100% found this document useful (5 votes)
31 views84 pages

Buy Ebook Groups Rings and Group Rings 1st Edition Antonio Giambruno Cheap Price

ebook

Uploaded by

meriamersn
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
100% found this document useful (5 votes)
31 views84 pages

Buy Ebook Groups Rings and Group Rings 1st Edition Antonio Giambruno Cheap Price

ebook

Uploaded by

meriamersn
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 84

Full download ebook at ebookname.

com

Groups rings and group rings 1st Edition Antonio


Giambruno

https://ebookname.com/product/groups-rings-and-group-
rings-1st-edition-antonio-giambruno/

OR CLICK BUTTON

DOWLOAD NOW

Download more ebook from https://ebookname.com


More products digital (pdf, epub, mobi) instant
download maybe you interests ...

Algebra Groups Rings and Fields 1st Edition Louis Halle


Rowen

https://ebookname.com/product/algebra-groups-rings-and-
fields-1st-edition-louis-halle-rowen/

A Guide to Groups Rings and Fields 1st Edition Fernando


Q. Gouvêa

https://ebookname.com/product/a-guide-to-groups-rings-and-
fields-1st-edition-fernando-q-gouvea/

Groups Rings and Galois Theory 2nd Revised ed. Edition


Victor P Snaith

https://ebookname.com/product/groups-rings-and-galois-theory-2nd-
revised-ed-edition-victor-p-snaith/

An Introduction to Abstract Algebra Sets Groups Rings


and Fields 1st Edition Steven H. Weintraub

https://ebookname.com/product/an-introduction-to-abstract-
algebra-sets-groups-rings-and-fields-1st-edition-steven-h-
weintraub/
Noncommutative Rings 5th printing Edition I. N.
Herstein

https://ebookname.com/product/noncommutative-rings-5th-printing-
edition-i-n-herstein/

Planetary Rings A Post Equinox View Esposito L.

https://ebookname.com/product/planetary-rings-a-post-equinox-
view-esposito-l/

Algebras Rings and Modules Lie Algebras and Hopf


Algebras 1st Edition Michiel Hazewinkel

https://ebookname.com/product/algebras-rings-and-modules-lie-
algebras-and-hopf-algebras-1st-edition-michiel-hazewinkel/

The Lord of the Rings Sketchbook First Edition Alan Lee

https://ebookname.com/product/the-lord-of-the-rings-sketchbook-
first-edition-alan-lee/

Arithmetical Properties of Commutative Rings and


Monoids 1st Edition Scott T. Chapman (Eds.)

https://ebookname.com/product/arithmetical-properties-of-
commutative-rings-and-monoids-1st-edition-scott-t-chapman-eds/
Groups, Rings
and Group Rings

© 2006 by Taylor & Francis Group, LLC


PURE AND APPLIED MATHEMATICS

A Program of Monographs, Textbooks, and Lecture Notes

EXECUTIVE EDITORS

Earl J. Taft Zuhair Nashed


Rutgers University University of Central Florida
Piscataway, New Jersey Orlando, Florida

EDITORIAL BOARD

M. S. Baouendi Freddy van Oystaeyen


University of California, University of Antwerp,
San Diego Belgium
Jane Cronin Donald Passman
Rutgers University University of Wisconsin,
Jack K. Hale Madison
Georgia Institute of Technology Fred S. Roberts
S. Kobayashi Rutgers University
University of California, David L. Russell
Berkeley Virginia Polytechnic Institute
Marvin Marcus and State University
University of California, Walter Schempp
Santa Barbara Universität Siegen
W. S. Massey Mark Teply
Yale University University of Wisconsin,
Anil Nerode Milwaukee
Cornell University

© 2006 by Taylor & Francis Group, LLC


LECTURE NOTES IN
PURE AND APPLIED MATHEMATICS

Recent Titles

J. Cagnol et al., Shape Optimization and Optimal Design


J. Herzog and G. Restuccia, Geometric and Combinatorial Aspects of Commutative Algebra
G. Chen et al., Control of Nonlinear Distributed Parameter Systems
F. Ali Mehmeti et al., Partial Differential Equations on Multistructures
D. D. Anderson and I. J. Papick, Ideal Theoretic Methods in Commutative Algebra
Á. Granja et al., Ring Theory and Algebraic Geometry
A. K. Katsaras et al., p-adic Functional Analysis
R. Salvi, The Navier-Stokes Equations
F. U. Coelho and H. A. Merklen, Representations of Algebras
S. Aizicovici and N. H. Pavel, Differential Equations and Control Theory
G. Lyubeznik, Local Cohomology and Its Applications
G. Da Prato and L. Tubaro, Stochastic Partial Differential Equations and Applications
W. A. Carnielli et al., Paraconsistency
A. Benkirane and A. Touzani, Partial Differential Equations
A. Illanes et al., Continuum Theory
M. Fontana et al., Commutative Ring Theory and Applications
D. Mond and M. J. Saia, Real and Complex Singularities
V. Ancona and J. Vaillant, Hyperbolic Differential Operators and Related Problems
G. R. Goldstein et al., Evolution Equations
A. Giambruno et al., Polynomial Identities and Combinatorial Methods
A. Facchini et al., Rings, Modules, Algebras, and Abelian Groups
J. Bergen et al., Hopf Algebras
A. C. Krinik and R. J. Swift, Stochastic Processes and Functional Analysis: A Volume of
Recent Advances in Honor of M. M. Rao
S. Caenepeel and F. van Oystaeyen, Hopf Algebras in Noncommutative Geometry and Physics
J. Cagnol and J.-P. Zolésio, Control and Boundary Analysis
S. T. Chapman, Arithmetical Properties of Commutative Rings and Monoids
O. Imanuvilov, et al., Control Theory of Partial Differential Equations
Corrado De Concini, et al., Noncommutative Algebra and Geometry
A. Corso, et al., Commutative Algebra: Geometric, Homological, Combinatorial and
Computational Aspects
G. Da Prato and L. Tubaro, Stochastic Partial Differential Equations and Applications – VII
L. Sabinin, et al., Non-Associative Algebra and Its Application
K. M. Furati, et al., Mathematical Models and Methods for Real World Systems
A. Giambruno, et al., Groups, Rings and Group Rings

© 2006 by Taylor & Francis Group, LLC


Groups, Rings
and Group Rings

Edited by
Antonio Giambruno
Università di Palermo
Palermo, Italy
César Polcino Milies
Universidade de São Paulo
São Paulo, Brazil

S.K. Sehgal
University of Alberta
Edmonton, Alberta, Canada

Boca Raton London New York

© 2006 by Taylor & Francis Group, LLC


C5815_Discl.fm Page 1 Tuesday, October 4, 2005 10:39 AM

Published in 2006 by
Chapman & Hall/CRC
Taylor & Francis Group
6000 Broken Sound Parkway NW, Suite 300
Boca Raton, FL 33487-2742

© 2006 by Taylor & Francis Group, LLC


Chapman & Hall/CRC is an imprint of Taylor & Francis Group

No claim to original U.S. Government works


Printed in the United States of America on acid-free paper
10 9 8 7 6 5 4 3 2 1

International Standard Book Number-10: 1-58488-581-5 (Softcover)


International Standard Book Number-13: 978-1-58488-581-8 (Softcover)
Library of Congress Card Number 2005053799

This book contains information obtained from authentic and highly regarded sources. Reprinted material is quoted with
permission, and sources are indicated. A wide variety of references are listed. Reasonable efforts have been made to publish
reliable data and information, but the author and the publisher cannot assume responsibility for the validity of all materials
or for the consequences of their use.
No part of this book may be reprinted, reproduced, transmitted, or utilized in any form by any electronic, mechanical, or
other means, now known or hereafter invented, including photocopying, microfilming, and recording, or in any information
storage or retrieval system, without written permission from the publishers.

For permission to photocopy or use material electronically from this work, please access www.copyright.com
(http://www.copyright.com/) or contact the Copyright Clearance Center, Inc. (CCC) 222 Rosewood Drive, Danvers, MA
01923, 978-750-8400. CCC is a not-for-profit organization that provides licenses and registration for a variety of users. For
organizations that have been granted a photocopy license by the CCC, a separate system of payment has been arranged.

Trademark Notice: Product or corporate names may be trademarks or registered trademarks, and are used only for
identification and explanation without intent to infringe.

Library of Congress Cataloging-in-Publication Data

Groups, rings and group rings / [edited by] Antonio Giambruno, César Polcino Milies, S.K. Sehgal.
p. cm. -- (Lecture notes in pure and applied mathematics ; 247)
Includes bibliographical references and index.
ISBN 1-58488-581-5
1. Group theory--Congresses. 2. Rings (Algebra)--Congresses. 3. Group rings--Congresses. I.
Giambruno, A. II. Milies, César Polcino. III. Sehgal, Sudarshan K.,
1936- IV. Lecture notes in pure and applied mathematics ; v 247.

QA174.G768 2006
512'.2--dc22 2005053799

Visit the Taylor & Francis Web site at


http://www.taylorandfrancis.com
Taylor & Francis Group and the CRC Press Web site at
is the Academic Division of Informa plc. http://www.crcpress.com

© 2006 by Taylor & Francis Group, LLC


Preface

This volume contains the proceedings of the Conference on “Groups,


Rings, and Group Rings” held in Ubatuba, Brazil, from July 26 to 31, 2004,
in honor of the sixtieth birthday of Prof. César Polcino Milies.
This was an occasion for bringing together experts from different areas of
algebra ranging from group theory to associative and nonassociative algebras
to coding theory.
The meeting was attended by more than 80 participants and experts from
several different countries including Argentina, Belgium, Brazil, Chile, Ca-
nada, Germany, Italy, Israel, Russia, Spain, Switzerland, Ukraine, Thailand
and USA.
The conference featured invited main lectures by E. Aljadeff, N. Andrus-
kiewitsch, Y. Bahturin, G. Cliff, A. del Rı́o, R. Exel, M. Ferrero, V. Futorny,
A. Giambruno, W. Herfort, W. Kimmerle, V. Kirichenko, B. Kühlshammer,
S. Lopez-Permouth, S. Sehgal, I. Shestakov, M. Shirvani, L. Shneerson, S.
Sidki, and E. Zelmanov and several other talks of shorter length. The papers
of most of the principal speakers and some other contributions are included
in the present volume.
The editors wish to express their appreciation to the agencies and the
institutions that contributed financial support: Capes, CNPq, FAPESP,
FINEP, Grupo Votorantim, Pró-Reitorias de Pesquisa e de Pós-Graduação,
and the Instituto de Matemática e Estatı́stica of the Universidade de São
Paulo.

Antonio Giambruno
Sudarshan K. Sehgal

© 2006 by Taylor & Francis Group, LLC


List of Participants

Eli Aljadeff Ángel del Rı́o


Technion, Universidad de Murcia,
Israel Spain
Nicolás Andruskiewitsch Michael Dokuchaev
Universidad Nacional de Córdoba, Universidade de São Paulo,
Argentina Brasil
Alfredo J. Aragona Vallejo Ann Dooms
Universidade de São Paulo, Vrije Universiteit Brussel,
Brasil Belgium
Yuri A. Bahturin Flaviana Dutra
Memorial University of Newfoundland, Universidade Federal de Minas Gerais,
Canada Brasil
Luiz G.X. de Barros Ruy Exel
Senac, Universidade Federal de Santa Catarina,
Brasil Brasil
Laurent Bartholdi Bernardo Felzenszwalb
École Polytechnique Fédérale, Universidade Federal do Rio de Janeiro,
Switzerland Brasil
Osnel Broche Cristo Raul A. Ferraz
Universidade de São Paulo, Universidade de São Paulo,
Brasil Brasil
Sheila Campos Chagas Vitor O. Ferreira
Universidade de Brası́lia, Universidade de São Paulo,
Brasil Brasil
Gladys Chalom Miguel Ferrero
Universidade de São Paulo, Universidade Federal do Rio Grande do
Brasil Sul, Brasil
Gerald Cliff Leila Maria Vasconcellos Figueiredo
University of Alberta, Universidade de São Paulo,
Canada Brasil
Mariana Cornelissen Vyacheslav Futorny
Universidade Federal de Minas Gerais, Universidade de São Paulo,
Brasil Brasil
Wagner Cortes Antonio Giambruno
Universidade Estadual de Maringá, Università di Palermo,
Brasil Italy

© 2006 by Taylor & Francis Group, LLC


André Gimenez Bueno Burkhard Külshammer
Universidade de São Paulo, Friedrich-Schiller-Universität,
Brasil Germany
Adilson Gonçalves Alicia Labra
Universidade Federal do Rio de Janeiro, Universidad de Chile,
Brasil Chile
Jairo Z. Gonçalves João Lazzarin
Universidade de São Paulo, Universidade Federal do Rio Grande do
Brasil Sul, Brasil
Carlos Grossi Yuanlin Li
Universidade Estadual de Campinas, Brock University,
Brasil Canada
Henrique Guzzo, Jr. Alexander Lichtman
Universidade de São Paulo, University of Wisconsin,
Brasil USA
Wolfgang Herfort Thierry Petit Lobão
Vienna University of Technology, Universidade Federal da Bahia,
Austria Brasil
Edson Iwaki Jonas Lopes
Universidade de São Paulo, Universidade de São Paulo,
Brasil Brasil
Sandra Jorge Pedro Lopes
Universidade Federal de Minas Gerais, Universidade Técnica de Lisboa,
Brasil Portugal
Stanley O. Juriaans Sergio R. López-Permouth
Universidade de São Paulo, Ohio University,
Brasil USA
Iryna Kashuba Valéria Ostete Jannis Luchetta
Universidade de São Paulo, Universidade de São Paulo,
Brasil Brasil
Marina Khibina Arnaldo Mandel
Institute of Engineering of Thermophy- Universidade de São Paulo,
sics of National Academy, Ukraine Brasil
Wolfgang Kimmerle Eduardo Marcos
Universität Stuttgart, Universidade de São Paulo,
Germany Brasil
Vladimir Kirichenko Lucia S.I. Murakami
Kyiv Taras Shevchenko University, Universidade de São Paulo,
Ukraine Brasil
Plamen Koshlukov Boris V. Novikov
Universidade Estadual de Campinas, Kharkov National University,
Brasil Ukraine

© 2006 by Taylor & Francis Group, LLC


Serge Ovsienko Lev Shneerson
Kiev Tars Shevchenko University, The City University of New York,
Ukraine USA
Henrique Panzarelli Said N. Sidki
Universidade de São Paulo, Universidade de Brası́lia,
Brasil Brasil
Luiz Antonio Peresi Maria Sobral Singer
Universidade de São Paulo, Universidade de São Paulo,
Brasil Brasil
César Polcino Milies Antonio Calixto Souza Filho
Universidade de São Paulo, Universidade de São Paulo,
Brasil Brasil
Noraı́ R. Rocco Maria Trushina
Universidade de Brası́lia, Universidade de São Paulo,
Brasil Brasil
Virgı́nia Rodrigues Ana Cristina Vieira
Universidade Federal do Rio Grande do Universidade Federal de Minas Gerais,
Sul, Brasil Brasil
Javier Sánchez Serdà Benjamin John Wilson
Universitat Autònoma de Barcelona, Universidade de São Paulo,
Spain Brasil
Nguyen Van Sanh Pavel Zalesskii
Mahidol University, Universidade de Brası́lia,
Thailand Brasil
Sudarshan K. Sehgal Andrei Zavarnitsine
University of Alberta, Universidade de São Paulo,
Canada Brasil
Ivan Shestakov Efim Zelmanov
Universidade de São Paulo, University of California,
Brasil USA
Mazi Shirvani Natalia Zhukavets
University of Alberta, Universidade de São Paulo,
Canada Brasil

© 2006 by Taylor & Francis Group, LLC


Contributors

Eli Aljadeff Gerald Cliff


Department of Mathematics Department of Mathematical and
Technion Statistical Sciences
Haifa, Israel University of Alberta
Edmonton, Alberta, Canada
Nicolás Andruskiewitsch
Facultad de Matemática, Astronomı́a y Ángel del Rı́o
Fı́sica Departamento de Matemáticas
Universidad Nacional de Córdoba Universidad de Murcia
Córdoba, Argentina Murcia, Spain

Yuri A. Bahturin Onofrio M. Di Vincenzo


Department of Mathematics and Dipartimento di Matematica
Statistics Università degli Studi di Bari
Memorial University of Newfoundland Bari, Italy
St. John’s, Newfoundland, Canada
Hai Q. Dinh
Gunther Bergauer Department of Mathematical Sciences
Institute for Analysis and Scientific Kent State University
Computation Warren, Ohio
University of Technology
Vienna, Austria Michael Dokuchaev
Instituto de Matemática e Estatı́stica
Osnel Broche Cristo Universidade de São Paulo
Instituto de Matemática e Estatı́stica São Paulo, Brasil
Universidade de São Paulo
São Paulo, Brasil Ann Dooms
Department of Mathematics
Gladys Chalom Vrije Universiteit Brussel
Instituto de Matemática e Estatı́stica Brussels, Belgium
Universidade de São Paulo
São Paulo, Brasil Raul A. Ferraz
Instituto de Matemática e Estatı́stica
Orin Chein Universidade de São Paulo
Temple University São Paulo, Brasil
Philadelphia, PA

© 2006 by Taylor & Francis Group, LLC


Vitor O. Ferreira Edgar G. Goodaire
Instituto de Matemática e Estatı́stica Memorial University of Newfoundland
Universidade de São Paulo St. John’s, Newfoundland, Canada
São Paulo, Brasil
Marcus Greferath
Walter R. Ferrer Santos Department of Mathematics
Centro de Matemática University College Dublin
Facultad de Ciencias Dublin, Ireland
Universidad de la República
Montevideo, Uruguay Nadiya M. Gubareni
Institute of Econometrics and Computer
Miguel Ferrero Science
Instituto de Matemática Technical University of Czȩstochowa
Universidade Federal do Rio Grande do Czȩstochowa, Poland
Sul
Porto Alegre, Brasil Henrique Guzzo, Jr.
Instituto de Matemática e Estatı́stica
Vyacheslav Futorny Universidade de São Paulo
Instituto de Matemática e Estatı́stica São Paulo, Brasil
Universidade de São Paulo
São Paulo, Brasil Darrell Haile
Department of Mathematics
Antonio Giambruno Indiana University
Dipartimento di Matematica e Bloomington, Indiana
Applicazioni
Università di Palermo Wolfgang Herfort
Palermo, Italy Institute for Analysis and Scientific
Computation
André Gimenez Bueno University of Technology
Instituto de Matemática e Estatı́stica Vienna, Austria
Universidade de São Paulo
São Paulo, Brasil Christian Höfert
Institut für Geometrie und Topologie
Carmen Rosa Giraldo Vergara Fachbereich Mathematik
Faculdade de Filosofia, Ciências e Letras Universität Stuttgart
Fundação Educacional Monsenhor Stuttgart, Germany
Messias
Sete Lagoas, Brasil Eric Jespers
Department of Mathematics
Jairo Z. Gonçalves Vrije Universiteit Brussel
Instituto de Matemática e Estatı́stica Brussels, Belgium
Universidade de São Paulo
São Paulo, Brasil Stanley O. Juriaans
Instituto de Matemática e Estatı́stica
Universidade de São Paulo
São Paulo, Brasil

© 2006 by Taylor & Francis Group, LLC


Wolfgang Kimmerle Valéria O. Luchetta
Institut für Geometrie und Topologie Instituto de Matemática e Estatı́stica
Fachbereich Mathematik Universidade de São Paulo
Universität Stuttgart São Paulo, Brasil
Stuttgart, Germany
Zbigniew Marciniak
Vladimir V. Kirichenko Institute of Mathematics
Department of Mechanics and Warsaw University
Mathematics Warsaw, Poland
Kyiv Taras Shevchenko University
Kiev, Ukraine David McNeilly
University of Alberta
Plamen Koshlukov Department of Mathematical and
Instituto de Matemática e Estatı́stica Statistical Sciences
Universidade Estadual de Campinas Edmonton, Alberta, Canada
Campinas, Brasil
Héctor Merklen
Burkhard Külshammer Instituto de Matemática e Estatı́stica
Mathematisches Institut Universidade de São Paulo
Friedrich-Schiller-Universität São Paulo, Brasil
Jena, Germany
Paola Misso
Daniela La Mattina Dipartimento di Matematica e
Dipartimento di Matematica e Applicazioni
Applicazioni Università di Palermo
Università di Palermo Palermo, Italy
Palermo, Italy
Lucia S.I. Murakami
Guilherme Leal Instituto de Matemática e Estatı́stica
Instituto de Matemática Universidade de São Paulo
Universidade Federal do Rio de Janeiro São Paulo, Brasil
Rio de Janeiro, Brasil
Michael Natapov
Yuanlin Li Department of Mathematics
Department of Mathematics Technion
Brock University Haifa, Israel
St. Catharines, Ontario, Canada
Boris V. Novikov
Sergio R. López-Permouth Kharkov National University
Department of Mathematics Kharkov, Ukraine
Ohio University
Athens, Ohio Konstantin Okunev
Novosibirsk State University
Novosibirsk, Russia

© 2006 by Taylor & Francis Group, LLC


Michael M. Parmenter Sudarshan K. Sehgal
Department of Mathematics and Department of Mathematical and
Statistics Statistical Sciences
Memorial University of Newfoundland University of Alberta
St. John’s, Newfoundland, Canada Edmonton, Alberta, Canada

Arturo Pianzola Ivan Shestakov


Department of Mathematical and Instituto de Matemática e Estatı́stica
Statistical Sciences Universidade de São Paulo
University of Alberta São Paulo, Brasil
Edmonton, Alberta, Canada and
Sobolev Institute of Mathematics
Antonio Pita Novosibirsk, Russia
Departamento de Matemáticas
Universidad de Murcia Mazi Shirvani
Murcia, Spain Department of Mathematical and
Statistical Sciences
César Polcino Milies University of Alberta
Instituto de Matemática e Estatı́stica Edmonton, Alberta, Canada
Universidade de São Paulo
São Paulo, Brasil Angela Valenti
Dipartimento di Metodi e Modelli
Manuel Ruiz Marı́n Matematici
Departamento de Métodos Università di Palermo
Cuantitativos e Informáticos Palermo, Italy
Universidad Politécnica de Cartagena
Cartagena, Spain Ana Cristina Vieira
Departamento de Matemática
Instituto de Ciências Exatas
Universidade Federal de Minas Gerais
Belo Horizonte, Brasil

© 2006 by Taylor & Francis Group, LLC


Contents

1. On fine gradings on central simple algebras 1


Eli Aljadeff, Darrell Haile, and Michael Natapov
2. On observable module categories 11
Nicolás Andruskiewitsch and Walter R. Ferrer Santos
3. Group gradings on integral group rings 25
Yuri A. Bahturin and Michael M. Parmenter
4. Profinite graphs – comparing notions 33
Gunther Bergauer and Wolfgang Herfort
5. Lie identities in symmetric elements in group rings: A survey 43
Osnel Broche Cristo and Manuel Ruiz Marı́n
6. Irreducible morphisms in subcategories 57
Gladys Chalom and Héctor Merklen
7. Bol loops with a unique nonidentity commutator/associator 73
Orin Chein and Edgar G. Goodaire
8. Weil representations of symplectic groups 81
Gerald Cliff and David McNeilly
9. Gradings and graded identities for the upper triangular matrices
over an infinite field 91
Onofrio M. Di Vincenzo, Plamen Koshlukov, and Angela Valenti
10. Structure of some classes of repeated-root constacyclic codes
over integers modulo 2m 105
Hai Q. Dinh
11. Units in noncommutative orders 119
Ann Dooms and Eric Jespers
12. Idempotents in group algebras and coding theory 137
Raul A. Ferraz, Valéria O. Luchetta, and César Polcino Milies
13. Finitely generated constants of free algebras 149
Vitor O. Ferreira and Lucia S.I. Murakami
14. Partial actions of groups on semiprime rings 155
Miguel Ferrero
15. Representations of affine Lie superalgebras 163
Vyacheslav Futorny
16. On algebras and superalgebras with linear codimension growth 173
Antonio Giambruno, Daniela La Mattina, and Paola Misso
17. On spectra of group rings of finite abelian groups 183
André Gimenez Bueno and Michael Dokuchaev

© 2006 by Taylor & Francis Group, LLC


18. Wedderburn decomposition of small rational group algebras 191
Carmen Rosa Giraldo Vergara
19. Some questions on skewfields 201
Jairo Z. Gonçalves
20. On the role of rings and modules in algebraic coding theory 205
Marcus Greferath and Sergio R. López-Permouth
21. Semiperfect rings with T -nilpotent prime radical 217
Nadiya M. Gubareni and Vladimir V. Kirichenko
22. The structure of the baric algebras 233
Henrique Guzzo, Jr.
23. On torsion units of integral group rings of groups of small
order 243
Christian Höfert and Wolfgang Kimmerle
24. On a conjecture of Zassenhaus for metacyclic groups 253
Stanley O. Juriaans and Sudarshan K. Sehgal
25. Nilpotent blocks revisited 263
Burkhard Külshammer
26. Decomposition of central units of integral group rings 275
Yuanlin Li and Michael M. Parmenter
27. Generic units in ZC 279
Zbigniew Marciniak and Sudarshan K. Sehgal
28. On quasi-Frobenius semigroup algebras 287
Boris V. Novikov
29. Twisted loop algebras and Galois cohomology 293
Arturo Pianzola

30. Presentation of the group of units of ZD16 305
Antonio Pita and Ángel del Rı́o
31. Engel theorem for Jordan superalgebras 315
Ivan Shestakov and Konstantin Okunev
32. A characterization of centre-by-finite subgroups of division
algebras 321
Mazi Shirvani
33. Isomorphic rational group algebras 331
Ana Cristina Vieira and Guilherme Leal

© 2006 by Taylor & Francis Group, LLC


On fine gradings on central simple algebras
Eli Aljadeff
Department of Mathematics, Technion, Haifa 32000, Israel
[email protected]
Darrell Haile
Department of Mathematics, Indiana University, Bloomington, IN 47405
[email protected]
Michael Natapov
Department of Mathematics, Technion, Haifa 32000, Israel
[email protected]

1. Introduction
Let G be a group and R an associative ring. A G−grading on R is a
decomposition of R as an additive group into the direct sum of subgroups
R = ⊕g∈G Rg such that Rg Rh ⊆ Rgh for any g, h ∈ G. The subgroup Rg
is called the homogeneous component of degree g and elements of Rg are
called homogeneous of degree g. The subgroup Re , where e is the identity
element of G, is called the identity component. The support of R is defined
by Supp R = {g ∈ G : Rg 6= 0}. If H = Supp R is a subgroup of G, we say
that the group H supports the G−grading on R. If R is an algebra over a
field F , a G−grading on R is called fine if dimF (Rg ) ≤ 1 for all g ∈ G. We
refer the reader to [5] for more details on G–gradings on algebra over a field
F.
Our first object in this paper is to observe that the notion of a fine grading
on a central simple algebra is basically the same as the notion of a projective
basis.
Recall that a finite dimensional F –algebra R is said to have a projective
basis if it has a basis {a1 , a2 , . . . , an } of invertible elements, such that for
every pair i, j there is an m such that ai aj = λij am for some λij ∈ F × . It
is often convenient to view such an algebra R as a twisted group algebra.
Let Γ be the subgroup of R× generated by a1 , a2 , . . . , an and F × , and let
G = Γ/F × , then R is isomorphic to the twisted group algebra F α G where
the cohomology class α ∈ H 2 (G, F × ) is defined by the central extension
1 −→ F × −→ Γ −→ G −→ 1
1

© 2006 by Taylor & Francis Group, LLC


2 E. Aljadeff, D. Haile, and M. Natapov

(Recall that for each finite group G and two-cocycle f in Z 2 (G, F × ) one can
form the twisted group algebra F f G = ⊕σ∈G F xσ where xσ xτ = f (σ, τ )xστ
for all σ, τ ∈ G. The resulting F –algebra depends (up to F –isomorphism)
only on the class α = [f ] in H 2 (G, F × ), and so we denote the algebra by
F α G.)
The precise statement that relates projective bases to fine gradings for
F –central simple algebras is as follows:
Theorem 1. Let F be a field. Let R be a finite dimensional central simple
algebra over a field F graded by a group G such that dimF (Rg ) ≤ 1 for all
g ∈ G. Then H = Supp R is a subgroup of G and R ∼ = F α H for some
2 ×
α ∈ H (H, F ).
The converse of Theorem 1 is clear. If R is an F –algebra isomorphic to a
twisted group algebra F α H, α ∈ H 2 (H, F × ), then H supports a fine grading
on R over F .
Theorem 1 is proved in Section 2.
We will call a group H, such that F α H is F –central simple for some F
and α ∈ H 2 (H, F × ), a group of central type (nonclassically). The class α,
such that F α H is F –central simple, is said to be nondegenerate. Recall
that classically a finite group Γ is of central type if it has anpirreducible
complex representation of the maximal possible degree, namely [Γ : Z(Γ)]
(where Z(Γ) denotes the center of Γ). It is easy to see that H is of central
type nonclassically if and only if there exists a classical central type group Γ
such that H ∼ = Γ/Z(Γ). Isaacs and Howlett proved, using the classification
of finite simple groups, that such Γ, and hence H, is solvable ([6, Theorem
7.3]). In this paper, we refer to “central type” only in the nonclassical sense.
Thus by Theorem 1 we obtain that the groups H that support fine grad-
ings of central simple algebras are precisely the groups of central type.
Theorem 1 allows us to give a precise description of the groups H that
support fine gradings of F –central division algebras. Let Λp be the following
list of p–groups:
Q
(1) H is abelian of symmetric type, that is H ∼ = (Zpni × Zpni )
(2) H ∼
n n
= H1 × H2 where H1 = Zpn o Zpn = h π, σ | σ p = π p =
s
1 and σπσ −1 = π p +1 i where 1 ≤ s < n and 1 6= s if p = 2 and
H2 is an abelian group of symmetric type of exponent ≤ ps
(3) H ∼
= H1 × H2 where ¿ ¯ À
¯ π 2n+1 = σ 2n = τ 2 = 1, στ = τ σ,
¯
H1 = Z2n+1 o (Z2n × Z2 ) = π, σ, τ ¯
σπσ −1 = π 3 , τ πτ −1 = π −1
and H2 is an abelian group of symmetric type of exponent ≤ 2
Combining Theorem 1 with [1, Theorem 1] and [2, Corollary 3] we have
Theorem 2. Let H be a finite group that is the support of a fine grading
on an F –central division algebra D. Then

© 2006 by Taylor & Francis Group, LLC


On fine gradings on central simple algebras 3

(1) H is a nilpotent group.


(2) If H ∼ = P1 × . . . × Pr is the decomposition of H into the product
of its Sylow pi –subgroups then Pi belongs to Λpi . In particular, the
commutator subgroup H 0 of H is cyclic.
Remark: In a future paper [8] Natapov shows that any group in Λp supports
a fine grading on an F –central division algebra for some F . Moreover, if
H∼ = P1 ×. . .×Pr , Pi are pi –groups in Λpi (pi are different primes), then there
exists a field F (independent of i) such that Pi supports a fine grading on an
F –central division algebra Di , 1 ≤ i ≤ r. Then, it is clear that H supports
a fine grading on an F –central division algebra, namely D = D1 ⊗F . . . ⊗ Dr .
Theorem 2 and the above remark give a complete classification of the
groups that support fine gradings on finite dimensional division algebras
over their center.
The main bulk of this paper is devoted to showing that if a nilpotent
group H supports
√ a fine grading on Mr (D) then the index of Mr (D) (=
ind(D) = dimF D) puts a strong constraint on H. More precisely:
Theorem 3. Let H be a nilpotent group that supports a fine grading on
an F –central simple algebra R = Mr (D), D a division algebra. Then there
exists a subquotient S of H of order ≥ ind(D)2 whose commutator subgroup
S 0 is cyclic.
Theorem 3 is proved in Section 3.

2. Fine gradings and projective bases


In this Section we prove Theorem 1. We follow the proof of [5, Theorem
5].
Claim 1: The identity component Re = F · 1.

Proof.
PLet aσ be a homogeneous element of degree σ, then Rσ = F aσ . Write
1 = σ∈H λσ aσ where λσ are scalars in F . We claim that λσ = 0 for all
σ 6= e. SupposePthere is τ 6= e, τ ∈ H, such that λτ 6= 0. Then writing
ag = ag · 1 = σ∈H λσ ag aσ for all g ∈ G we get that λτ ag aτ = 0, or,
equivalently, ag aτ = 0 for all g ∈ G. It follows Raτ = 0, a contradiction.
Hence, 1 = λe ae . In other words 1 ∈ Re and Claim 1 follows. ¤

Claim 2: Let g ∈ H and 0 6= ag ∈ Rg . Then ag is invertible.

Proof. Consider the two-sided ideal Rag R in R. It is a nonzero ideal, hence


Rag R = R. Write the unity element 1 as an element of Rag R:
à ! à !
X X X X
1= λiσ aσ ag µiτ aτ = λiσ µiτ aσ ag aτ
i σ∈H τ ∈H i,σ,τ

© 2006 by Taylor & Francis Group, LLC


4 E. Aljadeff, D. Haile, and M. Natapov

where λiσ and µiτ are in F . Since 1 is a homogeneous element of degree e,


there exist σ0 and τ0 in G such that σ0 gτ0 = e and aσ0 ag aτ0 is a nonzero
and hence invertible element of Re = F . This proves the claim. ¤
It follows that Rg Rh 6= 0 for any g, h ∈ H and therefore Rgh 6= 0 or,
equivalently, gh ∈ Supp R = H. That is H is a finite subset of G closed under
the multiplication. Therefore H is a subgroup of G. Since dim(Rg ) = 1 for
all g ∈ H it follows that ord(H) = dim(R).
Now pick homogeneous elements ai ∈ Rgi for every gi ∈ H. Then the
set {ai } is a basis of R and ai aj = λij am for some λij ∈ F × . Thus R is
isomorphic to F α H where α = [f ] ∈ H 2 (H, F × ) is defined by f (gi , gj ) = λij .

3. Fine gradings on central simple algebras supported by nilpotent


subgroups
In this Section we prove Theorem 3.
Observe first that if H is a nilpotent group that supports a fine grading
on an F –central simple algebra R then R ∼ = F α H is isomorphic to F α1 P1 ⊗
α α
F 2 P2 ⊗ · · · ⊗ F m Pm , where P1 , P2 , . . . , Pm are the Sylow pi –subgroups of
H and αi is the restriction α to Pi . Furthermore, if Pi has a subquotient
Si of order ≥ ind(F αi Pi ) with cyclic commutator then S = S1 × · · · × Sm
is a subquotient of H such that ord(S) ≥ ind(F α H) and the commutator
subgroup S 0 is cyclic. So, this reduces the analysis to p–groups.
Let G be a finite p–group which supports the grading on an F –central
simple algebra R. Write R = F α G. By the proof of [4, Theorem 1.2, Claim
II] F contains a nontrivial p–th root of unity. In particular char(F ) 6= p.
Let Γ be the subgroup of (F α G)× given by the central extension
π
α : 1 −→ F × −→ Γ −→ G −→ 1
Clearly F α G is spanned as an F -vector space (and even as an abelian group)
by Γ. We often write F α G = F (Γ). For every σ ∈ G we choose an element
uσ in Γ such that π(uσ ) = σ.
We start with the following reduction lemma.
Lemma 4. Let G be a p–group and F α G an T F –central simple algebra. Sup-
pose that G0 is not trivial. Let U ⊆ Z(G) G0 be a subgroup of order p. If
F α U is not simple then
a. F α U ∼
= F α U e1 ⊕...⊕F α U ep ∼
= F e1 ⊕...⊕F ep where ei ’s are mutually
orthogonal idempotents.
b. There is a subgroup T , U ⊂ T ⊂ G of index p such that
resG 2 × 2 ×
T (α) ∈ Im(inf : H (T /U, F ) → H (T, F ))

c. F α G ∼= Mp (F β H) where H ∼
= T /U and β ∈ H 2 (H, F × ) with
T /U
infT (β) = resG T (α).

© 2006 by Taylor & Francis Group, LLC


On fine gradings on central simple algebras 5

Proof. Let σ be a generator of U√and let upσ = a ∈ F × . Since F contains a


nontrivial p–th root of unity F ( p a)/F is a √ Kummer extension. Clearly, if
a is not a p–th power in F × then F α U ∼ = F ( p a) which is simple. Therefore
a = bp , b ∈ F × , and hence F α U = F U b where U b = hb−1 uσ i. We obtain
FUb∼ = FU b e1 ⊕ ... ⊕ F U
b ep = F e1 ⊕ ... ⊕ F ep where e1 = 1/p(1 + b−1 uσ + ... +
−1 p−1
(b uσ ) ).
To see (b) and (c) note that since U is normal in G, the group G acts on
F α U by conjugation. Moreover, since F α G is F –central the action permutes
transitively the simple components of F α U , [3, Lemma 1.2]. Let T ⊂ G be
the stabilizer of F α U e1 . Clearly, T is of index p in G and centralizes U b.
b −1 α ∼ b
Let T = π (T ) ⊂ Γ. By [3, Theorem 1.4], F G = Mp (F e1 (T e1 )) where
F e1 (Tbe1 ) is the algebra spanned by Tbe1 over F e1 ∼ = F . We show that
b ∼ β T /U G
F e1 (T e1 ) = F (T /U ) where infT (β) = resT (α).
Splicing together the short exact sequences
b −→ Tb −→ Tbe1 −→ 1
1 −→ U
and
res(α) : 1 −→ F × −→ Tb −→ T −→ 1,
we obtain a commutative diagram with exact rows and columns:

1 1
 
 
y y
F× F×
 
 
y y
b −−−−→
1 −−−−→ U Tb −−−−→ Tbe1 −−−−→ 1
  
  
y y y
1 −−−−→ U −−−−→ T −−−−→ T /U −−−−→ 1
 
 
y y
1 1
If we let β be the cohomology class which corresponds to the right column
T /U
then we get infT (β) = resG T (α). Finally, identifying F and F e1 we get a
surjective map of F –algebras
F β (T /U ) −→ F e1 (Tbe1 ).
The proof is completed by a dimension argument. ¤
We continue with a factorization lemma. It generalizes a similar statement
for division algebras (see [1]). The proof is identical and hence is omitted.

© 2006 by Taylor & Francis Group, LLC


6 E. Aljadeff, D. Haile, and M. Natapov

Lemma 5. Let P be a normal subgroup of G and assume F α P is F –central


simple. Then F α G ∼
= F α P ⊗ F β G/P for some β ∈ H 2 (G/P, F × ).
Next we consider the restriction of the cohomology class α ∈ H 2 (G, F × )
to the commutator subgroup G0 .
Lemma 6. Let F α G be F –central simple. Then resG G0 (α) has a representa-
tive 2–cocycle whose values are p–power roots of unity.
Proof. Consider the group Γ ⊆ (F α G)× , which is generated by represen-
tatives of G and F × in F α G. Observe that Γ is center by finite, so by a
theorem of Schur [10, Chapter 2, Theorem 9.8] its commutator subgroup
Γ0 is finite. It is easy to see that F × Γ0 /F × = G0 . It follows that F (Γ0 ),
the subalgebra generated by Γ0 , is isomorphic to the twisted group algebra
F α G0 . Since Γ0 is finite there exists a representative f ∈ α that takes finite
values on G0 , that is, for every σ, τ in G0 , f (σ, τ ) ∈ µ ⊂ F × , where µ denotes
the group of roots of unity in k. More than that, by the [10, Lemma of
Chapter 2, §9], since G is a p-group, then values of f are p-power roots of
unity. ¤
We say that a cohomology class is of finite type if it can be represented
by a cocycle which takes finite values in F × .
Lemma 7. Let E = Zp × Zp be an elementary abelian group of order p2 .
Let F α E be a twisted group algebra over F . Let σ, τ be generators of E and
uσ , uτ be representatives of σ, τ in F α E. Then
a. The algebra F α E is F –central simple if and only if uσ and uτ do
not commute.
b. Suppose uσ and uτ do not commute. If α ∈ H 2 (E, F × ) is of finite
type then
1. If p is odd then F α E ∼
= Mp (F )
2. If p = 2 then F α E ∼ = M2 (F ) or F α E ∼= H where H is the
nonsplit quaternion algebra (−1, −1)F .
c. Suppose uσ and uτ commute. If α ∈ H 2 (E, F × ) is of finite type then
F α E is not a field.
Proof. Consider first the case where uσ and uτ commute. Then the algebra
F α E is commutative, in particular it is not F –central. Furthermore, if α
is of finite type then F α E contains a finite noncyclic group generated by
uσ , uτ and hence F α E is not a field.
Assume now that uσ and uτ do not commute. Since σ and τ are commut-
ing elements of order p in E, the commutator (uσ1 , uσ2 ) = ξp - a p–primitive
root of unity. Write upσ1 = x ∈ F × and upσ2 = y ∈ F × , then, clearly, F α E is
isomorphic to the symbol algebra (x, y)p over F , in particular it is F –central
simple.
Now, if α ∈ H 2 (E, F × ) is of finite type, then F α E has the form (ζ1 , ζ2 )p
where ζ1 , ζ2 are p–power roots of unity. It is well known that such a symbol

© 2006 by Taylor & Francis Group, LLC


On fine gradings on central simple algebras 7

is a split algebra except the quaternion algebra (−1, −1)F , which may be
nonsplit in the case where F does not contain a 4th root of unity. ¤
The following proposition is the key for the proof of the Theorem.
Proposition 8. Let G be a p–group and F α G an F –central simple algebra.
If the commutator G0 is not cyclic then there exist a subquotient H of G and
a class β ∈ H 2 (H, F × ) such that F α G ∼
= Mp (F β H) or F α G ∼ = H ⊗ F β H.
T 0
Proof. Consider the nontrivial subgroup N = Z(G) G of G. Let Np
denote the p–elementary abelian subgroup of N .
By Lemma 4 we may assume that the algebra F α G is reduced, that is for
every subgroup U of N of order p, the corresponding subalgebra F α U is a
field. We proceed as in the proof of [4, Theorem 1.2].
Case (1): Np is not cyclic. Let P = hσ1 , σ2 i be any subgroup of order p2
in Np . Every cyclic subgroup of P generates a cyclotomic field extension.
In particular, for i = 1, 2, upσi = ζi where ζi is a p–power root of unity. We
may assume that ζ1 = ζ2t for some integer t. We claim that uσ1 and uσ2 do
not commute. Otherwise, we get (uσ1 u−t )p = 1, that is, the restriction of α
−t T 0 σ2
to the subgroup hσ1 σ2 i ⊆ Z(G) G is trivial, a contradiction. Hence, by
Lemma 7a, F α P is simple and by Lemma 5 we have
F αG ∼= F αP ⊗ F β H
where H is isomorphic to G/P . Since the restriction of α on P is of finite
type, by Lemma 7b we obtain that
– If p is an odd prime or p = 2 and the quaternion algebra (−1, −1)F
is split then F α G ∼= Mp (F β H).
– If p = 2 and H = (−1, −1)F is not split then F α G ∼ = H ⊗ F β H.
Case (2): Np is cyclic. A close inspection of the proof of [4, Theorem 1.2]
shows that there is a subgroup P in G0 , which is normal in G and isomorphic
to Zp × Zp . Let σ1 , σ2 be generators of P . Since P is not central, we cannot
proceed as in Case (1). We have two possibilities.
Case (2.1): F α P is a simple algebra. We claim that uσ1 and uσ2 do not
commute, for otherwise F α P is a commutative simple algebra that is, by
Lemma 7c, not a field, a contradiction. Recall that the restriction of α to
G0 ⊇ P is of finite type. Therefore, as in the Case (1), applying Lemmas 5
and 7b we get either F α G ∼ = Mp (F β H) or F α G ∼
= H ⊗ F β H with H ∼= G/P .
Case T α
(2.2): F P is not simple. Let us show that this case is impossible.
Since P Z(G) is not trivial, we may assume that σ1 ∈ Z(G). Then hσ1 i
is the unique subgroup of P of order p, which is normal in G. Since F α G is
reduced, F α hσ1 i is a field.
Note that there exists a subgroup K of P such that resG K (α) = 1, for if
not, arguing as in Case (1), we have that F α P is isomorphic to a symbol
algebra, contradicting our assumption. Of course, we may assume that K is
generated by σ2 , and hence there exists a representing 2–cocycle c : G×G →

© 2006 by Taylor & Francis Group, LLC


8 E. Aljadeff, D. Haile, and M. Natapov

F × of α such that c(σ2i , σ2j ) = 1 for all i, j. Furthermore, since (uσ1 , uσ2 ) = 1
(Lemma 7a) we may assume also c(σ2i , x) = c(x, σ2i ) = 1 for all x ∈ P , and
all i.
Consider the decomposition
(1) F α P = F α P e1 ⊕ ... ⊕ F α P ep
P
where ei = 1/p( k ξik ukσ2 ) are central mutually orthogonal idempotents
(here ξi are different p-roots of unity and ξ1 = 1). Note that F α P e1 is
isomorphic to the field F α hσ1 i.
The group G acts on F α P by conjugation and hence on the decomposition
(1). Since the algebra F α G is F –central simple, the group G permutes
transitively the simple components of F α P . Let T be the stabilizer of the
component F α P e1 , or equivalently the centralizer of e1 . Note that T is
of index p in G. On the other hand, since P is normal in G and hσ2 i is
not normal, the centralizer of hσ2 i in G is of index p and in fact is equal
to T . This implies that all elements uτ , τ ∈ CG (σ2 ) commute with uσ2 ,
contradicting a well-known characterization of a nondegenerate 2–cocycle
on a central type group, see e.g. [7, Problems 11.5, 11.8]. ¤

We now turn to the proof of Theorem 3.

Proof of Theorem 3. As mentioned above it is sufficient to consider the case


where H is a p–group. By repeated application of the proposition the result
follows easily in cases where p is odd or p = 2 and the quaternion algebra
(−1, −1)F is split. Of course, if after s(< n) applications of the proposition
we obtain a subquotient Hs such that Hs0 is cyclic, there is nothing to prove.
Otherwise F α H ∼ = Mpn (D) is isomorphic to Mpn (F βn Hn ) which says that
F βn Hn ∼= D. By [2, Corollary 3] we have that Hn ∈ Λp and in particular
Hn0 is cyclic.
We therefore
√ assume that p = 2 and H = (−1, −1)F is not split. In
particular −1 ∈ / F × . We claim that it is sufficient to show that if H ⊗
β ∼
F Hn = M2 (D) then Hn has a cyclic commutator subgroup (as above we
n

may assume that after s < n applications of the proposition the group Hs
has a noncyclic commutator): Since H ⊗F H ∼ = M4 (F ) it follows that if the
quaternion algebra H is factored out an even number of times we obtain
F αH ∼ = M2n (F βn Hn ) ∼= M2n (D) and hence Hn ∈ Λ2 . The remaining case
yields H ⊗ M2n−1 (F βn Hn ) ∼= M2n (D) and hence H ⊗ F βn Hn ∼ = M2 (D).
Assume Hn is not cyclic. If F βn Hn is reduced, it follows from the proof
0

of Proposition 8 that a quaternion algebra can be factored out from F βn Hn .


Clearly, this is impossible by an index argument. We conclude that F βn Hn
is not reduced. T
Let U ⊆ Z(Hn ) Hn0 be a subgroup of order 2 such that F βn U is not
simple. By Lemma 4 there is a subgroup T ⊃ U of index 2 in Hn such that
F βn Hn ∼= M2 (F β T /U ). It follows that H ⊗ F βn Hn ∼= M2 (H ⊗ F β T /U ) ∼
=

© 2006 by Taylor & Francis Group, LLC


On fine gradings on central simple algebras 9

M2 (D). In particular H ⊗ F β T /U is a division algebra over F . Now, since


Hn0 is not cyclic, the subgroup Hn0 /U ⊂ T /U is not trivial. We claim that
β restricted to Hn0 /U is of finite type. To see this let µ = {±1} ⊂ F × and
consider the commutative diagram with exact columns:
inf
H 2 (Hn0 /U, µ) −−−−→ H 2 (Hn0 , µ)
 
i i
y∗ y∗
inf
β ∈ H 2 (Hn0 /U, F × ) −−−−→ H 2 (Hn0 , F × )
 
p∗ p∗
y y
inf
H 2 (Hn0 /U, F × /µ) −−−−→ H 2 (Hn0 , F × /µ)
We know that inf (β) ∈ H 2 (Hn0 , F × ) is of finite type and hence it is
in the image of i∗ . This implies p∗ inf (β)
√ = inf (p ∗ (β)) = 0. But since
1 × ×
H (U, F /µ) = Hom(U, F /µ) = 0 ( −1 ∈ ×
/ F ) the map inf in the
bottom line is injective [9, Chapter VII, §6, Proposition 5]. This implies
p∗ (β) = 0 and hence β restricted to Hn0 /U has finite values and the claim is
proved. Now, β is not trivial on Hn0 /U since F β T /U is a √ division algebra.
β
It follows that F T /U contains a field isomorphic to F ( −1) and hence
H ⊗ F β T /U cannot be a division algebra.
This completes the proof of the Theorem. ¤

Acknowledgments
The first author is grateful to Plamen Koshlukov for useful discussions
held during the meeting in Ubatuba, Brazil.

References
[1] E. Aljadeff and D. Haile, Division algebras with a projective basis, Israel J. Math. 121
(2001) 173–198.
[2] E. Aljadeff, D. Haile, and M. Natapov, Projective bases of division algebras and groups
of central type, to appear in Israel J. Math.
[3] E. Aljadeff and J. Sonn, Projective Schur algebras have abelian splitting fields, J.
Algebra 175 (1995) 179–187.
[4] E. Aljadeff and J. Sonn, Projective Schur algebras of nilpotent type are Brauer equiv-
alent to radical algebras, J. Algebra 220 (1999) 401–414.
[5] Yu. Bahturin, S. Sehgal, and M. Zaicev, Group gradings on associative algebras, J.
Algebra 241 (2001) 677–698.
[6] R. Howlett and I. Isaacs, On groups of central type, Mathematische Zeitschrift 179
(1982) 555–569.
[7] I. Isaacs, Character Theory of Finite Groups, Dover Publications, New York, 1994.
[8] M. Natapov, Projective bases of division algebras and groups of central type II,
preprint.
[9] J.-P. Serre, Local Fields, Springer-Verlag, New York, 1979.
[10] M. Suzuki, Group Theory I, Springer–Verlag, New York, 1982.

© 2006 by Taylor & Francis Group, LLC


On observable module categories
Nicolás Andruskiewitsch
Facultad de Matemática, Astronomı́a y Fı́sica, Universidad Nacional de Córdoba,
CIEM – CONICET, (5000) Ciudad Universitaria, Córdoba, Argentina
[email protected]
Walter R. Ferrer Santos
Centro de Matemática, Facultad de Ciencias, Universidad de la República,
Iguá 4225 esq. Mataojo, CP 11400, Montevideo, Uruguay
[email protected]

¡Feliz cumpleaños César!

Introduction
Module categories should play the same fundamental role in the theory of
tensor categories, as representations do in the theory of groups. They were
introduced in [Be] and studied in several papers on category theory; they
appeared also in mathematical physics; see for example [FK]. Recently a
systematic study of module categories over fusion, or more generally finite,
tensor categories was undertaken in [O1, O2, ENO, EO]. In particular,
indecomposable module categories over the category of representations of
a finite group were classified in [O1] (characteristic 0) and [EO] (arbitrary
characteristic).
In the papers [ENO, EO], the authors consider rigid tensor categories with
appropriate finiteness conditions. One of the motivations for the discussions
in the present paper is the study of module categories over the tensor cate-
gory G M of rational modules over an algebraic group G. Furthermore, we
are interested in the induction functor from the category H M of represen-
tations of a closed subgroup H of G to G M. For this, one needs to consider
any rational module, not only the rigid (= finite-dimensional) ones. We are
naturally led to the notions of ind-rigid and geometric tensor categories, see
definition 1.2.
The purpose of the present paper is to begin the study of a class of module
categories over a tensor category C that we call observable module categories.
These module categories are simple in a suitable sense (that we introduce

1991 Mathematics Subject Classification. 16W30.


11

© 2006 by Taylor & Francis Group, LLC


12 N. Andruskiewitsch and W.R. Ferrer Santos

in this paper). If C = G M then the archetypical example is the module


category H M where H is an observable subgroup of G. We extend some
well-known results on observable subgroups to the setting of quotients of
Hopf algebras.
The notion of observable subgroups has the following geometric charac-
terization: a closed subgroup H of an algebraic group G is observable iff
the homogeneous space G/H is quasi-affine. This suggests that the study
of observable module categories could have a “non-commutative algebra”
flavor: they should correspond to “non-commutative quasi-affine varieties.”
Throughout, k denotes an arbitrary algebraically closed field, and all vec-
tor spaces, algebras, varieties, etc. are considered over k.

1. Preliminaries
In this section we recall some basic definitions of the theory of tensor
categories and of the theory of affine algebraic groups. We also introduce
the main definitions we shall work with.

1.1. Serre subcategories.


Let A be an abelian category and let A0 be a full abelian subcategory of
A. Recall that
• A0 is a Serre subcategory of A if subobjects, quotients, and finite
direct sums of objects in A0 are again in A0 .
• A0 is a closed subcategory of A if any object in A that is a colimit
of objects in A0 is itself in A0 .
• A0 generates A if any object in A is colimit of objects in A0 .
An example of the use of these abstract notions in a very general version
of duality theory is the following result. See also references in [NT].
Lemma 1.1. [NT, Lemma 3.2]. Let A and B be two Grothendieck cate-
gories. Let A0 and B0 be two generating Serre subcategories of A and B,
respectively. If F : A0 → B0 , G : B0 → A0 is an equivalence (resp. duality)
then there exists a unique extension F : A → B, G : B → A which is an
equivalence (resp. a pair of contravariant right adjoint functors). ¤
1.2. Tensor categories.
In this paper, by tensor category we understand an abelian k-linear tensor
category such that the unit object 1 is simple as in [BK], except that we
do not assume rigidity. This is because we are primarily interested in the
category of rational modules over an algebraic group, which is not rigid.
Given an object X in a tensor category C, a right dual of X is an object
X ∗ provided with maps
ev : X ∗ ⊗ X → 1, coev : 1 → X ⊗ X ∗ ,

© 2006 by Taylor & Francis Group, LLC


On observable module categories 13

called “evaluation” and “coevaluation,” such that the compositions


coev ⊗ id id ⊗ ev
X −−−−−−→ X ⊗ X ∗ ⊗ X −−−−−→ X (1.1)
id ⊗ coev ev ⊗ id
X ∗ −−−−−−→ X ∗ ⊗ X ⊗ X ∗ −−−−−→ X ∗ (1.2)
are the identities of X and of X ∗ , respectively.
Note that if X 6= 0 has a right dual then coev : 1 → X ⊗ X ∗ is monic.
Indeed, if coev is not monic then it is 0, as 1 is simple. But coev ⊗ id is
monic, thus X = 0. Analogously, ev : X ∗ ⊗ X → 1 is epi.
Similarly, a left dual of Y ∈ C is an object ∗ Y provided with maps
ev : Y ⊗ ∗ Y → 1, coev : 1 → ∗ Y ⊗ Y,
still called “evaluation and coevaluation,” such that the compositions
∗ coev ⊗ id id ⊗ ev
Y −−−−−−→ ∗ Y ⊗ Y ⊗ ∗ Y −−−−−→ ∗ Y (1.3)
id ⊗ coev ev ⊗ id
Y −−−−−−→ Y ⊗ ∗ Y ⊗ Y −−−−−→ Y (1.4)
are the identities of ∗ Y and of Y , respectively.
An object is rigid if it has a right and left dual.
If X ∈ C is rigid and Y, Z ∈ C then
C(Y, X⊗Z) ' C(X ∗ ⊗Y, Z), (1.5)

C(Y, Z⊗X ) ' C(Y ⊗X, Z). (1.6)
See [BK, Lemma 2.1.6].
A tensor category is rigid if every object is rigid. Our main interest here
is in a weaker notion of “ind-rigid tensor categories.”
Let C be a tensor category and let Crig be the full subcategory of rigid
objects. Then Crig is a rigid monoidal subcategory of C, but not necessarily
a Serre subcategory.
A subobject or a quotient object of a rigid object need not be rigid. For
example, let R be a commutative k-algebra (resp. any associative k-algebra
with unit) and C = R M the category of R-modules (resp. C = R MR the
category of R-bimodules) with ⊗ = ⊗R . Then an object M is rigid if and
only if it is finitely generated and projective as R-module (resp. both as a
left and right module); but quotients and submodules of finitely generated
projective modules in general do not inherit these properties.
On the other hand, if X and Y are rigid objects then X ⊕ Y is rigid.
Definition 1.2. An object in a tensor category is ind-rigid if it is the colimit
of a family of rigid objects.
A tensor category is ind-rigid if
(a) any object is ind-rigid,
(b) any subobject or quotient object of a rigid object is rigid.
An ind-rigid tensor category is geometric if in addition

© 2006 by Taylor & Francis Group, LLC


14 N. Andruskiewitsch and W.R. Ferrer Santos

(c) the tensor product functor is exact in both variables, and


(d) preserves colimits (on both sides).
Remark 1.3. If C is an ind-rigid tensor category then any object of C is the
colimit of a family of rigid sub-objects - take the epi-monic decomposition
of the arrows of an arbitrary colimit of rigid objects.
Remark 1.4. If C is an ind-rigid tensor category then Crig is a Serre subcat-
egory generating C. Thus, if C is a Grothendieck category then the duality
(contravariant) functor ∗ : Crig → Crig extends uniquely to ∗ : C → C.
Remark 1.5. In an ind-rigid tensor category, the simple objects are neces-
sarily rigid. Indeed, let S be a simple object and express it as a colimit of
rigid objects {dj : j ∈ J}. Then at least one of the arrows dj → S of the
associated cone is nonzero - otherwise S = 0 - hence S = Im dj is rigid.
Thus, not every tensor category is ind-rigid. For instance, the category
of representations of a Hopf algebra H is not ind-rigid if H has a simple
infinite-dimensional module, e.g., H = U (sl(2, k)).
Remark 1.6. If C is an ind-rigid tensor category and
(d’) the tensor product functor preserves colimits on one side,
then (d) holds. Indeed, assume that it preserves colimits on the left side.
Given x ∈ C, write it as a colimit x = colimj xj of rigid objects xj . If
y = colimi yi in C then
¡ ¢
x⊗y = (colimj xj )⊗(colimi yi ) ' colimj xj ⊗(colimi yi )
¡ ¢ ¡ ¢
' colimj colimi xj ⊗yi ' colimi colimj xj ⊗yi
¡ ¢ ¡ ¢
' colimi (colimj xj )⊗yi ' colimi x⊗yi ,
the first isomorphism since xj ⊗ is exact because xj is rigid.
Remark 1.7. Condition (c) fails in general, e.g., in R M or R MR . It is true
in rigid tensor categories, see [BK, Prop. 2.1.8]. Actually, the proof of [BK,
Prop. 2.1.8] shows that X ⊗ and ⊗ X are exact in an arbitrary tensor
category if X is rigid.
We do not know if (c) and (d) follow from (a) and (b), i.e., if any ind-rigid
tensor category is geometric.
A natural example of geometric tensor category is the category of comod-
ules over a Hopf algebra with bijective antipode.
1.3. Tensor functors and module categories.
Recall that a tensor functor between tensor categories is an additive func-
tor that preserves the tensor and the unit “up to a natural isomorphism.”
Technically, a tensor functor from a tensor category C to a tensor category
D is a triple (F, ζ, φ) where F : C → D is an additive functor, φ : 1 → F (1)
is an isomorphism in D, and ζX,Y : F (X)⊗F (Y ) → F (X⊗Y ) is a natural

© 2006 by Taylor & Francis Group, LLC


On observable module categories 15

transformation such that


F (aX,Y,Z )ζX⊗Y,Y (ζX,Y ⊗ idF (Z) ) = ζX,Y ⊗Z (idF (X) ⊗ζY,Z )aF (X),F (Y ),F (Z) ,
(1.7)
F (lX )ζ1,X (φ⊗ idF (X) ) = lF (X) , (1.8)
F (rX )ζX,1 (idF (X) ⊗φ) = rF (X) , (1.9)
for all X, Y, Z ∈ C.
If F : C → C 0 is a tensor functor and X is a rigid object of C then F (X)
is a rigid object of C 0 and F preserves duals. Thus F induces a monoidal
0 .
functor Frig : Crig → Crig
A module category over a tensor category C is an abelian k-linear category
M provided with an exact bifunctor ⊗ : C × M → M (the action) and
natural isomorphisms
mX,Y,M : (X ⊗ Y ) ⊗ M → X ⊗ (Y ⊗ M ), lM : 1 ⊗ M → M,
X, Y ∈ C, M ∈ M, such that the following diagrams commute:
((X ⊗ Y ) ⊗ Z) ⊗ M
j UUUU
aX,Y,Z ⊗idjjjjj UUmUX⊗Y,Z,M
jj UUUU
jjjj UUUU
jt jjj U*
(X ⊗ (Y ⊗ Z)) ⊗ M (X ⊗ Y ) ⊗ (Z ⊗ M )
mX,Y ⊗Z,M mX,Y,Z⊗M
² id⊗mY,Z,M ²
X ⊗ ((Y ⊗ Z) ⊗ M ) / X ⊗ (Y ⊗ (Z ⊗ M ))
(1.10)
and
mX,1,Y
(X ⊗ 1) ⊗ M / X ⊗ (1 ⊗ M ) (1.11)
OOO o
OOrOX ⊗id id⊗lM ooo
OOO oooo
OO' ow oo
X ⊗M
for all X, Y, Z ∈ C and M ∈ M.
In the definition, “exact bifunctor” means “it is exact in both variables.”
By analogy with the axioms (c) and (d) in the definition of geometric tensor
category, we shall also assume that the action satisfies the following axiom:
• the action functor ⊗ preserves colimits (on both sides).
Let M, M0 be two module categories over a tensor category C. A module
functor from M to M0 is a triple (F, b, u) where F : M → M0 is an
additive functor, bX,Y : F (X⊗Y ) → F (X)⊗F (Y ) is a natural isomorphism
and u : F (1) → 1 is an isomorphism, satisfying some natural compatibilities,
see [O2].
By a submodule category of a module M we understand a Serre subcate-
gory stable under the action.

© 2006 by Taylor & Francis Group, LLC


16 N. Andruskiewitsch and W.R. Ferrer Santos

Remark 1.8. As explained in [EO], it is worth restricting attention to


the class of exact module categories. Recall that M is an exact tensor
category over a finite tensor category C if P ⊗X ∈ M is projective for any
P ∈ C projective an any X ∈ M. Module functors from an exact module
category are always exact. Now, the inclusion functor from a full abelian
subcategory of an abelian category is exact precisely when the domain is a
Serre subcategory. This justifies our definition of submodule category.
The trivial module category is M = 0.
A module category is indecomposable if it is not isomorphic to a direct
sum of two nontrivial submodule categories [O1].
Definition 1.9. A nontrivial module category over a tensor category C is
simple if any proper submodule category is trivial.
1.4. Algebraic groups.
In this paper, we consider affine algebraic groups defined over k by “sub-
group” we mean a closed subgroup. If G is an affine algebraic group, we
denote as k[G] the Hopf algebra of polynomial functions on G.
Also we denote as G M, the category of all rational left G-modules (for
the basic properties of this category see [FR]). In particular it is well known
that G M is an abelian ind-rigid tensor category, where G Mrig is the full
subcategory of finite-dimensional rational G-modules. Clearly k[G] ∈ G M
when it is equipped with the left translation: (x, f ) 7→ x · f , for x ∈ G and
f ∈ k[G].
If G is an affine algebraic group and H a closed subgroup, then the coset
space G/H is naturally an algebraic G–variety. In the above situation the
restriction functor ResG H : G M → H M is a tensor functor. Then we can
consider H M as module category over G M, that is X ⊗ N = X|H ⊗ N , for
a rational H-module N and a rational G-module X. The restriction functor
has an adjoint, that we denote as IndG H : H M → G M, called the induction
functor. Explicitly,
IndG H
H (M ) = (k[G]⊗M ), M ∈ H M,
where the superscript H means that we are taking H-invariants. Note that
IndG
H has usually infinite dimension; this is why we can not work directly
with rigid tensor categories. The counit of this adjunction EM : IndG
H (M ) →
M for M ∈ H M, is called the evaluation map.
1.5. Observable subgroups.
Let G be an algebraic group and H a closed subgroup. Recall that a
rational character η : H → k is extendible, if there is a nonzero polynomial
f ∈ k[G] such that for all x ∈ H, x · f = η(x)f .
The following properties are equivalent:
(a) G/H is a quasi–affine variety.
(b) All rational characters χ : H → k are extendible.

© 2006 by Taylor & Francis Group, LLC


On observable module categories 17

(c) For all H–modules M ∈ H M, the evaluation map EM is surjective.


See [FR, Ch. 10] for the relevant definitions and proofs. If any of (a), (b),
(c) holds then one says that H is observable in G.
Here are some examples of subgroups that are observable (or not):
• If the only rational character χ : H → k is the trivial one, then H
is observable in any algebraic group containing it. Thus, semisimple
and unipotent algebraic subgroups are always observable.
• Finite subgroups are observable.
• Normal subgroups are observable.
• Finite-index subgroups are observable.
• A parabolic subgroup of a semisimple algebraic group is not observ-
able.

2. Observable tensor categories


2.1. Definition and basic properties.
Let F : C → D be a tensor functor between arbitrary tensor categories C
and D. Consider D as module category over C via F ; that is,
c ⊗ d := F (c) ⊗ d, c ∈ C, d ∈ D.
To illustrate the concepts in action, we prove the following result.
Lemma 2.1. Assume that F has a right adjoint I : D → C. Then the full
subcategory D0 of D with objects
D0 = {d ∈ D : I(d) = 0}
is stable under action by objects of Crig . If I preserves colimits then D0 is a
full abelian subcategory with the same kernels and cokernels as in D.
Proof. Let d ∈ D0 . If c ∈ Crig , x ∈ C, then
C(x, I(c⊗d)) ' C(x, I(F (c)⊗d)) ' D(F (x), F (c)⊗d)
' D(F (c)∗ ⊗F (x), d) ' D(F (c∗ ⊗x), d)
' C(c∗ ⊗x, I(d)) = 0,
where we have used (1.5). Thus c⊗d ∈ D0 for any c ∈ Crig as claimed. Since
I is a right adjoint, it preserves limits [M, Ch. V, §5, Th. 1].
If I preserves colimits then D0 is closed under colimits. Thus D0 is closed
under kernels and cokernels and is a full abelian subcategory of D. ¤
Here is one of the main definitions in the present paper.
Definition 2.2. Let F : C → D be a tensor functor between arbitrary tensor
categories C and D and assume that F has a right adjoint I : D → C. We
say that the category D is observable over C if the counit for the adjunction
εd : F (I(d)) → d is an epimorphism for all d ∈ D.

© 2006 by Taylor & Francis Group, LLC


18 N. Andruskiewitsch and W.R. Ferrer Santos

Here recall that εd corresponds to idI(d) via D(F (I(d)), d) ' C(I(d), I(d)).
The motivation for the above definition comes from the theory of algebraic
groups, see Section 1.4. Indeed, a subgroup H is observable in an algebraic
group G if and only if H M is observable in G M.
Theorem 2.3. In the notations above, if D is ind-rigid and observable over
C, then it is simple as a module category over C.
Proof. Let N ⊂ D be a nonzero submodule category. Take 0 6= m ∈ N and
write it as a colimit in D of rigid objects {dj : j ∈ J}, dj ∈ Drig . One of the
arrows of the cone fj : dj → m is nonzero, since otherwise m = 0. Using the
standard epi, monic factorization of fj we obtain a pair of morphisms in D
and an object 0 6= n ∈ D, h : dj → n, g : n → m with fj = gh, h epi and
g monic. As N is closed under subobjects, it follows that n ∈ N and as n
is a quotient of dj , it is rigid. In other words, N contains at least one rigid
object n.
In this situation the morphism εn∗ ⊗id : I(n∗ )⊗n = F (I(n∗ ))⊗n → n∗ ⊗n
is an epimorphism by the observability hypothesis. Since N is a submodule
category, I(n∗ ) ⊗ n ∈ N and we conclude that n∗ ⊗ n ∈ N , because N is
closed under quotients. Using the monomorphism given by the coevaluation
map coev : 1 → n∗ ⊗ n, we conclude that 1 ∈ N .
Consider now an arbitrary object d ∈ D; in this situation the morphism
εd ⊗ id : I(d) ⊗ 1 = F (I(d)) ⊗ 1 → d ⊗ 1 ' d is epi and we conclude that
d ∈ N . Thus N = D. ¤
2.2. Observable quotients of Hopf algebras.
Now let A be a Hopf algebra with bijective antipode and let MA be the
tensor category of right A-comodules. It is well known that MA satisfies
the properties required to be geometric.
Let π : A → B be a Hopf algebra quotient and let Res = ResB A : M →
A
B
M be the corresponding “restriction” functor; that is, if M is a right
comodule over A via ρ : M → M ⊗ A then Res M = M|B is the vector space
M considered as a right comodule over B via (id ⊗π)ρ. Clearly, Res is a
tensor functor.
If g ∈ G(B) is a group-like and M is a right B-comodule then M g denotes
the isotypic component of type g, that is
M g = {m ∈ M : ρ(m) = m ⊗ g}.
If g = 1, we denote M B = M 1 the subcomodule of invariant elements.
Let M ∈ MB . If we consider A ∈ MA via the comultiplication, then
A|B ⊗M ∈ MB with coaction ρ1 ; explicitly,
ρ1 (a⊗m) = a(1) ⊗m(0) ⊗π(a(2) )m(1) , a ∈ A, m ∈ M.
There is another right coaction of A on A⊗M by
ρ2 (a⊗m) = a(2) ⊗m⊗S(a(1) ), a ∈ A, m ∈ M.

© 2006 by Taylor & Francis Group, LLC


On observable module categories 19

It is easy to see that these two coactions commute. Then the space of
invariants (A|B ⊗M )B with respect to ρ1 is a right comodule over A via ρ2 .
In this situation we can define a map EM : (A|B ⊗M )B → M - the evaluation
map - by
¡X ¢ X X
EM fi ⊗mi = ε(fi )mi , fi ⊗mi ∈ A⊗M.
i i i
Then EM is a map of B-comodules, natural on M .
The following result is well known; see [F]. We sketch a proof for com-
pleteness. The proof does not differ from the analogous proof for algebraic
groups, see [FR, Th. 6.6.11].
Proposition 2.4. The “induction” functor Ind = IndA
B : M
B → MA ,

given by
IndA B
B (M ) = (A|B ⊗M ) , M ∈ MB ,
is a right adjoint to ResB
A . The counit for the adjunction is the map EM .

Proof. (Sketch). Let N ∈ MA . Given ϕ ∈ HomB (N, M ), define ϕ


e ∈
Hom(N, A|B ⊗M ) by
X
ϕ(n)
e = S −1 (n(1) )⊗n(0) , n ∈ N.
Straightforward calculations show that:
(a) The image of ϕe is contained in (A|B ⊗M )B .
(b) ϕe is a morphism of A-comodules and EM ϕ e = ϕ.

ϕ
e
N JJ / (A|B ⊗M )B
JJ
JJ
J
ϕ JJJ
EM
J$ ²
M
(c) A morphism of A-comodules N → (A|B ⊗M )B such that EM ϕ
e=ϕ
coincides necessarily with ϕ.
e
This proves the Proposition. ¤
We are then in the situation of definition 2.2; it is natural to introduce
the following notion.
Definition 2.5. A Hopf algebra quotient π : A → B is observable over A
(or B is observable over A) if the evaluation map EM : (A|B ⊗M )B → M is
an epimorphism for all M ∈ MB .
As for algebraic groups, there are alternative characterizations of observ-
ability. The proof of the following result mimics the proof of the analogous
result for algebraic groups, see [FR, Ch. 10].

© 2006 by Taylor & Francis Group, LLC


20 N. Andruskiewitsch and W.R. Ferrer Santos

Theorem 2.6. Let π : A → B be a Hopf algebra quotient. The following


are equivalent:
(a) B is observable over A.
(b) For every finite-dimensional M ∈ MB there exists a finite-dimensio-
nal N ∈ MA and an epimorphism of B-comodules ϕ : N|B → M .
(c) For every finite-dimensional M ∈ MB there exists a finite-dimensio-
nal N ∈ MA and a monomorphism of B-comodules M → N|B .
Furthermore, if any of these conditions hold then
(d) For every g ∈ G(B), there exists f ∈ (A|B )g such that π(f ) = g.

Proof. (a) =⇒ (b): Let s : M → (A|B ⊗M )B → M be a linear section


of the evaluation map EM . The A-comodule N generated by the image
s(M ) is finite-dimensional and ϕ = π|N : N|B → M is an epimorphism of
B-comodules.
(b) =⇒ (a): Assume first that M is finite-dimensional. Let ϕ : N|B → M
be an epimorphism of B-comodules and let ϕ e ∈ Hom(N, (A|B ⊗M )B ) be the
morphism of A-comodules such that EM ϕ e = ϕ. Since ϕ is surjective, EM is
surjective.
If M has arbitrary dimension, then it is the union of its finite-dimensional
subcomodules and the claim follows from the naturality of the evaluation
map EM .
(b) ⇐⇒ (c): by duality.
(b) =⇒ (d): Let M = kg and let N ∈ MA finite-dimensional provided
with an epimorphism of B-comodules ϕ : N|B → kg. If n ∈ N satisfies
ϕ(n) = g and α ∈ N ∗ satisfies α(n) = 1 then the matrix coefficient f =
φα,n ∈ A satisfies π(f ) = g. ¤
Remark 2.7. As we have seen, the implication (d) =⇒ (b) is also valid for
algebraic groups. It is likely that this implication is not always true. The
proof in [FR] makes use of the modules of antisymmetric tensors of a given
rational module, not a comodule in the case of a general Hopf algebra.
As for algebraic groups, cosemisimple Hopf algebra quotients are observ-
able.
Theorem 2.8. Let π : A → B be a Hopf algebra quotient. If B is cose-
misimple then B is observable over A.
Proof. Note that the following diagram commutes:
π⊗ id
(A|B ⊗M )B / (B⊗M )B
OOO
OOO
O eM
EM OOOO
E
O' ²
M

© 2006 by Taylor & Francis Group, LLC


On observable module categories 21
¡ ¢
where E eM : (B⊗M )B → M is again given by E eM P fi ⊗mi = P ε(fi )mi ,
i i
P e
f
i i ⊗m i ∈ B⊗M . Now it is clear that EM is an isomorphism, since
B B
(B⊗M ) = IndB (M ) is adjoint to the identity functor.
Thus, we have to prove that π⊗ id : (A|B ⊗M )B → (B⊗M )B is surjective.
But A|B ⊗M ' ker(π⊗ id)⊕(B⊗M ) since B is cosemisimple, and the needed
surjectivity follows taking the invariant submodules at both sides of this
isomorphism. ¤

3. Linearized Sheaves
Let G be an affine algebraic group. Let X be a G–variety and let G MX
be the abelian category of G–linearized sheaves on X, see [CPS, p. 453]. If
F is a G–linearized and if x ∈ X, we denote by Fx the stalk of the sheaf
at the point x. Then for every pair x ∈ X, g ∈ G there exists a linear
isomorphism ϕg,x : Fx → Fg·x , regular on the pair (g, x), in such a way that
if h ∈ G, the following diagram commutes:
ϕg,x
Fx F / Fg·x
FFF
FF ϕh,g·x
ϕhg,x FF
" ²
F(hg)·x

In particular, notice that the structure sheaf OX ∈ G MX .

Then G MX is a module category over G M. Indeed, if M ∈ G M and


F ∈ G MX , we define M ⊗ F ∈ G MX as the presheaf associated to the
sheaf (M ⊗ F)(U ) = M ⊗ F(U ), U open in X. Hence, for all x ∈ X,
(M ⊗ F)x = M ⊗ Fx and ψg,x : M ⊗ Fx → M ⊗ Fg·x is ψg,x = g · ⊗ ϕg,x .
If T : Y → X is a morphism of algebraic varieties and G is a sheaf
on Y , then we denote as T∗ (G) the “direct image” sheaf on X defined as
T∗ (G)(U ) = G(T −1 U ), for U ⊂ X open; see [H, p. 65]. Assume that X and
Y are G–varieties and that T intertwines the G–actions. If the sheaf G is
G–linearized, then T∗ (G) is also G–linearized. Moreover, in this situation,
we have a module functor T∗ : G MY → G MX .
Lemma 3.1. Let X be a G–variety and let Y be a closed G–stable subvariety
of X. Then we can identify G MY with a submodule category of G MX .
Proof. If Y is a closed subvariety of an arbitrary variety X and ι is the
inclusion, then ι∗ (G)x = Gx if x ∈ Y and zero otherwise, for any sheaf G on
Y [H, Ex. 1.19, p. 68]. This implies that we can identify G MY with the full
abelian subcategory of G MX consisting of those G–linearized sheaves F on
X such that FP for any P ∈ / Y . It is clear that subobjects and quotients of
objects in G MY are again in G MY . Thus, G MY can be identified with a
submodule category of G MX . ¤

© 2006 by Taylor & Francis Group, LLC


22 N. Andruskiewitsch and W.R. Ferrer Santos

Remark 3.2. Actually, G MY can be identified with a closed submodule


category of G MX , cf. subsection 1.1, because of [H, Ex. 1.12, p. 67]. We
wonder whether it is important to distinguish closed submodule categories.
Lemma 3.3. Let X be a G–variety. If the the category G MX is a simple
module category, then the action of G on X is transitive.
Assume further that the orbit map G → X is separable (this is automatic
if k has characteristic 0). Then X ' G/H for a closed subgroup H of G
with the left regular action.
Proof. Let Y be a closed orbit for the action of G on X. The submodule
category G MY is not zero since OY ∈ G MY . Thus, G MY = G MX . Hence,
the sheaf OX is in G MY , and this means that OX,x = 0 for all x ∈ X \ Y ,
therefore X = Y . The second statement is a well-known fact in the theory
of algebraic groups. ¤
Let H be a closed subgroup of G and let π : G → G/H be the canonical
projection.
Theorem 3.4. (a) The functor L : H M → G MG/H given by
¡ ¢
LN (U ) = H OG (π −1 (U )) ⊗ N
for a rational H-module N and U open subset of G/H, is an equivalence of
categories, whose inverse is given by taking the fiber at the identity:
F 7→ FeH , F ∈ G MG/H .
(b) L is a module functor. If K is a subgroup of H then the following
diagram commutes:
L
HM −−−−→ G MG/H
 
 π
Resy y ∗
L
KM −−−−→ G MG/K ,
where π : G/H → G/K is the canonical projection.
Proof. (a) is [CPS, Th. 2.7]. The proof of (b) is straightforward. ¤
Theorem 3.5. Assume that k has characteristic 0. If X is a G–variety
then the following are equivalent:
(a) G MX is isomorphic to an observable module category.
(b) X ' G/H for a closed subgroup H observable in G.
Proof. By Theorem 2.3, if such an isomorphism exists then G MX should
be simple. Then Lemma 3.3 says that X ' G/H, and Theorem 3.4, that
G MX ' H M. The claim now follows from the characterization of observable
subgroups, see subsection 1.4. ¤

© 2006 by Taylor & Francis Group, LLC


On observable module categories 23

The following problems arise naturally: Is the converse to Lemma 3.3


true? That is, what are the closed subgroups H such that H M is simple?
More generally, is any simple module category over G M twist-equivalent (in
some appropriate sense) to H M with H closed subgroup of G?

Acknowledgments
Results in this paper were obtained during visits of W.F.S. to the Uni-
versidad de Córdoba, November 2003, and of N.A. to the Universidad de la
República, November 2004. These visits had partial funding, respectively, by
the Fundación Antorchas and PEDECIBA. The work of N.A. was partially
supported by CONICET, Agencia Córdoba Ciencia, ANPCyT-FONCyT,
TWAS (Trieste) and Secyt (UNC). N.A. thanks Sonia Natale and Blas Tor-
recillas for interesting conversations. The work of W.F.S. was partially sup-
ported by CSIC-Udelar and Dinacyt-MEC, Uruguay. W.F.S. thanks Ignacio
López for many exchanges concerning category theory.

References
[BK] B. Bakalov and A. Kirillov, Jr., Lectures on tensor categories and modular func-
tors, University Lecture Series 21, Amer. Math. Soc., Providence, RI, 2001.
[Be] J. Bénabou, Introduction to bicategories, Reports of the Midwest Category Sem-
inar (1967), Lecture notes in Math. 47, pp. 1–77, Springer, Berlin.
[CPS] E. Cline, B. Parshall, and L. Scott, A Mackey imprimitivity theory for algebraic
groups, Math. Z. 182, pp. 447–471 (1983).
[ENO] P. Etingof, D. Nikshych, and V. Ostrik, On fusion categories, Annals Math., to
appear, math.QA/0203060.
[EO] P. Etingof and V. Ostrik, Finite tensor categories, Mosc. Math. J. 4 (2004), no.
3, 627–654, 782–783. math.QA/0301027.
[F] W. Ferrer Santos, Cohomology of comodules, Pacific J. Math. 109 (1983), 179–
213.
[FR] W. Ferrer Santos and A. Rittatore, Actions and invariants of algebraic groups,
Series: Pure and Applied Mathematics, 268, Dekker-CRC Press, Florida (2005).
[FK] J. Fröhlich and T. Kerler, Quantum groups, quantum categories and quantum
field theory, Lecture Notes in Math. 1542, Springer-Verlag, Berlin (1993).
[H] R. Hartshorne, Algebraic Geometry, 6th. corrected printing, Springer Verlag,
(1993).
[M] S. Mac Lane, Categories for the Working Mathematician, 2nd. edition, Springer
Verlag (1998).
[NT] C. Nastasescu and B. Torrecillas, Morita duality for Grothendieck categories
with applicatinos to coalgebras, Commun. Alg., to appear.
[O1] V. Ostrik, Module categories over the Drinfeld double of a finite Group, Int.
Math. Res. Not. 27, 1507–1520 (2003).
[O2] V. Ostrik, Module categories, weak Hopf algebras and modular invariants, Trans-
form. Groups, 2 8, 177–206 (2003).

© 2006 by Taylor & Francis Group, LLC


Group gradings on integral group rings
Yuri A. Bahturin
Department of Mathematics and Statistics,
Memorial University of Newfoundland, St. John’s, NL, A1C 5S7, Canada
[email protected]
Michael M. Parmenter
Department of Mathematics and Statistics,
Memorial University of Newfoundland, St. John’s, NL, A1C 5S7, Canada
[email protected]

To Professor César on his 60th birthday.

Abstract. In this paper we study group gradings on an integral group


ring Z G. We show that if G abelian and the grading group is any finite
abelian 2-group then any grading is “trivial,” that is, all elements of G
are homogeneous. In the case where G is nonabelian, we deal only with
grading by the cyclic group C2 of order 2. We indicate certain “natural”
gradings and show that if G is finite nilpotent, any C2 -grading of Z G is
one of them.

1. Introduction
Let A be an associative algebra over a commutative ring LR with 1. We
say that A is graded by a finite abelian group H if A = h∈H Ah where
each Ah is an R-submodule of A and Ah1 Ah2 ⊂ Ah1 h2 for all h1 , h2 ∈ H.
Our main concern in this paper will be the determination of all possible
gradings by abelian groups on integral group rings, that is, when A = ZG
and H is abelian.
This paper continues our research on gradings in groups rings starter in
[BP].

2. Abelian groups
The main significance of this section is to show that if G is a finite abelian
group and ZG is graded by C2 , the cyclic group of order 2, then all elements
of G are homogeneous. It follows that the only nontrivial C2 -gradings of ZG
in this case are of the form ZG = ZH ⊕ (ZH)g where H is a subgroup of
index 2 in G and g ∈ / H.
25

© 2006 by Taylor & Francis Group, LLC


26 Yu. A. Bahturin and M.M. Parmenter

The result we actually prove (Theorem 2.3) is quite a bit more general.
We start by recalling two results. The first lemma is well known but we
include it for completeness.
Lemma 2.1. Assume that an R-algebra A (R a commutative ring with 1)
is graded by a finite abelian group H. Then
(a) If 1 is the multiplicative identity of A then 1 is in Ae (where e is the
identity of H). P
(b) If z ∈ Z(A), the center of A, and z = zg where zg is in Ag for all
g in H, then zg is in Z(A) for all g in H.
P
Proof. (a) Assume that 1 = ag where ag is inPAg for each g. Let y ∈ Ah
be any homogeneous element of A. Then y = ag y and ag y is in Agh for
each g. It follows that y = ae y, and similarly y = yae . We conclude that ae
is a multiplicative identity for A, as desired. P
(b)PLet y ∈ Ah be any homogeneous element of A. Then yzg = yz =
zy = zg y. Since H is abelian and yzg is in Ahg = Agh , and zg y is in Agh ,
we have yzg = zg y for all g ∈ H. Hence zg is in Z(A) as desired. ¤
We will need the following important result about central units in integral
group rings.
Theorem 2.2. [MS, Corollary 7.3.3] If α is a central unit of finite order in
ZG, then α belongs to ±Z(G). ¤
Theorem 2.3. Assume ZG is graded by a finite abelian 2-group K. If
g ∈ Z(G) is of finite order then g is homogeneous.
Proof. We proceed by induction on |K|. First assume |K| = 2, that is
K = {e, x}, and ZG = Ae ⊕ Ax . Say g = αe + αx where αe ∈ Ae and
αx ∈ Ax . Since g is central in ZG, according to Lemma 2.1, αe and αx
are both central in ZG. Note that 1 = g n = (αe + αx )n . In the binomial
expansion of (αe + αx )n terms are alternately in Ae and Ax . Since 1 ∈ Ae ,
it follows that (αe − αx )n = 1 also. Since αe − αx is a central unit of finite
order in ZG, αe − αx must belong to ± G so that g − 2αx ∈ ± G. This can
only happen if either αx = 0 or αx = g and in either case we have that g is
homogeneous, as desired.
Next assume |K| > 2 and the result is true if ZG is graded
L by any finite
abelian 2-group K1 with 2 ≤ |K1 | < |K|. We have ZG = k∈K Ak .
Assume to the contrary that
P g ∈ G is not homogeneous so in the homo-
geneous decomposition g = k∈K αk of g there exist αx 6= 0 and αy 6= 0 for
some x 6= y in K. Note that if L is any subgroup of K we can grade ZG by
K/L in the usual way. It follows by the inductive hypothesis that xy −1 ∈ L
for all nonidentity subgroups L of K. This implies that K must be cyclic
and xy −1 = z where z is the unique element of order 2 in K. It also implies
(since this works for any such x, y) that g = αy + αyz . But now we can
conclude using exactly the same reasoning as seen before that αy − αyz is

© 2006 by Taylor & Francis Group, LLC


Group gradings on integral group rings 27

also a central unit of finite order in ZG and hence belongs to ±G. This gives
the desired contradiction, so g is homogeneous, as desired. ¤

Remarks.
(1) Theorem 2.3 says that if G is a finite abelian group and ZG is graded
by a finite abelian 2-group K then all elements of G are homoge-
neous. In addition, if G is a Hamiltonian 2-group then all units of
finite order in ZG are trivial and the proof of Theorem 2.3 shows
that the same result holds.
(2) The proof of Theorem 2.3 can also be extended to the case where ZG
is graded by Cp , the cyclic group of order an odd prime p, as long as
we know that central units of finite order in the group ring Z[ε](G)
are trivial (where ε is a primitive pth root of 1). Berman’s proof of the
result uses the fact that |r| ≥ 1 for all nonzero r in Z[ξ], and this √still
holds in Z[ξ](G) when p = 3. Indeed, a typical element in Z[− 21 + 23 i]
√ p
looks like a = z0 + z1 (− 21 + 23 i) and |a| = z02 + z12 − z0 z1 , proving

our claim. If we set ξ = − 21 + 23 i then the proof of Theorem 2.3 in
the case when ZG is graded by C3 will look like the following.
Say ZG is graded by {e, x, x2 }. Then also Z[ξ]G is graded by
{e, x, x2 }. Let g = α0 + α1 + α2 with αi ∈ Axi . Then α0 , α1 , α2
are central units and g n = (α0 + α1 + α2 )n = 1 for some natural
number n. If we expand the previous relation then it splits into
three components
X X X
αi αj αk = 1, αi αj αk = 0, αi αj αk = 0.
i+j+k=0 i+j+k=1 i+j+k=2

If we replace αi by αi ξ j then these relations will hold and so α0 +


α1 ξ + α2 ξ 2 and α0 + α1 ξ 2 + α2 ξ are also central units of finite order.
So α0 + α1 ξ + α2 ξ 2 = ξ i g1 and α0 + α1 ξ 2 + α2 ξ = ξ j g2 for some
g1 , g2 ∈ G and some integers i, j.
Hence (1 − ξ)(α1 + (1 + ξ)α2 ) = g − ξ i g1 . Since 1 − ξ is a nonunit
in Z[ξ] this forces g = g1 . Similarly g = g2 . Now we have the system

 α0 + α1 + α2 = g
α0 + α1 ξ + α2 ξ 2 = ξ i g

α0 + α1 ξ 2 + α2 ξ = ξ j g.

If α0 6= 0 then ξ i = ξ j = 1 and we easily derive α1 = α2 = 0 hence


g = α0 ∈ Ae . If α0 = 0 but α1 6= 0 then ξ i = ξ, ξ j = ξ 2 and α2 = 0
hence g = α1 ∈ Ax . Otherwise g = α2 ∈ Ax2 .
(3) Continuing on from remark 2, induction can also be applied showing
that Theorem 2.3 holds if ZG is graded by a finite abelian group of
order 2α 3β .

© 2006 by Taylor & Francis Group, LLC


28 Yu. A. Bahturin and M.M. Parmenter

(4) Again continuing on from remark 2, Berman’s proof fails for any
p ≥ 5 (if p ≥ 7 then | − 1 + ε| < 1 and if p = 5 then |1 + ε2 | < 1).
However we do not know whether Theorem 2.3 holds in this case.
We do know that if G is such that ZG has no nonzero nilpotent
elements, for example if G is abelian, and ZG is graded by Cp = hxip
and Axj 6= 0 for some 0 < j < p then Axi 6= 0 for all i. Indeed, if
α 6= 0 is in Axj then αs ∈ Axjs for all s. But αs 6= 0 because ZG has
no nonzero nilpotent elements. So if ZG has no nonzero nilpotent
elements and |G| < p then any grading of ZG by Cp is trivial.

3. General Results
In this section we will investigate C2 -gradings of ZG for arbitrary finite
groups G. If G has a subgroup H of index 2 and g ∈ G \ H then an
obvious C2 -grading is given by ZG = ZH ⊕ (ZH)g. Additionally, ZG =
γ(ZH) ⊕ γ((ZH)g) is a C2 -grading for any automorphism γ of ZG.
Question. Do the above represent all possible C2 -gradings of ZG?
The answer to this question is still not clear. Shortly we will give two
results, Theorem 3.2 and (especially) Corollary 3.3, which describe what
appears to be a more general method of constructing C2 -gradings. In fact,
Theorem 3.4 shows that this method characterizes all C2 -gradings of ZG
for a large class of finite groups G. A number of specific examples to this
construction are presented later in Section 4, but we have been unable to
verify that gradings arising from this method could not also be obtained by
a suitable choice of H and γ in the example mentioned above.
Our first result will be applied to integral group rings but is stated in a
much more general setting.
Proposition 3.1. Let R be a ring with 1 such that (R, +) is 2-torsion
free. Then there is a 1 − 1 correspondence between gradings of R by C2 and
automorphisms θ of R such that θ2 = I (the identity automorphism) and
θ(r) − r ∈ 2R for all r ∈ R.
Proof. First assume R = R0 ⊕R1 is a C2 -grading. Let θ be the automorphism
of R defined by θ(x + y) = x − y where x ∈ R0 , y ∈ R1 . Then θ satisfies the
given conditions.
Conversely assume θ is any automorphism of R satisfying the given con-
ditions. Define R0 = {r ∈ R | θ(r) = r} and R1 = {r ∈ R | θ(r) = −r}.
Clearly, R02 ⊂ R0 , R12 ⊂ R0 R0 R1 ⊂ R1 , and R1 R0 ⊂ R1 . The second
assumption tells us that if r ∈ R then r + θ(r) = 2r0 and r − θ(r) = 2r1 for
some r0 , r1 ∈ R.
By the first assumption, we know that r + θ(r) is in R0 , so θ(2r0 ) = 2r0 .
Similarly, r − θ(r) is in R1 so θ(2r1 ) = −2r1 . Since (R, +) has no 2-torsion,
r0 ∈ R0 and r1 ∈ R1 . Also, 2r = 2r0 + 2r1 , which implies (as above)

© 2006 by Taylor & Francis Group, LLC


Group gradings on integral group rings 29

that r = r0 + r1 . Since R0 ∩ R1 = 0 (again using no 2-torsion), we have


R = R0 ⊕ R1 is a C2 -grading.
Finally, observe that the lack of 2-torsion in (R, +) tells us that the rela-
tions just described are inverses of each other. ¤
In addition to the statement of Proposition 3.1, the precise nature of the
1 − 1 correspondence will be used later on.
Examples.
(1) Let H be a subgroup of index 2 in a group G. Define ιH to be
the automorphism of ZG induced by ιH (g) = g if g ∈ H, ιH (g) =
−g if g ∈/ H. Then ιH satisfies the required conditions, and the
corresponding grading is ZG = ZH ⊕ (ZH)g where g ∈ / H.
(2) More generally, if H is a subgroup of index 2 in G and γ is any
automorphism of ZG then θ = γ ◦ ιH ◦ γ −1 satisfies the required
conditions. The corresponding grading is ZG = γ(ZH) ⊕ γ((ZH)g),
noted earlier.
We now proceed to show that if R = ZG then the automorphisms of the
type described in Proposition 3.1 can be obtained from certain types of units
in the rational group algebra QG. Specific examples of such construction
are given in Section 4. For a large class of finite groups, including finite
nilpotent groups, Theorems 3.2 and 3.4 together give a characterization of
all C2 -gradings of ZG.
In the following results ιH has the same meaning as in the previous ex-
ample. In addition, when H = G ιH is the identity automorphism.
Theorem 3.2. Let H be a subgroup of index ≤ 2 in G and β an auto-
morphism of G such that β(H) = H (we will also use β to represent its
extension to an automorphism of ZG). Assume that a unit u in QG satis-
fies the following two conditions:
(i) For any g ∈ G, (u(β ◦ ιH (u)))β 2 (g)(u(β ◦ ιH (u)))−1 = g;
(ii) Inn u, the inner automorphism of QG induced by u, is an automor-
phism of ZG when restricted to ZG. Moreover, uβ(g)u−1 − g ∈ 2ZG
for all g ∈ G.
Then, restricted to ZG, the automorphism (Inn u) ◦ β ◦ ιH satisfies the con-
ditions of Proposition 3.1.
Proof. Let θ = (Inn u) ◦ β ◦ ιH . For any s ∈ ZG,

θ2 (s) = θ(u((β ◦ ιH )(s))u−1 )


= u((β ◦ ιH )(u((β ◦ ιH )(s))u−1 ))u−1
= u((β ◦ ιH )(u))((β ◦ ιH ◦ β ◦ ιH )(s))((β ◦ ιH )(u−1 ))u−1
= (u((β ◦ ιH )(u)))(β 2 (s))(u((β ◦ ιH )(u)))−1

© 2006 by Taylor & Francis Group, LLC


30 Yu. A. Bahturin and M.M. Parmenter

since β(H) = H. Property (i) now tells us that θ2 (s) = s, so θ2 = I.


Also, if g ∈ G then θ(g) equals uβ(g)u−1 or −uβ(g)u−1 . In either case
property (ii) tells us that θ(g) − g ∈ 2ZG. ¤
Later we will give examples of units satisfying the conditions of Theorem
3.2. In all such examples β will be the identity automorphism and u will be
in ZG, so we restate Theorem 3.2 in this simpler setting.
Corollary 3.3. Let H be a subgroup of index ≤ 2 in G. Assume that a unit
u in ZG satisfies the following two conditions.
(i) uιH (u) is central in ZG;
(ii) ugu−1 − g ∈ 2ZG for all g ∈ G.
Then, if Inn u denotes the inner automorphism of ZG, the automorphism
θ = (Inn u) ◦ ιH satisfies the conditions of Proposition 3.1.
We now show that, for a large class of groups, the conditions in Theorem
3.2 characterize C2 -gradings of ZG.
We recall first one of the Zassenhaus Conjectures (see [S, S2, p. 205]).
AUT. If θ is an augmentation preserving automorphism of ZG, then there
exist β ∈ Aut G and w ∈ QG such that θ(g) = wβ(g)w−1 for all g ∈ G.
Roggenkamp and Scott showed that AUT holds for finite nilpotent groups,
and also give an example of a finite metabelian group where AUT does not
hold (see S2, Section 37, for more information).
Theorem 3.4. Let G be a group such that AUT holds in ZG and assume
θ is an automorphism of ZG satisfying the conditions of Proposition 3.1.
Then θ = (Inn u) ◦ β ◦ ιH for some subgroup H of index ≤ 2 in G, some
automorphism β of G (extended to ZG) satisfying β(H) = H and some unit
u of QG satisfying conditions (i) and (ii) of Theorem 3.2.
Proof. Let ZG = R0 ⊕ R1 be the C2 -grading associated with θ, θ(x0 + x1 ) =
x0 − x1 where xi ∈ Ri for each i. If ε is the augmentation map in ZG, define
H = {g ∈ G | ε(θ(g)) = 1}.
It is easily checked that H is a subgroup of index ≤ 2 in G. Moreover,
θ ◦ ιH is an augmentation preserving automorphism of ZG, so the fact that
AUT holds tells us that θ ◦ ιH = Innu ◦ β where β is an automorphism of
ZG induced by an automorphism of G, u is a unit of QG and Innu is as
described in the first sentence of condition (ii) of Theorem 3.2. We then get
θ = Innu ◦ β ◦ ιH . As in the proof of Theorem 3.2 we see that θ2 = I implies
(u((β ◦ ιH )(u)))((β ◦ ιH ◦ β ◦ ιH )(g))(u((β ◦ ιH )(u)))−1 = g.
for all g ∈ G. This means that the augmentation of the lefthand side is 1 for
all g and this implies β(H) = H. It is easy to see that this means condition
(i) is satisfied. Also θ(g)−g ∈ 2ZG for all g implies condition (ii) is satisfied,
and the proof is complete. ¤

© 2006 by Taylor & Francis Group, LLC


Group gradings on integral group rings 31

As noted earlier, finite nilpotent groups are known to satisfy AUT, so


Theorem 3.4 applies to this case. However, we can get some additional
information.

Proposition 3.5. Let G be finite nilpotent and assume ZG = R0 ⊕ R1 ,


graded by C2 . If R1 ⊂ ∆(G) then R1 = 0.

Proof. We proceed by induction on |G|. If G is abelian then Theorem 2.3


tells us all group elements are homogeneous. Since R1 ⊂ ∆(G) we must
have G ⊂ R0 and so R0 = ZG and R1 = 0.
Assume the result holds for all nilpotent groups of order less than |G| and
choose z ∈ Z(G) of prime order p. Theorem 2.3 tells us z is homogeneous,
and since R1 ⊂ ∆(G) this means z ∈ R0 . Thus (1 − z)ZG is a graded
ideal of ZG, and so the grading on ZG induces a C2 -grading on ZG/(1 −
z)ZG ∼ = Z(G/hzi) = R̄0 ⊕ R̄1 . By the induction hypothesis R̄1 = 0, so
R1 ⊂ (1 − z)ZG.
Let r1 ∈ R1 , so r1 = s(1 − z) for some s ∈ ZG. Now s = s0 + s1 ,
si ∈ Ri for each i. Since 1 − z is in R0 this implies r1 = s1 (1 − z). Since
s1 ∈ R1 , s1 = t(1 − z) and, again, if t = t0 + t1 , then s1 = t1 (1 − z) giving
r1 = t1 (1 − z)2 . Continuing we get r1 is in (1 − z)n ZG for all n. Since
(1 − z)p is in p ZG, r1 = 0. ¤

Returning to the proof of Theorem 3.4, when G is finite nilpotent Propo-


sition 3.5 tells us that if θ is not the identity automorphism then θ cannot be
augmentation preserving (since R1 6= 0). It follows that when G is nilpotent
the subgroup H obtained in this theorem cannot be equal to G, a fact that
is not at all clear without Proposition 3.5 (an may not be true for general
finite G).

4. Examples
We now present some examples of units which satisfy the conditions in
Corollary 3.3.
(a) If u is a central unit in ZG, both conditions of Corollary 3.3 hold.
But then θ = ιH so we have nothing new.
(b) Let G = ha, b | a8 = 1, b2 = 1, ba = a5 bi. Take H = ha2 , bi. If
u = (1 + a + a2 )2 − â, then

uιH (u) = ((1 + a + a2 )2 − â)((1 − a + a2 )2 − ιH (â))


= (1 + a2 + a4 )2 − 2(1 + a2 + a4 + a6 )

© 2006 by Taylor & Francis Group, LLC


32 Yu. A. Bahturin and M.M. Parmenter

which is central. To check condition (ii) we need only consider


uau−1 − a and ubu−1 − b. The first equals 0 and
ubu−1 − b = [(1 + a + a2 )2 − â]b[(1 + a3 + a6 )2 − â] − b
= [(1 + a + a2 )2 − â][(1 + a7 + a6 )2 − â]b − b
= [(1 + a + a2 )2 − â]2 a4 b − b
= (−8 − 6a + 6a3 + 9a4 + 6a5 − 6a7 )a4 b − b
which is in 2ZG.
(c) Let G = ha, b | a8 = 1, b2 = 1, ba = a3 bi and take H = ha2 , bi. If
u = (1 + a + a2 )4 − 10â, then it is not hard to see that conditions (i)
and (ii) are satisfied.
(d) Let G = ha, b | a16 = 1, b2 = 1, ba = a7 bi and take H = ha2 , bi. If
u = (1 + a + a2 )8 − 410â, again conditions (i) and (ii) are satisfied.
Cautioning Example.
Finally here is an example when a unit satisfies conditions (i) and (ii) of
Corollary 3.3 but does NOT give a new grading.
Let G be the dihedral group of order 8, that is, G = ha, b | a4 = 1, b2 =
1, ba = a3 bi. Let H = ha2 , bi. If u = 1 + (1 − b)a(1 + b) then it is easily
checked that u satisfies conditions (i) and (ii). But the resulting grading is
the same as
1
ZG = v(ZH)v −1 ⊕ v((ZH)a)v −1 where v = 1 + (1 − b)a(1 + b) in QG.
2
Acknowledgments
This research was supported in part by NSERC.

References
[BP] Yu. A. Bahturin and M.M. Parmenter, Generalized commutativity in group al-
gebras, Canad. Math. Bull. 46 (2003), 14-25.
[MS] C. Polcino Milies and S.K. Sehgal, An Introduction to Group Rings, Algebras
and Applications, Kluwer, 2002.
[S] S.K. Sehgal, Topics in Group Rings, Marcel Dekker, 1978.
[S2] S.K. Sehgal, Units in Integral Group Rings, Longmans, 1999.

© 2006 by Taylor & Francis Group, LLC


Profinite graphs – comparing notions
Gunther Bergauer and Wolfgang Herfort
Institute for Analysis and Scientific Computation, University of Technology,
Vienna, Austria
[email protected]
[email protected]

Dedicated to C. Polcino Milies, S. Sehgal, and S. Sidki.

Abstract. Defining a profinite graph to be a projective limit of an


inverse system of finite graphs can lead to different objects depending
upon which of the (usually) equivalent concepts of a finite graph is
employed.

1. Categories of (profinite) graphs


Graphs can be oriented - each edge has an origin and a terminal vertex -
or nonoriented. In the first situation we shall talk about digraphs, otherwise
about graphs.
Starting with precise definitions, let us recall Serre’s definition of a graph
[5], which we would like to term a Serre digraph.

Definition 1.1. A digraph consists of a set Γ together with two functions


o, t : Γ → Γ satisfying ot = t and to = o. The vertices V (Γ) of the digraph
is the set o(Γ) ∪ t(Γ) and the set of (directed) edges E(Γ) is Γ \ V (Γ).
A quadruple (Γ, o, t, ι) where (Γ, o, t) is a digraph and ι an involution
acting on Γ with set of fixed points V (Γ) is a Serre digraph. Morphisms in
this category are morphisms of digraphs φ : Γ → ∆ with φι = ιφ. Serre uses
bar ¯ for ι.
A morphism is strict if φ(E(Γ)) ⊆ E(φ(Γ)).

Any concept of a profinite graph, known to the authors, arises by

• replacing the sets Γ, V (Γ) by boolean spaces (i.e., projective limits of


finite sets; see Definition 1.2 below) and letting o, t (and, after having
given Definition 1.5, δ) be continuous with analogous properties as
in the respective finite situation.
33

© 2006 by Taylor & Francis Group, LLC


34 G. Bergauer and W. Herfort

• characterizing any such profinite graph as a projective limit of an


inverse system of finite graphs and morphisms in the respective cat-
egory.
Definition 1.2. A boolean or profinite space is a compact totally discon-
nected Hausdorff space. Alternatively it is the projective limit of an inverse
system of finite sets [4].
A category of boolean spaces has objects, boolean spaces, and morphisms
continuous functions, termed maps.
For a finite group G a boolean space X is a G-space, if G acts continuously
upon X. Morphisms in the category of G-spaces are maps respecting the
action.
The above recipe leads to profinite graphs as defined in [4].
Definition 1.3. A profinite digraph is a boolean space Γ which is a digraph
with o and t maps. A morphism φ : Γ → ∆ is a morphism of digraphs which
is a map.
Alternatively, Γ is the projective limit of an inverse system of finite di-
graphs.
When E(Γ) is closed the graph is boolean in the sense of [2], where it was
shown that every such graph is the projective limit of an inverse system of
finite digraphs with strict morphisms.
Then morphisms are strict maps.
A profinite Serre digraph is a Serre digraph which at the same time is a
hιi-space. Such a graph is the projective limit of finite Serre digraphs.
A different concept of graph is found, e.g., in [6], although no definition for
morphisms is given. We ¡find ¢ it convenient to denote for any set X the set of
X
all n-element subsets by n and at most n-element subsets by F≤n (X). For
the profinite situation the Vietoris topology on the set of all closed subsets
of a boolean space appears to be suitable.
Definition 1.4. For a topological space X, the Vietoris-topology on the set
C(X) of all nonempty closed subsets of X has a basis of open sets
k
[
W (O1 , . . . , Ok ) := {A ∈ C(X) | ∀j = 1, . . . , k : A ∩ Oj 6= ∅ ∧ A ⊆ Oj }
j=1

with k ∈ NI and Oj running through all clopen subsets of X.


S
Since, for clopen sets O1 , . . . , Ok , U1 , . . . , Ul , setting O := ki=1 Oi and
S
U := lj=1 Uj , as a consequence of the definition one has
W (O1 ∩ U, . . . , Ok ∩ U, O ∩ U1 , . . . , O ∩ Ul ) = W (O1 , . . . , Ok ) ∩ W (U1 , . . . , Ul ),
these sets indeed form a basis for the topology. We claim that every
W (O1 , . . . , Ok ) is also closed. Indeed, again as an immediate consequence

© 2006 by Taylor & Francis Group, LLC


Profinite graphs – comparing notions 35

S
of the definition, setting O := ki=1 Oi , and letting dash denote the comple-
ment of a subset of X in X, one finds
k
[
0
C(X) \ W (O1 , . . . , Ok ) = W (O , O1 , . . . , Ok ) ∪ W (Oi0 ),
i=1

to be a finite union of open sets, i.e., it is closed as well.


When X is finite and discrete, then C(X) consists of all nonempty subsets
and it is not hard to see that the Vietoris-topology coincides with the discrete
topology on C(X). When (Xα , φαβ ) is an inverse system of finite sets,
then there is an inverse system (C(Xα ), φ∗αβ ), where φ∗αβ sends each set
A ∈ C(X) to {φαβ (a) | a ∈ A}. Then, when X = lim Xα , it turns out
←−α
that C(X) = lim C(Xα ) and the Vietoris-topology coincides with the one,
←−α
when C(X) is considered as a profinite space.
The definitions show that for any boolean space X the sets F≤n (X) for
all n ∈ NI are closed subsets of C(X).
Definition 1.5. A graph (Γ, δ) consists of the following data:
(1) a set Γ and a subset V (Γ) (the vertices);
(2) a function δ : Γ → F≤2 (V (Γ)) satisfying δ(v) = {v} for all v ∈ V (Γ).
Then E(Γ) := Γ \ V (Γ) is the set of its edges.
When v ∈ δ(e) we say that v and e are adjacent. A function φ : Γ → Γ0
is a morphism if δ 0 φ = φ∗ δ. Here φ∗ denotes the function from F≤2 (Γ) →
F≤2 (Γ0 ) induced by φ, i.e., φ∗ ({a, b}) := {φ(a), φ(b)} for any a, b ∈ V (Γ).
A Serre graph is a triple (Γ, δ, ι) with (Γ, δ) a graph and ι an involution
acting on Γ with fixed point set V (Γ).
A profinite graph is a graph (Γ, δ) with Γ a boolean space, V (Γ) a closed
subset of Γ, F≤2 (Γ) equipped with the Vietoris topology (Definition 1.4)
and δ a map. Morphisms in the category of profinite graphs are morphisms
of the underlying graphs which at the same time are maps.
A profinite Serre graph is a Serre graph (Γ, δ, ι) with (Γ, δ) a profinite
graph and Γ at the same time a hιi-space.
An edge e with δ(e) containing a single element only is a loop.
Let us remark:
(1) There is the notion of ¡simple
¢ graph [3]. It is stipulated that δ is
V
injective with values in 2 . When δ is not injective, the graph has
multiple edges and is sometimes called a multigraph.
(2) It is common to use strict morphisms in graph theory [3], i.e., edges
go to edges and vertices to vertices - no mixing. During the present
article we allow collapsing edges for better coherence with Defini-
tion 1.1.

© 2006 by Taylor & Francis Group, LLC


36 G. Bergauer and W. Herfort

(3) In finite graph theory these concepts are equivalent at least for ob-
jects in the respective categories in the following sense: given a di-
graph, one easily constructs a Serre digraph by introducing an in-
verse edge ē for every e. Conversely for a Serre digraph select from
each pair of edges exactly one. Every digraph can be produced from
a Serre digraph in this manner.
Next, given a digraph Γ, define δ : E → F≤2 (V ) by setting δ(e) :=
{o(e), t(e)}. Similarly as before one can see that each digraph gives
rise to a unique graph and conversely, every graph arises in this way.
Our main concern is showing that in the profinite situation,
(1) from each digraph (Γ, o, t) one can produce a profinite Serre digraph
by “doubling” the space of edges with Γ naturally embedded, but
there exists a profinite Serre digraph not containing a naturally em-
bedded profinite digraph.
(2) from every profinite digraph one can construct a profinite graph by
“forgetting” orientation, but not every profinite graph arises in this
way.
(3) every profinite Serre graph gives rise to a profinite graph and
conversely (Lemma 2.6).
(4) every Serre digraph gives rise to a profinite graph by “identifying”
each edge e with ē but not every profinite graph arises in this way.
Taking (3) into account, our comments can be displayed in a diagram:

Lemma 2.4
Example 3.6
-
digraph Serre digraph
@ ¡
Lemma 2.3 @ ¡Lemmata 2.5 & 2.6
Corollary 3.5 Example 3.4
@ ¡
R
@ ª
¡
graph
When the underlying boolean space(s) are second countable, then, simi-
larly as in the discrete situation, the concepts are equivalent, i.e., the arrows
in the diagram can be “reverted” (see Lemmata 2.3 and 2.4).

2. Relations between the concepts


Definition 2.1. Given a profinite graph (Γ, δ), then a profinite digraph
(Γ, o, t) is a coherent digraph presentation provided δ(m) = {o(m), t(m)},
whenever m ∈ Γ.
Lemma 2.2. Let G ∼ = C2 and X a G-space. When X is second countable
or C2 acts freely there is a continuous section from X/G to X.
Proof. This follows from Lemmata 5.6.5 and 5.6.7 in [4]. ¤

© 2006 by Taylor & Francis Group, LLC


Profinite graphs – comparing notions 37

Lemma 2.3. Given a digraph (Γ, o, t), let the graph (Γ̃, δ) be defined by
setting E(Γ̃) := E(Γ), V (Γ̃) := V (Γ) and δ(m) := {o(m), t(m)}. Then
(Γ, o, t) is a coherent digraph presentation of (Γ̃, δ).
Every second countable profinite graph (Γ̃, δ) has a coherent digraph pre-
sentation.
Proof. The continuity of δ is a consequence of F≤2 (V (Γ)) being homeomor-
phic to Γ × Γ/hαi, where α permutes the coordinates.
For the second statement, let (Γ̃, δ) be a second countable profinite graph.
Then Y := Γ̃ × Γ̃ is a C2 -space where the involution permutes coordinates
with Y /C2 homeomorphic to F≤2 (Γ̃). By Lemma 2.2 there exists a contin-
uous section σ : Y /C2 → Y . For every edge define o(e) := σ(δ(e)C2 /C2 )
and let t(e) be the edge in δ(e) \ {o(e)} if the latter is nonempty and o(e)
else. ¤
Lemma 2.4. When (Γ, o, t) are the data of a profinite digraph then data
(Γ0 , δ 0 , ι) of a profinite Serre digraph can be constructed as follows: Set Γ0 :=
ΓqV (Γ) Γopp and let ι denote the evident involution sending each edge e ∈
E(Γ) to eopp ∈ Γopp and each vertex v ∈ V (Γ) to v = v opp ∈ V (Γopp ).
Here (Γopp , δ opp ) is the digraph with (oopp , topp ) := (t, o).
Every second countable profinite Serre graph arises in this manner.
Proof. Observe the existence of a homeomorphism ψ : Γ → Γopp sending
v ∈ V (Γ) to v and e ∈ E(Γ) to eopp ∈ E(Γopp ). Then ψ (and its inverse)
give rise to an involutory homeomorphism ι of ΓqΓopp which factors via the
identification of the identical copies of V (Γ) in the fixed set V (Γ)qV (Γopp )
under the involution inside ΓqΓopp . Finally set δ 0 (e) is equal to δ 0 (eopp ) :=
{o(e), t(e)}.
The proof of the last statement is similar to the one in Lemma 2.3. ¤
In the situation of the previous Lemma we say that (Γ, o, t) is an oriented
shadow of (Γ0 , δ 0 , ι).
Lemma 2.5. Given a profinite Serre digraph (Γ, o, t, ι) a profinite Serre
graph (Γ0 , δ 0 , ι) can be constructed by setting Γ0 := Γ/hιi and δ 0 (ē) :=
{o(e), t(e)}. Here bar denotes passing to the hιi-orbit.
Proof. We omit the easy proof. ¤
Lemma 2.6. For a profinite Serre graph (Γ, δ, ι) let Γ0 be the quotient of Γ
under the action of the involution and define δ 0 by setting δ 0 (m) := {m, m̄}.
Then (Γ0 , δ 0 ) are the data of a profinite graph.
Conversely, given a profinite graph (Γ, δ) let h be an isomorphism of Γ
onto a copy Γ0 of Γ. Form the profinite graph ΓqΓ0 and amalgamate by
identifying v with h(v) for all v in V (Γ). For all e ∈ E(Γ) define δ 0 (e) :=
δ(e) =: δ 0 (h(e)). Then (Γ0 , δ 0 , ι) is a profinite Serre graph whose quotient
according to the first part of the Lemma is (Γ, δ).

© 2006 by Taylor & Francis Group, LLC


Another random document with
no related content on Scribd:
hänen vuoteensa vierestä.

Paitsi kulunutta pukua merkittiin se vainajan ainoaksi


jälkeenjättämäksi omaisuudeksi.

ELÄKE
<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<
<<<<<<<<<<44

— Kas niin, vanha veikko! tervehti lääkäri istahtaessaan Esko Kiilan


sairasvuoteen ääreen. — Unta on tullut sentään — vai?

Sairas murahti jurosti ja lyhyeen, jonka lääkäri vanhana toverina


ymmärsi myöntymykseksi ja johon hän oli täydellisesti tyytyväinen.
Monena muuna päivänä oli hän saanut tyytyä niukempaankin
vastaukseen.

Sairaan kädet, jotka velttoina lepäsivät peitteellä, alkoivat liikkua,


ja toisen niistä hän siirsi vitkalleen tyynylle päänsä taakse, katseen
sillä aikaa harhaillessa ympäri huonetta, jonka seinillä riippui pyssyjä,
laukkuja ja kaikenlaisia metsämiehen kapistuksia.

— Saisit sen sanoa suoraan ja peittelemättä, Riska.

— No niin, kyllä minä luulen…

— Sinä tiedät sen.

— Tiedän, veli, tiedän.


Sairas ummisti silmänsä ja henkäisi pitkään, henkäisi toisenkin
kerran ja lausui sitten:

— Se helpoitti se — kiitos.

Lääkäri oli koko ajan tarkannut potilaansa parroittuneita kasvoja,


joista korkea nenä kohosi kuin kivi kaislikosta ja joihin silmät,
raukeudestaan huolimatta, loivat sentään jonkun verran eloa.

Moneen päivään ei tämän sairasvuoteen ääressä oltu vaihdettu


näin lukuisia sanoja.

— Sinä otat sen rauhallisesti vastaan, vanha, uljas toveri. Vaikka


minä pelkäsin…

— Pelkäsit. Ei kai tämä sinun ensimäinen kuolemantuomiosi ollut?


Lienet niitä jo sadoittain tautivuoteille kyydinnyt…

— Tovereista olet ensimäinen…

Sairaan ja lääkärin katseet yhtyivät hetkeksi. Kumpaisenkin


mieleen välähti kuvia elämän aikaisemmilta ajoilta. Yhteisten
muistojen runsaudesta olisi saattanut koota paksun kirjan, joka ei
varmaankaan olisi jäänyt omaa ominaista viehätystään vaille.

Nyt oli elämän syyspuoli kumpaisellakin. Hänelläkin, joka toi


kuoleman viestin veljelleen, oli aika lailla harmaata ohimoilla.

— Laura Paatero on käynyt täällä, virkahti Esko Kiila


äänettömyyden jälkeen, eikä lääkäri kummastellut enää
vähäpuheisen potilaansa puheliaisuutta. Kun hän oli jo näin paljon
puhunut, saattoi hän kai puhua enemmänkin…
— Laura Paatero… Se tahtoo sanoa?…

— Sinä et muista. Ollin sisar, jonka kanssa olin kihloissa muinaisina


päivinä.

— Kas, nythän sen muistankin. Veli, sinähän sait sitten omituisen


vastenmielisyyden morsiantasi kohtaan, ja kaikki raukesi.

— Selittämättömän vastenmielisyyden — ja ilman pienintäkään


syytä. Sitä tapahtuu jos jonkinlaista avarassa maailmassa.

— No — ja hän? Mitä hän nyt tahtoi?

Sairas rypisti kulmiaan ja sulki silmänsä kuin päästäkseen jotakin


kiusallista näkemästä. Vasta pitkän hetken perästä hän puhui:

— Sinä muistanet ehkä kuinka menettelyni masensi hänet. Mutta


minä en voinut toisin. Enkä tahtonut tietää enkä kuullakaan mitään
hänestä. Ja kaikki mitä sitten pitkien vuosien kuluessa on tietooni
tullut, supistuu siihen, että hän on elänyt yksin kuten minäkin.

— Ja tämän kertoi hän itsekin käydessään sinulle?

— Niin, ja myöskin sen, että oli hyljännyt enemmän kuin yhden


arvokkaan kosijan — hänellähän oli näköä kylläkin, hänellä — ja että
oli hukannut terveytensä suruunsa… ja… ja että toimeentulonsa on
nykyisin kaikkea muuta kuin tyydyttävä.

Seurasi taas äänettömyys.

Sen kuluessa otti lääkäri sairaan käden peitteeltä painaen


sormensa valtimolle, mutta katseensa antoi hän kulkea ympäri
huonetta ja kuvitteli mielessään miltä täällä mahtoi näyttää silloin,
kun Laura Paatero seisoi tämän vuoteen ääressä kertoen hukkaan
menneestä elämästään. Mahtoiko Eskolta riittää kolmeakaan sanaa
hänelle —? Pyysikö istumaankaan —?

— Milloinka hän kävi täällä?

— Eilen.

— Olitko edes siedettävä hänelle?

— Luulisin… Veli, minä… minä tahtoisin…

— Mitä sitten?

— Hyvittää. Sovittaa hiukan minkä voin.

Lääkäri hymähti surullisesti. Hän ei voinut muuta, muistaessaan


minkä viestin vasta oli tuonut.

— Ja miten luulet voivasi — nyt — enää — niin lohdullista kuin se


olisikin elämän viimeisinä hetkinä.

— Kiitos siitä, että annoit minulle kuoleman varmuuden. Vain sen


avulla voinkin tehdä mitä aion.

— Merkillinen mies…

— Mutta onko kuolema ehdottoman varma?

— Valitettavasti, veli. Sinähän tahdoit tietää alastoman totuuden,


ja minä annoin sen sinulle sellaisenaan.

— Kepeä totuus. Tieto, joka antaa minulle tarvittavat voimat.


Kunhan elinpäiväni eivät sittenkin venyisi liian pitkiksi?
— Korkeintaan kolme, neljä viikkoa on jäljellä.

Sairas käännähti ja katsoi tutkivasti lääkäriin. Hänen hetkeksi


elostuva katseensa tahtoi aivan lävistää toisen kasvot.

— Siinä on liikaakin, sanoi hän sitten kylmästi.

Lääkäri istui vielä hetkisen taaksepäin nojautuneena katsellen


harmaita, tylyjä kasvoja. Niiltä näytti yhtäkkiä pyyhkiytyneen pois
äskeinen odottamaton avomielisyys, ja tiukkaan puristautuva suu
ilmoitti puhelun täksi kerraksi loppuneen.

Mitä mahtoi kuolemaansa kohti kulkija suunnitella, mietti Riska.


Mikä oli oleva hänen sovituksensa? Vajaampana tai täydellisempänä
täytyi sen kuitenkin tapahtua aineellisuuden puitteissa, ainoastaan
siinä. Sielullinen, rakkauden vaatima sovitus — ei, sellainen olisi
tässä tapauksessa ollut kuulumaton. Ehkä tekee hän testamentin,
luovuttaa kaiken jäljelle jäävän omaisuutensa entiselle
morsiamellensa. Niin, sehän oli ainoa mahdollinen. Mutta vähän tuo
toinen siitä kostuisi — hyvin vähän. Esko Kiilan omaisuus sisältyi
tämän vanhanpojan asumuksen seinien sisään, näihin kolmeen
yksinkertaiseen huoneeseen, joissa hän oli asunut pian
parikymmentä vuotta koirinensa, vanhan emännöitsijän hoivaamana.
Ei tästä suuria lähtisi, ainakaan rahallisesti. Mahdollisesti kyllä hän,
joka nämä kapistukset saisi, antaisi niille suuren aarteen arvon,
rakkauden ja muistojen hopeoiman — ja niin tietysti tekisikin nainen,
joka parhaimmat ikävuotensa oli epätoivoista rakkautta povessaan
kantanut. Mutta puutteellisiksi hänen päivänsä silti jäisivät ja
leipänsä yhtä niukaksi kuin ennenkin.

Lääkäri nousi verkalleen tuoliltaan toivottaen rauhallista lepoa,


johon sairas vastasi silmiään vain hiukan raoittamalla.
Huomenna, huomenna hän aikoi jatkaa keskustelua ja huomiseen
mennessä miettiä ja punnita kaikilta mahdollisilta puolilta
aikomaansa askelta.

Arkana ja epämääräisenä ensin mieleen herättyään oli se aie


tuntunut oudolta ja mahdottomalta ja — mikä pahinta — tuntunut
naurettavalta. Mutta siitä huolimatta ei Esko ollut saanut sitä
hyljätyksi mielestään, vaan oli hitain, varovaisin ajatuksin askarrellut
tunti tunnilta sen kimpussa. Nyt, kun hän oli, joskin täysin peitetyssä
muodossa, ottanut sen puheeksi lääkäri Riskan kanssa, tunsi hän
selvästi, että tuo salainen aikomus oli siten ikäänkuin polkenut
varmemman jalansijan itselleen hänen sielussaan. Lujittautunut
siellä. Naurettavuus haihtui pois. Teon, keski-ikäisen, kuolemaansa
kohti kulkevan miehen teon täytyi toki olla jotakin muuta kuin
naurettavaa. Sellaisia askeleita oli maailmassa tai paremminkin
maailmasta erotessa astuttu ennenkin.

Pikainen ja epäämätön kuolema!

Se oli nyt siis varma, se…

Hän katseli elämäänsä sen valossa, sen voimakkaan, kirkastavan


tunnepiirin vaikutuksessa. Tuleva, tuntematon elämä, jonne mentiin
haudan kautta, siinsi hänen edessään kuin rannaton ulappa, suuri ja
summaton, illan hämärä tuntu äärettömyytensä yllä… Sen helmaan
hän liukuisi pian, näin, niinkuin työnnetään vene vesille aavalta
hietarannalta. Mutta ei yhtään, ei ainuttakaan ihmistä saanut jäädä
sille rannalle katkerasti hänen jälkeensä katsomaan.

Ei Laura Paaterokaan… se ainoa, jolla vain olisi siihen ollutkin


todellinen syynsä.
Seuraavana päivänä sai lääkäri Riska tietää, ettei hyvityksen
muoto ollutkaan testamentti.

Saavuttuaan tavallisella määrääjällään Kiilan kotiin tapasi hän


portaissa pastorin, jolla oli sama matkanmäärä kuin hänellä
itselläänkin.

— Ka terve! Olipa kaunista, että tulit katsomaan vanhaa toveria,


ennenkuin hän jättää meidät.

— Tulen hänen pyynnöstään. Sain kutsun myöhään illalla.


Luuletko… luuletko, että se on tehty tulevaista ajatellen?

— Päinvastoin, veli pastori. Oletan ennemmin, että nykyinen, tai


paremminkin se, minkä hän tänne jälkeensä jättää, aiheuttaa hänen
pyyntönsä.

Pastori pudisti päätään.

— Ei, veli Riska. Tai ehkä sekin. Mutta monet niistä, jotka eivät
enää katsoneet tarvitsevansa sinua, ovat lähettäneet kutsunsa
minulle vielä senkin jälkeen. Näes, kun seistään iäisen portilla…

— Kenties. Siihenhän tullaan lopullisesti vain kerran. Kuitenkin


luulisin — ja sinähän tunnet hänet yhtä hyvin kuin minäkin — että
sellainen luonne kuin Esko pyytää apua vain niissä asioissa, joissa
välttämättä tarvitaan välikäsiä, kuten monissa maallisissa. Suo
anteeksi, pastori, eikö iäisen portilla jokainen ole oma pappinsa…

Pastori katsoi miettiväisesti lääkäriä, ja hänen suurilla, avonaisilla


kasvoillaan näyttäytyi surumielinen ilme.
— No niin, veli. Me emme kiistele niistä asioista… Kaikki on Herran
hallussa — Eskommekin.

He astuivat varovin askelin peräkkäin sairaan huoneeseen


istuutuen kahden puolen vuodetta ja lausuen jonkun tervehtivän,
toverillisen alkajaissanan kumpainenkin.

Vanha, kunnon toveri makasi siinä heidän keskellään hiljaisena ja


harvapuheisena kuten ainakin, tällä kertaa erikoista mielenkiintoa
herättävänä sen vuoksi, että oli saanut kaukaisen lähtökutsun, ja
vielä senkin vuoksi, että hänellä oli jotain tärkeätä sanottavaa juuri
tällä hetkellä.

Ja hänellä näytti olevan todellinen tarve sanoa se pian,


siekailematta.

— Älä hämmästy, vanha veikko, sanoi hän kiinnittäen katseensa


pastoriin, äläkä ainakaan tee vastaväitteitä, ellei virkasääntösi ja
kirkkokäsikirjasi sitä ehdottomasti vaadi. Tämä tässä — hän laski
kätensä lääkärin polvelle — on sanonut rehellisesti milloin minun on
lähdettävä, ja se, että minun on lähdettävä, asettaa minulle erään
velvollisuuden ja tekee samalla sen täyttämisen mahdolliseksi. Te
tunnette kumpainenkin suhteeni Laura Paateroon — entiseen
morsiameeni, muistathan, pastori? Ei kovin kauan sitten puhuimme
hänestä, poiketessasi täällä muutaman kuolevan luota tullessasi. No
niin, mutta muistatko, että sinun oli määrä vihkiä meidät silloin
muinaisina päivinä. Tahdotko nyt tehdä sen, mikä silloin jäi
tekemättä?

Pappi ja lääkäri katsahtivat hämmästyneinä toisiinsa. Tätä he eivät


voineet ajatellakaan.
— Sekö siis onkin sinun sovituksesi? virkahti Riska.

— Miksei sekin. Katsokaa, tässä ei ole kysymys rakkaudesta —


avioliiton oikeasta edellyttäjästä — eikä tunteenomaisesta
mielenmuutoksesta. Minun puoleltani. Tosin se, että tunnen ja tiedän
seisovani haudan partaalla, hioo olemuksestani paljon pois sellaista,
mitä maailmaan kuuluva mies pitäisi oikeutettuna ominaisuutenaan
— taikka omituisuutenaan. En ole hetkeäkään ajatellut sitä,
tuottaisiko tämä haudanlikeinen yhdistyminen minkäänlaista
onnentunnetta sille naiselle, joka läpi elämänsä on kulkenut minulle
uskollisena. Enkä haluaisi hänen sitä minulle ilmaisevankaan… Ainoa
ja kaikki mitä tahdon ja tarkoitan on, että saisin turvatuksi hänen
puutteenalaiset päivänsä sillä elinkautisella eläkkeellä, jonka hän
kuoltuani tulisi saamaan valtion virkamiehen leskenä.

Sairas oli puhunut tyynellä, miltei kylmällä äänellä ja lopetettuaan


jäi ummistetuin silmin odottamaan vastausta.

Viipyi kotvanen, ennenkuin pastori hiukan järkytetyllä äänellä


kysyi:

— Oletko puhunut kaikesta tästä morsiamesi… neiti Paateron


kanssa? Ja onko hän täysin suostuvainen sekä tietoinen siitä,
minkätähden sinä haluat ottaa tämän askeleen?

— Ei ole mitään syytä epäillä Eskon avomielisyyttä tässä suhteessa


Laura Paateroa kohtaan, sanoi lääkäri nopeasti, ennenkuin
asianomainen itse ennätti vastata.

— Kiitos, veli. Sinä uskot siihen ainoaan hyveeseen, mikä minulla


on, rehellisyyteeni. Ja pelkään sinun tietävän, että ehkä liiankin
koruttomasti olen sanonut sen hänelle. Siis pastori: voitko täyttää
pyyntöni heti?

— En aivan heti. Sinähän tiedät edellä kulkevat valmistavat


toimenpiteet.

— Mutta poikkeustapauksessa. Eivätkö kirkolliset säädöksesi sisällä


sellaisesta mitään? Oletetaan tämän päivän olevan viimeiseni…

— Kyllä — mutta —

Pastori empi hetkisen. Sitten hän kääntyi Riskan puoleen aikoen


kysyä, oliko sairaan tila niin heikko, ettei hänen luultu voivan elää
niiden viikkojen yli, jotka tavallinen, laillinen järjestys edellytti
kuulutuksen ja vihkimisen välille, mutta samassa hetkessä tuntui
hänestä sellainen kysymys kovalta ja säädyttömältä kuolevan itsensä
lähettyvillä, ja hän vaihtoi sen toisiin, ylimalkaisiin sanoihin, jotka
eivät mitenkään haavoittaneet, jos kohta eivät varmaa, pikaista
lupaustakaan antaneet.

— Minä teen kaiken voitavani, veli, ja täytän pyyntösi niin pian


kuin mahdollista, lopetti hän. — Herramme on kyllä jatkava viimeisiä
elinpäiviäsi niin, ettei sovituksesi työ jää keskeneräiseksi.

Heti siitä päivästä lähtien, jolloin vihkiminen tapahtui, kävi Esko


Kiila ylenmäärin levottomaksi ja rasittuneeksi. Hänen sairautensa
viimeinen kehitysaste esiintyi nopeammin kuin oli voitu aavistaakaan.
Todennäköisesti oli hänellä vain hiukkasen aikaa jäljellä.

Jo vihkimistilaisuudessa — tässä harvinaisessa ja


mieleenpainuvassa, jossa morsian esiintyi hunnutta ja kruunutta ja
sulhanen tuskin jaksoi päätänsä tyynyltä nostaa, — huomasi lääkäri,
kuinka jokin ennentuntematon kiusallisuudansävy laskeutui miehen
koko olemukseen ja täydellisesti valtasi sen. Eikä hänen tarvinnut
kauan arvella muutoksen syytä, vielä vähemmin kysyä sitä, sillä jo
seuraavana päivänä pyysi Kiila:

— Siirrä minut sairashuoneeseesi, kunnon Riska, siirrä vielä tänä


päivänä. Pyydän sitä sielullisten, en ruumiillisten kipujeni takia. Nyt,
kun olen antanut hänelle, Lauralle, minkä saatoin antaa, tahdon
päästä vanhaan yksinäisyyteeni kuolemaan. Sinä ymmärrät. Tätä
taloa en tunne kodikseni enää… Näiden huoneiden ilma on käynyt
vieraaksi minulle…

Lääkäri pyyhkäisi silmiään salaamatta liikutustansa. Ja sairas


jatkoi:

— Selitä taudin käänteen vaativan siirtoa … kiellä kaikilta


katsomaan pääsykin… Sinä vanha, kunnon toveri ymmärrät, ettei
minulta saata liian paljon vaatia, ainakaan tällaisessa asiassa…

Hän ymmärsi sen hyvinkin, hän.

Ja astuessaan hetkistä myöhemmin sairashuonetta kohti


tehdäkseen siellä ystävänsä viimeisen pyynnön aiheuttamat
valmistukset tuli hän vakuutetuksi siitä, että Esko Kiila oli
harvinaisuus niin hyvin eläessään kuin kuollessaankin.

KUN ONNI MURTUU


<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<51
Komea Martta Herlevi — seurapiirin keskus ja ihanne — istui
kyyrysissään pienellä nurkkasohvalla häpeän ja epätoivon
murtamana. Vielä toissa päivänä hän oli iloiten ja onnea sädehtien
emännöinyt tässä rikkaasti sisustetussa huoneessa kaupungin
parhaimmiston ympäröimänä.

Pankinjohtaja Jonas Herlevin pidot — kas, nehän olivat jotain


erikoista. Kutsu niihin oli samanarvoinen kuin pääsiäistähtisateen
osallisuuteen pääseminen ja poissulkeminen niistä katkeraa kuin
pakollinen virkaero. Olihan joka taholla päivänselvää, että ne, jotka
seurustelivat tässä suuresti kunnioitetussa talossa, olivat
paikkakunnan moitteettominta valioväkeä, sen hienointa ja —
kadehdituinta säätyä. Isännän miehekäs, ehdotonta luottamusta
herättävä olemus ja emännän henkevä sulo vetivät puoleensa jo
sinänsä, puhumatta siitä yhteiskunnallisesta kiillosta, joka ympäröi
taloa.

Ja nyt —!

Komea Martta Herlevi painoi käsillään ohimoitaan viskautuen


epätoivoisena sohvalle, josta hän oli noussut ja johon taas oli
laskeutunut monen monta kertaa kuluneen, hirvittävän
kadotuksenyön kuluessa. Hänen miehensä oli vangittu
miljoonakavalluksesta johtamassaan pankissa, tieto siitä oli salamana
kiitänyt läpi kaupungin — ooh! — kaikkialle ympäri koko maan.
Sukulaiset, ystävät ja tuttavat olivat jo saaneet tietoonsa tuon
uskomattoman sanoman…

Pitkänä, syksyisenä yönä, jolloin kuu ja tähdet kurkistivat


akkunoista sisään, oli hän iskun huumeessa käynyt läpi nuoren,
voitokkaan ja onnellisen elämänsä. Pää pystyssä oli hän tietään
kulkenut, ylpeänä ja onnellisena Jonas Herlevin rinnalla, joka aina,
heidän tutustumisensa ensi hetkestä asti, oli omistanut hänen
ehdottoman luottamuksensa ja lämpimän ihailunsa. Hän oli itse yksi
niistä kristallikirkkaista luonteista, jotka jokaisena hetkenään voisivat
antaa katsoa sielunsa syvimpään soppeen tarvitsematta painaa
silmiään alas kenenkään edessä. Ja niin piti hän suurimpana
onnenaan sitä, että hänen miehensä oli paikkakunnan luotetuin
henkilö, vilpittömin olento kaikessa, olipa sitten kysymyksessä
rahalliset ja yhteiskunnalliset asiat tai puhtaasti tunteellinen puoli.
Vielä eilen, aamupuoleen, olisi hän syvimmällä inholla torjunut
luotaan kenen tahansa, joka vain pienimmänkään tahran olisi
viskannut Jonas Herlevin kunnialle. Nyt olivat häneltä kaikki aseet
riistetyt. Kalliopohja, jolla hän turvallisena oli seisonut, luisui pois
tahmean, likaisen liejun lainehtiessa rannattomana hänen
ympärillään…

Mutta jos jotkut niistä, jotka nyt puolella ja toisella selvitteleivät


tästä lähintä yhteiskuntaakin ankarasti kohdanneesta kolauksesta,
luulivat, että hän hylkää onnettoman miehensä tai edes vierookaan
häntä, niin he erehtyvät suuresti. Olisiko hän oikea vaimo, jos
mustimmalla hetkellä jättäisi hänet, jolle hyvinä päivinä tiesi olleensa
kaikki kaikessa. Hänet, joka ehkä odottamattoman onnettomuuden,
vielä selvittämättömän seikan pakoittamana oli tullut tehneeksi sen,
josta häntä nyt syytettiin.

Ei, viimeiseen asti tuli hänen seisoa uskollisen puolison


kaikkivoittavalla kärsivällisyydellä hairahtuneen miehensä rinnalla.
Olihan se vain velvollisuus, johon rakkaus ja uskollisuus yhtyneinä
antoivat voimaa.

Mutta näinä lähipäivinä, ennenkuin järkytetyn mielen rauha ehti


hiukkasen palautua, ei hän voinut eikä tahtonut nähdä ketään
tuttavistaan. Jonkinlainen vaistomainen pelko sanoi hänelle, että
heillä, kaikesta huolimatta, voisi olla jotain salattua pahaa
katseissaan… Mielethän muuttuvat niin pian… Ja monasti ilman
syytäkin osoittautuu tänään särkyneeksi se, mikä eilen vielä oli lujaa
kuin ystävyys.

Hän kulki aatoksissaan perheestä toiseen, ystävän luota ystävän


luo. Mutta niin läheisiltä kuin he kaikki tähän asti olivat tuntuneetkin,
ei hän nyt saanut sitä luottamusta syntymään heitä kohtaan, jota
onnettomuuteen joutunut helpoituksekseen tuntee ystävän
uskollisuudessa. Joku oli käynyt häntä tapaamassa, mutta hän ei
ollut ottanut vastaan. Taaskin kuului ovikellon soitto. Turhaan —
turhaan. Hän pysyi järkähtämättömänä. Sillä aivan varmasti keksisi
hän sen salatun osaaottavimmankin katseen pohjalta…

— Minun täytyy saada olla yksin, Hanna, sanoi hän palvelustytölle.


— Ei kukaan muu kuin pikku Ossi saa tulla luokseni, ymmärrätkö…

*****

Mutta paria tuntia myöhemmin tuli palvelijatar arkaillen jälleen, ja


hänen äänensä, aivan lyhyesti ilmoittaessaan neiti Erna Reinin, oli
nöyrää rukousta täynnänsä. Martta loi häneen nopean katseen,
jolloin tytön silmät kyyneltyivät ja väreet hänen kasvoillaan sanoivat:
»ettehän voi olla pitempään yksinänne… tehän menehdytte…» Ja
niin purskahti hän itsekin itkuun, tyrskyvään, valtavaan, lausuttuaan
hiljaa, Erna, Erna —!

Eikö siinä kuitenkin ollut yksi, yksi ainoa niistä monista, jonka nimi
herätti hiljaista, vapauttavaa luottamusta. Sen saattoi nähdä hänen
koko olemuksestaan, kun hän hetkisen kuluttua astui kevein askelin
hänen luokseen ja istuen sohvalle hänen viereensä kietoi mitään
virkkamatta kätensä hänen ympärilleen.

Ja niin kului pitkä, hiljainen hetki.

Martta painoi kuumat, itkettyneet kasvonsa ystävän olkapäätä


vasten, ja viileän päällysvaatteen kosketus toi jo sinänsä ulkonaisesti
helpottavan tunteen.

Mutta hänellä ei ollut mitään sanomista, ainoastaan nyyhkytyksiä.


Kaikki se raskas ja pimeä, jota sielu oli tulvillaan, kieltäytyi
sanallakaan purkautumasta kuuluville. Ja äänetönnä kuin kivinen
patsas istui Ernakin.

Vasta sitten, kun viereisestä huoneesta alkoi kuulua pikku Ossin


iloista leperrystä, käännähti Erna hieman tavatakseen Martan
katseen ja virkkoi ainaisella rauhallisella äänellään:

— Kuulehan, me lähdemme autolla Saarnivaaraan. Sinun on hyvä


päästä joksikin aikaa täältä pois… Ja pikkumies varmaan oikein
riemastuu matkasta.

Martta kohottautui kiihkeästi. Mutta ennenkuin vastasi, katsoi hän


ja näki rauhoituksekseen, ettei Ernan katse suinkaan kuvastanut sitä
pelättyä pahaa — ei hiekkajyväsen vertaa. Sekunnissa ehti hän sen
todeta sisimmässään.

— Ei, Erna. Tyydy ajamaan Ossin kanssa, sitä pyydän sinulta. Minä
en voi, en saata lähteä…

— Martta kulta, ajattelin, että viipyisimme muutaman päivän


Saarnivaarassa. Järkytys on sinulle liian suuri tässä ympäristössä…
Ja syyspäivä on mitä ihanin, — jatkoi hän kiireesti, — niin kevyttä,
raikasta kaikkialla ulkona.

Martta intti vastaan.

— Minä en voi… toisti hän. — Minun paikkani on täällä nyt.


Tuntisin itseni pelkuriksi ja petturiksikin, jos poistuisin
kaupunginportin ulkopuolellekaan. Erna, minkätähden hän lienee
tämän tehnytkin, täytyy minun kuitenkin tukea häntä, edes
läsnäolollani… käynneilläni, jos sallitaan.

Hän oli kuivannut kyyneleet tummista, hehkuvista silmistään,


joissa paloi kiihkeä, kärsimätön tuli. Ja puhuessaan hän aivan naulasi
katseensa Ernan kasvoihin, jännityksellä, päästämättä pienintäkään
elettä huomaamattaan raukeamaan.

Ja kun Martta nousi katsoen häneen läpitunkevan pyytävästi, ei


hän voinut muuta kuin totella. Tuossa tuokiossa oli pikkumies puettu
rikkaisiin, lämpimiin tamineihinsa ja kannettu alas oven edessä
odottavaan autoon.

Myöhemmin päivällä, Ossin ja Erna Reinin palattua pitkältä


retkeltään keltapunaisiin metsiin, joissa kesäinen vihreys oli
vaihtunut syksyn moninaiseen väriasteikkoon ja lintujen ainainen
liverrys muuttunut ihmeteltäväksi hiljaisuudeksi, vaati muuan vieras
nainen huomattavan kiihkeästi puheillepääsyä rouva Herlevin luokse.
Hän ei ottanut kuullakseenkaan palvelijattaren vastausta, ettei rouva
ota ketään vastaan. Itsepäisesti hän vetosi asiansa tärkeyteen ja
kiireellisyyteen eikä poistunut, ennenkuin hänelle oli myönnetty
puheillepääsy.
Vieras oli nähtävästi toivonut tapaavansa rouva Herlevin
kahdenkesken, sillä hänen katseensa viipyi pitkään ja
tyytymättömänä Erna Reinin kasvoilla, joka seisoi Martan vieressä,
tämän tuolinselustaan nojaten.

Hän näytti tuntevan Martan, vaikkei Martta tuntenut häntä.

Molemminpuoliset mittailevat silmäykset kävivät samassa


jonkinlaisesta taisteluunvaatimisesta, ja Martta tunsi heti voimakasta
vastenmielisyyttä itsepäistä vierastaan kohtaan. Hänen omituisesta
olemuksestaan, josta ei suinkaan ulkonaista kauneutta puuttunut,
huokui silminnähtävää röyhkeyttä — luontaista vaiko tehtyä — johon
mitä kummallisimmin sekaantui kaiken läpi kuultavaa noloutta. Ja
kumarrus, jonka hän teki tervehdykseksi, ilmaisi hänen voivan
alentua matalimpaankin imarteluun.

Martta Herleviä puistatti, hän ei ollut koskaan tavannut


senkaltaista ihmistä.

Minkäänlaista lähestymistä ei tapahtunut kummaltakaan puolen.


Ainoastaan kädenliikkeellä osoitti rouva Herlevi ovenviereistä tuolia
vieraalle tämän esiteltyä itsensä:

— Line Kjäll.

Seurasi sitten hetken hiljaisuus, jonka tukaluus rasitti jokaista


huoneessa-olijaa.

— Tahtoisin kahdenkesken puhua pari sanaa rouva Herleville, jos


saan luvan…

Martan katse pysäytti pyytäen Ernan, joka aikoi poistua.


— Voitte yhtä hyvin puhua asianne ystäväni kuullen. Minulla ei ole
mitään salattavaa häneltä.

— Niinkuin tahdotte.

Vieraan äänessä oli pisara myrkkyä, ja omituinen hymy ailahti


hänen kasvoillaan.

Mutta ennenkuin hän sai sananpäästä kiinni, kiskoi hän toista


riisuttua hansikastaan sormiensa välissä, neuvotonna, kuten näytti,
erilaisten ilmeiden risteillessä nopeasti hänen voimakkaan,
kaunispiirteisen suunsa ympärillä.

— Toimeentuloni pakoittamana — alkoi hän sitten epävarmasti,


alasluoduin katsein — — täytyi minun tulla luoksenne näin
sopimattomaan aikaan. Miehenne on hoitanut eräänlaisia asioitani, ja
nyt olen saanut häneltä tiedon, että hänen kirjoituspöytänsä
laatikossa on minulle osoitettu kirje, raha-asioitani koskeva.

Line Kjällin sanat ja ääni olivat lisänneet rouva Herlevin sisäistä


kammoa. Hänestä tuntui, kuin piilisi vieraan asiassa kaikkein
hirvittävintä, ja hän tunsi jalkainsa yhtäkkiä hervahtavan
voimattomiksi. Muuten hän olisi heti noussut etsiäkseen pyydettyä
kirjettä ja antanut sen hänelle päästäkseen niin pian kuin mahdollista
erilleen tästä naisesta, jonka olemus nostatti ennentuntemattoman
vastenmielisyyden hänen sydämeensä.

Nyt hän ei uskaltanut nousta.

Hetkinen ja vielä sittenkin hetkinen täytyi hänen koota voimiaan ja


varmistautua siitä, ettei horjahtaisi lattian poikki astuessaan.

— Ja minuako mieheni on pyytänyt antamaan teille tuon kirjeen?


— Hän sanoi, luullakseni, että asessori Reimerkin voisi sen tehdä,
mutta hänhän on matkoilla tänä päivänä.

— Vai niin. Ja teidän nimenne oli —?

— Line Kjäll.

— Rouva?

— Neiti.

Vieras lausui sen taas alasluoduin katsein, ja Martta Herlevi nousi


mennäkseen miehensä työhuoneeseen, jossa kaikki oli vielä samassa
kunnossa kuin sinä hetkenä, jolloin hänet sieltä riistettiin.

Keskustelun aikana ei Erna Reini ollut hetkeksikään kääntänyt


katsettaan vieraasta. Syrjäisenä ja vaieten oli hänen ollut helppo
tehdä havaintojaan ja päätelmiään, joissa hän piankin, tavattoman
teräväsilmäisenä ja älykkäänä, ehti tulla karmaisevaan
lopputulokseen.

— Onko mahdollista, että Jonas on käskenyt Martan antamaan


kirjeen tuolle… myllersi hänen tulistuneessa mielessään. — Kuinka
kummassa hän on julennut laittaakin tämän tänne… Taivaan isä, eikö
niissä vieraissa miljoonissa ollut kylliksi!

Ylpeällä päänheitolla siirsi hän katseensa naisesta pois ja tähysti,


näköjään välinpitämättömänä, ulos punaiselle peltikatolle, jossa tuuli
tanssitti kellastunutta vaahteranlehteä.

Siinä voi siis sittenkin olla perää, mitä hän kaupungilla oli kuullut,
niin uskomattomalta kuin se hänestä talon läheisimpänä ystävänä
tuntuikin… Eikö siis herttaisimmastakaan aviomiehestä ollut takeita…
Eikö ainutkaan heistä ollut vapaa tuosta mustalaisverestä… eikö…
eikö…

Sillä välin selaili Martta miehensä laatikoita, ensin turhaan, kuten


näytti. Mutta viimein sattuivat hänen silmänsä pieneen
nahkasalkkuun, joka ei ollut paljon kookkaampi tavallista kirjekuorta,
ja hän veti sieltä esiin sulkemattoman kirjeen, jonka kuoreen, aivan
oikein, oli kirjoitettu: »Neiti Line Kjäll. Täällä».

Nopeasti, kuin olisi paperi polttanut hänen sormiaan, aikoi hän


sulkea kotelon, mutta samassa paisui herännyt epäilys
tuhatkertaiseksi, ja unhoittaen aikaisemman itsensä ja tapansa
vetäisi hän kuoressa olevan paperin esiin, levittäen sen eteensä
pöydälle. Paperiarkilta, jonka sisään oli pistetty tuhannen markan
seteli, luki hän miehensä tutulla käsialalla kirjoitetut sanat:

»Luultavasti on tämä viimeinen summa, jonka pitkiin aikoihin


voin sinulle antaa. Mutta jos joudut puutteeseen, käänny August
Reimerin puoleen. En tahdo, että jäät kärsimään. J. H.»

Martta Herlevi jäykistyi hetkeksi näiden rivien eteen. Ne


muodostivat ruman portin, jonka takaa häämöitti pitkä, kaamea
holvi, josta ei voinut eroittaa muuta kuin pimeyttä — pimeyttä —
loppumattomiin!

Oliko hän enää oma itsensä —?

Oliko tämä hänen ja hänen lapsensa koti —?

Ja oliko se hänen Jonaksensa, joka oli kirjoittanut nuo sanat tuolle


paperille —?
Yhtäkkiä hän tunsi hurjan repäisyn rinnassaan, korvissa humisi, ja
jostain syvästä rotkosta, oman järkytetyn sielunsa pohjalta, hän kuuli
kamalan naurun: Jonas… miksei! Miljoonakavaltaja… Mikä estää sitä
pettämästä omaa vaimoansa, joka pettää koko yhteiskunnan,
tuhannet luottavat ihmiset!

Ja nyt seisoi tuolla toinen julkea ja alhainen ihminen odottamassa


hänen avuliasta palvelustaan saadakseen käsiinsä häpeän setelin…
verirahan…

Niin suuri kuin vastakohdannut häpeä ja nöyryytys olikin, kävi se


tämän kautta tuhansin kerroin kammottavammaksi. Kaikesta
huolimatta oli hänellä äsken vielä tuki, jalansija, millä seistä, ja se
tuki oli ollut hänen miehensä luuloteltu uskollisuus.

Luuloteltu — niin — sillä tarvitsihan sukeltautua esiin vain jonkun


hämärän Linen — ja kaikki, kaikki oli hukassa!

Ylpeyden ja omanarvontunnon esiinpyrkivä välähdys herätti hänet


tästä musertavan pettymyksen tuottamasta huumauksesta.

Ei ainakaan tuo tuolla saanut hänen masennustaan huomata!

Ja niinkuin ainakin jalo ihminen, joka jonkun erityisen syyntähden


on kulkenut kappaleen matkaa toverinsa rinnalla rikoksen tietä,
mutta kääntyy heti pois, kun yhä suurempi rikos avaa hänen
silmänsä, nousi Marttakin, vaikka jalat horjahtivat, sulki kirjeen, vei
sen odottavalle Line Kjällille sanoen tyynesti kuin tavarakuittia
antaessaan:

— Tässä se näkyy olevan.

*****
Vieraan mentyä valtasi hänet polttava halu juosta heti pois tästä
talosta.

Monet uhkeat, rikkaasti sisustetut huoneet, jotka olivat olleet


hänelle äärettömän rakkaat, kasvit, joita hän hellin käsin oli hoitanut,
ja monet kymmenet pikkutavarat, jotka hän oli kasannut tänne kuin
lintu korsia pesäänsä, kävivät vieraiksi ja sietämättömiksi yhtäkkiä.
Ne eivät enää olleet hänen.

Ah, vähät siitä, jos ne olisi pitänyt uhrata Jonaksen rikoksen


tähden — siitä hän oli jo viimeisen unettoman yön kuluessa ehtinyt
itsensä kanssa sopia. Mutta tämä, että tuntematon, saastainen käsi
pyyhkäisee kaiken yli ja asettaa kaikki eletyt onnelliset päivät aivan
toiseen valoon kuin siihen auringonkirkkaaseen, jossa ne vuodesta
vuoteen olivat hohtaneet!

Oliko hän enää Jonas Herlevin vaimo —?

Tuskin, tuskin aikoihinkaan, vaikkei hän itse ollut siitä tiennyt.


Hänen avio-onnensa, josta hän oli iloinnut, kuten iloitaan harvinaisen
timantin omistamisesta, olikin vain kurja hely, niin taitavasti
väärennetty, että vasta väärentäjien oma varomaton menettely
saattoi sen ilmaista hänelle.

Hän kulki kuin unessa huoneesta huoneeseen, kertomusten


aaveiden lailla, jotka eivät saa rauhaa haudassaan. Jos hän pysähtyi
missä tahansa, kiinnitti katseensa mihin tahansa, aina kurkistivat
esiin iloiset, huolettomat muistot. Tämä koti oli kuin puisto, jossa
kukki, tuoksui ja soi… oli ollut…

Nyt kattoi valkea härmä sen ihanuudet, ja talvinen sää uhkasi


nietostaa joka sopen.
*****

Hyvä Jumala, hänen täytyi mennä Jonaksen luokse — ehkä vielä


jotakin oli pelastettavissa.

Mutta hän ei voinut tulla lämpimänä kuten edellisellä käynnillään.


Epätoivoisesta sydämestään täytyi hänen kysyä heti — heti:

— Kuka on Line Kjäll, ja mitä sinulla on hänen kanssansa?

Mies, jolta kaikki portit olivat suljetut ja joka oli saanut tuhansien
ihmisien kunnioituksen itseänsä kohtaan vaihtumaan
halveksumiseksi, säpsähti. Hänellä ei ollut voimaa kätkeä
rikollisuuden ilmettä siltä ainoalta, jolta hän vielä odotti lämpöä.
Vajaan vuorokauden oli hän istunut täällä telkien takana jaksamatta
ajatella muuta kuin menetettyä kunniaansa. Tämä aika, niin lyhyt
kuin se olikin, oli ankaruudellaan ehtinyt riisua häneltä kaiken
ylpeyden ja vilpin — ne eivät enää auttaneet. Hän oli valmis
tekemään tiliä jokaisesta teostaan ja avoimesti vastaamaan kaikkiin
tehtyihin kysymyksiin. Hänellä ei ollut enää mitään muuta
kaunistuksen keinoa.

Kuitenkin nyt, Martan seisoessa hänen edessään tämä kysymys


huulillaan, lysähti hän kokoon, painoi päänsä alas ja vaikeni. Kuinka
hän jaksaisi vastata siihen… kuinka luopua parhaimmasta, mitä vielä
jäljellä oli…

— Ole rehellinen, Jonas. Ainoastaan sillä voit jotain pelastaa.

Jonas Herlevi istui kauan ääneti. Hikipisarat kihosivat hänen


otsalleen, ja tuontuostakin hän väänteli käsiään tuskallisesti. Viimein
hän nosti kyyneltyneet silmänsä ja koetti puhua, mutta kaikki, mitä
hän sitten tahtoikin sanoa, supistui kahteen tukahtuneeseen sanaan:

— Anna anteeksi!

Miehen nöyryys hellytti Martan mieltä, ja hänen äänensä ei ollut


enää ankara kuten äsken, hänen sanoessaan:

— Onko sinulla enää oikeutta anteeksipyyntöönkään…

— Kaikilla on siihen oikeus. Joskaan ne, joilta sitä pyydetään, eivät


läheskään aina ole velvolliset anteeksi antamaan… Martta, on
kauheata seisoa sinun edessäsi kaksinkertaisena rikollisena…

— Joskin vasta sitten, kun rikokset ovat tulleet ilmi…

— En tahdo puolustautua — millä sen tekisinkään. Mutta sinä olit


parhain puoliso, ja siksi uskallan —

— On oikein sanoa olit. Nyt kai en sitä enää saata olla.

— Martta! Niin et sanonut eilen. Mikset voi, mikset saattaisi?

— Siksi että sitä on myöskin Line Kjäll.

— Line Kjäll — on paljasta roskaa vain. Semmoista on maailman


meno.

— Hänestä sinä kuitenkin huolehdit viime hetkelläsi. Oi hyvä, hyvä


Jumala, olisit edes sulkenut tuon kirjeen!

— Olinko sen sitten jättänyt sulkematta?

Rouva Herlevi kohautti olkapäitänsä.


— Tämä on kohtaloa, Martta, tämä kuten sekin, että viimeisten
pitojamme aikana tarkastaja saapui pankkiin… Sinä ainoa, älä jätä
minua!

Jonas Herlevi tapaili vaimonsa kättä — turhaan. Kyyneleet estivät


häntä näkemästä, kuinka Martta kätki kätensä selkänsä taa,
peräytyen miltei tiedottomasti askeleen taaksepäin.

— Olen luottanut sinuun kuin taivaalliseen isäämme. Voitko


käsittää miltä tuntuu nyt…

— Miltä tuntuneekaan, anna anteeksi — anna!

— Jos en voi…

— Sinä voit!

Se tuli varmaan luottavimmasta sydämestä, niin toivorikkaalta se


kuulosti.

Rouva Herlevi painoi rajusti päänsä käsiinsä. Sinä voit! Sinä voit!
humisi hänen korvissaan. — Sinä et voi! kiisti toisaalta vastaan.

Ja vasta tällä hetkellä selvisi hänelle täydellisesti, kuinka särjetty ja


rikkirevitty hän oikeastaan oli… kuinka epätietoinen ja vaappuva, kun
tosi-taistelu tuli kysymykseen. Ja että taistelu oli alkava nyt…
epätoivoinen ja raju, sen hän tunsi. Nämä kauheat olosuhteet
järkyttivät hänen sieluansa niin suuresti, ettei hän pitkään aikaan
pääsisi täysin selville ratkaisusta — senkin hän tunsi. Hän vihasi
itseään tuntiessaan säälistä Jonakseen joutuneensa sisimmässään
ristiriitaan. Ja pyrkiessään selviytymään ristiriidastaan sääli hän
omaa heikkouttaan.
Jos hän olisi muutamaa päivää aikaisemmin saanut miehensä
uskottomuuden tietoonsa — olisi hän nyt kaukana täältä, iäksi
poissa!

Ei, hän takertui siihen ajatukseen sittenkin katsahtaessaan


Jonaksen masentuneihin kasvoihin.

Iäksi — niinkö?

Ja taas hän ei ollut varma, etteikö sittenkin kavalluksen ja


vangitsemisen viesti olisi palauttanut häntä — säälistä —
tuhoutuneen miehensä luokse.

— Martta, sinä et sano mitään… Etkö voi … etkö tahdokaan…

— En tiedä, Jonas… Siltä, joka on lyöty tainnoksiin, ei voi vaatia


vastausta. Minun täytyy ensin tointua tästä. Ja sitten, kun olen
selvillä itsestäni… sitten Jonas…

Miehen murtuneet kasvot sävähtivät epätoivosta. Tämä hänen


ensimäinen tuomionsa oli varmaan ankarin. Eikä ollut mihinkä
vedota siitä…

— Jätä minulle edes toivon kipinä! huudahti hän. — Martta,


rakkahin, älä vie kaikkea… Kuinka minä muuten jaksaisin tämän
pimeän ajan ylitse…

Sehän se juuri oli Martankin paula. Siitä silmukasta hän ei päässyt


irralleen.

Toivon kipinän — sen hän saattoi antaa. Mutta enempään, hyvä


Jumala, enempään ei hänellä ollut voimaa…
KARKURI

Hiljainen koputus kuului ovelta.

Sitä tuskin eroittikaan tuulen ruskinalta ja aaltojen pauhulta, jotka


lakkaamatta löivät rantapenkereeseen tuvan akkunan alla. Vasta
tarkan kuulostamisen jälkeen pääsi varmuuteen siitä, että ihminen
siellä liikuskeli.

Ja silloin karkasi tuvassa yksin istuva tyttö ylös työnsä äärestä,


viskasi tukun päreitä tuleen ja hiipi ovelle kuulostamaan.

— Kuka siellä?

Kysymykseen ei kuulunut hiljaisintakaan vastausta. Laineet vain


meurusivat rannalla täydentäen pauhinallaan pimeän syysillan
kolkkoutta.

— Kuka siellä — vastatkaa!

Kysyjän ääni oli tuikea, mutta pelon värettä se ei kyennyt tyyten


salaamaan.

— Ystävä on — avaa!
Vapisevin käsin raoitti tyttö ovea, vetäisi sen kiinni ja raoitti taas.
Siten hän taisteli hetken oman arkuutensa ja sisäänpyrkijän äänen
herättämän turvallisuuden välillä. Mutta ulkopuolella-olija ei jaksanut
odottaa. Hän tarttui oveen suurella, tahraisella kädellään ja työnnälsi
huoneeseen.

Takassa loimottavan tulen valossa tyttö näki kookkaan miehen,


jonka parroittuneet kasvot näyttivät väsyneiltä. Koko hänen
jättimäisestä olennostaan kuvastui syvä uupumus ja raukeus.

Mutta silmissä oli eloa. Ne miltei leimusivat levotonta sielun


kuohua, ja nopeasti kuin tulen liekki kulki niiden katse ympäri tuvan,
katto-orsien, permannon, lieden ja peittämättömien akkunoiden
välillä.

Sitten vasta hän katsahti tyttöön.

— Oletko sinä Anna Mari?

— Olen.

— Ja niin olet tullutkin äitiisi…

Vieras ojensi arkaillen kätensä, ja arkaillen otti tyttökin sen


omaansa.

— Kuka te sitten olette?

Miehen katse kiersi uudelleen tuvan, kiersi tarkemmin kuin äsken,


ja hänen kasvonsa vavahtivat omituisesti. Vastaamatta tytön
kysymykseen hän sanoi:

— Lienetkö tullut äitiisi luonteesikin puolesta.


— Ja jos olen… mitä sitten?

— Sitten sen sanoisin…

— Minkä?

— Nimeni.

Anna Mari tuijotti hämmästyneenä vieraaseen, jonka puhetapa ja


katse olivat niin kummalliset. Ensi kertaa näki hän tuon miehen
edessään, ja kuitenkin tuntui, kuin tapaisi hän hänessä vanhan
tuttavan. Ja niin hänen pelkonsa alkoi haihtua.

— Mutta sanokaa, sanokaahan toki! huudahti hän välittömästi. —


Eihän teille voi siitä mitään vahinkoa koitua.

Jättiläisen kasvot vavahtivat jälleen. Niissä kuvastui suuri


mielenliikutus. Sitä peittääkseen hän kumartui lieden kulman
pimentoon ja vetäisi sieltä kuin vanhasta tottumuksesta esiin lyhyen
penkin, jolle istahti.

— Anna Mari, sanoi hän sitten, etkö sinä tunne minua?

— En.

— Etkö muista koskaan nähneesikään?

— En.

— Ei ihmekään, olithan siihen aikaan vielä pieni tyttö.

— Mihin aikaan sitten?

— Siihenpä siihen, kun isäsi… vangittiin.


Anna Marilta putosi puu, jota hän juuri oli asettamassa lieteen, ja
hän kuiskasi pelokkaasti:

— Isä! Olisitteko…

Mies käänsi hetkeksi katseensa toisaalle, ja Anna Mari huomasi,


kuinka hänen hartiansa hytkähtelivät. Mutta pian hän jälleen
oikaisihe ja tytön puoleen kääntyen virkkoi:

— Kas niin, älä nyt pelkää, Anna Mari, ethän? Ettehän te minua
vielä odottaneet… eikä minun aikani ollut vielä tullutkaan… Mutta
minä en kestänyt siellä kauempaa… totisesti, minä en jaksanut… Ja
kun viimein sattui tilaisuus, jolloin saatoin karata, tein sen, vaikka
henkeni uhalla. Ehkä olet lukenut siitä sanomalehdistä?

Anna Mari tuijotti sanatonna eteensä. Hänen jalkansa vapisivat


niin, että hän töintuskin kesti seisaallaan.

Ja muistot, jotka vuosien kuluessa olivat haaltumistaan


haaltuneet, leimahtivat ilmi eläviksi jälleen. Hän muisti päivän, jona
isä vangittiin… muisti kirstun, johon ruhjottu mies pantiin… ja muisti
äidin, joka siitä päivin kulki kuin varjo maan päällä, kunnes hänet —
kolme vuotta sitten — peitettiin kirkon multaan.

Tuoko nyt oli isä —?

Niin, suuren ja voimakkaan hän muisti hänen olleen.

Mutta tämä tässä oli hiljainen ja hyvä…

Isä oli aina ollut kuin tulenliekki, eikä hänen suustaan oltu
lempeitä sanoja totuttu kuulemaan.

You might also like