \downdots  \ar[l]_-{g_1} &

2720: s_q \ar[d]_{h_q} \ar[l]_-{\,g_{q-1}} &

2721: \ar@{}[d]^{\longmapsto} \\

2722: s_0 &

2723: s_1 \ar[l]^-{g_0} &

2724: \downdots \ar[l]^-{g_1} &

2725: s_q \ar[l]^-{\,g_{q-1}}  &

2726:          }}

2727: \qquad\;

2728: \vcenter{\xymatrix@R=1.5em@C=1.9em{

2729: \!\!s_0  \ar@(dl,dr)[rrr]_{ h_0(g_0 \cdots g_{q-1})^{-1} }  & s_1 \ar[l]_-{g_0} & \downdots \ar[l]_-{g_1} &

2730: s_q \ar[l]_-{\,g_{q-1}}\,.\!\!\!\!\! \\

2731:          }}

2732: \]

2733: The compatibility with the maps to $\con G$ is clear.

2734: \end{proof}

2735: 

2736: The following is the linear analogue of Proposition~\ref{keyproposition}.

2737: 

2738: \begin{corollary} \label{keypropositionlinear}

2739: For every conjugacy class $\conjclass{c} \in \con G$ there is

2740: natural transformation of functors from the orbit category

2741: $\Or G$ to the category of simplicial $k$-modules,

2742: \[

2743: k [E_{\bullet} Z_G \langle c \rangle]

2744: \otimes_{k Z_G\langle c \rangle }

2745: k \map( G/\langle c \rangle , ? )

2746: \otimes_k

2747: \CN_{\bullet}^{\otimes_k} R

2748: \to

2749: \CN_{\bullet \conjclass{c}}^{\otimes_k} R \calg^G( ? )

2750: \]

2751: which is objectwise a homotopy equivalence.

2752: \end{corollary}

2753: \begin{proof}

2754: Apply the functor free $k$-module $k(-)$ to the map in Proposition~\ref{keyproposition} and recall the identification

2755: \eqref{decompose-k-R}.

2756: \end{proof}

2757: 

2758: Observe that we have a decomposition of $G$-homology theories

2759: \begin{eqnarray} \label{decomp}

2760: H_\ast^G \big( - ; C_{\bullet}^{\HH^{\otimes_{\!k}}\!R} \big)

2761: %\cong H_\ast^G \Big( - ; \bigoplus_{\conjclass{c} \in \con G } C_{\bullet  \conjclass{c}}^{\HH^{\otimes_{\!k}}\!R} \Big)

2762: \cong \bigoplus_{\conjclass{c} \in \con G } H_\ast^G \big( - ; C_{\bullet  \conjclass{c}}^{\HH^{\otimes_{\!k}}\!R} \big)\,,

2763: \end{eqnarray}

2764: because the tensor product over the orbit category and homology both commute with direct sums.

2765: For each of the summands, we have the following computation.

2766: 

2767: \begin{proposition} \label{computation-one-summand}

2768: For every $G$-$CW$-complex $X$ and every $\conjclass{c} \in \con G$, there is a natural isomorphism

2769: \[

2770: H_{\ast}^G \big( X ; C_{\bullet \conjclass{c}}^{\HH^{\otimes_{\!k}}\!R} \big) \; \cong \;

2771: H_\ast \big( X^{\langle c \rangle} \times_{Z_G \langle c \rangle} E Z_G {\langle c \rangle} ;

2772: \CN_{\bullet}^{\otimes_k} R \big)\,.

2773: \]

2774: \end{proposition}

2775: 

2776: \begin{proof}

2777: On the level of simplicial abelian groups, Corollary~\ref{keypropositionlinear}, in combination with

2778: Lemma~\ref{coyoneda}, yields

2779: $$

2780: \renewcommand{\arraystretch}{1.7}

2781: \begin{array}{l}

2782: \widetilde{\IZ} [S_{\bullet} X^?_+ ] \otimes_{\IZ \Or G} C_{\bullet \conjclass{c}}^{\HH^{\otimes_{\!k}}\!R} (?) \; \simeq

2783: \qquad\qquad\quad \\

2784: \qquad\qquad\quad \simeq \; \widetilde{\IZ} [ S_{\bullet} X^?_+ ] \otimes_{\IZ \Or G } k[ E_{\bullet} Z_G \langle c \rangle ]

2785: \otimes_{k Z_G \langle c \rangle} k \map ( G / \langle c \rangle , ? ) \otimes_k \CN_{\bullet}^{\otimes_k} R \\

2786: \qquad\qquad\quad \cong \; \widetilde{\IZ} [ S_{\bullet} X^{\langle c \rangle}_+ ] \otimes_{\IZ Z_G \langle c \rangle }

2787: \IZ [ E_{\bullet} Z_G \langle c \rangle ] \otimes_{\IZ} \CN_{\bullet}^{\otimes_k} R\,,

2788: \end{array}

2789: $$

2790: hence the result.

2791: \end{proof}

2792: 

2793: \begin{lemma} \label{coyoneda}

2794: Let $F$ be a contravariant functor from $\Or G$ to simplicial $k$-modules. Then, for every subgroup $H\leq G$,

2795: there is a natural isomorphism

2796: \[

2797: F(?) \otimes_{k \Or G} k \map_G( G/H , ? ) \; \cong \; F(G/H)\,.

2798: \]

2799: \end{lemma}

2800: 

2801: We can now finish the proof of the part of Theorem~\ref{the:splitHochschild} concerned with Hochschild homology.

2802: 

2803: \begin{proof}[Proof of Lemmas~\ref{leftvertical} and \ref{lowerhorizontal}]

2804: Compute the relevant maps in diagram \eqref{proof-diagram} using \eqref{decomp} and Proposition~\ref{computation-one-summand}.

2805: Observe that by the very definition of $\EGF{G}{\calf}$, we have $\EGF{G}{\calf}^{\langle c \rangle} = \varnothing$

2806: if and only if $\langle c \rangle \in \calf$. So the projection $\id \otimes pr_{\calf}$ is the zero map

2807: exactly on those summand which are anyway trivial. This proves Lemma~\ref{leftvertical}. For

2808: $\langle c \rangle \in \calf$, the map $\EGF{G}{\calf}^{\langle c \rangle} \to \pt$ is a homotopy equivalence. Therefore,

2809: $\EGF{G}{\calf}^{\langle c \rangle} \times EZ_G \langle c \rangle \to\pt \times EZ_G \langle c \rangle$ is an equivalence

2810: of free $Z_G \langle c \rangle$-spaces and hence remains an equivalence if we quotient out the $Z_G \langle c \rangle$-action.

2811: This establishes Lemma~\ref{lowerhorizontal}.

2812: \end{proof}

2813: 

2814: The following example gives a further illustration of the computation achieved above.

2815: 

2816: \begin{example} \label{rem:comparison_with_standard_splitting}

2817: Combining \eqref{some-isos}, the isomorphism \eqref{decomp} and Proposition~\ref{computation-one-summand},

2818: we get, for every $G$-$CW$-complex $X$, a decomposition

2819: $$H_\ast^G(X;\JH \JH^{\otimes_k}R) \; \cong \;  \bigoplus_{\conjclass{c} \in \con G }

2820: H_\ast\big( X^{\langle c \rangle} \times_{Z_G \langle c \rangle} E Z_G {\langle c \rangle} ;

2821: C_\ast^{\HH^{\otimes_{\!k}}}( R) \big)\,,$$

2822: where each direct summand on the right-hand side is the (non-equivariant) hyper\-homology of the space

2823: $X^{\langle c \rangle} \times_{Z_G \langle c \rangle} E Z_G {\langle c \rangle}$

2824: with coefficients in the Hochschild complex, i.e.\ in the $k$-chain complex

2825: $C_\ast^{\HH^{\otimes_{\!k}}} (R) = \DK_{\ast} ( \CN_{\bullet}^{\otimes_k} R )$.  In the case $R = k$, the degree zero inclusion

2826: $k \to C^{\HH^{\otimes_{\!k}}} (k)$ is a homology equivalence and hence a chain homotopy equivalence,

2827: because both complexes are bounded below and consist of projective $k$-modules. Thus, we infer

2828: \[

2829: H_\ast\big( X^{\langle c \rangle} \times_{Z_G \langle c \rangle} E Z_G {\langle c \rangle} ; C_\ast^{\HH^{\otimes_{\!k}}} (k) \big)

2830: \; \cong \; H_\ast\big(X^{\langle c \rangle} \times_{Z_G \langle c \rangle} E Z_G {\langle c \rangle};k\big)

2831: \]

2832: for each conjugacy class $\conjclass{c}$, and therefore

2833: $$H_\ast^G \big( X;\JH \JH^{\otimes_k}R \big) \; \cong \;

2834:  \bigoplus_{\conjclass{c} \in \con G }

2835: H_\ast \big( X^{\langle c \rangle} \times_{Z_G \langle c \rangle} E Z_G {\langle c \rangle};k \big)\,.$$

2836: In the special case where $X = \pt$ and $R = k$, we rediscover the well-known decomposition

2837: of $k$-modules

2838: \begin{eqnarray}

2839: \label{standard-decomposition_of_HH(KG)}

2840: \HH^{\otimes_k}(kG) &  \cong &

2841:  \bigoplus_{\conjclass{c} \in \con G }  H_\ast \big( B Z_G {\langle c \rangle};k \big)\,.

2842: \end{eqnarray}

2843: If we insert $X = \EGF{G}{\calf}$ for an arbitrary family of subgroups $\calf$, we obtain

2844: \[

2845: H_\ast^G \big( \EGF{G}{\calf};\JH \JH^{\otimes_k}R \big) \; \cong \;

2846:  \bigoplus_{ \substack{\conjclass{c} \in \con G \\ \langle c \rangle \in \calf}}  H_\ast \big( B Z_G {\langle c \rangle};k \big)\,,

2847: \]

2848: because $\EGF{G}{\calf}^{\langle c \rangle} \times EZ_G \langle c \rangle$ is a model for $E Z_G \langle c \rangle$ if

2849: $\langle c \rangle \in \calf$ and is empty otherwise. The map $\EGF{G}{\calf} \to \pt$ induces the obvious inclusion.

2850: \end{example}

2851: 

2852: \begin{remark}

2853: Of course one does not need the elaborate setup using spectra, nor Theorem~\ref{the:splitHochschild},

2854: in order to prove the well-known decomposition \eqref{standard-decomposition_of_HH(KG)}. But our aim was to compare

2855: the Hochschild assembly map with the one for K-theory. There is no chain complex version of the assembly map

2856: on the level of $K$-theory. Furthermore, our effort has the pay off that it can be generalized to topological

2857: Hochschild homology and its refinements as explained in \cite{Lueck-Reich-Rognes-Varisco(2005)}.

2858: \end{remark}

2859: 

2860: 

2861: 

2862: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

2863: \subsection{Splitting cyclic, periodic cyclic and negative cyclic assembly maps} \label{subsec:splitting-HN-HP-assembly}

2864: 

2865: 

2866: Observe that the sum decomposition \eqref{linear-splitting} is compatible with the cyclic structure.

2867: Keeping notation as in \eqref{def-C-ast-HXkR}, we hence obtain a decomposition

2868: \begin{eqnarray*}

2869: C_{\ast}^{\HX^{\otimes_{\!k}}\!R}(?) & = & \bigoplus_{\conjclass{c} \in \con G }

2870: C_{\ast \conjclass{c}}^{\HX^{\otimes_{\!k}}\!R}(?)

2871: \end{eqnarray*}

2872: of $\IZ\Or G$-chain complexes. Compare with the splitting \eqref{decomposition of C_bullet^HH}.

2873: There is consequently a version of diagram \eqref{proof-diagram}

2874: with $C_{\bullet \conjclass{c}}^{\HH^{\otimes_{\!k}}\!R}$ replaced everywhere by $C_{\ast \conjclass{c}}^

2875: {\HX^{\otimes_{\!k}}\!R}$, and where the upper horizontal map corresponds to the generalized assembly map

2876: for $\HX$-homology.

2877: In order to prove the cyclic homology

2878: part of Theorem~\ref{the:splitHochschild} and to establish Addendum~\ref{add:split-HN},

2879: it suffices to obtain the analogues of Lemmas~\ref{leftvertical} and \ref{lowerhorizontal} with $C_{\bullet

2880: \conjclass{c}}^{\HH^{\otimes_{\!k}}\!R}$ replaced everywhere by $C_{\ast \conjclass{c}}^{\HX^{\otimes_{\!k}}\!R}$.

2881: However, this follows immediately from the following proposition.

2882: 

2883: 

2884: \begin{proposition}

2885: Let $X \to X^{\prime}$ be a map of $G$-$CW$-complexes and $Z_{\bullet}(?) \to Z_{\bullet}^{\prime}(?)$

2886: a map of \emph{cyclic $\IZ\Or G$-modules}, i.e.\ a natural transformation between functors from the orbit

2887: category $\Or G$ to the category of cyclic abelian groups. Keep notation as at the

2888: beginning of Subsection~\ref{define-bicomplexes}.

2889: 

2890: If the induced map

2891: \begin{eqnarray*}

2892: H^G_{\ast} \big( X ; C_{\ast}^{\HH} ( Z_{\bullet} ) \big) \to H^G_{\ast} \big( X^{\prime} ;

2893: C_{\ast}^{\HH} ( Z_{\bullet}^{\prime} ) \big)

2894: \end{eqnarray*}

2895: is an isomorphism, then the map

2896: \begin{eqnarray*}

2897: H^G_{\ast} \big( X ; C_{\ast}^{\HC} ( Z_{\bullet} ) \big) \to

2898: H^G_{\ast} \big( X^{\prime} ; C_{\ast}^{\HC} ( Z_{\bullet}^{\prime} ) \big)

2899: \end{eqnarray*}

2900: is an isomorphism. If moreover $X$ and $X^{\prime}$ are finite $G$-$CW$-complexes, then also

2901: the maps

2902: \begin{eqnarray*}

2903: H^G_{\ast} \big( X ; C_{\ast}^{\HP} ( Z_{\bullet} ) \big) \xrightarrow{\cong}

2904: H^G_{\ast} \big( X^{\prime} ; C_{\ast}^{\HP} ( Z_{\bullet}^{\prime} ) \big), \\

2905: H^G_{\ast} \big( X ; C_{\ast}^{\HN} ( Z_{\bullet} ) \big) \xrightarrow{\cong}

2906: H^G_{\ast} \big( X^{\prime} ; C_{\ast}^{\HN} ( Z_{\bullet}^{\prime} ) \big)\,

2907: \end{eqnarray*}

2908: are isomorphisms.

2909: \end{proposition}

2910: \begin{proof}

2911: There is a short exact sequence of chain complexes

2912: \[

2913: 0 \to C_{\ast}^{\HH} (Z_{\bullet}) \to C_{\ast}^{\HC} (Z_{\bullet}) \to

2914: C_{\ast}^{\HC} (Z_{\bullet}) [-2] \to 0\,,

2915: \]

2916: which is natural in $Z_{\bullet}$, see \cite[2.5.10 on pages 78--79]{Loday(1992)}. We use here the notation $C_{\ast}[r]$

2917: for the chain complex which is shifted down $r$ steps, i.e.\ $(C_{\ast}[r])_n = C_{n+r}$. Since $\Tot^{\oplus}$ and

2918: $\widetilde{C}_{\ast}^{\cell} ( X^?_+) \otimes_{\IZ \Or G} (-) $ are exact functors (we use here that $\widetilde

2919: {C}_{\ast}^{\cell} ( X^?_+)$ is a free $\IZ \Or G$-module), the maps induced by $X \to X^{\prime}$ and $Z_\bullet

2920: \to Z_\bullet^{\prime}$ lead to a short exact ladder diagram of chain complexes. The corresponding long

2921: exact ladder in homology, the fact that $H_{\ast}^G ( X ; C_{\ast}^{\HH}( Z_{\bullet} ))$ and $H_{\ast}^G ( X ;

2922: C_{\ast}^{\HC}( Z_{\bullet} ))$ are concentrated in non-negative degrees

2923: and an easy inductive argument based on the Five-Lemma finish the proof for cyclic homology.

2924: 

2925: \smallskip

2926: 

2927: In order to prove the statement for periodic cyclic homology, one uses that the periodic cyclic complex can be considered

2928: as the inverse limit of the tower of cyclic complexes

2929: \begin{eqnarray} \label{tower}

2930: \ldots \to C_{\ast}^{\HC}( Z_{\bullet} ) [4] \to C_{\ast}^{\HC}( Z_{\bullet} ) [2] \to

2931: C_{\ast}^{\HC}( Z_{\bullet} ) [0]\,.

2932: \end{eqnarray}

2933: For $n\geq 0$, we have the following natural maps\,:

2934: \begin{eqnarray*}

2935: H_{n}^{G} \big(X ; C_{\ast}^{\HP} ( Z_{\bullet} ) \big) & \cong &

2936: H_n \Big( \Tot^{\oplus} \big( \widetilde{C}_{\ast}^{\cell} ( X^?_+) \otimes_{\IZ \Or G} \lim_r C_{\ast}^{\HC}

2937: (Z_{\bullet}) [2r] \big) \Big) \\[.3em]

2938: & \to &

2939: H_n \Big( \lim_r \, \Tot^{\oplus} \big( \widetilde{C}_{\ast}^{\cell} ( X^?_+) \otimes_{\IZ \Or G} C_{\ast}^{\HC}

2940: (Z_{\bullet}) [2r] \big) \Big) \\[.3em]

2941: & \to & \lim_r \,

2942: H_n \Big( \Tot^{\oplus} \big( \widetilde{C}_{\ast}^{\cell} ( X^?_+) \otimes_{\IZ \Or G} C_{\ast}^{\HC} (Z_{\bullet})

2943: [2r] \big) \Big) \\[.3em]

2944: & \cong & \lim_r \, H_{n + 2r}^G \big( X ; C_{\ast}^{\HP} ( Z_{\bullet} ) \big)\,.

2945: \end{eqnarray*}

2946: In Lemma~\ref{lim-lemma} below, we show that the first map is an isomorphism if $X$ is a finite $G$-$CW$-complex.

2947: The second map sits in a short exact $\lim^1$-$\lim$-sequence, because the maps in the tower \eqref{tower}

2948: above are all surjective and the functors

2949: $\widetilde{C}_{\ast}^{\cell}(X^?_+) \otimes_{\IZ \Or G} (-)$ and $\Tot^{\oplus}$ preserve surjectivity, compare

2950: \cite[Theorem~3.5.8 on page~83]{Weibel(1994)}.

2951: Since we already know the comparison result for the $\lim$- and $\lim^1$-terms involving cyclic homology, a Five-Lemma argument

2952: yields the result for periodic cyclic homology.

2953: 

2954: \smallskip

2955: 

2956: It remains to prove the statement about negative cyclic homology.

2957: There is a natural exact sequence of chain complexes \cite[5.1.4 on page~160]{Loday(1992)}

2958: \[

2959: 0 \to

2960: C_{\ast}^{\HN} ( Z_{\bullet} ) \to

2961: C_{\ast}^{\HP} ( Z_{\bullet} ) \to

2962: C_{\ast}^{\HC} ( Z_{\bullet} )[-2] \to

2963: 0\,.

2964: \]

2965: Again, one uses that $\widetilde{C}_{\ast}^{\cell}(X^?_+) \otimes_{\IZ \Or G} (-)$ and $\Tot^{\oplus}$ are exact

2966: functors to produce a long exact ladder in homology and uses the Five-Lemma.

2967: \end{proof}

2968: 

2969: In the previous proof we used the following statement.

2970: 

2971: \begin{lemma} \label{lim-lemma}

2972: Suppose that $X$ is a finite $G$-$CW$-complex.

2973: Then the natural map

2974: {\small\begin{eqnarray*}

2975: \;\;\Tot^{\oplus} \Big( \widetilde{C}_{\ast}^{\cell}(X^?_+) \otimes_{\IZ \Or G} \lim_r C_{\ast}^{\HC} (Z_{\bullet}) [2r] \Big)

2976: \xrightarrow{\cong}\;

2977: \lim_r\, \Tot^{\oplus} \big( \widetilde{C}_{\ast}^{\cell}(X^?_+) \otimes_{\IZ \Or G} C_{\ast}^{\HC} (Z_{\bullet}) [2r] \big)

2978: \end{eqnarray*}}

2979: \hspace*{-.3em}is an isomorphism.

2980: \end{lemma}

2981: 

2982: \begin{proof}

2983: There exists an exact sequence (by explicit construction of an inverse limit)

2984: \[

2985: 0 \to

2986: \lim_r C_{\ast}^{\HC}( Z_{\bullet} ) [2r]

2987: \to

2988: \prod_{r = 0}^{\infty} C_{\ast}^{\HC}( Z_{\bullet} ) [2r]

2989: \to

2990: \prod_{r = 0}^{\infty} C_{\ast}^{\HC}( Z_{\bullet} ) [2r]\,.

2991: \]

2992: As $\widetilde{C}_{\ast}^{\cell}(X^?_+) \otimes_{\IZ \Or G} (-)$ and $\Tot^{\oplus}$ are exact

2993: functors, we see that it suffices to study the natural map

2994: \[

2995: \xymatrix{

2996: \Tot^{\oplus} \big( \widetilde{C}_{\ast}^{\cell}(X^?_+) \otimes_{\IZ \Or G}

2997: \prod_{r = 0}^{\infty} C_{\ast}^{\HC}( Z_{\bullet} ) [2r] \big)

2998: \ar[d] \\

2999: \prod_{r = 0}^{\infty} \Tot^{\oplus} \big( \widetilde{C}_{\ast}^{\cell}(X^?_+) \otimes_{\IZ \Or G}

3000:  C_{\ast}^{\HC}( Z_{\bullet} ) [2r] \big)

3001: }

3002: \]

3003: Let $C_\ast^{{\scriptscriptstyle\leq p}} \subseteq \widetilde{C}_{\ast}^{\cell}(X^?_+)$ be the $\Or G$-sub-complex

3004: which agrees with $\widetilde{C}_{\ast}^{\cell}(X^?_+)$ up to dimension $p$ and is trivial in dimension $>p$. This

3005: yields a finite filtration by our assumption on $X$. There is an induced map of filtered chain complexes

3006: \[

3007: \xymatrix{

3008: F_\ast^p = \Tot^{\oplus} \big( C_\ast^{{\scriptscriptstyle\leq p}} \otimes_{\IZ \Or G}

3009: \prod_{r = 0}^{\infty} C_{\ast}^{\HC}( Z_{\bullet} ) [2r] \big)

3010: \ar@<1em>[d] \\

3011: {}^{\prime}\!F_\ast^p = \prod_{r = 0}^{\infty} \Tot^{\oplus} \big( C_\ast^{{\scriptscriptstyle\leq p}} \otimes_{\IZ \Or G}

3012:  C_{\ast}^{\HC}( Z_{\bullet} ) [2r] \big)

3013: }

3014: \]

3015: and the induced chain map of filtration quotients $F_\ast^p / F_\ast^{p-1} \to {}^{\prime}\!F_\ast^p \big/

3016: {}^{\prime}\!F_\ast^{p-1}$ can be identified with the composition

3017: \[

3018: \xymatrix{

3019: \Tot^{\oplus} \big( \widetilde{C}_{p}^{\cell}(X^?_+) \otimes_{\IZ \Or G}

3020: \prod_{r = 0}^{\infty} C_{\ast}^{\HC}( Z_{\bullet} ) [2r] \big)

3021: \ar[d] \\

3022: \Tot^{\oplus} \big( \prod_{r = 0}^{\infty} \widetilde{C}_{p}^{\cell}(X^?_+) \otimes_{\IZ \Or G}

3023: C_{\ast}^{\HC}( Z_{\bullet} ) [2r] \big)

3024: \ar[d] \\

3025: \prod_{r = 0}^{\infty} \Tot^{\oplus} \big( \widetilde{C}_{p}^{\cell}(X^?_+) \otimes_{\IZ \Or G}

3026:  C_{\ast}^{\HC}( Z_{\bullet} ) [2r] \big)

3027: }

3028: \]

3029: because the tensor product with a fixed module over the orbit category, $\Tot^{\oplus}$ and $\prod_{r=0}^{\infty}$ all ``behave

3030: well'' (in an obvious sense) with respect to taking quotients. The second map in the composition above is clearly an isomorphism.

3031: The first map is an isomorphism because the assumption on $X$ implies that each $\widetilde{C}_{p}^{\cell}(X^?_+)$ is

3032: a finitely generated free $\IZ \Or G$-module, compare \cite[page~167]{Lueck(1989)}. Since the filtrations are finite, this

3033: concludes the proof.

3034: \end{proof}

3035: 

3036: \begin{remark}

3037: If we would only assume that $X$ is a $G$-$CW$-complex of finite type instead of being finite, then one would have the same

3038: conclusion that the induced map of filtration quotients is an isomorphism for each $p$, as in the proof of Lemma~\ref{lim-lemma},

3039: but the second filtration would not necessarily be exhaustive. The $0$th module of the complex

3040: \[

3041: \prod_{r = 0}^{\infty} \Tot^{\oplus} \big( \widetilde{C}_{\ast}^{\cell}(X^?_+) \otimes_{\IZ \Or G}

3042:  C_{\ast}^{\HC}( Z_{\bullet} ) [2r] \big)

3043: \]

3044: would for instance contain the infinite product $\prod_{r=0}^{\infty} \widetilde{C}_{2r}^{\cell}(X^?_+)

3045: \otimes_{\IZ \Or G} C_0^{\HC} ( Z_{\bullet} )$, whereas an element that is contained in ${}^{\prime}\!F_\ast^p$

3046: for some $p$ has to be contained in the corresponding infinite direct sum.

3047: \end{remark}

3048: 

3049: 

3050: \bibliographystyle{plain}

3051: \bibliography{dbdef,dbpub,dbpre,dbkhhextra}

3052: 

3053: 

3054: \end{document}

3055: